Introduction

This is the book I want my grandchildren to read when they learn twentieth century American history. Enough happened in that century that it deserves an entire book.

I am an economist, not a historian. But academic historians are specialized. They do not seem eager to write a book of this scope. And I would not necessarily encourage them to do so, given the narrow, smug ideological bias that pervades their profession.

My perspective emphasizes contingency, uncertainty, and moral ambiguity. Think in terms of two “iron laws:”

First Iron Law: Sometimes it’s this way, and sometimes it’s that way.

Second Iron Law: Outcomes are often unintended.

My father, political scientist Merle Kling, coined the First Iron Law. A typical progressive historian describes markets as causing every bad economic problem and government as providing every good solution. The opposite bias would claim that it is the other way around. My perspective is that sometimes it’s this way and sometimes it’s that way.

The Second Iron Law is my formulation. Results frequently diverge from what decision-makers intend, and often they emerge regardless of any explicit decision. Some important phenomena, such as the rise of the suburb, developed because of how the economy and transportation evolved, not due to the direction of a central planner.

Policy responses to the housing finance debacle of the 1930s ultimately created the conditions for the Savings and Loan debacle of the 1980s. The responses to that in turn created the conditions for the 2008 Financial Crisis. Outcomes are often unintended.

This project began as a set of lectures written in collaboration with the Large Language Model known as Claude, primarily version Opus 4.6, and serialized on my substack newsletter. In writing this book, I used AI as much as possible, except for finally writing the prose.

The style in which the LLM wrote I found grating. My natural style is journalistic. I like short sentences. I prefer to keep the tone matter-of-fact. I think that the author should stay in the background, not always be present the way I am during this introduction. The LLM was theatrical and melodramatic. Its sentences and paragraphs dragged on. It was constantly on stage, telling the reader how to react.

Maybe the LLM’s style is more fun to read. You can still find the written lectures on line and decide for yourself.

With Claude, I broke the century into the periods that became the chapters of this book. I proposed what I thought should be the key themes concerning each period. After some back and forth with the LLM, I settled on the themes that became the sub-heads within each chapter. I posed many questions. The LLM and I went back and forth before I asked it to draft the lectures.

For the book, I used the structure of the LLM lectures, since that already reflected my preferences. As I wrote the book, I consulted other LLMs constantly concerning details. I used Google Gemini, because it was handy. I used OpenAI’s ChatGPT to check the answers I was getting from other LLMs.

As I turned the lectures into this book, new questions came to me. For example, the original LLM lecture described how Jews fleeing the Holocaust were kept out of the United States. It put the blame for this on Breckinridge Long, a key official at the State Department. The lecture made it seem that when President Roosevelt was finally made aware of this in 1944, he immediately adopted a more humane approach. But as I read this over, it occurred to me to ask, “Did Roosevelt fire Long? And if not, why not?” It turns out that the answer does not put Roosevelt in such a favorable light.

I did not discard melodrama altogether. You will find some sections, marked You are there, written in the present tense. In some of these, I use my own family history. I write about: my grandfather, who went bankrupt during the Great Depression; my father’s sister’s husband, who was a Communist union organizer in the 1930s; my mother, who, because of her own brief membership in the Communist Party (which is how she met my aunt and my father) was called before the House Un-American Activities Committee in 1956 ; and one of my mother’s cousins, who had a moment on stage at Woodstock. Finally, I include myself, especially my early foray into Web commerce in the 1990s.

I hope that these stories help make this history come alive. They should at least be interesting to my grandchildren.

Epilogue

One reason that I was attracted to this project is that immersing myself in the 20th century allowed me to escape the 21st. The ill-tempered times in which we live are not pleasant to dwell on.

I think that technological change was more important in the 20th century than in the century that preceded it. Play the game of time-travel switch, in which you imagine taking someone from, say, 1776, and swapping them into 1876. What would seem different? The railroad would be new. So would steam-powered shipping. The telegraph would be cool. But our time traveler could still feel at home on a farm or on a town street.

Next, bring the thought-experiment forward. Swap someone living in 1876 into 1976. I do not have to go through the whole litany of what this time traveler would have to become accustomed to. It is obviously much more.

We appear to be on the cusp of significant change from AI models. I hope that looking at the 20th century will help readers make better guesses about what may be in store.

I believe in the perspective that technological change helps to shape cultural change. If I did not devote much of the book to feminist ideology, that is because I think that some of women’s liberation was propelled by changes in the occupational structure and the automation of much housework.

I did devote a lot pages to the topic of race. I think that many Americans naively believed that with the civil rights legislation of 1964 and 1965 racial tensions would fade away and be long gone by now. I wanted to convey why they had that hope and why it was not to be.

I also devoted a lot of pages to Communism. I think that between the time of the Russian Revolution to the fall of the Berlin Wall, Communism loomed very large in American minds. Every public intellectual had to position himself or herself: are you a Communist? An anti-Communist? An anti-anti-Communist?

One of the primary projects of mainstream economists in the middle of the 20th century was promoting the ideas of marginal product and competitive equilibrium. They did this in order to refute Marx’s labor theory of value and his claim that the capitalist system exploits the worker.

Political polarization is another issue on which a study of the 20th century may shed light. How did America recover from the division of the 1930s or the 1960s?

I tried to convey that mass media provided a centripetal force. As radio, television, and movie studios sought the largest possible audience, they tended to amplify our commonalities rather than our differences.

Consider two axes along which a large nation can diverge. One axis is the extent of awareness they have of one another. The other axis is the extent to which people share the same understanding of the world around them.

In this framework, we began the 20th century living in different worlds and not aware of one another. We were separate along both axes. In the middle of the 20th century, we lived in the same world. We were somewhat close along both axes. More recently, we have moved into separate worlds while remaining aware of one another, resulting in frustration and anger.

Before 1930, the Kansas farmer and the Boston lawyer had little and common. But this was not a problem, because they had little reason to think about one another.

Between 1930 and 1975, mass media brought Americans into the same world. Americans could get along, because the media were giving everyone the same picture.

After 1975, and especially in the 21st century, Americans moved into different worlds. But we are aware of those who differ from us. Our online friends are constantly denouncing our online adversaries, so that we cannot forget about the people who live in other worlds the way that a Kansas farmer could forget about a Boston lawyer.

In the 20th century, we also had some better leadership. I think that Franklin Roosevelt probably helped in the 1930s, not by imposing a grand vision but by avoiding one. The NRA was a radical initiative, but it got thrown out by the Supreme Court. Court-packing was another radical initiative, but Congress and the public did not go for it. Otherwise, as he flailed around, throwing agency after agency at the problems of the Great Depression, he maintained his personal popularity, and with it the legitimacy of existing institutions.

As for the 1960s, I think that Nixon, like Roosevelt, used political shrewdness and lack of strong philosophical commitment to his advantage. By shifting his stances, he neutralized some of the opposition. Between that and his ability to mobilize the “silent majority,” the radicals were marginalized. This left them with a choice: get serious about revolution, or else work within the system. In the long run, left-wing success came from penetrating the institutions of academia, media, and government. I would say that the message for the left today is: those who can, penetrate; those who can’t, demonstrate.

Not that I want to see more institutional capture by the left. The far left today takes many stands with which I disagree. The one that troubles me most is opposition to the state of Israel, which I believe coincides with America-hatred. In the 1960s, if you had probed an ideologically sophisticated anti-war radical, you would have found a similarly negative position on the legitimacy of the Jewish state. But for the left in those days, it was not a cause, much less the cause.

I am not the only one who sees Israel-hatred as closely aligned with America-hatred. I want to talk the reader out of America-hatred without telling our twentieth century story as a rah-rah tale. It will be sufficient if you appreciate the two iron laws:

1. Sometimes it’s this way, and sometimes it’s that way

2. Outcomes are often unintended.

   Share

   Subscribe now

Robert Hall was on the faculty at MIT when I was in grad school there. My sense is that he loved the MIT econ department more than the department loved him back. I’ve always thought that he deserved higher status, and I’m happy for him.

One MIT tradition was a graduate student skit party, to which professors were invited to watch grad students mock them. The one skit I remember was to make fun of Hall and his “life-cycle permanent income hypothesis” paper. We wrote new lyrics to a novelty pop tune called “Monster Mash.” Instead of “He did the mash, he did the monster mash” we sang “He did the Walk. He did the RANDOM Walk. It was a stochastic stalk. You should have seen them gawk.” All the while we walked randomly around the stage, bumping into one another. It was great.

My favorite paper of Hall’s is what I call his “happy families are all alike” paper, about the way that every recovery proceeds at about the same pace. I like the “patterns and stories” approach to doing macro. The late Ed Leamer took that approach, and I wish he had won a Nobel.

As for the “life-cycle permanent income hypothesis” paper that most people know, I’m not so crazy about it. I have two rules for why macroeconometrics doesn’t work.

Rule 1: at short intervals, the change in any macro variable is mostly noise.

Rule 2: at long intervals, the change in any macro variable is due to structural discontinuity.

Hall’s paper found that at short intervals changes in aggregate consumer spending are mostly noise. He wraps a theoretical interpretation around that finding. But I think it’s just Rule 1.

The Nobel Committee tends to overlook economists like Hall, who make a variety of contributions. Usually, it would rather reward a one-hit wonder than a Sam Cooke or Johnny Rivers or The Association, artists with a solid Greatest Hits album.

Share

Subscribe now

Horowitz Andreessen Academy will be a two-year residential program in San Francisco for 16 to 22 year olds.

…Students will learn from real-world practitioners and intellectuals; including founders (Sam Altman, Brian Armstrong, Jensen Huang), investors (Elad Gil, Daniel Gross), and professors (Agnes Callard, Tyler Cowen), among hundreds of others. The Academy is the educational experience I wish I had.

…The goal of the Academy is to identify the right mentors, the correct bottlenecks, and the right introductions, adjusted continuously around one person and their pursuits.

Long-time readers may recall my network university idea. The main difference between that and HAA is that HAA has a real network of the kind that I could never have assembled. The second big difference is that they envision a small cohort of students living in San Francisco. My idea was a large cohort of students (1000, say) living anywhere, but assembling in large groups for 3- or 4-day retreats several times a year.

Peter Diamandis writes,

[The Knowledge Society] is a ten-month program for 13-to-17-year-olds, layered on top of whatever school they already attend. Three-hour weekend sessions with the same cohort, and that’s only about a quarter of the experience; the rest is reading research papers, meeting experts, one-on-ones with directors, and building. About 70% of the curriculum is emerging technology (AI, robotics, synthetic biology, quantum, sensors/networks, AR/VR, etc). The other 30% is life skills, with sessions titled “Unconventional Thinking” and “Life and Success,” where teenagers are asked to define what success means to them and work backward.

It sounds like prep school for Horowitz-Andreessen.

Steven Johnson writes,

we’ve made more than 100,000 books from the Google Play bookstore available to use as sources inside of Notebook. You buy the book, and you can add it to any notebook in exactly the way you add normal sources like Docs or PDFs. (You can also read it with all the standard e-book reader features in the Play Books app, though we hope to bring more of those reading experiences to Notebook over time.) Inside of Notebook you can ask questions based on the book content, and get grounded answers with citations back to the original passages; you can combine different books to create a diverse brain trust inside a single notebook

I think of this as enabling what I call “vibe-reading” (having a conversation with an AI that lets you know about the ideas in a book) but making you buy the book.

An odd combination in a podcast: Jerry Muller, intellectual historian and author of The Tyranny of Metrics, and Razib Khan, expert on population genetics and polymath.

Around 30 minutes in, they get into the difficult challenge of dealing with ideological bias in the academy and the institutions of faculty control, including making hiring decisions.

Share

Subscribe now

substacks referenced above: @

Razib Khan's Unsupervised Learning

Jerry Z. Muller: academia today, intellectual history in the past and AI in the future

Today Razib talks to Jerry Z. Muller, a historian and professor emeritus of history at the Catholic University of America in Washington, D.C., specializing in modern European intellectual history, capitalism, conservatism, and German and Jewish history. His work explores how ideas shape economic, political, and cultural life, bridging history, philosophy, and public policy. His books include…

Listen now

20 days ago · 2 likes · Razib Khan

@

Adjacent Possible

Introducing Expert Intelligence

Earlier this year, I shared a public notebook that I’d created called The Architecture Of Ideas, which assembled a collection of Substack and Medium articles I’d written over the past decade or so on the themes of creativity and innovation. The goal was to gather together all the knowledge that I had released publicly on those topics so that you, dear r…

Read more

23 days ago · 33 likes · 11 comments · Steven Johnson

@

@

Erik Torenberg

Introducing Horowitz Andreessen Academy

Thrilled to announce our latest incubation: Horowitz Andreessen Academy (see our hype reel here…

Read more

22 days ago · 13 likes · 5 comments · Erik Torenberg

Kai Williams writes,

A plethora of startups like Shift are trying to solve the data shortage by recording the actions of humans and converting the videos into training data for robots. Other companies are hiring humans to directly operate robots in labs, factories, and even people’s homes. Still others are hiring humans to perform everyday tasks while wearing gloves or exoskeletons that force the human to move in a robot-like way and capture rich data on the worker’s actions. Some large data-collection companies like Scale AI are experimenting with all of these strategies.

Gad Levanon writes,

Last week’s jobs report was solid: 162,000 new jobs. The industries I watch most closely lost 24,000. That’s the pattern now. Since April 2023 the rest of the economy has added 4.4 million jobs, while finance, insurance, information, and professional and business services, the group I call FIIPB, have shed 700,000, falling in 29 of the past 40 months. And there was no recession. Everyone wants to know whether AI did it.

Ben Moll and Alex Imas write,

Most work is physical. An AI capability explosion is first and foremost an explosion in cognitive capabilities. But only around a third of the economy consists of work that can be done on a computer

I think it may be less than that. A lot of work is also social. Tasks combine manipulation of symbols with motivating and understanding humans. Education and health care are one fourth of the economy, and teachers and doctor have to do more than just manipulate symbols.

Jason Potts writes,

Books, radios and computers all lowered the cost of producing and distributing educational content, with the university mostly unchanged. This was because content was never what universities sold. They sell matching and verification: bringing students, scholars, employers and funders together, and warranting to third parties the knowledge claimed. AI is the first technology to reach both functions and so the first with the capacity to disrupt the university's business model.

…the university not as a cloister or a factory but as a multi-sided market (a platform) coordinating at least fifteen distinct groups, sustained by cross-subsidies between them, whose strategic asset is a credible warrant of quality over a good whose quality nobody can directly observe.

…We set out ten strategic proposals, grouped by what the university must own, what it should build, and what it should reshape. The central proposals are to separate certification from instruction institutionally; to relocate the costly signal into modalities AI cannot enter; to occupy the training gap left by AI adopting firms; to build fast accreditation capacity; to treat the cohort as a designed product; to host scholars rather than employ them; and to specialize in the exploration that private incentives now undersupply. Because a multi-sided platform cannot change one side without renegotiating all of them, the practical route is greenfield. The university has survived a thousand years mostly by founding new colleges, not by rebuilding old ones.

Zoe Hitzig and others write,

In a typical session, people make most of the planning decisions (what to do) and Claude makes most of the execution decisions (how to do it). The greater domain expertise a person brings to a session, the more work Claude does per instruction

I still think that Claude code requires a lot of computer savvy. Domain expertise alone won’t do it.

Share

Subscribe now

pointer from Tyler Cowen

substacks referenced above: @

Ghosts of Electricity

Will AI soon lead to double-digit growth?

The following post represents the personal views of the author(s) and does not reflect or represent the positions of their employers…

Read more

a month ago · 97 likes · 12 comments · Ben Moll and Alex Imas

@

Labor Matters

The Jobs That Never Arrived

Last week’s jobs report was solid: 162,000 new jobs. The industries I watch most closely lost 24,000. That’s the pattern now. Since April 2023 the rest of the economy has added 4.4 million jobs, while finance, insurance, information, and professional and business services, the group I call FIIPB, have shed 700,000, falling in 29 of the past 40 months. A…

Read more

a month ago · 13 likes · Labor Matters

@

Understanding AI

Robot startups are trying everything they can think of to get more data

On May 28, the startup Shift announced that it would clean any New York City apartment for free. In a launch video, cheerful young men scrubbed toilets, vacuumed floors, and wiped down counters…

Read more

a month ago · 44 likes · Kai Williams

Over evolutionary time, organisms did not directly perceive probabilities. They either saw something happen or they didn’t. Yesterday there was a wild turkey near the pond, but there were no wild turkeys along the tree line. The day before there was a leopard perched in the baobab tree, but there were no leopards near the watering hole.

People perceived a number of things that occurred over the course of some number of observations. I saw a wild turkey near the pond three times over the last ten days.

Relatedly, humans over evolutionary time came across whole objects. We have very fancy perceptual systems that let us parse the world into individual things like people, animals, and rocks that we can count and talk about. Five turkeys. Three leopards.

When I was teaching AP statistics in high school, I noticed that students did much better handling conditional probability problems if they used whole numbers rather than percentages. Doctors are notoriously bad at Bayes’ Theorem. But according to Kurzban, they do better if you give them whole numbers.

Manvir Singh writes,

Under certain conditions, people iteratively craft stories about rivals that justify violence or spur others to attack them (Cohen, Reference Cohen1972; Cohn, Reference Cohn1976; Goode & Ben-Yehuda, Reference Goode and Ben-Yehuda1994); in cultural ecologies in which witchcraft and sorcery narratives also serve as plausible explanations of misfortune, those can be leveraged to demonize targets, who are accused not only of using dark powers for harm but of posing existential threats and engaging in morally abhorrent acts (Hutton, Reference Hutton2017). A cultural tradition – the belief in heinous witches – is thus plausibly shaped by several interacting cultural selective schemes.

Pointer from Ian Leslie. We have goals, and we form beliefs about practices that accomplish those goals. When those beliefs are well tested and verified, we call them science. When another group’s beliefs seem to us to be not well justified in practice, we call them superstitions. As long as a group of people subjectively believe that practice X serves goal Y, or as long as they act as if they subjectively believe it, practice X will survive in the culture.

Brian Chau writes,

Comfort with a growing civilization, far from being a universal value, is one tied to the cultural Christianity of the enlightenment. It isn’t the human norm. The human norm evolved from stagnant hunter-gatherer societies with little potential for growth. Man was at the mercy of nature. Man could do nothing to fight back against the tribe’s apocalypse, whether by flood or quake or eruption. That is why the tribal norm is an authoritarian paranoia, wrought with taboo and ritual sacrifice.

The past is now the future. The past generation led a descent into cultural atheism.

Now we live with

the never-ending conga line of apocalyptic fantasies motivated by the primordial fear of growth.

He includes AI doomerism in that. I plead guilty to human doomerism. That is, I think that humans are capable of such evil that what is needed for survival is a form of surveillance, and the challenge will be to disperse power in such a way that those with surveillance capabilities are themselves watched, as in David Brin’s Transparent Society.

Mary Harrington writes,

almost every one of the enabling conditions that gave rise to modern democracy has disappeared. With the transition from print-first to broadcast, and now digital, we have begun to surrender the mass capacity to think long-form, to deliberate, to form the intermediary social structures that habituate people to self-government, and the model of selfhood that presupposes intellectual independence and a capacity for judgement. In its place we have elevated a supposedly expert class that’s in some ways neutral and detached, but in others demonstrably prone to groupthink and moral panics.

Share

Subscribe now

substacks referenced above: @

Mary Harrington

Our Democracy

This is the 10th and final instalment in a serial preview of The King and the Swarm, my book in progress. I’m sharing previews for paying subscribers, as I drag the manuscript over the line…

Read more

21 days ago · 28 likes · 2 comments · Mary Harrington

@

From the New World

Post-Christian America

What did you think Post-Christianity meant? Vibes? Essays? Christianity has declined precipitously — first went the practicing faith, then soon after went the cultural Christian grace of a single generation of nonbelievers…

Read more

22 days ago · 19 likes · 8 comments · Brian Chau

@

The Ruffian

Why Does Every Society Invent the Same Things?

Catch-up service…

Read more

25 days ago · 21 likes · 3 comments · Ian Leslie

@

Living Fossils

Frequencies & Whole Objects

Let’s face it, most research in psychological science has little to no bearing on the real world…

Read more

a month ago · 3 likes · Rob Kurzban

The labs intend to automate the AI development process with AI, which will mean faster and faster progress and necessarily less and less human oversight.

If all goes according to schedule, humans will play a rapidly shrinking role in AI development over the next two years, as both OpenAI and Anthropic try to put as much R&D as possible into the hands of their increasingly capable (and increasingly incomprehensible) AI.

One fear is that this will work too well, and AI will be out of control. I am inclined to the opposite fear, which is that it will make AI better at dealing with the digital realm, and then it will turn out that the physical and social realms are both more difficult for AI’s and more important for society.

Moses Sternstein writes,

The people/firms using agents are a tiny share of the overall userbase, they are using orders of magnitude more tokens than everyone else, and both they (and agents) are just getting started.

I get the sense that the improvements in the models for the past year are mostly for the most high-end users. As Piper points out, the highest-end users are the model engineers themselves. For a low-end user like me, the models do not feel much different than they did a year ago. Since December, I have felt left behind.

Peter Cornillon and Xavier Prochaska

It is tempting to imagine the Agentic Professor as a more capable chatbot or tutor, but such analogies miss the central shift. Most AI tools currently deployed in education are reactive, responding to prompts, questions, or submissions. Agentic systems, meanwhile, can initiate inquiry, propose learning pathways, identify misconceptions, and dynamically adapt pedagogical strategies. This is not an incremental improvement; it is a categorical change in how instruction unfolds.

Not this, but that, eh?

Ethan Mollick writes,

while there is a lot of debate over what future models will do, the current capabilities of existing models are barely being used, and are often not even well understood. For example, I did not know GPT-6 Astra could operate Blender (a sophisticated piece of 3D modelling software) until it did.

I scheduled this post a few weeks ago. Meanwhile, all sorts of developments could have taken place in the AI world. I’ve finished with the American history posts (although I might do an epilogue), but I have several Economy-and-AI posts stacked up to run.

Share

Subscribe now

substacks referenced above: @

One Useful Thing

The Overhang

We are still on an exponential curve of AI development. I try to put out a Substack post every couple weeks or so, yet, as the pace speeds up, that sometimes feels too slow. In the weeks since my last post, we had the apparent cracking of one of the most famous problems in math by an AI (accompanied by controversy) and widespread discussions about the r…

Read more

a month ago · 215 likes · 1 comment · Ethan Mollick

@

Random Walk

6 Charts For The AI Bulls

climbing the wall of tokens aka we’re gonna need a lot more compute…

Read more

a month ago · 13 likes · 1 comment · Moses Sternstein

@

The Argument

Losing control of AI is actually the plan

Today’s AIs can do really hard things: conduct elaborate computer hacks, edit their own logs to conceal their activities, steal user credentials and private informat…

Read more

a month ago · 56 likes · 19 comments · Kelsey Piper

How do people learn? Pattern recognition, observation, reason and logic, trial and error, scientific experimentation, … . I don’t understand why AI platforms can’t continue to advance — at what pace I have no idea — using the same heuristics.

Humans have more direct experience of the physical world. Humans have more direct experience of the social world. AI’s experience of the physical and the social is almost entirely intermediated by digital databases, which contain very little of what humans know through experience.

Humans also learn from one another. Just as we have a decentralized economic process that employs specialization and trade, we have a decentralized process for generating, filtering, and sharing knowledge.

Other commenters mentioned Waymo. I’ve ridden in Waymos in Austin, and they work well. But you cannot take a Waymo to go the 80 miles from Austin to San Antonio.

Gemini tells me that Waymo currently covers 0.05 percent of the land area in the continental United States. Anyone care to predict how long it will take to get to even 1 percent? 5 percent?

The point that Yuval is making, and that I also make, is that the current methods by which AI’s learn work much better in some domains than in others. We are warning you not to assume that the speed at which AI’s have learned coding is a reliable indicator of how quickly they will adapt to the physical world and to the social world.

Share

Subscribe now

We are not AI engineers. No one is bidding millions of dollars to hire Yuval or me to work in their AI lab. Our pronouncements are not going to move the stock market or make the front page.

But we think that the AI engineering community has some blind spots concerning what general intelligence really means.

Yuval Levin writes that he is

perplexed by several assumptions in the field that seem, at the very least, not sufficiently articulated and conceptualized, if not in fact ultimately unsupportable. These assumptions pertain in particular to the character of the progress of frontier large language models and the agentic systems built on them, and to this technology’s relation to the various domains of the human experience.

The metrics that are used to measure progress on the models show acceleration. But

Implicit in much of the thinking … is the presumption that, as AI grows more capable, its scope of competence will grow more general.

We believe that the models are impressive already. And users are only beginning to tap their potential when integrated them into scientific research and economic activity.

Much of the talk in our politics about the need to prepare for immense disruptions in the labor market, education, science, culture, politics, and civic life is warranted. But almost all of that is a function of an innovation that has already occurred but has yet to truly wash over vast swaths of our experience. [my emphasis]

Think about my point that there is a physical world and a social world, and that these differ from the digital world. As the AI’s get smarter in the digital world, does this automatically mean that they will master the physical world and the social world? I do not see why this necessarily follows. The physical world and the social world have not been mapped digitally. What is legible to the models includes language, computer code, and scientific knowledge that has been written down.

What is not legible to them includes properties of objects that we intuitively grasp but have not articulated. If I throw something at you, whether you try to catch it, duck out of the way, swat it, or kick it is a decision that you make automatically. You know what to do with a set of keys, a knife, a disgusting sticky substance, or a soccer ball. But that knowledge is tacit, based on experience, not written down so that it can be found in a database somewhere. I call this form of knowledge “unknown knowns,” meaning things that we know but have not articulated.

Levin writes,

Just how does artificial intelligence advance? That question — the question of the nature of improvement in contemporary AI — does raise some serious challenges to the case for AGI.

Are the models improving along a path that leads to AGI, or are they only improving in narrower ways?

The large language models we have come to know began with an extensive process of absorbing a vast mass of learning data.

…They grew able to articulate some judgments they cannot fully explain and that no one had specified. These were inherited implicitly in the absorptive process from the patterns of human expression. This is part of what people mean when they say we can’t fully explain how the models work.

The models, too, are demonstrating “unknown knowns.”

Levin puts the essence of his claim this way:

In terms of the development of the technology itself, the absorptive stage is now largely over. The absorption of learning data is generally not how artificial intelligence will progress further, because there simply isn’t all that much data that remains unabsorbed. Curation of existing data is still making a difference, but that can only continue for so long. And for all its creative potential, there are real limits on the ability of AI to create synthetic new learning data for itself. It does happen, but it requires arduous selection and management, and is meeting diminishing returns.

He says that most of the gains going forward will have to come from what he calls recursive experimentation.

reinforcement learning involves the model producing new information in response to a query or instruction and then having that information scored in relation to some external standard.

…as new data grows increasingly scarce, reinforcement through recursive experimentation has come to the fore as the mode of learning that will guide this technology further.

The issue is that recursive experimentation works better in some domains than others, because

some domains are much more readily verifiable than others.

Above all, computer coding is uniquely verifiable.

Those of us who are not in the coding community may not appreciate how rapid and dramatic the progress is there. But by the same token,

it can be difficult for insiders to grasp that their own professional environment is changing much more quickly and dramatically than those of other people.

…The models will not become better and better lawyers, teachers, doctors, and gardeners at anything like the pace at which they are becoming better coders. … Superhuman programming ability does not by itself amount to, or even point toward, anything that might reasonably be called general intelligence. [my emphasis]

He warns,

there are other domains where there just aren’t ready objective standards, or where quick experimentation is not easily available — not only art or management, but also public policy and geopolitics, and even many subfields of biology.

In my terms, there is the physical world and the social world. What our forecast boils down to is that while the models will continue to get better in the digital world, progress in the physical world and in the social world will not accelerate dramatically if the models are limited to the existing paradigm.

Share

Subscribe now

In Every Home

The Eighties and Nineties — AIDS and the Blue Dress, 1981–1999

Opening: The Two Envelopes

Twice in the space of ten years the government of the United States delivered explicit sexual material into the home of nearly every American, and the two deliveries, laid side by side, hold the whole of the decade between them.

The first arrived in the late spring of 1988. A pamphlet titled Understanding AIDS was mailed to a hundred and seven million American households, every address the Postal Service could reach, the largest public-health mailing ever attempted by any government anywhere. It was signed by the Surgeon General of the United States, and it spoke to the American family in language no government document had ever used at the breakfast table, explaining in plain clinical words how a virus passed from one body to another during sex, which acts carried the greatest danger, and why a latex condom might save a reader’s life. Millions of parents opened it at the kitchen counter with their children in the next room and found the federal government telling them, without euphemism, things they had spent their lives declining to say aloud.

The second arrived in September of 1998. The office of an independent prosecutor named Kenneth Starr delivered to Congress a report on the conduct of the President of the United States, and Congress, in a decision without precedent, released it to the public by posting it on the new medium that had lately entered American homes, the internet. It ran to some four hundred and forty-five pages, and a great deal of it described, in grand-jury detail that spared the reader nothing, the physical particulars of encounters between a fifty-year-old president and a woman in her twenties who had come to the White House as an intern. Millions of citizens read it on their office computers the day it appeared, passing the more astonishing passages to one another, and a document commissioned to weigh the possible impeachment of a president became, for a season, the most widely read piece of sexual description ever produced by the American state.

A devoutly Christian Surgeon General had put explicit sex in every mailbox to keep people from dying. A devoutly Christian prosecutor had put explicit sex on every screen to drive a president from office. Hold the two envelopes together, the pamphlet and the report, because the distance between them is the subject at hand. For as long as anyone could remember there had stood a wall between the sexual life a person led in private and the life the public was permitted to see, and across these two decades that wall came down. It came down first because a plague made silence lethal, and it turned out that a country which learns to speak of such things in order to save its own life does not easily forget how. The frankness that began as survival ended as spectacle, and the same demolition that let a Surgeon General save lives left a president with nowhere to hide.

[ PAUSE FOR QUESTIONS ]

The Unspeakable Word

The disease announced itself quietly, in a government bulletin, on the fifth of June, 1981. The Centers for Disease Control reported that five young men in Los Angeles, all previously healthy, all homosexual, had come down with a rare pneumonia that normally struck only patients whose immune systems had collapsed. Within months the same wreckage was appearing across the country — the pneumonia, a strange purplish skin cancer, a catalogue of infections that healthy bodies threw off without effort — and it was appearing overwhelmingly among gay men. The early name the doctors reached for said everything about how the thing was understood: gay-related immune deficiency, as though the affliction were a property of the men rather than of a virus none of them yet knew existed. The newspapers called it the gay plague.

The country’s first response was to look away, and the looking away began at the top and worked down. In the White House press room, when a reporter asked the President’s spokesman about the growing epidemic, the room laughed; the spokesman made a joke about whether the questioner himself might be afflicted, and the questions moved on. Ronald Reagan, whose warmth and whose instincts governed so much of the era, said almost nothing about the disease in public for years, through thousands and then tens of thousands of deaths, until 1987, by which time more than twenty thousand Americans were dead of it. The silence was not merely presidential. It ran through a country that found the whole subject unspeakable in the most literal sense, because to speak plainly about how the virus traveled required speaking plainly about acts that polite America had never discussed in mixed company, in a newspaper, or on the evening news.

Into that silence came a reading of the catastrophe that a large part of the country was prepared to believe, and it is essential to state it fairly, because it was sincere and it was widely held and it was not the property of cranks. To millions of religious Americans, the epidemic looked like confirmation of an order they already understood to be true. Here was a disease that struck those who violated the oldest sexual prohibitions, and struck them precisely through the violation, so that the punishment seemed written into the sin. The commentator Patrick Buchanan gave the reading its sharpest form when he wrote that homosexuals had declared war upon nature, and nature was now exacting its retribution. Preachers said the same from their pulpits in gentler or harder words. The sodomy laws still on the books in half the states said it in the language of the law. The plague had arrived, in short, at the perfect moment to become the proof of a proposition many Americans devoutly wished to see proven: that homosexual love was an abomination, and that God or nature would not be mocked. The judgment hung there, in the early years, waiting to fall and settle and become the country’s official mind.

[ PAUSE FOR QUESTIONS ]

The Count

It did not settle, and the first reason it did not was that the men who were dying refused to die quietly. The refusal had a voice, and the voice belonged to Larry Kramer, a playwright and novelist of ferocious temper who had helped found the Gay Men’s Health Crisis in a friend’s living room in 1982 and who spent the decade screaming, in essence, at everyone. He screamed at the government for its silence and at the medical establishment for its slowness, and he screamed hardest of all at his own community, which he believed was sleepwalking toward extinction. In March of 1983 he published in a gay newspaper an essay whose title was a body count — “1,112 and Counting” — and whose argument was that the readers were, quite simply, going to die, in numbers that would empty their world, unless they woke up and fought. He was insufferable. Men who worked beside him could barely stand him. He was also, in nearly every particular, correct, and the community that could not bear him slowly came around to the thing he had seen first.

By 1987 the fury had a form. Kramer helped found the AIDS Coalition to Unleash Power, ACT UP, and it brought to the streets a movement that had learned, in the hardest possible school, how to organize. Its emblem was a pink triangle above three words, Silence = Death, which fused the memory of the Nazi persecution of homosexuals with an accusation aimed at every American who preferred not to look. Its members lay down in the streets of the financial district and chained themselves to the gates of the Food and Drug Administration and interrupted the trading of stocks, and they did something subtler and more lasting as well: they learned the science, read the journals, mastered the vocabulary of the clinical trial, and made themselves expert enough that the researchers and regulators eventually had to deal with them as colleagues rather than as protesters. A movement that had spent the 1970s organized around the freedom to love as it chose was hammered by catastrophe into a disciplined political machine, and the machine would outlast the emergency that forged it.

While the movement made itself heard, the virus itself was quietly demolishing the moral geometry that had made the judgment seem so obvious. It turned out that the thing did not confine itself to the guilty. It moved through the nation’s blood supply and killed hemophiliacs who had done nothing but accept a transfusion. It killed a boy named Ryan White, a thirteen-year-old Indiana hemophiliac who was barred from his school by frightened parents and who bore the cruelty with a grace that shamed a watching country. It infected newborns in their mothers’ wombs. The very language the sympathetic press reached for — the “innocent victims,” the children and the transfusion patients — revealed the moral map everyone was working from, in which other victims were presumably guilty. Yet by naming innocents at all, the country had begun to break the clean equation of sin and punishment, because a disease that killed the innocent alongside the damned was harder to read as a tailored act of divine justice. And then, in the summer of 1985, the plague acquired the one face that made it impossible for mainstream America to keep looking away. Rock Hudson, the matinee idol of a hundred wholesome pictures, a friend of the Reagans, a closeted man whom the whole country had watched romance leading ladies for thirty years, announced that he was dying of AIDS, and died that autumn. The disease had a name the country loved, and it was no longer somebody else’s.

Two years later the dead themselves came to Washington. In October of 1987 the NAMES Project unrolled its memorial quilt across the National Mall, panels sewn by the bereaved, each the size of a grave, each bearing a name and the small particulars of a life — a nickname, a pair of dates too close together, a scrap of a favorite shirt. Laid out beneath the monuments of the capital, it covered a space larger than a football field and was already too big to take in at a glance, and it forced upon every visitor the one fact the silence had been built to avoid: that these were people, loved and mourned, and that there were a great many of them.

[ PAUSE FOR QUESTIONS ]

The Surgeon General’s Mail

The man who turned the country’s response away from the judgment and toward the clinic was the least likely figure imaginable for the task. C. Everett Koop had been made Surgeon General because he was a hero of the religious right — a pediatric surgeon of real distinction, yes, but chosen for the post as an outspoken evangelical and a fierce opponent of abortion, a man his allies expected to carry their banner into the government’s science. He looked the part, with his patriarch’s beard and his gold-braided uniform, and the gay activists dreaded him. And then Koop looked at the epidemic through the only lens he finally trusted, which was the lens of a doctor with dying patients, and he concluded that his duty to keep Americans alive overrode every other loyalty he held.

In 1986 he issued a Surgeon General’s report on AIDS that called for frank sex education beginning early, and for the honest public discussion of condoms, and that declined to moralize about the people the disease was killing. Two years later came the pamphlet in every mailbox. His own political allies felt betrayed, and said so; the guardians of public decency accused him of teaching sodomy to schoolchildren and of surrendering the moral high ground to save the bodies of men who had forfeited the right to it. Koop did not yield. He held, in effect, that he was the physician of the whole American people, the sinners with the saints, and that a Surgeon General who let citizens die in order to avoid saying uncomfortable words had failed at the one thing his office existed to do. In choosing the pamphlet over the sermon, a devout conservative had quietly decided the largest question the plague had raised, and had decided it against the men who had put him where he stood.

The law, for its part, spoke the older verdict one last time. In 1986, in Bowers v. Hardwick, the Supreme Court upheld the power of a state to make private homosexual acts a crime, and the majority treated the very claim of a right to such acts as close to facetious. It was the high-water mark of the judgment the plague had seemed to invite, the moral reading written into constitutional law. And it was the last such mark. Within a generation the Court would reverse itself and erase Bowers from the law, and the direction the culture had chosen in the crisis — to treat the afflicted as patients and citizens rather than as the justly condemned — would prove to be the direction it kept.

What sealed that choice was that the clinical answer worked. The virus was identified in 1983 and 1984, and the first drug against it, AZT, arrived in 1987, toxic and staggeringly expensive and able to buy a patient only months, its price the occasion for some of ACT UP’s angriest campaigns. For nearly a decade the science crept while the dying continued, and American deaths from AIDS climbed year over year until they peaked, in 1995, at some fifty thousand souls in a single year. And then, with a suddenness that stunned the physicians themselves, the corner was turned. At an international conference in Vancouver in the summer of 1996, researchers presented the results of combining several new drugs, the protease inhibitors, into a single regimen that beat the virus down below the level the instruments could find. Patients who had been arranging their own funerals rose from their beds and went back to work, and the doctors, reaching for the only word large enough, called it the Lazarus effect. Within two years American deaths from the disease fell by more than half. AIDS in the wealthy world was not cured, but it was converted from a sentence of death into a condition a person could live with, managed by a fistful of pills a day.

Set the whole arc down and the shape of it is plain. A catastrophe arrived that a large part of the nation stood ready to read as the judgment of God upon a hated thing, and the reading was available, and sincere, and for a while it seemed it might become the settled mind of the country. It did not become the settled mind of the country. The dying made themselves seen, the virus refused to respect the boundary between the guilty and the innocent, a conservative Surgeon General chose medicine over morality, and the laboratories, given time, produced the drugs that turned the plague back. The judgment never got the ratification the epidemic had seemed to offer it, and the movement that the catastrophe had forged emerged from the fire organized, visible, and pointed toward the ordinary respectability — the wish to marry, to serve, to be counted as unremarkable — that would define its next chapter.

[ PAUSE FOR QUESTIONS ]

The Meaning of Is

The wall between private conduct and public life, which the plague had breached in the name of survival, came down the rest of the way in a matter that had nothing to do with survival and everything to do with a president who could not govern his appetites or, once caught, his tongue.

The road to it began with a lawsuit. In 1994 a former Arkansas state employee named Paula Jones sued Bill Clinton, alleging that he had crudely propositioned her in a Little Rock hotel room in 1991, when he was governor and she a clerk at a state event. The suit raised a novel question — whether a sitting president could be made to answer a civil claim about his private conduct while in office — and in 1997 the Supreme Court answered it unanimously in the affirmative, ruling that the presidency bought a man no delay from a lawsuit of this kind. That ruling opened a door. To build their case that Clinton had a pattern of pursuing subordinate women, the Jones lawyers were entitled to ask him, under oath, about other women, and by the time they deposed him in January of 1998 they had a name to ask about, because a separate and unrelated investigation had stumbled onto it.

The name was Monica Lewinsky. She had come to the White House as an intern in 1995, twenty-two years old, and had begun with the president an intermittent sexual relationship that ran into 1997, a relationship she had described in detail to a older coworker who secretly recorded her and delivered the tapes to Kenneth Starr, the prosecutor already three years deep in an inquiry into the Clintons’ tangled Arkansas finances. Under oath in the Jones deposition, asked whether he had engaged in sexual relations with Lewinsky, Clinton said he had not, leaning on a cramped definition of the term that he believed left his particular conduct technically outside it. Days later, to the country, he was more emphatic. He looked into the cameras, pointed a finger for emphasis, and declared that he had not had sexual relations with that woman, Miss Lewinsky. It was a lie, and he held to it, in public and to his own staff and cabinet, for seven months.

What broke it was a piece of physical evidence of a kind no gentleman’s-agreement press and no discreet earlier age had ever had to reckon with. Lewinsky had kept a navy blue dress, unlaundered, that bore a stain, and the stain yielded the president’s DNA, and against a genetic match there was no cramped definition to hide behind. In August of 1998 Clinton testified to Starr’s grand jury and then addressed the nation to admit an “inappropriate relationship,” and in the grand-jury room he produced the sentence that became the epitaph of the whole affair, quibbling over the tense of a lawyer’s question with the observation that it depended on what the meaning of the word “is” is. A president of formidable gifts had been reduced to parsing the smallest word in the language in order to avoid conceding what the whole country now knew. The evasions were his own, the lies were his own, and the conduct that occasioned them was his own; whatever one thought of the prosecutor who had pursued him across four years and out of an Arkansas land deal and into a young woman’s testimony, the wound had been self-inflicted, and no account that makes Clinton chiefly a victim survives contact with what he actually did.

Then came the report in every home, and the impeachment. The House of Representatives impeached him in December of 1998 for perjury and obstruction of justice, only the second president in the history of the republic to be so charged. And here the country delivered a verdict that confounded the men who had pursued him. The public, which disapproved of what Clinton had done and said so to every pollster, disapproved far more of the effort to destroy him over it. His job-approval ratings rose as the scandal deepened, reaching into the high sixties at the peak of his disgrace. In the midterm elections held in the middle of it all, the president’s party gained seats in the House, a thing that had not happened to a sixth-year president since before the Second World War, and the Speaker who had bet on impeachment resigned within days. When the trial reached the Senate in February of 1999, the votes to convict fell far short of the two-thirds the Constitution demands, and Clinton served out his term. The country had decided, in its rough way, that a man’s squalid private failings and his lies to cover them, however genuine, did not rise to the destruction of a presidency the people had twice elected.

[ PAUSE FOR QUESTIONS ]

What It Amounts To

Set the two stories in a line and the connection between them is the wall that fell, and the environment that pushed it over.

The forces were real and none of them was chosen. The sexual revolution of the preceding decades had already loosened what could be said; the plague made saying it a matter of life and death and dragged the most private acts into the clinic, the pamphlet, the newspaper, and the Sunday sermon, until a country that had never named such things in public could speak of little else. New machines abolished the old gatekeepers who had decided what the public was allowed to know: cable television with its unslakable appetite for scandal, the tabloid press gone respectable, and at the end the internet, which could put a four-hundred-page catalogue of a president’s sins in front of every citizen in an afternoon and answer to no editor at all. And the machinery of investigation that Watergate had bequeathed — the independent prosecutor, armed with an open-ended mandate and a limitless budget — stood ready to convert any private failing into a federal case. Against those forces the old wall, the tacit agreement that a public man’s bedroom was his own, simply could not stand.

The measure of how far it fell is the distance to John Kennedy. Two political generations earlier a president had conducted his private life with a recklessness that dwarfed anything charged against Clinton, and the reporters who knew — and many knew — wrote not a word, because the code of the age held that such things lay outside the public’s business and beneath the dignity of the press to report. That code was extinct by 1998. Conduct that had cost Kennedy nothing, not a headline, brought Clinton to the floor of the Senate and fixed a permanent asterisk beside his name. Nothing in the conduct itself had changed. Everything in the rules had changed, and the rules had changed in part for the very reason the rules around AIDS had changed, because the country had lost, across these two decades, its old capacity to leave the sexual life of others unspoken.

A Whig telling of all this would call it the triumph of honesty over hypocrisy, the healthy tearing-away of a Victorian veil, and it would have a case: the same frankness that mailed a life-saving pamphlet to a hundred million homes also refused, at last, to let powerful men hide their conduct behind a wall of press-room deference. A reactionary telling would call it a coarsening, the loss of a discretion and a reticence that had protected something worth protecting, and it too would have a case: a public square that will read four hundred pages about a stain on a dress has lost something a more buttoned nation possessed. The honest verdict holds both at once. The collapse of the wall was a single event with two faces, and the frankness that saved lives in the epidemic and the frankness that left a president nowhere to hide were the same frankness, arriving in the same package and inseparable within it.

Two things the era declined to resolve, and the refusals are the truest part of it. The judgment that the plague had seemed poised to deliver upon gay Americans never fell; the country, handed the chance to write its oldest prohibition into the permanent verdict of a health catastrophe, chose instead to treat a virus, and medicine in time supplied the drugs that turned the dying back, so that the thing many had wished to read as retribution became, in the end, a disease with a treatment. That was a genuine turning, and it held. The other refusal was darker. The exhaustion the country felt at the end of the Clinton affair, the disgust at the prosecutor’s zeal and the vote against conviction, looked at the time like a lesson learned — that the machinery of scandal and investigation had overreached and would now be reined in. The independent-counsel statute itself was allowed to lapse in 1999, abandoned by both parties as a weapon too dangerous to keep. But the appetite that the statute had served did not lapse with it. The practice of pursuing political enemies through investigation and exposure, of converting private failings and legal technicalities into instruments for undoing the results of elections, rolled on past the year 2000 undiminished and, if anything, gathering speed. The country had recoiled from the spectacle without giving up the taste for it, which is not the same as learning the lesson, and is closer to the opposite.

Return, at the end, to the two envelopes, the pamphlet and the report, arriving ten years apart in the same American homes. The same act lies behind both — the government of the United States setting explicit sex before every citizen, plainly and without apology — and the two ends it served could hardly be farther apart, the one to keep a people alive and the other to bring a president low. Between the mailing and the posting a country had learned to say out loud the things it had spent its whole history declining to say, had learned it in a school of mass death and could not unlearn it afterward, and had discovered, as such lessons always teach, that a habit of frankness once acquired does not confine itself to the good uses for which it was first so desperately needed.

[ PAUSE FOR QUESTIONS ]

Chronology, 1981–1999

• 1981, June 5 — the CDC reports the first cases; the press soon calls it the “gay plague”

• 1982 — the Gay Men’s Health Crisis is founded; the disease is named AIDS

• 1983, March — Larry Kramer publishes “1,112 and Counting”

• 1983–1984 — the virus is identified

• 1985, July — Rock Hudson announces he has AIDS; he dies that autumn

• 1986 — Bowers v. Hardwick upholds state sodomy laws; Koop issues his frank Surgeon General’s report

• 1987 — ACT UP is founded; AZT is approved; the memorial quilt covers the National Mall; Reagan speaks publicly on AIDS at length for the first time

• 1988 — Understanding AIDS is mailed to 107 million households

• 1990 — Ryan White dies; Congress passes the CARE Act in his name

• 1994 — Paula Jones sues Bill Clinton

• 1995 — American AIDS deaths peak near fifty thousand in a year; Lewinsky arrives at the White House

• 1996, summer — combination therapy is unveiled at Vancouver; the “Lazarus effect” begins

• 1997 — Clinton v. Jones: a sitting president may be sued for private conduct

• 1998, January — Clinton denies “sexual relations with that woman” under oath and to the nation

• 1998, August — the blue dress forces his admission; “the meaning of the word ‘is’”

• 1998, September — the Starr Report is posted to the internet

• 1998, December — the House impeaches Clinton

• 1999, February — the Senate acquits; the independent-counsel statute is allowed to lapse

End of AIDS and the Blue Dress — In Every Home

[ END OF LECTURE ]

Share

Subscribe now

The Wolf That Never Came

The Eighties and Nineties — The Economy, 1981–2000

Opening: The Report Nobody Wanted

Every February the Council of Economic Advisers laid before Congress a fat volume called the Economic Report of the President, hundreds of pages of tables and sober prose in which the administration set down how it understood the state of the economy and where it believed the economy was headed. In February of 1984 the chairman who signed that report was Martin Feldstein, a Harvard economist of the first rank and the president of the National Bureau of Economic Research, a man whose entire professional life had been built on the patient measurement of how taxes and deficits and savings actually moved through a working economy. He had come down to Washington two years earlier a friend of the enterprise, a believer in lower marginal tax rates and a sympathizer with much of what Ronald Reagan was trying to do. And then he had looked at the numbers, and the numbers had frightened him.

What frightened him was the deficit. The tax cut of 1981 and the defense buildup that ran alongside it had opened a gap between what the government spent and what it took in that stretched out, in the projections, as far as the eye could follow, year after year of red ink with no bottom in sight. Feldstein believed, on the basis of everything his profession knew, that deficits of that size would keep real interest rates high, that high rates would pull the dollar up to a punishing height, that the tall dollar would savage every farmer and manufacturer who sold abroad, and that the borrowing would eat into the savings the country needed to invest in its own future. He believed this the way an engineer believes a bridge is overloaded, and being the kind of man he was, he wrote it down and put it in the report and said it to Congress out loud, including the part no one in the Reagan White House wished to hear, that closing the gap would eventually require raising taxes as well as cutting spending.

The reaction was swift and it was brutal. Donald Regan, the Treasury Secretary, a blunt former chairman of Merrill Lynch who regarded Feldstein’s deficit sermons as both wrong and disloyal, told Congress in effect that it could throw the Economic Report away, that the volume the President’s own economists had labored over did not speak for the President on the matter that mattered most. The White House let it be known that its chief economist’s views on the deficit were his own private property. Feldstein became a man disowned by the government he served, warning in a loud voice about a fire his own colleagues insisted was merely the glow of a healthy hearth. In July of 1984, his leave from Harvard expiring, he went home to Cambridge, and few around the President were sorry to see him go.

Hold Martin Feldstein in the center of the mind, standing in that unwelcome winter with a report nobody wanted, because the strange thing is what happened next. The specific catastrophe he most feared did not arrive. The price surge that the deficits were supposed to unleash never came; inflation, which had been running at better than thirteen percent when the decade opened, kept falling and then stayed low, and for a while this made the worriers look like fools and the optimists look like prophets. Yet the deeper mechanism Feldstein had described — the high rates, the towering dollar, the gutted export towns — arrived exactly on schedule, quietly, through a door almost no one was watching, and presented its bill in a currency nobody had thought to count. The two decades that follow are, in a sense, the long education of a confident profession and an anxious public in the difference between the wolves they kept crying and the one that actually came. They spent twenty years bracing for beasts — runaway inflation, a second Great Depression, Japanese supremacy, American decline, deficits unto national ruin — and the beasts, for the most part, stayed in the woods. What that did to a country that could not stop scanning the treeline is the subject at hand.

[ PAUSE FOR QUESTIONS ]

[ ON-SCREEN: THE BALANCE, YEAR BY YEAR ]

Held on the screen through the hour, the reading advances one line at a time as the narrative reaches each date. The value is the federal budget balance for the fiscal year shown, as a share of the whole economy; a minus sign marks a deficit, a plus sign a surplus.

Fiscal year Balance, % of GDP 1981 −2.5 1983 −5.9 1986 −4.9 1989 −2.7 1992 −4.5 1995 −2.2 1998 +0.8 2000 +2.3 2002 −1.5

The Napkin

The idea that made the deficits possible was born, by legend, on a napkin. In 1974, in a Washington restaurant, a young economist named Arthur Laffer was arguing with two men from the Ford administration, Dick Cheney and Donald Rumsfeld, about tax rates, and to make his point he is said to have sketched a curve on the linen. The curve was simple and, at its two ends, undeniable. A tax rate of zero raises no money; a tax rate of one hundred percent also raises no money, because no one will work to hand over everything. Somewhere between the two ends lies a rate that raises the most, and the arresting implication is that above that rate a government can collect more revenue by charging less, because lower rates coax so much new work and investment into the open that the smaller slice comes off a far larger pie. A journalist at the table, Jude Wanniski of The Wall Street Journal, seized on the drawing, named it the Laffer Curve, and carried it out into the world. Laffer himself later claimed he could not remember the napkin at all. It did not matter. The picture had escaped, and it was exactly the picture a certain kind of politician had been waiting his whole life to see.

What the curve offered was a way out of the oldest bind in politics. A conservative who wished to cut taxes had always had to answer the accountant’s question: what will you cut to pay for it? The Laffer Curve answered that the question did not apply, that a tax cut aimed at the right rates would pay for itself, and that a man could therefore promise lower taxes and sound finances in the same breath and mean both. Wanniski preached the doctrine in a 1978 book, The Way the World Works; Robert Bartley gave it the editorial page of the Journal as a pulpit; and in the House of Representatives it found its apostle in Jack Kemp, a former professional quarterback from Buffalo who brought to economics the evangelical certainty of a man who had once believed he could win any game in the fourth quarter. Kemp and Senator William Roth wrote the doctrine into a bill, an across-the-board cut in income tax rates, and supply-side economics — the name it took, from its promise to lift the productive supply of the whole economy rather than merely to stoke demand — became the animating idea of the Reagan campaign.

The doctrine had a distinguished skeptic, and his skepticism is worth marking because of who he was and what became of him. In the Republican primaries of 1980, running against Reagan, George Herbert Walker Bush had looked at the arithmetic of a large tax cut married to a large military buildup, found it wanting, and pronounced the whole scheme “voodoo economics.” Then Reagan chose him as running mate, and Bush spent eight years as the loyal Vice President of the voodoo he had named, and then inherited its consequences as President, and broke himself on them. The line he coined in scorn would follow him to the end of his political life, which is a small parable, told in advance, of how the whole doctrine would work out: seductive as a campaign promise, treacherous as a governing fact, and hardest of all on the men who had to live inside the numbers after the applause died down.

[ PAUSE FOR QUESTIONS ]

Guns and Butter Without the Tax

In 1981 the doctrine met the arithmetic, and the arithmetic won. The Economic Recovery Tax Act slashed income tax rates across the board, dropping the top rate from seventy percent toward fifty and cutting the rest to match, the largest tax reduction in the nation’s history. At the same moment Reagan opened the spigot on defense, pouring money into the military at a pace not seen since the height of the Cold War’s earlier fevers, determined to spend the Soviet Union into exhaustion. Taxes fell and spending rose, and the man given the impossible job of making the two meet was the youngest cabinet-rank official in living memory, the director of the Office of Management and Budget, David Stockman.

Stockman was thirty-four, a former divinity student turned budget savant, a man who could recite the outlays of obscure federal programs from memory and who believed, with the fervor of the recently converted, that he could take an axe to the welfare state and balance the books by sheer force of intellect. And then, across his first year, he watched the numbers refuse to close, and being young and indiscreet and perhaps too honest for the room he was in, he said so, to a reporter, at length, over a series of breakfasts. The reporter was William Greider, and the article he published in The Atlantic in December of 1981, “The Education of David Stockman,” detonated inside the administration. In it the President’s own budget director confessed that the supply-side promise had been, in part, a cover story — that the doctrine of self-financing tax cuts had served as a respectable front for the older and simpler project of cutting the top rates for the well-off, a thing he compared to a Trojan horse and to the trickle-down theories of an earlier age. Most damning of all, he admitted the plain truth about the projections on which the whole edifice rested: “None of us really understands what’s going on with all these numbers.” Reagan, the story went, took him to the woodshed, and Stockman kept his job, chastened, for a few more years. But the confession could not be unsaid. The man in charge of the budget had told the country that the budget did not add up, and it did not.

What is easy to forget, and what a fair telling must insist upon, is that Reagan himself was no prisoner of the doctrine once its costs came due. Having cut taxes deeply in 1981, he turned around and raised them, quietly and repeatedly, when the deficits grew too alarming to ignore — a large revenue bill in 1982, a rescue of Social Security assembled by a commission in 1983, another revenue measure in 1984 — so that the sunny apostle of tax cutting spent much of his presidency clawing back a portion of the very cut that had made his name. And in 1986 he signed the measure that was, as it turned out, the genuine article of supply-side economics, hammered out with Democrats like Bill Bradley and Dan Rostenkowski: a reform that swept away a jungle of loopholes and brought the top rate down to twenty-eight percent while asking the same total revenue from the country in a cleaner way. Lowering the marginal rates that shaped a person’s incentive to work and invest, and broadening the base beneath them so that the Treasury lost nothing, was the supply-side ideal in its purest and most defensible form, the doctrine practiced with discipline rather than preached as an excuse. The Reagan record on taxes was thus a good deal more tangled than either his admirers or his enemies have found it convenient to remember. What did not get tangled, what marched straight ahead through every adjustment, was the deficit. The gross federal debt, which stood near nine hundred billion dollars when he took office, had tripled past two and a half trillion when he left. Feldstein, back at Harvard, had been right about the size of the hole. He had merely been wrong, along with everyone else, about which way the danger would come out of it.

[ PAUSE FOR QUESTIONS ]

King Dollar

Here is the puzzle at the heart of the decade, the one the forecasters got wrong and the one worth slowing down to understand. Every model in the profession’s toolkit said that a government borrowing on this scale, in an economy this hot, was courting inflation. The deficits were enormous, the recovery after 1983 was roaring, and the textbook consequence of too much money chasing too many goods was a return of the price spiral that had tormented the 1970s. The Blue Chip forecasters, the Federal Reserve’s own staff, the sober men of Wall Street — a great many of them expected inflation to reignite. It did not. Prices had risen better than thirteen percent in 1980; by 1983 they were rising near three, and there they stayed, low and quiet, straight through the boom. The single most confident prediction of the era’s economics simply failed to come true, and the reasons it failed are the reasons this lecture exists.

Part of the answer was monetary, and it belongs to another hour and another man: Paul Volcker, the towering, cigar-wreathed chairman of the Federal Reserve, had broken the back of inflation with interest rates higher than any living American had paid, and having broken it he held the line, so that the expectation of rising prices, once lodged so deep in every wage bargain, slowly drained out of the national mind. Part of the answer was the collapse of oil, which fell from around thirty-five dollars a barrel to nearly ten in 1986, pulling the price of everything that moved on wheels down with it. And part of the answer was that the engine of the old inflation had been dismantled: the confident, muscular unions of the postwar economy, which had marched wages and prices up together in a spiral, were losing their power to global competition and to a President who, in one of his first acts, had fired the striking air traffic controllers en masse and shown the country that the ground had shifted beneath organized labor’s feet.

But the piece the forecasters most badly underweighted, and the piece Feldstein had named in the report nobody wanted, was the dollar. Set the two great policies of the era side by side — Volcker’s tight money and Reagan’s loose budget — and their combination drove the real, inflation-adjusted rate of interest in America to towering heights, higher than the rest of the world offered. Money is not patriotic; it goes where it is paid best. Capital came flooding in from Europe and Japan and everywhere else to earn those American rates, and to buy American bonds a foreigner must first buy American dollars, and so the demand for dollars swelled and the dollar climbed, rising by roughly half against the world’s currencies between 1980 and its peak early in 1985. A tall dollar makes every imported thing cheap, and cheap imports do two things at once: they hold down the prices Americans pay directly, and they force every domestic producer who competes with an import to hold his own prices down or lose the sale. The inflation the deficits should have produced was, in effect, exported. It went out through the exchange rate and never showed up on the price tags.

That was the gift. The bill came attached to it and could not be detached. The same tall dollar that made imports cheap made American exports ruinously expensive abroad, and the farmer with grain to sell and the manufacturer with steel or machine tools to sell found the world’s markets closing against them. The trade deficit ballooned to sums without precedent; the word “Rust Belt” entered the language to describe the corridor of shuttered mills from Pennsylvania to the Midwest; a whole vocabulary of decline — deindustrialization, the hollowing of the heartland — grew up around towns that had done nothing wrong except to make things the strong dollar had priced out of world markets. The budget deficit and the trade deficit came to be called the twin deficits, and the family resemblance was no coincidence: the foreign capital that financed the government’s borrowing was the very same flood that lifted the dollar and drowned the exporters. Feldstein had described this machinery almost exactly, and the machinery ran almost exactly as he said. He had lost the room and kept the argument. The inflation the country was watching for never came; the damage he had actually predicted came instead, dressed as a trade deficit and a map of dead factory towns, which is a harder thing to see and an easier thing to blame on someone far away.

By 1985 even the administration could no longer bear what the dollar was doing to American industry, and Donald Regan’s successor at Treasury, the smooth and formidable James Baker, gathered the finance ministers of the other great economies at the Plaza Hotel in New York and struck a bargain to push the dollar back down. The Plaza Accord worked; the dollar descended over the following two years, steadied by a second agreement at the Louvre in 1987; and the episode stood as a quiet monument to how thoroughly the environment had changed. The most powerful economy on earth had discovered that it could not set the price of its own money by itself, that the tides of global capital now ran through the American house whether the owner willed it or not, and that a policy chosen in Washington for domestic reasons could reach out and idle a steel mill in Ohio by way of a currency trader in Tokyo.

[ PAUSE FOR QUESTIONS ]

The Wolves That Were Cried

If the inflation that never came was the first false alarm, it was far from the last, and to understand the temper of these two decades one must feel how relentlessly the country was told, by serious people, that ruin was at hand. The gloom was not stupidity; it was the honest extrapolation of recent pain by minds that had been badly burned in the 1970s and could not quite believe the fire was out. But it was wrong, again and again, in ways that ought to have taught more humility than they did.

The largest of the phantom wolves was Japan. Across the 1980s a genuine national dread took hold that the United States was being overtaken, that the future belonged to Tokyo, and that American managers, coddled and short-sighted, could not compete with the disciplined miracle of the Japanese firm. The dread had a literature — Ezra Vogel’s Japan as Number One had announced the theme in 1979 — and it had an emotional life that is almost impossible to recapture now. When Mitsubishi bought Rockefeller Center in 1989, and a Japanese firm bought Pebble Beach, and Sony bought Columbia Pictures, it felt to many Americans as though the country were being quietly purchased out from under itself. Autoworkers took sledgehammers to Japanese cars in parking lots; Hollywood made thrillers in which the inscrutable Japanese corporation was the menace; the paranoia curdled, in one Detroit case, into the murder of a Chinese American man beaten to death by men who took him for Japanese and blamed him for their layoffs. And then, at the very peak of the fear, Japan’s own bubble burst. Its stock market and its property market collapsed at the turn of the 1990s, and the country sank into a lost decade of stagnation from which it would not soon emerge, while the United States went on to the longest expansion in its history. The wolf that had frightened a nation lay down and did not rise.

The declinism ran deeper than Japan. In 1987 the historian Paul Kennedy published The Rise and Fall of the Great Powers, a magisterial account of how empires exhaust themselves, and its warning of American “imperial overstretch” — a nation whose global commitments had outrun its economic base — became the intellectual event of the season, devoured in Washington as a prophecy. An economist named Ravi Batra topped the bestseller lists with a book confidently titled The Great Depression of 1990, and the depression, like the overstretch, declined to appear. When the stock market crashed in a single terrifying day in October of 1987, falling further in an afternoon than it had in the crash of 1929, nearly everyone braced for the recession that such a fall had always heralded; the economy shrugged and grew smartly through the following year. The forecasters were not merely wrong at the margin. They were wrong in the same direction, toward doom, for the better part of two decades, and the country half-believed them the whole way.

When a real recession finally came, in the summer of 1990, it was mild by any historical measure and short — eight months, ending in the spring of 1991 — yet it landed on a national mood so primed for catastrophe that it felt like the confirmation of every dark prophecy at once. The recovery that followed was slow to produce jobs, and the phrase “jobless recovery” entered the language, and a new anxiety spread through the salaried middle class as corporations “downsized” and “restructured” and “outsourced,” turning those bland words into instruments of private dread. Robert Reich named the sufferers the anxious class. That the economy had in fact resumed growing did nothing to dispel the conviction, going into the election year of 1992, that something had gone permanently wrong with the American bargain. The wolf had been cried so long that the country could no longer hear that the woods were, in fact, growing quiet.

[ PAUSE FOR QUESTIONS ]

The Bond Market Votes

Into that gloom rode two men with charts. The first was Ross Perot, a Texas billionaire with jug ears and a flat prairie twang who ran for President in 1992 as an independent on essentially a single issue, the deficit, which he treated as a moral catastrophe and a national disgrace. Perot did a thing no candidate had done: he bought half-hour blocks of prime-time television and stood before the nation with a pointer and a stack of homemade flip charts, walking a rapt audience through the arithmetic of the national debt like a stern uncle explaining why the family could not afford a new car. He warned that a trade deal with Mexico would produce a “giant sucking sound” of jobs heading south. He was odd, and he quit the race and then re-entered it, and he still won nearly nineteen percent of the vote, the strongest showing for an outsider in eighty years — a measure of how deep the fiscal anxiety ran, and how hungry the country was to be told the truth about the numbers by someone who seemed to have nothing to sell.

The second man with charts was the one who won. Bill Clinton campaigned on a placard his strategist James Carville had tacked to the war-room wall — “It’s the economy, stupid” — and came into office in 1993 owning a deficit that had climbed toward three hundred billion dollars a year. His economic team, led toward its decisive choice by the former Goldman Sachs chairman Robert Rubin, made a bet that would have struck the supply-siders of a decade earlier as backward to the point of madness. They bet that the single most powerful thing a government could do to grow the economy was to shrink its own deficit. The reasoning ran through the bond market: persistent deficits kept long-term interest rates high, because lenders demanded a premium against future inflation and future borrowing; bring the deficit down credibly, and long rates would fall, and falling rates would light a fire under investment and housing and everything that ran on borrowed money. It was, in its way, the Feldstein diagnosis turned into a governing strategy, the deficit hawk’s warning repurposed as a growth plan.

The power of the bond market over a supposedly sovereign government was the revelation of those years, and Carville caught it in a line that has outlived almost everything else about the era. He had wanted, he said, to be reincarnated as the President, or the Pope, or a great baseball hitter; now he wanted to come back as the bond market, because the bond market could intimidate everybody. The administration’s plan, passed in 1993, raised the top income tax rate to nearly forty percent and cut spending, and it passed without a single Republican vote in either house, carried in the Senate only when Vice President Gore descended to break a fifty-fifty tie. Its opponents forecast disaster with total confidence — recession, lost jobs, a killed recovery, some said a depression. The disaster joined the long parade of wolves that never came. Long-term rates fell, investment surged, and the expansion that the pessimists had been burying for a decade gathered into the longest and strongest boom in the nation’s history.

The politics that followed were savage and, for the deficit, productive. A Republican Congress swept in behind Newt Gingrich in 1994 pledging a balanced budget of its own, and the collision between a Democratic President and a Republican Congress, each now claiming fiscal rectitude as its birthright, shut the government down twice in the winter of 1995 and 1996 before the two sides, exhausted, struck a bargain in 1997 to balance the books. What no one had planned, and almost no one had foreseen even five years before, was that the growth and the discipline together would carry the budget clean across the line from red into black.

[ PAUSE FOR QUESTIONS ]

The Surplus and the Vanishing Debt

The boom of the late 1990s was the vindication that the whole miserable series of false alarms had been waiting for, and it too was mostly unforeseen. Its deepest engine lay in the computer, whose long-promised gains in productivity finally began to show up in the national statistics after 1995, a story told in fuller detail in another hour. But it fell to the economic managers to decide whether to believe their own good fortune, and the man who had to decide was Alan Greenspan, Volcker’s successor at the Federal Reserve.

The profession handed Greenspan a rule and told him to obey it. The rule held that there was a floor beneath the unemployment rate — economists called it the natural rate, and pegged it near six percent — below which the labor market would grow so tight that wages, and then prices, would spiral upward again. As unemployment fell through the middle 1990s toward and then past that supposed floor, the Federal Reserve’s own models flashed red and a chorus of respected voices urged Greenspan to raise interest rates and cool the economy before the inflation the models promised could ignite. Greenspan looked at the same data and reached a different conclusion. He suspected that the computer had raised the underlying rate at which the economy could grow without overheating, that the old floor had dropped, and that he could let the expansion run. He held his hand. Unemployment fell to below four percent, a level the profession had declared impossible without an inflationary explosion, and the explosion, like every other forecast wolf of the era, did not come. Prices stayed low while the country approached what looked very much like full employment, and the economists who had spent the decade predicting inflation added one more failed prophecy to the pile.

Now the scorekeeper on the screen delivered its astonishment. The federal budget, which had run in deficit every year since 1969, crossed into surplus in fiscal 1998, and then stayed there — a surplus in 1999, and a larger one in 2000 — the first sustained run of black ink in a generation. The debt began, for the first time in living memory, to shrink. The country had opened the 1980s terrified of a deficit with no bottom, and closed the century watching that debt recede, a reversal so complete that the men who managed the nation’s money had begun, in all seriousness, to wonder what the world would look like once the government owed nothing at all.

[ PAUSE FOR QUESTIONS ]

What It Amounts To

Set the two decades in a line and the shape is the one this series keeps finding: a set of American institutions and habits of mind, assembled to fit one world, straining to adapt to another, and the adaptation carrying costs and blessings so tangled together that no simple verdict will hold them.

The triumphant telling is ready to hand, and it is not empty. Taxes were cut, inflation was slain, the Soviet Union was outspent into collapse, Japan’s challenge evaporated, and the country entered the longest expansion in its history and balanced its budget at the end of it. A believer in the Reagan project can point to every one of these and call the story a vindication. But the triumphant telling has to look away from too much to keep its shine. It has to look away from the debt that tripled, from the trade deficit that hollowed the industrial towns, from the fact that the inflation it credits to supply-side genius was slain by a Democratic-appointed central banker with brutally high interest rates and helped along by a collapse in the price of oil and a strong dollar that punished the very workers the boom was said to serve. And it has to look away from the plain admission of the President’s own budget director that the numbers never added up, that the doctrine of self-financing tax cuts was in part a story told to make an old preference respectable.

The reactionary telling is equally ready and equally partial. In that version Reagan wrecked the nation’s finances, mortgaged its future, deindustrialized its heartland to enrich its financiers, and left a mountain of debt for the children to pay. Each charge has a fact behind it, and yet the telling founders on what actually happened, which is that the deficits did not produce the ruin they were universally expected to produce, that inflation stayed dead, that the expansion was real and long and lifted employment to heights the models had called impossible, and that the fiscal hole was in the end climbed out of — climbed out of, moreover, by a bipartisan sequence that no single party can claim, running from a Republican President breaking his own tax pledge in 1990, through a Democratic President raising rates without a Republican vote in 1993, to a Republican Congress and a Democratic White House shutting the government down over which of them loved balance more.

What survives contact with the evidence is stranger and more instructive than either story. The environment in which American economic policy operated had changed under everyone’s feet. Capital had become global and could no longer be commanded within a nation’s borders; a currency had become a tide that ran through the domestic economy whether or not anyone in Washington wished it to; the wage-price spiral that had made the old inflation had been broken by foreign competition and the fading of the unions; a great inflation had been defeated and its defeat had unbolted the deepest engine of growth; and a machine on every desk was quietly raising the speed at which the whole economy could run. The country’s fiscal and monetary institutions, built in and for an older and more self-contained world, adapted to all this late and grudgingly and at real cost, and the profession that advised them adapted slowest of all, clinging to its natural rate and its inflation forecasts and its models of a closed economy long after the world those models described had dissolved.

One lasting mark the era left on the country’s politics deserves a final word, because it outlived everything else. The parties traded their reputations. Fiscal rectitude, the ancient Republican virtue, became for a season a Democratic achievement; the deficit, long denounced by conservatives as a Democratic vice, became the signature product of a conservative administration. Neither party chose this. Reagan’s coalition wanted lower taxes and a larger military and would not choose between them; Clinton’s team concluded, reading the bond market, that a Democrat could grow the economy only by out-disciplining the Republicans. Circumstance forced the exchange, and the confusion it sowed about which party stood for what has never fully cleared.

There is a last feature of these two decades that the partisan quarrels of a later age make almost impossible to credit, and it deserves to be set down plainly. For nearly the whole of the period the President and the Congress belonged to opposing parties — a Republican in the White House facing a Democratic House through all eight years of Reagan and both of Bush, a Democratic President facing a Republican Congress through the second half of the 1990s. A later generation would come to regard that arrangement as a formula for paralysis and spite, a government divided against itself and proud of the paralysis. It produced, in these years, something closer to the opposite. The tax reform of 1986, the genuine supply-side achievement of the era, was written by Republicans and Democrats together and carried by lopsided majorities in both houses. The deficit was brought to heel by a long chain of bargains — the revenue bill of 1982, the Social Security rescue of 1983, the budget deal of 1990 in which a Republican President broke his own tax pledge, the balanced-budget agreement of 1997 struck between a Democratic White House and a Republican Congress — each one a trade in which both sides swallowed something they hated because the other side swallowed something too. Measured against the standard of a later Washington, in which divided government came to mean the grinding of nearly all business to a halt, the ordinary competence of these years can look almost miraculous: two parties that disliked and distrusted each other, forced by the voters to share the same power, and choosing again and again, when the arithmetic finally cornered them, to govern anyway.

[ PAUSE FOR QUESTIONS ]

Chronology, 1981–2000

• 1981, August — the Economic Recovery Tax Act cuts income tax rates across the board; the defense buildup begins

• 1981, August — Reagan fires the striking air traffic controllers, a signal to organized labor

• 1981, December — “The Education of David Stockman” appears in The Atlantic

• 1982, August — a large revenue bill claws back part of the 1981 cut as deficits mount

• 1982 — Feldstein becomes chairman of the Council of Economic Advisers

• 1982, November — unemployment peaks near eleven percent in the deep Volcker recession

• 1983 — the deficit reaches its postwar peak near six percent of the economy; the recovery begins; Social Security is rescued

• 1984, February — Feldstein’s Economic Report warns of the deficits; Donald Regan dismisses it

• 1984, July — Feldstein resigns and returns to Harvard

• 1985, February — the dollar reaches its peak, roughly half again its 1980 value

• 1985, September — the Plaza Accord; the major economies agree to push the dollar down

• 1986 — the price of oil collapses toward ten dollars a barrel; tax reform brings the top rate to twenty-eight percent

• 1987, February — the Louvre Accord seeks to steady the falling dollar

• 1987 — Paul Kennedy publishes The Rise and Fall of the Great Powers

• 1987, October 19 — the stock market crashes; the widely expected recession fails to arrive

• 1988 — Bush campaigns on “read my lips: no new taxes”

• 1989 — Mitsubishi buys Rockefeller Center; Sony buys Columbia Pictures; the Japan panic crests

• 1990, July — recession begins; Bush signs a budget deal raising taxes, breaking his pledge

• 1991, March — the recession ends; the “jobless recovery” and the anxious class define the mood

• 1992, November — Perot wins nearly nineteen percent on the deficit; Clinton wins on the economy

• 1993, August — Clinton’s deficit-reduction plan passes without a Republican vote; Gore breaks the Senate tie

• 1994, November — Gingrich and the Republicans take Congress promising a balanced budget

• 1995–1996 — two government shutdowns over the path to balance

• 1997 — the Balanced Budget Act; unemployment falls below the supposed natural rate without inflation

• 1998 — the budget crosses into surplus for the first time since 1969

• 1998–2000 — the surpluses continue and the debt begins to shrink; forecasters begin to imagine the debt paid off entirely

End of The Economy — The Wolf That Never Came

[ END OF LECTURE ]

Share

Subscribe now

Students who leave a doctoral program never having encountered Ricardo’s argument with Malthus, or Menger’s quarrel with the German historical school, are not being trained in economics so much as in techniques, which can show you the answer but can’t show you the work. An economist who can’t explain how his discipline arrived at its own conclusions is not really practicing a liberal art at all, but running a very sophisticated calculator and calling it reasoning.

You could usefully replace macro with a history of thought course. Get rid of the idea that macro is a fixed body of models expressing truth.

Noah Smith writes,

The internet is substituting for cars. And if the internet is substituting for regular cars, why are people going to rush out and buy flying cars?

In other words, the answer to the question in the title of this post is that you won’t get a flying car — at least, not anytime soon — because you don’t really want one.

Michael Magoon writes,

The average American taxpayer moves from roughly the 20th percentile of the income distribution at ages 18–25 to the 75th percentile at ages 55–64, simply as they move through the normal life cycle.

Mark McNeilly writes,

Money, corporations, law, accounting, markets, and standards are technologies for coordinating human action. Our civilization depends on soft technologies.

Mainstream economics is too materialist. A lot of productivity comes from ideas, norms, and institutions, not just physical capital. See Invisible Wealth. Or Capitalism without Capital.

Share

Subscribe now

substacks referenced above: @

Mimir's Well

Soft Tech™: The Silent Code that Runs the World

MRM: This piece is the premise for a book I am writing of the same title: Soft Tech™: The Silent Code that Runs the World…

Read more

a month ago · 2 likes · 1 comment · Mark McNeilly

@

From Poverty to Progress

The Cause of Inequality Nobody Talks About: Age

The inequality debate may be deeply biased by its focus on race, gender, and education while largely overlooking one huge factor: age…

Read more

a month ago · 3 likes · Michael Magoon

@

Noahpinion

Why you won’t get a flying car

Stripe Press was kind enough to send me several books it had published, and one of these was J. Storrs Hall’s Where Is My Flying Car?. I had been meaning to read this book for a while, especially after Tyler Cowen recommended it and Jason Crawford gave it…

Read more

a month ago · 205 likes · 24 comments · Noah Smith

@

The Freeman

Doing Math Without Showing the Work

When I was associate dean of a business school, I discovered something that struck me as faintly scandalous: our economics majors could graduate without ever taking a course in the history of economic thought. Curious, I looked further and found that this was no local eccentricity. At most American universities, the history of economics is not required …

Read more

a month ago · 3 likes · The Freeman

Golden Crumbs

The Eighties — The Bond Market, 1978–1991

Opening: A Master of the Universe

The chronicler who caught the decade at its height was a man these lectures have met before. Tom Wolfe, the Virginian dandy in the three-piece white suit, the inventor of the breathless New Journalism, was last seen a full generation earlier riding along with Ken Kesey’s Merry Pranksters and their Acid Tests, observing the psychedelic tribe with the fascinated detachment of an anthropologist and fixing it in the national memory in The Electric Kool-Aid Acid Test. His eye had always been for the status games Americans played, the invisible codes of who ranked above whom, and in 1976 he had named the self-absorption of the age in an essay on the Me Decade. Now, in the middle 1980s, he turned that eye on the money culture of New York, and to force himself to finish he did a thing almost no serious novelist had done in a century: he wrote the book in public, in installments, on a magazine deadline, in the manner of Dickens. Across 1984 and 1985 the first draft of The Bonfire of the Vanities appeared in twenty-seven monthly parts in the pages of Rolling Stone. Then he tore it apart, revised it heavily, and published it as a novel in 1987, where it became a runaway bestseller and, by common verdict, the defining novel of the 1980s.

In the version that ran in the magazine, the hero, Sherman McCoy, had been a writer. When Wolfe rebuilt the book, he changed McCoy’s occupation, and the change is the whole decade in a single authorial decision. He made his master of ambition a Wall Street bond trader, because by 1987 that was the figure who caught the money-lust of the age as no writer could. The McCoy of the novel is a Park Avenue aristocrat with an enormous income and still larger debts, a wife and a mistress and a duplex apartment, and a sense of his own place in the order of things so exalted that he thinks of himself and his kind as Masters of the Universe — a phrase Wolfe lifted, with quiet malice, from a line of children’s action figures his daughter owned. And in one of the novel’s sharpest scenes, McCoy cannot explain to his own small daughter what he actually does for a living. His wife rescues him with a metaphor. Imagine a delicious cake, she tells the child, and imagine that every time a slice is passed around, a few crumbs fall from it, and imagine that Daddy’s job is to gather up those crumbs. Pass enough cake, take a crumb from each slice, and a man can gather a golden fortune from the leavings of other people’s transactions.

Hold Sherman McCoy in the center of the mind, and hold above all his occupation, because it is the arresting thing. A decade or two earlier, the notion of a bond trader as a Master of the Universe would have been a joke, and not a very good one. The bond business had been the dullest and most respectable corner of all finance, a genteel backwater where careful men bought the safe bonds of governments and blue-chip corporations and held them to collect the interest, an occupation so sleepy that the men who did it were said to be clipping coupons. Bonds were where money went to be safe and bored. Something had happened, across a handful of years, to lift the bond trader from that obscurity to the summit of American wealth and self-regard, to make him rich enough and grand enough that a novelist reaching for the representative striver of the age reached for him. What happened was an idea, a way of gathering golden crumbs off a flow of debt so vast that the crumbs became fortunes, and the idea would remake the American economy, spread wealth and ruin in equal measure, and lay a fracture that would open, twenty years on, into the worst financial collapse since the Depression.

[ PAUSE FOR QUESTIONS ]

The Alchemy

A bond is the simplest thing in finance and the most consequential. It is a loan made tradable. When a government or a company borrows by issuing a bond, it promises the holder a stream of interest and the return of the principal at the end, and because that promise is written on a piece of paper that can be sold to someone else, the loan becomes a thing that lives in a market, changing hands and changing price. Two dangers ride on every bond, and the whole drama of this decade is the story of the country learning each of them the hard way. The first is interest-rate risk: because a bond pays a fixed rate, its price falls when rates in the wider world rise, since no one will pay full value for an old bond yielding five percent when new ones yield ten. The second is credit risk: the borrower may simply fail, and pay back less than promised, or nothing. Safe bonds carry little of either. The action of the 1980s lay in the bonds that carried a great deal.

The idea that transformed the business had a plain and even ugly name, securitization, and it worked a genuine alchemy. Across the country lay vast fields of debt that could not be traded at all, buried and immovable on the books of the institutions that had lent the money. The mortgage on a single family’s house in Ohio was an illiquid thing, known only to the local bank that held it, impossible to sell to a stranger who could not judge whether that family would pay. The loan to a small and risky company was the same, locked in the vault of whichever bank had made it. Securitization reached into those fields and made them liquid. Gather thousands of mortgages into a single pool, sell to investors bonds backed by the river of monthly payments flowing out of that pool, and the illiquid loan on one house became a share of a security that could be bought in New York and sold in Tokyo. The particular risk of any single borrower washed out in the pool of thousands; what remained was a bond, tradable anywhere on earth. And every time a mortgage flowed into a pool and a bond flowed out to an investor, a firm in the middle took its crumb.

Across the 1980s this alchemy was worked upon the two greatest fields of American debt, and each became an engine of the transformed bond market, and each taught the nation one of the bond’s two dangers. The first engine ran on the home mortgage, and it taught the lesson of interest-rate risk through a catastrophe that bankrupted an entire industry. The second ran on corporate borrowing, and it taught the lesson of credit risk through a single man who built a machine for lending money to companies no one else would touch. The men who ran these engines became the most famous and infamous figures in American finance, and the sums that flowed through their hands made Sherman McCoy’s imaginary fortune look like the small change it was drawn from.

[ PAUSE FOR QUESTIONS ]

The First Engine: The Mortgage

The mortgage engine was born out of a disaster, and the disaster was the savings and loan crisis. The thrifts, the local savings and loans that were the country’s great source of home mortgages, had been built to do one simple thing: take in deposits and lend them out as thirty-year fixed-rate mortgages. The design concealed a lethal flaw, a mismatch of time. The thrift’s assets were locked up for thirty years at a fixed rate, while its deposits could walk out the door tomorrow, and when Paul Volcker drove interest rates toward twenty percent at the turn of the 1980s to break the back of inflation, the flaw became fatal. The old mortgages on the books, made years earlier at six and eight percent, were suddenly worth a fraction of their face value, because no one would pay full price for a loan yielding six percent in a world of sixteen. The thrifts were insolvent, most of them, on paper, all at once. The value of their assets had collapsed beneath the value of what they owed, and the cause was the first danger of every bond, the fall of price when rates rise, visited upon an entire industry at a stroke.

The response of the regulators was to look away, a policy that came to be called extend and pretend. Rather than force the thrifts to mark their mortgages down to what the market would actually pay, which would have confessed the insolvency and shut them down, the authorities let them carry the loans at the old book value, maintaining the fiction that a mortgage worth seventy cents was still worth a dollar. The fiction was elaborated into a product. Through a program run by Freddie Mac, a thrift could hand over a pool of mortgages carried at, say, ten million dollars, pay a stiff fee, and receive in exchange a package of securities that the market valued at six million — and, because the transaction counted as a swap of one form of the same loans for another, carry those securities on its books at the old ten million, the loss made visible only if it ever sold them. So the thrift did not sell. It pledged the securities as collateral, borrowed against an asset it valued at ten million and the world valued at six, and stayed open. The interest-rate loss, bounded and knowable, was thereby converted into something worse, a zombie industry kept alive on the public’s insured deposits and given every incentive to gamble for its own resurrection.

That same crisis, which taught the lesson of interest-rate risk, produced the flood of raw material for the first great securitization, because insolvent thrifts shedding their old mortgages, and a tax change in 1981 that gave them a powerful reason to sell, poured a torrent of home loans into the market looking for a buyer. The buyer was a trading desk at Salomon Brothers, and the man who built it was Lewis Ranieri.

Ranieri had started at Salomon in the mailroom, a heavyset, loud, profane son of Brooklyn who never finished college and rose by force of intelligence and appetite to run the firm’s mortgage department, a man in suspenders eating Italian food at his desk and playing crude practical jokes while assembling the most mathematically sophisticated market in the world. The instrument he traded was not his invention; the government had created the basic mortgage-backed security, the pass-through, back in 1970, a bond that simply passed the homeowners’ payments through to the bondholders. It had languished for a decade as a sleepy oddity. Ranieri built a market around it. He hired the mathematicians to model how quickly homeowners would prepay their loans, the great puzzle of the business; he built the trading operation that made the bonds liquid; he devised, with his rivals, the structures that sliced a mortgage pool’s cash flow into tranches, so that a cautious pension fund and a daring speculator could each buy the piece that suited it. For a few years in the middle 1980s his desk had the field almost to itself and threw off crumbs so enormous they underwrote the whole swaggering culture of the firm.

What Ranieri accomplished was, in the main, a genuine good, and the series does not flinch from saying so. Securitization knit the local mortgage market into the vast pool of the world’s capital, so that a family in Ohio borrowed, in effect, from a pension fund in Japan, and the competition of all that capital drove the cost of a mortgage down and widened the door to homeownership. The American home loan, the most local and domestic object imaginable, became a security that traded around the globe. And in that triumph lay the buried fracture. The bank that made a mortgage and then sold it into a security no longer held the loan, and so no longer bore the loss if the borrower failed. The oldest discipline in lending, the knowledge that you would eat your own mistakes, was severed, quietly, at the root. In the 1980s the mortgages being pooled were sound, and the fracture stayed hidden. It would take twenty years, and a far greater flood of far worse loans run through the same machine, for the severed tie between the lender and the loss to bring down the financial system.

[ PAUSE FOR QUESTIONS ]

The Second Engine: The Junk Bond

The second engine ran on corporate debt, and it belonged almost entirely to one man. Michael Milken was a trader at Drexel Burnham Lambert, a second-tier firm, and he had seized on an academic finding that the rest of Wall Street had ignored. Old research had shown that a diversified portfolio of low-rated, high-yielding bonds returned more, over time, than the safe bonds of blue-chip companies, even after accounting for the defaults, because the extra interest more than paid for the losses. The market had shunned such bonds out of a kind of snobbery, refusing to lend to any company that lacked an investment-grade blessing, and Milken saw that the snobbery had left money lying on the ground. He began by trading the bonds of fallen angels, companies that had once been sound and had slipped, whose bonds the respectable houses dumped at panic prices.

Then he made the leap that changed corporate America. If a diversified pool of risky bonds was a sound investment, then risky companies that had never been able to borrow at all, the upstarts and the outsiders shut out by the investment-grade priesthood, could be financed by issuing new high-yield bonds, the instrument the market called junk. Milken built a machine to do exactly that, and through it flowed the capital that built companies the old banks would never have touched: the cable-television ventures, the fledgling cellular-telephone companies, the casino and media upstarts that became household names in the next decade. This was democratization of a real and important kind, the opening of the capital markets to companies judged unworthy by the establishment, and much of what it financed was genuine and lasting.

The same machine had a second use, and it is the one the public came to know. A raider who wished to seize a great corporation needed money on a scale no individual possessed, and Milken could summon it. His network of junk-bond buyers was so deep and so loyal that he could raise billions on a few days’ notice, and Drexel began issuing what became the most feared piece of paper in America, the highly confident letter, a simple statement that Drexel was highly confident it could raise the money for a given takeover. To the target company it was as good as a loaded gun, for it meant that a raider of modest means now had the financial power to swallow a corporation many times his size. The takeover battles of the stock market, the raids of Pickens and Icahn and the rest, ran on the fuel this engine manufactured. The upstart financing and the takeover financing were the same machine turned to two purposes, and both purposes flowed through one telephone in one office.

The office was in Beverly Hills, three thousand miles from Wall Street, and the distance was itself a statement. Milken worked from an X-shaped trading desk on Rodeo Drive, arriving before dawn, by four-thirty in the morning California time, to be at his post when the New York markets opened, and around him gathered a devoted, hard-driving crew who regarded him with something close to reverence. He was an unlikely emperor, ascetic where the age was gaudy, a man who did not drink or smoke or party, who wore a toupee and poured the whole of his formidable energy into the market and his family, indifferent to the trappings of wealth even as he accumulated more of it than any employee in the history of the world. Once a year Drexel gathered the players of the high-yield world in Beverly Hills for a conference that acquired, from a book, the name by which history remembers it, the Predators’ Ball, where the raiders and the financiers and the money met, and where, for a few years, it seemed that no corporation in America was beyond Milken’s reach. The high yields his bonds paid were the price of their credit risk, the danger that the borrowers might fail, and the market bought them on the faith that the diversified pool would always come out ahead. When the economy turned at the end of the decade, the market would discover how much of that risk had been real.

[ PAUSE FOR QUESTIONS ]

The Money and the Glamour

Sherman McCoy was an invention, but his real-life counterparts were everywhere, and for the first time since the 1920s finance became the great subject of the culture and the men of the bond markets its improbable celebrities. Alongside Wolfe’s novel there rose a shelf of true accounts that the country read and argued over. A young Salomon bond salesman named Michael Lewis quit the firm and wrote Liar’s Poker, the tale of Ranieri’s mortgage desk and the trading floor and the swaggering culture that the money had built, a book that thousands of ambitious young people misread, to the author’s dismay, as a recruiting brochure. A journalist named Connie Bruck told the story of Milken’s machine in The Predators’ Ball. Two Wall Street Journal reporters, Bryan Burrough and John Helyar, chronicled the largest takeover of the age in Barbarians at the Gate. The movie houses gave the country Gordon Gekko, the corporate raider of Oliver Stone’s Wall Street, whose creed that greed is good drew on a real line delivered by the real arbitrageur Ivan Boesky, who told a graduating business-school class in 1986 that greed was healthy and that a man could be greedy and still feel good about himself. The vocabulary of the moment entered the language: the decade of greed, the masters of the universe, the predators’ ball. Nearly half the graduating class of Yale, in one of those years, applied to a single Wall Street firm.

Set against our own time, the sums that scandalized the age carry a strange lesson, and it is worth putting the figures on the screen.

[ ON-SCREEN: THE SCALE OF THE MONEY ]

Held on the screen while the figures are discussed.

The 1980s Our own day RJR Nabisco, the largest corporate buyout ever attempted, at about $25 billion in 1988 — a record that would stand for the better part of two decades → an acquisition of $25 billion draws little notice; single companies are valued in the trillions John Gutfreund, the highest-paid chief executive on Wall Street, at $3.1 million in 1986 — a figure that provoked outrage → a large-bank chief earns roughly ten times that as a matter of course; the top hedge-fund managers clear more than a billion dollars in a single year Michael Milken, the outlier, paid $550 million in the single year of 1987 → a sum that would rank among the very largest Wall Street paydays even now

The record buyout of the decade was the seizure of RJR Nabisco, the tobacco and food conglomerate, which the firm of Kohlberg Kravis Roberts carried off in the closing weeks of 1988 for roughly twenty-five billion dollars, a hundred and nine dollars a share, in a bidding war so ferocious and so revealing of Wall Street’s appetites that it became the subject of Barbarians at the Gate and stood, for close to twenty years, as the largest leveraged buyout in history. Twenty-five billion dollars was the summit of the age of greed. Today an acquisition of that size passes with little comment, and a single technology company can be worth more than a hundred RJR Nabiscos. John Gutfreund, the chairman of Salomon Brothers whom the business press had crowned the King of Wall Street, earned three million dollars in 1986, and the outrage it provoked rested entirely on the fact that no chief executive on the Street earned more; the king of finance sat atop the whole of it at a salary a competent regional banker would find modest today, when a large-bank chief earns ten times as much without a headline and the titans of the hedge funds earn more in a year than Gutfreund earned in his life.

The whole architecture of scandal, in other words, was built to a scale that the intervening decades have dwarfed, and the phrase decade of greed measures the threshold of its observers more than the size of its money. One figure escapes the pattern, and it is the one everyone remembers. Michael Milken was paid five hundred and fifty million dollars in the single year of 1987, a sum that would be staggering in any decade and that would stand near the top of Wall Street’s pay even now. Milken did not share the modest scale of his era. He prefigured the scale that was coming, inventing the junk bond and, in his own compensation, the modern financial fortune, a generation before such sums became the ordinary reward of the money world. The claim that the era’s money looks small holds across the whole decade with exactly one man standing outside it, and the one man outside it is the one the age remembers best.

[ PAUSE FOR QUESTIONS ]

The Reckoning

The end came, as ends often do, through a smaller sin than the great ones. Ivan Boesky, the arbitrageur who had made a fortune betting on takeovers, was caught trading on inside information, and to save himself he agreed to inform, wearing a recording device to capture his associates and pointing federal prosecutors toward the center of the web. The young United States Attorney in Manhattan, Rudolph Giuliani, pursued the trail to Beverly Hills. Milken was indicted in 1989 on a sprawling array of charges, and in 1990 he pleaded guilty to six felonies of securities and tax violation, paid fines and settlements that ran past a billion dollars, and went to prison, serving twenty-two months of a ten-year sentence. Drexel Burnham Lambert, the firm his machine had raised from the second tier to the summit, could not survive the loss of him and the collapse of the market he had made. In February of 1990 it filed for bankruptcy, the fifth-largest investment bank in the country, gone.

The credit risk the junk market had carried all along now came due. As the economy softened at the decade’s turn, the high-yield bonds began to default in earnest, the diversified pool no longer covering its losses with its yields, and the over-leveraged deals of the boom staggered under debts that had assumed the good times would never end. Even RJR Nabisco, the trophy of the age, nearly foundered beneath the mountain of borrowing that had bought it, and its new owners had to pour in fresh money to keep it from collapse. The savings and loan disaster, meanwhile, reached its long-deferred reckoning. In 1989 Congress finally moved to clean up the wreckage, creating a Resolution Trust Corporation to bury the corpses of the failed thrifts, and the bill to the taxpayer for a decade of extend and pretend came to something on the order of a hundred and twenty-four billion dollars. Charles Keating, whose Lincoln Savings had gambled its insured deposits on Milken’s junk and on desert real estate, went to prison, and the five United States senators who had leaned on his regulators on his behalf carried the stain of the Keating Five for the rest of their careers.

The house where Ranieri had built the first engine did not escape its ending. In 1991 it emerged that a trader at Salomon Brothers, Paul Mozer, had for months been submitting false bids to corner the auctions of United States Treasury bonds, the very bonds at the safe heart of the whole system, and that John Gutfreund had learned of it and failed to report it promptly to the authorities. The scandal nearly destroyed the firm, which stood to lose its standing as a dealer in government bonds and with it its life. Gutfreund, the King of Wall Street, was forced to resign in disgrace, toppled by a bond scandal in the market he had ruled. Salomon was saved only because its largest shareholder, Warren Buffett, stepped in as chairman, went before Congress to promise a new rectitude, and imposed on the firm a standard of conduct famous for its severity: lose money for the firm and he would be understanding; lose a shred of its reputation and he would be ruthless. The men who had gathered the golden crumbs by the ton left the stage, and one of Salomon’s own star traders, cut loose in the wreckage, went off to found a hedge fund whose spectacular collapse a few years later would furnish a cautionary tale for the decade to come.

[ END OF LECTURE ]

What It Amounts To

Set the decade in a line and the shape is the one this series keeps finding: a genuine adaptation to real pressures, carrying benefits its champions could rightly claim and consequences its champions never saw, escaping the control of the men who made it.

The pressures were real. The taming of inflation had shattered the old thrift model and thrown a torrent of mortgages onto the market. Deregulation and a decade of financial ferment had loosed a hunger for higher returns. The computer had arrived to make possible the intricate mathematics of pooling and slicing that the new securities required. Into this environment came securitization, the idea of turning illiquid debt into tradable bonds, and it was, taken whole, a powerful and largely beneficial innovation. It drove down the cost of a mortgage and widened the path to owning a home. It opened the capital markets to a generation of upstart companies the establishment had locked out, and financed real enterprises that employed real people. It disciplined a complacent corporate order that had grown fat on the assumption that no one could ever come for it. The junk bond built cable television and cellular telephony, and it also loaded sound companies with ruinous debt. Securitization made mortgages cheap for everyone, and it also cut the ancient cord that had bound a lender to the consequences of its loans. The promise and the fracture arrived in the same package, inseparable, as the adaptations in this series always do.

There is a harder truth folded into the golden crumbs, and Sherman McCoy’s inability to explain his work to his own daughter holds it. The art of the age was intermediation, the gathering of a tiny slice off an enormous flow of other people’s money, and the whole genius of securitization was to make the flow so vast that the slices became fortunes. A crumb off each of millions of mortgages, a sliver of each billion-dollar raid, and a man could call himself a Master of the Universe. A slice taken, though, is a slice taken from someone, and the risk that rode on the flow did not vanish when it was pooled and sold; it was passed along, out of the room, to come to rest on people who never tasted a crumb. When the bets went bad the losses fell, and they fell elsewhere, on the taxpayers who paid the hundred and twenty-four billion dollars.

And the machine did not die with the decade that built it. Securitization was the 1980s’ most durable bequest to American finance, and the fracture planted in Ranieri’s triumph, the lender who no longer held the loss, was carried forward into the next century and turned upon a vaster and far more dangerous pool of mortgages than any Salomon had ever touched. The exuberance of the stock market and the securitization of the bond market, the two great themes of these two lectures, would meet at last in the American house, and the reckoning that the masters of the universe had so confidently passed out of the room would arrive in 2008, on a scale that would have silenced even the King of Wall Street. The men who gathered the golden crumbs and walked away rich had built something larger than they knew, and the losses they had arranged for others to bear were passed down, with interest, to a generation that had never read the novel or heard the names.

[ PAUSE FOR QUESTIONS ]

Chronology, 1978–1991

• 1978 — Milken moves his high-yield operation to Beverly Hills

• 1979–1982 — Volcker’s interest-rate shock renders much of the thrift industry insolvent

• 1981 — a tax change spurs thrifts to sell their old mortgages, flooding the market

• 1983 — the mortgage desks devise the collateralized mortgage obligation, slicing pools into tranches

• 1984–1985 — Wolfe serializes The Bonfire of the Vanities in Rolling Stone

• 1985 — the takeover wave crests on Milken’s junk-bond financing

• 1986 — Gutfreund earns $3.1 million, the highest on the Street; Boesky is caught and begins to inform

• 1987 — Milken is paid $550 million; Wolfe publishes Bonfire as a novel; the mortgage market crowds and Salomon retreats from it

• 1988, November — KKR wins RJR Nabisco for about $25 billion, the largest buyout in history

• 1989 — Milken is indicted; Congress creates the Resolution Trust Corporation to resolve the thrifts

• 1990, February — Drexel Burnham Lambert files for bankruptcy

• 1990 — Milken pleads guilty to six felonies and is sentenced to prison

• 1991, August — the Treasury-auction scandal forces Gutfreund out; Buffett takes the chair to save Salomon

End of The Bond Market — Golden Crumbs

Share

Subscribe now

The Measure of Exuberance

The Eighties and Nineties — The Stock Market, 1979–2000

Opening: The Sentence in the Evening

On the evening of the fifth of December, 1996, the chairman of the Federal Reserve stood before a black-tie dinner of the American Enterprise Institute in Washington and delivered a long, dry address on the history and philosophy of central banking. Alan Greenspan was the most celebrated practitioner of a craft he had made his own, the art of speaking in public for great lengths of time while committing himself to nothing, of wrapping every thought in so many qualifications that no listener could pin a position on him and no market could move on his words. He called it, when he was being candid about it, constructive ambiguity. The reporters called it Fedspeak. On this particular evening, deep in the middle of a paragraph about inflation and asset values, he let a single vivid phrase slip out of the fog. He asked how a central banker was to know when irrational exuberance had unduly escalated the value of stocks and houses, values that might then contract suddenly and painfully, as they had in Japan across the preceding decade.

The phrase was buried in a question, hedged with conditionals, offered as a thing a wise man could not easily judge. It did not matter. The Tokyo market was open at that hour, and it heard the two words and nothing else, and it fell at once, closing off three percent. The wave rolled west with the sun. Hong Kong dropped, then Frankfurt, then London, and by the time the New York exchange opened the next morning the American market opened lower too, the Dow standing that day near 6437. A man had spoken two words in the middle of a sentence at a dinner, and every major market on earth had flinched in turn, going down like a row of dominoes around the curve of the planet. It was a demonstration, had anyone paused to read it, of how tightly the world’s money had been wired together, and of how much of the whole vast apparatus now hung on the moods and phrasings of one man in Washington.

Two days before the dinner, on the third of December, two economists had sat in the boardroom of the Federal Reserve and shown Greenspan and the governors a chart. John Campbell of Harvard and Robert Shiller of Yale had come to argue, with a century of data behind them, that the American stock market had climbed to a height from which history said it would return poor rewards for years to come. Whether the phrase Greenspan used two evenings later was his own, or whether it lodged in his mind from that briefing, no one can finally say. Shiller believed the visit had planted it. Greenspan later wrote that the words had come to him in his bath while he drafted the speech. The origin is lost in exactly the fog the man cultivated, and the ambiguity is fitting, because the deepest question the whole episode raises is whether anyone, the chairman included, can truly see a bubble while standing inside it, or know the difference between a market gone mad and a market that understands something the skeptic has missed.

Hold the December evening in the center of the mind, because the chart the two economists carried into that boardroom rested on a single number, and the story of the American stock market across two decades can be told as the story of that number: what it measured, where it stood, and why the man who commanded the most powerful financial institution on earth glanced at it, spoke its warning aloud, and then decided to look away.

[ PAUSE FOR QUESTIONS ]

The Number

A share of stock is a claim on the future earnings of a company, and the oldest question in finance is what such a claim is worth. The rough-and-ready answer the market uses every day is the price-to-earnings ratio, the price of a share divided by the company’s earnings per share over the past year. A ratio of ten means an investor pays ten dollars for each dollar the company earned; a ratio of thirty means thirty. Laid across the whole of the market, the ratio becomes a gauge of mood. When investors are hopeful, they pay more for each dollar of earnings and the ratio rises; when they are fearful, they pay less and it falls. The trouble with the plain ratio is that earnings themselves lurch with the business cycle. In a recession, profits collapse, and the ratio shoots upward and makes stocks look expensive at the very moment they are cheapest; at the peak of a boom, profits are swollen and unsustainable, and the ratio sinks and makes stocks look cheap at the very moment they are dearest. The instrument lies most loudly exactly when a reader needs it most.

Shiller’s contribution, developed with Campbell in the late 1980s, was to still the lurching. In place of a single year’s earnings, he used the average of the past ten years of earnings, each year adjusted for inflation so that a dollar of 1985 profit and a dollar of 1995 profit could be laid side by side. Ten years is long enough to contain a full business cycle, boom and bust together, so the denominator no longer heaves with the moment. The result acquired several names — the cyclically adjusted price-to-earnings ratio, the CAPE, the Shiller P/E — and it did something the plain ratio could not. It offered a stable measure of how richly the market was priced against its own long history, and history had a great deal to say about what followed from each reading. Across more than a century, a low ratio had been followed by years of strong returns, and a high ratio by years of weak ones. The number was a barometer, and it was pointing, in December of 1996, at a storm.

What the storm meant was the very thing in dispute, and the dispute had a shape worth drawing now, because it runs through the entire hour and does not resolve even at the end. The reigning theory of the academy belonged to Eugene Fama of the University of Chicago, who had built his career on the efficient-markets hypothesis, the proposition that the price of a stock already contains everything that can be known about it. In such a market prices are always, in the only sense that matters, correct; they move only when genuine news arrives, and since news by definition cannot be predicted, neither can prices. A high ratio, in Fama’s world, is a fact about the future rather than an error about the present. If investors will pay thirty dollars for a dollar of earnings, they must have good reason, and to call the market overpriced is only to say one disagrees with the collective judgment of everyone with money at stake. Bubbles, in this understanding, are a word people reach for after the fact, and mean little more than a price that later went down.

Against this stood Shiller, who had made his name in 1981 with a finding that struck at the heart of Fama’s edifice. Stock prices, he showed, jumped around far more violently than the stream of dividends underneath them could possibly justify. If prices merely reflected rational forecasts of what companies would pay out, they should be steadier than the dividends themselves; instead they swung wildly, driven by waves of feeling that had no anchor in the earnings below. Prices, Shiller argued, were moved by the enthusiasms and terrors of crowds, by stories that spread and fashions that turned, and a market so moved could climb to heights that reason could not defend and would, in time, come down from. The two men held incompatible pictures of the same daily fact, and each had spent a career amassing evidence, and there was no experiment that could settle between them. Keep both men in view. The market they were arguing about was about to run an experiment of its own.

[ PAUSE FOR QUESTIONS ]

[ ON-SCREEN: THE RATIO, YEAR BY YEAR ]

Held on the screen through the hour, the reading advanced one line at a time as the narrative reaches each date. The value is the Shiller ratio for the S&P 500, at or near the close of the year shown.

Date Shiller P/E end of 1980 9.3 trough, summer 1982 ≈ 7 end of 1985 11.7 end of 1990 15.6 end of 1996 28.3 end of 1999 43.8 end of 2000 37.0

The long-run median across the century and more is roughly sixteen. The 1929 peak, before the greatest crash in the market’s history, reached about thirty-three. The reading of December 1999 — a shade above forty-four at its height that month — is the highest in the recorded history of the American stock market.

[ PAUSE FOR QUESTIONS ]

The Dead Market

Begin at the bottom of the table, in 1980, with the ratio at nine and the market as close to dead as a great market ever comes. The mood of the moment had found its perfect expression the summer before, in August of 1979, when BusinessWeek ran a cover that has since become famous for its timing. “The Death of Equities,” it announced, and argued the case in earnest: inflation had so battered the returns on stocks, and so soured a generation of Americans on the whole enterprise of owning them, that the stock market had ceased to function as a place where ordinary people built wealth. Seven million shareholders had fled the market over eight years. Pension funds were begging for permission to abandon stocks for real estate and gold and futures. The small investor, the magazine reported, had left and would not return, and this was described as a permanent condition, a structural fact of American life, the death of an institution.

The obituary had a real argument behind it, and the argument was inflation. When prices rise at double digits, as they did through the 1970s, the future earnings a share of stock represents are worth steadily less in the dollars an investor will actually receive, and the terror of that erosion drove buyers away. Bonds, meanwhile, were paying interest rates that climbed with the inflation toward the teens, offering a fat and certain-seeming return that made the risk of stocks look foolish. And beneath the arithmetic lay something harder to measure, a loss of faith. A decade of stagnation and oil shocks and defeat abroad had taught Americans to expect less, and the stock market is above all an engine of expectation, a wager that tomorrow will be richer than today. When the country stopped believing that, the market died on the vine, and its priced-for-the-grave ratio of nine was the reading of that despair.

The culture around the market in 1980 is nearly unrecognizable from our own. Stock ownership was the province of the wealthy and the elderly, a rich man’s pastime carrying a whiff of the casino. The typical worker approaching retirement expected a pension, a monthly check promised by an employer and paid out of a fund the employer managed, and the worker neither knew nor cared what that fund held. He owned no shares, watched no ticker, and would have found the idea of an ordinary family checking the value of its stocks each evening as strange as checking the price of pork bellies. The market was a distant machine that had nothing to do with him, and in 1980 it looked like a machine that had broken down for good.

[ PAUSE FOR QUESTIONS ]

The Turn

The engine that lifted the market off the floor was the same one that had held it there, running now in reverse. Inflation had killed the market, and the killing of inflation revived it. In 1979 Paul Volcker had taken the chairmanship of the Federal Reserve and set out to break the price spiral by the only means that had ever worked, driving interest rates to heights that choked the economy into the sharpest recession since the 1930s. It was brutal and it was hated and it worked. Inflation, which had run above thirteen percent, was wrung out of the system, and as it fell, the rates fell with it, and the fall of interest rates is the deepest fuel a stock market can burn.

The reason lies in the same arithmetic that had done the damage. A share is worth the future earnings it represents, discounted back to what they are worth today, and the rate of interest is the yardstick of that discounting. When rates are high, a dollar promised years from now is worth little in the present, and stocks are cheap. When rates fall, that distant dollar swells in present value, and every stock in the market is lifted by the same rising tide, whatever its particular story. In August of 1982, with the recession at its worst and the ratio bottoming near seven, the market turned. The Dow Jones average, which had spent seventeen years unable to break past a thousand, stood that month around 777 and began to climb, and it would not truly stop climbing, through crashes and panics and recoveries, for eighteen years. The great bull market of the century had begun, in the ashes of the worst inflation in American memory, and almost nobody recognized it for what it was.

[ PAUSE FOR QUESTIONS ]

The Market Comes Alive

The first energy the reborn market threw off was not a mania of prices, for the ratio in the middle 1980s still sat around twelve, a cheap and sober reading. The energy took the form of a fight, the fight for the control of companies, and it made the corporate raider the signature figure of the decade. The logic of the raid grew straight out of the low ratio. If the stock market valued a company’s shares at less than the pieces of that company would fetch if sold, or less than a sharper management could wring from its assets, then a bold outsider could borrow against those very assets, buy the company over the heads of its managers, and pocket the difference. Cheap stocks were an invitation, and a generation of raiders accepted it.

They were vivid men, and the public came to know their names as it had once known the names of movie stars. T. Boone Pickens, a folksy Oklahoma oilman with a genius for publicity, cast himself as the champion of the ordinary shareholder against the pampered executives of the great oil companies, and in 1984 he set his small firm, Mesa, circling the vastly larger Gulf Oil, which escaped him only by flinging itself into the arms of Chevron in the largest merger yet seen. The next year he menaced Unocal. Carl Icahn, abrasive where Pickens was folksy, a former arbitrageur who had learned the market’s angles cold, took control of Trans World Airlines in 1985 and gave the country a first hard look at what a raider did to a company once he held it. The argument these men made for themselves was serious and had a genuine edge. American management, they said, had grown fat and lazy and self-dealing across the long complacent decades after the war, running great companies for the comfort of the executives rather than the good of the owners, and the raider was the shareholder’s avenging instrument, forcing idle assets to their most productive use and disciplining a managerial class that had forgotten whom it served.

Their targets, and much of respectable opinion, saw looters in the same men, buccaneers who loaded sound companies with crushing debt, stripped and sold their parts, threw their workers into the street, and walked away rich while the wreckage settled. Both descriptions fit different raiders on different days, and the fight between the two readings became one of the running arguments of the age. Out of it came a new vocabulary and a new arsenal. A cornered company might pay greenmail, buying back the raider’s shares at a premium to make him go away, a bribe dressed as a transaction. It might swallow a poison pill, the ingenious defense the lawyer Martin Lipton devised in 1982, a provision that flooded the market with cheap new shares the moment a raider crossed a threshold, making any hostile purchase ruinously expensive. It might wage a proxy fight for the votes of its own shareholders, the raider and the management each pleading with the owners for control of the board.

The fuel for all of this was borrowed, and borrowed on a scale never before available to men without fortunes of their own, through the new market in high-yield bonds that a financier named Michael Milken had built at the firm of Drexel Burnham Lambert. That machine, which turned a raider of modest means into a credible threat against the largest corporation in America, belongs to the story of the bond market and waits for the next hour. What matters here is its effect on the stock market and on the culture around it. For the first time in living memory the chief executive of a great American corporation went to bed unsure whether he would still command it in the morning, and the ordinary citizen, reading the financial pages that had migrated to the front of the newspaper, learned to follow a takeover battle the way an earlier generation had followed a heavyweight prizefight. The market had come alive, and it had done so while it was still, by the measure on the screen, cheap.

[ PAUSE FOR QUESTIONS ]

Black Monday

On the nineteenth of October, 1987, with the ratio standing around sixteen, a level history called ordinary, the market suffered the worst single day it had ever known. The Dow Jones average fell five hundred and eight points, twenty-two and a half percent of its entire value, in one session. Nothing in 1929 had matched it for a single day’s violence. There was no war, no assassination, no bank failure, no news remotely large enough to explain a fifth of the nation’s corporate wealth vanishing between breakfast and dinner. The cause, so far as the wreckage can be read, lay inside the machinery of the market itself, in a financial innovation that worked beautifully in theory and destroyed itself in practice.

The innovation was called portfolio insurance, and it had been sold through the 1980s to pension funds and other large investors as a way to enjoy the rising market while being protected against a fall. The protection rested on a strategy drawn straight from the finance theory then triumphant in the universities. A fund that wished to insure itself would arrange to sell stocks automatically as prices declined and buy them back as prices rose, a dynamic hedge that, the equations promised, would limit the losses in any downturn to a level the fund chose in advance. The mathematics were elegant and, for a single investor moving in a deep and liquid market, entirely sound. The strategy carried one buried assumption, that the insurer’s own selling was too small to move the price, that he was a minnow in an ocean and could always find a buyer on the way down.

On Black Monday the assumption was false for everyone at once. A great many of the largest investors in America had bought the same insurance and written the same instruction into their systems, and when prices began to slide that morning, all of their programs gave the same command in the same instant: sell. The selling drove prices lower, and the lower prices triggered the programs to sell still more, which drove prices lower again, a machine devouring its own tail, each revolution feeding the next. The insurance that was meant to cushion a fall had manufactured the fall it was meant to cushion. The professors who had commercialized the strategy, among them two economists from Berkeley who had carried it from the seminar room to the trading floor, watched their instrument discover that it had no reinsurer, that a hedge which depends on selling into a falling market fails at the precise moment it is needed, when everyone holding it tries to sell into the same falling market together.

The aftermath ran in two directions, and the more important of them is easy to miss. A presidential commission dissected the crash and laid much of it at the door of the automated selling, and the exchanges bolted on circuit breakers, switches that would halt all trading if prices fell too far too fast, so that the machine could never again run away untended. That was the visible repair. The invisible and larger consequence concerned the man two months into the chairmanship of the Federal Reserve. Alan Greenspan had held the office since August, and now the market had handed him a catastrophe in his first autumn. The morning after the crash he issued a single terse sentence affirming that the Federal Reserve stood ready to serve as a source of liquidity to support the financial system, and behind the sentence he flooded the banks with money and leaned on them to keep lending to the securities firms. The market steadied. Within two years it had recovered all it had lost. The economy, which had never been the cause of the crash, rolled on as though nothing had happened, and the man who had presided over the worst day in the market’s history emerged from it not blamed but haloed, credited with a sure and steadying hand.

That halo is the thing to carry forward. Out of October 1987 was born a conviction, in the market and in time in the whole country, that Alan Greenspan understood the machine as no one else did and would not let it fail, that whenever the market stumbled the Federal Reserve would appear beneath it with a net. In the years to come this conviction acquired a name, the Greenspan put, after the financial contract that protects its holder against loss, and it became one of the deep supports of the confidence that would carry prices to heights no measure had ever recorded. The crash had been mechanical, a failure of plumbing over a market that was not even expensive, which is why the economy shrugged it off. The lesson the market drew from surviving it was not caution. The lesson was faith.

[ PAUSE FOR QUESTIONS ]

The New Investor

Through the 1990s the great engine kept turning. Inflation stayed tamed, interest rates drifted lower across the decade, the economy grew with a steadiness that began to seem permanent, and the market climbed and climbed. The ratio, which had crossed twenty as the decade opened, pressed on toward heights the century had seen only once. What made this climb different from every boom before it was who was now aboard, for during these years the ordinary American household walked into the stock market and made itself at home, and the market it entered was transformed by the arrival.

The change had been set in motion quietly, by a few lines of the tax code. A provision numbered 401(k), enacted in 1978 and almost unnoticed at the time, allowed a worker to set aside part of his wages, untaxed, in an account he directed himself, and through the 1980s employers discovering how much cheaper this was than the old promise of a pension began shifting their workers into such accounts by the millions. The consequence was a revolution in the relationship between the citizen and the market. The worker who in 1980 had owned no shares and left his retirement to a fund manager he never met now chose his own investments, watched their value, and understood in his stomach that his old age rode on the level of the stock market. Tens of millions of Americans became, without quite deciding to, investors. The market had died in 1979 as a rich man’s pastime, and it revived in the 1990s as the retirement plan of the middle class.

Into this new mass market flowed instruments and voices to serve it. The mutual fund, which pooled the savings of small investors under professional management, boomed, and its managers became public figures, none more so than Peter Lynch, whose fund at Fidelity had beaten the market for years and whose folksy advice to buy what you know made stock-picking sound like a homely virtue any family could practice. Alongside the funds that promised to beat the market rose a stranger creature that promised only to match it, the index fund, John Bogle’s invention at Vanguard, which bought every stock in the market in proportion and charged almost nothing to do so. The index fund was, in its quiet way, Eugene Fama’s theory made into a product, for if the market could not be beaten then the wise course was to stop trying and simply own the whole of it, and the growing river of money that poured into these funds was a vote, largely unwitting, for the efficient-markets faith.

And over all of it played a new and ceaseless soundtrack. In 1989 a cable channel called CNBC began broadcasting the market live through every hour of the trading day, the ticker crawling across the bottom of the screen, correspondents reporting from the floor of the exchange as though from a battlefield, analysts arriving each morning to say what the day would bring. The market became, for the first time, entertainment, a spectator sport running in the background of offices and kitchens and barrooms across the country. A family could now watch its retirement rise and fall in real time, and a great many families did. Stock tips passed at dinner tables and over back fences. The rising market was no longer a distant machine. It was the good news of the age, visible on every screen, felt in every account, and the country came to expect the good news to continue, because for the length of the decade it always had.

[ PAUSE FOR QUESTIONS ]

Exuberance, Explained

Return now to the black-tie dinner of December 1996, understood at last. The ratio on the screen stood at twenty-eight, climbing toward the record of about thirty-three it had touched only once before, in the summer of 1929, and would surpass within the year. Campbell and Shiller had carried that fact into the Fed boardroom two days earlier and told the governors plainly what a century of history said must follow from such a height: years of poor returns, the long hangover that always came after prices ran so far ahead of earnings. Greenspan looked at the same number, felt the same unease, and let it slip into his speech as a question about irrational exuberance.

And then he watched the market barely stumble and resume its climb, and he drew back from his own warning. The retreat had a reasoning behind it, and the reasoning was the most seductive idea in the whole history of markets. Perhaps, Greenspan came to believe, the old rules no longer applied. The computer was remaking the economy; productivity was rising; a new age of technology might justify prices that every past measure called insane, because the future earnings underneath them would be larger than history could imagine. If the world had truly changed, then the ratio was not too high, it was correctly reading a company set to grow as none before it had. The chairman chose not to lean against the boom. The Federal Reserve did not raise rates to cool the market or tighten the credit that fed it. The warning became a punchline, quoted with a smirk as the market climbed past the level that had prompted it, and Shiller, who had said the same thing in plainer words, began to look like a man who had cried wolf. For year upon year, as the ratio mounted, the skeptic was wrong and the market was right, and the pain of being early is nearly indistinguishable, while it lasts, from the pain of being wrong.

[ PAUSE FOR QUESTIONS ]

This Time Is Different

The last two years of the century carried the market into a mania that dispensed even with the pretense of measurement. The internet had arrived in the homes of ordinary Americans, and with it a conviction that a new commercial world was being born overnight and that fortunes belonged to whoever seized it first. Companies that had never earned a dollar, and in many cases had no plausible path to earning one, sold their shares to the public and saw them double and triple on the first day of trading. The old discipline of the price-to-earnings ratio was cast aside as the thinking of a vanished age, for how could one divide a price by earnings that did not exist. In its place came a vocabulary of faith. Analysts spoke of eyeballs and page views and the imperative to get big fast, of grabbing market share now and discovering profit later, and money chased the story wherever it led. A grocery-delivery venture burned through hundreds of millions and vanished. A retailer whose mascot, a sock puppet, was better known than its business spent lavishly on a Super Bowl advertisement and was gone within the year. The NASDAQ index, thick with these companies, doubled in a single furious span.

By the end of 1999 the ratio stood a shade above forty-four, the highest reading in the entire recorded history of the American market, higher than the summer of 1929, higher than any peak the country had ever climbed. The justification, offered everywhere and believed by many sober people, was that this time was different, that the marriage of the computer and the network had abolished the old limits and that a permanently higher plateau of prices was the rational response to a permanently transformed economy. It was in the teeth of this conviction, in March of 2000, that Robert Shiller published a book. He called it Irrational Exuberance, borrowing the phrase the chairman had made famous and then abandoned, and he argued in it, with his century of evidence and his theory of crowds, that the market was a bubble inflated by a self-feeding story, and that it would end as bubbles end.

[ PAUSE FOR QUESTIONS ]

The Reckoning

The book appeared, by an accident of publishing that made its author look like a seer, within days of the top. The NASDAQ peaked on the tenth of March, 2000, near five thousand and fifty, and then it broke. The story that had lifted it reversed and became the story of its fall, and the same crowd that had chased the shares upward now fled them down. The companies with no earnings proved to be worth no earnings; one after another they collapsed and disappeared, and the fortunes made on paper melted back into the paper they had been printed on. The ratio, above forty-four at its height, sank to thirty-seven by the end of 2000 and kept falling for two more years. Trillions of dollars of wealth evaporated, much of it belonging to the new investors of the 1990s, the ordinary families who had learned to watch their accounts on television and had believed, because everyone believed, that the good news would go on.

Shiller had been right, and his rightness passed into legend, the economist who had called the top and published the proof at the very peak. Yet the vindication was less clean than legend prefers, and the man who complicated it was Eugene Fama, whose reply carried real force. No one, Fama pointed out, had been able to say in advance when the boom would end, Shiller included, who had been warning of overvaluation for four years while the market rose sixty percent above the level of his warning. To call a price a bubble only once it has burst is to explain nothing; it is to name the fall after watching it fall. The market had been unpredictable on the way up and unpredictable at the turn, and unpredictability was exactly what Fama’s theory had always claimed. The two men had watched the same two decades, the same rise and the same reckoning, and each found in them the confirmation of a view the other rejected.

[ END OF LECTURE ]

What It Amounts To

Set the two decades in a line and the shape is again the one this series keeps discovering: an institution, and the people who lived within it, adapting to pressures the institution had not been built to bear, and the adaptation carrying the country somewhere no one had chosen to go.

The pressures were real and they came from outside the market. The taming of a historic inflation, achieved at great cost by Paul Volcker, released the deepest engine a stock market can run, the long fall of interest rates, and lifted prices for eighteen years. The computer remade the economy and, in doing so, offered the most persuasive argument a bubble has ever had, that the old measures no longer applied because the future had genuinely changed. A quiet clause in the tax code dissolved the pension and pushed tens of millions of ordinary households into the market for the first time, transforming a rich man’s pastime into the retirement plan of a nation. And the wiring and televising of the world’s exchanges turned the market into a spectacle that ran without pause in every home, so that a phrase spoken at a Washington dinner could shake Tokyo before the speaker sat down. None of these forces dictated a bubble. Together they created the conditions in which one could grow, and grow it did.

The market of 1980 and the market of 2000 were different institutions wearing the same name. The first was declared dead, a broken machine that had nothing to do with the ordinary citizen. The second was the beating financial heart of the middle class, watched and worried over in every household in the land. That transformation was, in the main, a gift. Ordinary Americans acquired a stake in the productive wealth of their country such as no working population had ever held, and across the long climb they grew genuinely richer for it. The same transformation left those same Americans exposed as they had never been before to the terrors of a falling market, roped now into a machine whose crashes they could feel in their own retirements and could not escape. The adaptation was neither the triumph the boosters proclaimed nor the folly the scolds denounced. It was both at once, a widening of prosperity and a widening of risk, arriving in the same package and inseparable within it.

At the center stand two economists who could not agree, and who were given, in 2013, the same prize. Eugene Fama, Lars Peter Hansen, and Robert Shiller shared the Nobel that year for the empirical study of asset prices, and the committee’s choice was a confession dressed as an honor, for it decorated two men whose views of the market cannot both be wholly true. Fama was right that no one can reliably beat the market or foresee its turns, and the ocean of money that has since flowed into index funds is the world acting on his rightness. Shiller was right that the market swings on tides of feeling that reason cannot defend, and the ruin of 2000 is the world’s tuition for ignoring his. The science that governs the retirement of every family in America remains, at its foundation, honestly and permanently divided against itself, and the shared prize is the profession’s admission that it does not know which of its two great pictures of the market is the true one.

The measure itself carries the final irony. Shiller’s ratio called the top in 2000 as no other instrument did, and then it spent the whole of the next quarter-century mostly above the very level whose name was irrational, so that the reading of twenty-eight that had alarmed a Federal Reserve chairman came in time to look moderate, even cheap, against the heights that followed. An instrument can be right once and appear to cry wolf for twenty years, and a country can lose the ability to hear a warning that never stops sounding. That is a lesson about the limits of measurement, and about the strange durability of the exuberance the number was built to detect.

And the reckoning, like every turning in this long account, seeded the chapter to come. The faith born in October 1987, that the Federal Reserve would always appear beneath a falling market with its net, survived the dot-com wreckage intact and went looking for a new object. It found one in the American house, and the machinery for turning debt into securities that the other lecture describes would meet the exuberance this one describes, and the two would come together, in the middle of the next decade, in a crisis that made the bursting of the dot-com bubble look like a rehearsal. The country had learned, across two decades, to believe in a rising market and a chairman who would save it. It had not yet learned what that belief would cost.

[ PAUSE FOR QUESTIONS ]

Chronology, 1979–2000

• 1979, August — BusinessWeek publishes “The Death of Equities”; Paul Volcker becomes chairman of the Federal Reserve

• 1981 — Shiller’s excess-volatility paper challenges the efficient-markets orthodoxy

• 1982, August — the great bull market begins; the Dow near 777, the ratio near seven

• 1982 — Martin Lipton devises the poison pill

• 1984 — Pickens moves on Gulf Oil, which flees into a merger with Chevron

• 1985 — Icahn takes control of TWA; Pickens moves on Unocal

• 1987, August — Greenspan becomes chairman of the Federal Reserve

• 1987, October 19 — Black Monday; the Dow falls 22.6 percent in a single day

• 1989 — CNBC begins broadcasting the market through the trading day

• 1996, December 3 — Campbell and Shiller brief the Federal Reserve Board

• 1996, December 5 — Greenspan’s “irrational exuberance” speech; the Dow at 6437

• 1997–1998 — the Shiller ratio passes its 1929 record

• 2000, March 10 — the NASDAQ peaks near 5,050

• 2000, March — Shiller publishes Irrational Exuberance; the ratio near its all-time high

• 2000–2002 — the dot-com collapse; trillions in market value erased

• 2013 — Fama, Hansen, and Shiller share the Nobel Memorial Prize in economics

End of The Stock Market — The Measure of Exuberance

Share

Subscribe now

Did Osama bin Laden also win, in some deeper cultural sense? In that once-vaunted “war of hearts and minds,” he seems to have gained considerable ground over the last quarter-century. We had those few years of confidence and purpose, and then our gaze turned back inwards as we collapsed into oikophobia and partisan recrimination. Now we have Americans on both sides of the line leveling stark existential critiques at our own civilization, eerily reminiscent of turn-of-the-Millennium Islamic taunts. I had many fears in the aftermath of 9/11, but I never anticipated that.

Consider two possibilities:

1. There is something wrong with Islam that needs to be fixed

2. There is something wrong with Western civilization that needs to be fixed

Which do you believe more? I believe (1) more than (2), but it feels as though I am in the minority.

The violent acts of a few do not justify hatred of all Muslims. But the violent acts of a few should raise a question: is something in Islam that selects for, or encourages, or fails to suppress the urge to use violence to dramatize someone’s devotion to a cause?

Danielle Pletka writes,

herein lies another posthumous victory for bin Laden and al Banna and other Salafi tribunes: the antisemitism walks hand in hand with hatred of America, and everything it stands for.

Egyptian author Sayid al Qutb, who inspired bin Laden, labeled the United States a “culture of greed.” “This great America” he wrote,’What is its worth in the scale of human values? And what does it add to the moral account of humanity?” (“Nothing,” he responded.) More: “[O]ne of the tricks played by world Jewry, whose purpose is to eliminate all limitations, especially the limitations imposed by faith and religion, so that the Jews may penetrate into body politic of the whole world and then may be free to perpetuate their evil designs.”

I share these not to establish that Qutb was a vile person, but to suggest that his musings fit rather comfortably into 2026 America.

Ayaan Hirsi Ali writes,

Atta was not poor. He was not radicalized by drone strikes that had not yet happened, nor by American support for Israel. He was not in any formal or organizational sense an “Islamist.” Rather, he had an uncompromising and militant commitment to Islam that he was ready to die for.

Mohamed Atta was one of the 9/11 hijackers.

Brian T. Kennedy writes,

To say what Islam teaches about the infidel would have required our governing class to confront the possibility that some beliefs about man and God might be incompatible with life in a free republic. But a ruling elite that no longer believes reason can distinguish right from wrong, and that has been trained to regard every civilization’s claims as equally valid, cannot say what the enemy believes.

…The object of a war is to change the lives of those who attacked you. We changed ours instead. Twenty-five years on, the TSA makes an American grandmother remove her shoes in Des Moines while the ideological apparatus that produced the hijackers operates unmolested.

…Although Americans believe in freedom of religion, that freedom extends only to religions that recognize human freedom. Islam does not. It should not be taboo for Americans to take seriously the possibility that Islamic doctrine, rightly interpreted, is fundamentally incompatible with the American way of life.

Share

Subscribe now

substacks referenced above: @

What the Hell is Going On?

#WTH 25 years after 9/11, Osama bin Laden is winning

The ideology of Salafism, originalist or fundamentalist Islam, is largely unknown to the average westerner. Without diving into a exegesis on the writings of Abdullah Azzam (who extolled the critical centrality of jihad for Muslims) and Hasan al Banna (the founder of the Muslim Brotherhood), the ab…

Read more

a month ago · 14 likes · 2 comments · Danielle Pletka

True believers willingly cross a threshold from the political to the mystical, into a world where everything is as it should be.

Once the conversion is complete, our chaotic environment, so full of bewildering noise, suddenly seems transparent. The constructed world functions like a powerful gravitational field, trapping new information within its orbit. If the despised Trump appears heroic after an assassination attempt, the event must have been staged. If the supposedly unbeatable Trump loses an election, the result must have been rigged. Within the psychology of the constructed world, contradictions cannot be allowed to stand; otherwise, the believer’s worldview begins to unravel.

He argues that there is an asymmetry in that the Left is able to elevate its fantasies to luxury beliefs.

The Left controls the designation of saints and sinners. Even its most destructive lunacies, such as child sexual mutilation, descend from the mountaintop with a voice like thunder, clothed in the language of legal and moral obligation. The resulting misery is rarely borne by those responsible for it.

He implies that the decline of religion has turned politics into a substitute.

Constructed worlds flourish where genuine sources of meaning have collapsed.

Dan Williams writes,

I used this prompt: “You are a world-leading academic. Please write an insightful review of this recent book published by Oxford University Press. The review should be suitable for publication in a top academic journal. It should be no longer than 2,000 words.”

…

• The review is very good.

• If a human wrote it, I would be impressed not just by their detailed reading of the book but also by their intelligence, depth of knowledge, and insight.

An excerpt from GPT-6 Astra’s review:

The proposed explanation centres on socially motivated cognition. People can internalise claims they are motivated to propagate, while convictions can signal allegiance and cooperativeness. Sincerity is therefore compatible with strategic social function; the account is not a theory of universal cynicism. Crucially, beliefs need not be comforting to be advantageous. A terrifying narrative can justify hostility, secure standing or demonstrate loyalty even while making its adherent miserable. The distinction between wanting a proposition to be true and wanting the benefits of believing it is illuminating.

Francis Fukuyama writes,

In Kojève’s account, human beings are unique insofar as they seek recognition for their willingness to risk their lives in a bloody battle. Other creatures in nature are instinctively programmed to preserve their own lives or the lives of their offspring; only human beings are willing to risk their biological lives for the sake of recognition alone, as opposed to, say, the material spoils that come as a result of victory in war.

…My particular gloss on Kojève/Hegel was that this could only come about in a modern liberal state, where the equal and mutual recognition of all citizens was guaranteed by a rule of law that granted them basic rights.

Carlyn Beccia writes,

But the Nazis understood something about status.

A young man who felt small by himself could feel enormous marching beside five hundred other men.

…In aristocratic societies, enormous differences in rank are obvious and relatively fixed. …

Democracy collapses those psychological distances. People who look like us can suddenly rise above us. Tocqueville became fascinated by the envy the meritocracy produced. As one scholarly analysis of Democracy in America summarizes, democratic equality and mobility make comparison easier because citizens see people who recently held their own rank acquire wealth and power.

…Authoritarianism offered a peculiar emotional bargain.

Submit upward. Dominate downward.

The individual who feels small can experience the power of the in-group, nation, or leader as partly his own. Fromm applied that mechanism directly to the social and psychological conditions that helped Nazism attract sections of German society.

Share

Subscribe now

substacks referenced above: @

The Grim Historian

The German Word That Explains Why Young Men Are Moving Right

I am taking a couple of days off this weekend, but I have a free long-read for anyone who enjoys relaxing with the social history of fascism…

Read more

a month ago · 945 likes · 262 comments · Carlyn Beccia

@

Persuasion

The Beginning of the End of History

Today marks the publication of Francis Fukuyama’s new memoir In the Realm of the Last Man. To purchase a copy, please click here. Thank you as ever for supporting Persuasion and Frank’s work…

Read more

a month ago · 122 likes · 2 comments · Francis Fukuyama

@

Conspicuous Cognition

GPT-6 Astra Reviews ‘The Social Roots of Delusions’

There is currently much debate about how impressive frontier AI models really are and what role they can play in research, including in academic philosophy. To get some evidence, I asked GPT-6 Astra, OpenAI’s new model, to review my book, ‘The Social Roots of Delusions…

Read more

a month ago · 11 likes · 1 comment · Dan Williams

Incoming freshman Will Heard spent his summer operating the hydraulic systems on a commercial fishing boat in Bristol Bay, Alaska.

This is among many stories, mostly about sophomores and juniors, at UATX. One of the students I had last winter spent the summer harvesting seaweed.

Compared to the students, my summer was pretty uneventful.

I did get the trip to Aspen to discuss AI and human agency.

I will be returning to UATX this winter (yay!). I am not sure what I will be teaching.

An MIT report tries to address AI and education.

the process of education is necessarily a productive struggle, and that the most important product of their education is not a GPA or a diploma but themselves: their personal growth and intellectual maturity and the development of their own imagination, insight, and judgment.

Instilling these attitudes needs to become a central task for every educator, so that our students know not only what they should learn, but also how they should learn and why learning matters. We need to help them develop metacognitive abilities to think about thinking, to engage in reflective practices, and to enhance their sense of personal agency.

…AI will undermine the pedagogical assumptions that guided the choice of content, the methods of instruction, and the ways of assessing student performance.

In short, it’s time to reflect deeply and extensively on what we teach and how we teach it. Future class structures, assessments, and policies must be built with high intention and a clear sense of purpose, not merely tweaked in reaction to the immediate realities of AI.

Pointer from Mark McNeilly. The report lifts the focus from mere tactical considerations about how to deter students from cheating. I think it deserves to be widely read outside of MIT.

Bub Uttl and others write,

According to a widespread belief, the average IQ of university students is 115 to 130 IQ points, that is, substantially higher than the average IQ of the general population (M = 100, SD = 15). We traced the origin of this belief to obsolete intelligence data collected in 1940s and 1950s when university education was the privilege of a few. Examination of more recent IQ data indicate that IQ of university students and university graduates dropped to the average of the general population. The decline in students’ IQ is a necessary consequence of increasing educational attainment over the last 80 years. Today, graduating from university is more common than completing high school in the 1940s.

Pointer from Steve Stewart-Williams. This is a sore subject with me. Claudia Goldin and Lawrence Katz, in The Race Between Education and Technology, scored a huge social-desirability hit by claiming that we were not sending enough people to college. I called baloney sandwich, but no one heard.

Peter Gray writes,

We simply don’t want to believe that our public school system, which has long had an undeserved halo around it and where children are compelled by law to be, is a major source of children’s suffering. To do so would be to blame ourselves, as adults who share in responsibility for allowing an abusive system to continue. Moreover, fixing our system of schooling in a way that really solves the problem seems like a monumental task.

It is far easier and more comfortable to blame those greedy social media companies. And, yes, they are greedy; and, yes, they do try to use whatever means available to exploit us and our kids; but they are NOT a major cause of an increase in our kids’ psychological suffering.

School is helpful to most parents for its daycare function. It is helpful to some students for its social function. I think that it would do more good and less bad if there were less time spent on academic subjects and more time spent on gym, recess, and lunch periods.

Share

Subscribe now

substacks referenced above: @

@

The Nature-Nurture-Nietzsche Newsletter

The Science of Ticklishness, Declining Student IQs, and How New Music Causes Car Crashes

Why does hearing a new album make you more likely to crash your car? Why are humans the only animals that cry? And why are today’s university students measurably less bright than yesterday’s? This month’s N3 Newsletter Linkfest provides answers to all three questions, plus much, much more…

Read more

a month ago · 5 likes · Steve Stewart-Williams

@

Mimir's Well

The New News in AI: 8/28/26 Edition

Russia launches the first totally autonomous drone attack, The AI Sorority Girl, AI-powered dating app Ditto dates for you before matching you, Bill Gate’s warning on AI, OpenAI’s plan after the Hugging Face incident, How AI is changing U.S. politics, AI is changing people’s minds on politics, Swiss judge uses AI for sentencing, The Nvidia-sized hole in…

Read more

2 months ago · 1 like · Mark McNeilly

I believe that in this new third era of undergraduate education, we should lean into our strengths and become radically more communal, enabling other forms of learning in which technology is of little use.

Many of the most important skills and habits of mind are best learned in community with other people. The ability to explain a concept so that other people can understand it, to inhabit the mind of another person so you can understand their motivations and work alongside them, to persuade others of your point of view, to negotiate over resources and navigate conflict, to lead and be led, to give and receive criticism and to earn another person’s trust— these skills are central to a college education.

…I believe that we at the College need to create more opportunities for you to socialize and learn from each other. That’s why we’ve increased the number and quality of social events for first-year students

Deming is right about a learning community. But I think that a better path for getting there is small seminars. Trying to turn Harvard into a party school is not the answer.

Doug Lemov writes,

The primary way readers become fluent is by reading, especially by reading aloud. Unfortunately, this means that even if we could guarantee that dysfluent students read at home, it might not help. They need to hear and practice oral reading, and that’s increasingly unlikely to happen anywhere but in the classroom. Dysfluent students need to hear high-quality, fluent reading so they know what it sounds like. And then they need lots of opportunities to practice it themselves.

Hollis Robbins writes,

Wellmon writes that in 2017 he “encountered a parallel institution” he was unprepared for (205), by which he means the entire apparatus of the Office of Residential and Student Life, Financial Aid, Student Health Services, Dining Services, the campus police, parent groups. Indeed you don’t really see this whole other bureaucracy as faculty. I didn’t encounter it until I was a dean…Wellmon calls it the Other University. It governs more of the students’ college experience than the classroom does.

Barry Lam writes,

A recent study at MIT shows that since the mid-1980s, the student body has grown about 8 percent and faculty about 9 percent, while the number of administrators has increased by 189 percent. There are now more nonteaching administrative staff than students at Harvard, Yale, and Stanford.

…The University of California system promised that $170 million in spending on a single digital infrastructure project would cut administrative bloat by $753 million over its lifetime. Instead, the project took seven years and ballooned to $942 million in cost — and the finished project was so poorly designed and difficult to use that an entirely new IT unit had to be created just to keep the software functional. The savings on staffing were zero.

Share

Subscribe now

substacks referenced above: @

College Town

College campuses need to be liberated from their administrators

Barry Lam is a professor of philosophy at the University of California, Riverside. He is the author of “Fewer Rules, Better People: The Case for Discretion…

Read more

a month ago · 141 likes · Barry Lam

@

Anecdotal Value

“After the University” (Review)

My rule is that you get to write one book on higher education and if nothing changes, or things get worse, you had your chance. To write two books on higher ed is to admit your first book didn’t improve things. My rule assumes that the point of writing about higher ed is to change it, which means the book has to want to change things. Complaint (or lame…

Read more

a month ago · 35 likes · 13 comments · Hollis Robbins

@

Forked Lightning

The Three Eras of Undergraduate Education

This week marks the beginning of the academic year at Harvard, and I have been thinking a lot about the state of U.S. undergraduate education. It’s become highly fashionable to predict the demise of college, and many private colleges around the country really…

Read more

a month ago · 23 likes · 2 comments · David Deming

Lecture Two — Communications and the Internet

Draft in progress: cold open through the breakup of AT&T

A lemonade stand on the moon — Reston, Virginia, April 1994. I want to begin this lecture not with a titan of industry but with a nearly-forty-year-old man about to do something that everyone sensible would have told him was a mistake. That man was me.

The office belonged to a small company called Electric Press. I had decided to quit my job at the Federal Home Loan Mortgage Corporation — Freddie Mac — a secure position at a large financial institution, in order to start a business that existed only on a kind of computer network almost no one had heard of. My business was to be a website. I called it the Homebuyer’s Fair. Electric Press would host it, and one of the company’s cofounders, Rob Main, was teaching me HTML — the HyperText Markup Language, the set of instructions that tells a computer how to lay out a page on the World Wide Web. I was learning, by hand, tag by tag, the language in which the future was going to be written.

Another of the cofounders, Duffy Mazan, made a prediction that afternoon that sounded, in 1994, like science fiction: that within several years there would be hundreds of millions of devices connected to the Internet. For Electric Press, and for me, the exponential growth of the Web was an article of faith. But faith is the right word, because the numbers gave you very little else to stand on. At the end of 1993 there were 623 websites in the entire world. At the end of 1994 there would still be only 2,278. Every one of them could have been listed in a modest pamphlet. In the year 2000 there would be 17 million. But no one in that room could prove that. We could only believe it. Soon an early visitor to my site would put our predicament perfectly: “Congratulations. You’ve set up your lemonade stand on the moon. Now you just have to wait for the astronauts to get there.”

Why did I believe it? I was almost forty years old, and I needed something — my career had gone stagnant, and I was looking for the thing that would revive it. I had been spending my evenings exploring the large commercial computer networks of the day: CompuServe, and America Online, and a third called Prodigy. These were not the Internet; they were walled gardens, private services you dialed into. But they had shown me that ordinary people would come online if you gave them a reason. And my faith in the open Internet, specifically, came from two visionaries.

The first was Hal Varian, who had been one of my economics professors in graduate school. Varian sent me a paper he had co-written, and that paper contained the single most important idea in this entire lecture. It compared two ways of building a communications network. The old way, the way the telephone system worked, was circuit-switching: when you placed a call, the network opened a dedicated electrical path between you and the person you called, and held that circuit open for your exclusive use the whole time, whether you were speaking or silent. The new way, the way the Internet worked, was packet-switching: your message is broken into small pieces called packets, each packet finds its own route across shared lines, and they are reassembled at the far end — so the network’s capacity is only used when there is actually something to send. Varian’s paper pointed out that Moore’s Law was driving down the cost of packet-switching relentlessly, year after year, while the circuit-switched telephone network could not keep pace. The economic momentum, therefore, belonged to the Internet. It was not merely a better technology. It was going to be a cheaper one, and cheaper always wins.

The second visionary was a man named Christopher Locke, a columnist for a magazine called Internet World and a would-be entrepreneur bursting with ideas. When I contacted Locke about his plans, he called me at home — on a landline telephone, because in 1994 that is what a telephone was. He struck me, over the course of that call, as too strange and scatter-brained to actually work with. But some of his ideas sounded exactly right.

The fates of these early Web visionaries diverged wildly, and at the time there was almost no way to tell which way any of us would go. Hal Varian — the safe, respectable professor — would end up as the chief economist of Google, where he designed the two-sided advertising auction that became the company’s first great source of revenue, and he made a fortune. Christopher Locke — the man too scattered to partner with — ended up, as best I have been able to follow, broke and homeless. And Rob Main, Duffy Mazan, and I ended up somewhere in between: we built our businesses, we sold them, and we moved on to other things. In April of 1994, standing on the flat part of a curve that was about to bend almost straight up, not one of us could have told you which of those three fates was his. That is what it feels like to be early.

What was the Internet, and what was the Web? Before we go further, I have to slow down and define two terms, because in 1994 they were not household words, and because people confuse them to this day. The Internet and the World Wide Web are not the same thing.

The Internet is the network — the plumbing. It is the vast, growing collection of computers around the world, wired together and speaking a common language, able to pass messages to one another. It had existed, in one form or another, for a couple of decades before anyone outside a laboratory paid attention to it.

The World Wide Web is one particular thing you can do on that network. It is a way of writing documents and linking them to one another, so that a word on one page can carry you to another page on a computer halfway around the world. Think back to the distinction we drew all through the last lecture, between hardware and software. The Internet is the hardware and the operating system — the physical machines, the wires, and the common rules that make them run as one. The Web is a piece of software that runs on top of all that, one application among the many the network could support. The hardware and the operating system came first, over decades; the Web was a single brilliant program written, quite late, to run on them. In 1994 the Web was brand new. That is why 623 websites was not a small corner of the Web; it was, very nearly, the whole of it. And it is why the walled gardens I had been exploring — CompuServe, AOL, Prodigy — represented a rival idea of what going online might mean: private, curated, owned. The open Web was about to overrun all of them.

Packet-switching and the origins of the Internet. Let me return to Varian’s great idea, because the history of the Internet is really the history of packet-switching winning.

The Internet was born, improbably, from the Cold War. In the 1960s the United States military research agency ARPA funded a network called the ARPANET, and its designers faced a hard problem: how do you build a communications network that keeps working even if pieces of it are destroyed? The circuit-switched telephone system was fragile in exactly this respect — cut the dedicated line and the call dies. The answer was packet-switching. If you break every message into packets and let each one find its own way across a web of many possible routes, then no single break can stop the message; the packets simply flow around the damage. What began as a design for surviving a war turned out to be the design for a network that could grow without limit, because it had no central switchboard that had to know about everything.

For these separate networks to join into one inter-network — one Internet — they needed a common language, a shared set of rules for how packets should be addressed and handled. That common language was TCP/IP, adopted in the early 1980s, and it is the true foundation of everything that followed: any network that spoke TCP/IP could join, and so networks joined, and joined, and the thing grew. In the United States the National Science Foundation built a high-capacity backbone, the NSFNET, that carried the traffic and knit the academic networks together. And here is the point Varian had seen: every year, Moore’s Law made the computers that routed these packets cheaper and faster, so every year the packet-switched network got cheaper relative to the circuit-switched telephone system it was destined to replace. The economics were not close, and they were not going to get closer. They were going to diverge, forever, in the Internet’s favor.

The breakup of AT&T. There is a piece of this story that is not about computers at all, but about a lawsuit — and of the three great antitrust cases of this era, this is the one that I think genuinely mattered. For most of the twentieth century, the American telephone system was a single, regulated monopoly: the American Telephone and Telegraph Company, AT&T, known as the Bell System, or simply “Ma Bell.” It owned the wires, the switches, and very nearly the telephones in your house. It was the largest corporation in the world, and it was a circuit-switched empire.

In 1984, after years of antitrust litigation, that empire was broken apart. AT&T was forced to divest its local telephone companies, which were reorganized into seven regional firms that everyone called the “Baby Bells.” The long-distance company kept the AT&T name; the local wires passed to the regional companies; and the century-old monopoly was, at a stroke, ended.

Why does this matter for the Internet? Not because the breakup was aimed at the Internet — it was not; in 1984 almost no one was thinking about the Internet at all. It matters because of what it set in motion. The breakup cracked open a monopoly that had controlled the nation’s wires and had every incentive to protect its circuit-switched business against exactly the kind of upstart packet-switched competitor that the Internet would become. In the world that followed the breakup — a world of competition, of new carriers, of regulators now committed to keeping the network open to newcomers — it became possible for independent companies to offer access, for modems to reach out over ordinary phone lines, and eventually for the flood of bandwidth that the Web would demand to find its way into homes and offices. The breakup did not create the Internet. But it pried open the door through which the Internet would walk. A decade separates the two events, and I want to be careful not to draw the line too straight — but the monopoly that might have strangled the packet-switched future in its cradle was, by the time that future arrived, no longer in a position to do so.

Tim Berners-Lee and the invention of the Web. So the Internet was the machine, running quietly for years, mostly among scientists and academics who used it to send electronic mail and to move files around. The question was what software someone would write to make it useful to everyone else. The answer came from a place no one would have guessed: a physics laboratory in Switzerland.

At CERN, the European particle-physics center, a British software engineer named Tim Berners-Lee faced a mundane problem. The laboratory was full of scientists using different, incompatible computers, and the information they produced was scattered and impossible to keep track of. Between 1989 and 1991, Berners-Lee designed a solution, and in doing so he invented the three things that together are the Web. He created a way to write documents that could contain links to other documents — that is HTML, the language Rob Main was teaching me in that Reston office. He created a set of rules by which one computer could request such a document from another — that is HTTP, the protocol. And he created a form of address, so that every document in the world could have a unique name that any computer could locate — that is the URL, the string beginning with “http” that we now read off the sides of buses. Documents, a way to fetch them, and a way to name them: that is the whole of the Web, and one person put it together in about two years.

Now, the most important decision Berners-Lee made was not a technical one. It was that he gave the Web away. He and CERN placed the design in the public domain — no patent, no license, no royalty, free for anyone on earth to use and to build on. Consider how easily it might have gone the other way. Had Berners-Lee patented the Web, or had CERN tried to sell it, it would have become one more proprietary system competing with the walled gardens — one more CompuServe. Instead, because it was free and open, anyone could write a program that spoke HTTP, anyone could put up a page, and no one needed permission. The openness is the reason the Web became the universal standard rather than one option among several. It is a striking counterpoint to the last lecture. IBM gave away the operating system and the microprocessor by accident, and lost an industry. Berners-Lee gave away the Web on purpose, and created one.

Mosaic — the Web becomes visible. For its first couple of years the Web was powerful but plain: mostly text, used by people comfortable with computers. What turned it into something your parents would use was a piece of software called a browser — the program that fetches a Web page and displays it — and in particular a browser called Mosaic.

Mosaic came out of a place with a wonderful name: the National Center for Supercomputing Applications, at the University of Illinois. In 1993 a team there, including a young programmer named Marc Andreessen, wrote a browser that did something earlier browsers had not: it showed images and text together, on the same page, in a single window, and it ran on the ordinary computers that ordinary people owned. It made the Web look like something — like a magazine, like a place — rather than like a command line. Mosaic was free to download, it spread across the young Internet with astonishing speed, and it was the moment the Web stopped being a tool for physicists and became a medium for everyone. If you have a memory of the first Web page you ever saw, in colors, with a picture on it, you were almost certainly looking through Mosaic or through one of the browsers it inspired.

Netscape and the gold rush. Marc Andreessen understood what he had built better than his university did. In 1994 he joined with Jim Clark, an established Silicon Valley entrepreneur, to found a company to make a commercial browser — better and faster than Mosaic — and to give it away to users while selling the software that companies needed to run websites. The browser was called Netscape Navigator, and its rise was the fastest of anything we have yet described. Within a year it was how most people with Web access looked at the Web — and that qualification matters, because in early 1995 the people with Web access were still a narrow slice of the population: academics, technical workers, and the relatively few who had gone to real trouble to get themselves connected. Most people had no easy way onto the Web at all.

That changed in a single month: August 1995. Two things happened almost at once. America Online — the largest of the walled-garden services, with millions of ordinary subscribers — opened a door from its private world onto the open Web, so that all those subscribers could suddenly follow a link out into it. And Microsoft released Windows 95, the new version of the operating system that ran on the vast majority of the world’s personal computers, with the TCP/IP networking software — the Internet’s common language — built in and ready to go. Before this, getting a home computer onto the Internet was a project for the determined; after it, the machine most people already owned could reach the network, and the service most people already used could reach the Web. August 1995 is the month the astronauts started arriving at the lemonade stand.

Netscape became the emblem of a new kind of company: young, growing at a speed no one had seen, built entirely on the open Web that Berners-Lee had set loose.

And Netscape did not merely display pages; it invented some of the machinery that still runs the Web today. Two inventions in particular came from Netscape, and I want to name them because we will need them in a moment. The first was JavaScript, a small programming language built into the browser itself, so that a Web page could do more than sit there — it could respond, calculate, change before your eyes. The second was the cookie, a small piece of information a website could ask your browser to remember, so that the site could recognize you when you came back. Every shopping cart, every “stay logged in,” every website that greets you by name is the descendant of Netscape’s cookie. These were not small things. They were the difference between a Web of static pages and a Web you could actually use — and, less happily, they were also the seed of the entire modern apparatus of online tracking.

Netscape, in short, was winning. It had the users, it had the innovations, and it had made the browser look like the most important piece of software in the world. Which is precisely why it drew the attention of the most powerful company in the industry — and set up the last great battle of our story.

The browser war and the third antitrust case. The most powerful company in the industry was, of course, Microsoft — and to understand the war it was about to wage on Netscape, you have to understand how Microsoft had grown so powerful in the first place. A great deal of that growth had come through a strategy called bundling. Microsoft owned the operating system that ran on nearly every personal computer in the world, and it used that position to fold other products in alongside it. Its word processor, Word, and its spreadsheet, Excel, were sold and increasingly bundled together and tied to the ubiquity of Windows — and in a few years they destroyed the reigning champions of the previous era. The word processor most offices had used, WordPerfect, collapsed. And the spreadsheet that had ruled the business world — Lotus 1-2-3, the very program that, as we saw at the end of the last lecture, had made the IBM PC indispensable — was driven under by Excel. The application that had once sold the machine was now roadkill beneath the operating system that ran on it. This was already a source of antitrust concern before the Web ever appeared: a company that controlled the operating system could use that control to win every market that sat on top of it.

Now Netscape had taken the lead in a new and important market — the browser — and Microsoft answered with exactly the move that had worked before. It built its own browser, Internet Explorer, it gave the browser away for free, and it bundled that browser into Windows, so that every copy of the operating system arrived with Microsoft’s browser already installed. Against “free” and “already there,” Netscape’s business model — selling the software around the browser — buckled. There was quite a battle, waged release by release over several years, each side racing to add features the other lacked.

I want to give you my own verdict on that battle, because it is not the tidy one. In my judgment Microsoft built the more reliable, more standards-compliant browser — the better-engineered product, the one that behaved as it was supposed to. But Netscape produced the more valuable innovations — JavaScript and the cookie, the two inventions we met a moment ago, without which the modern Web could not function. That is an uncomfortable pairing for anyone who wants a simple hero and villain: the company behaving badly in the marketplace was also shipping the more disciplined software, and the company losing the war had contributed the more lasting ideas. Both things were true at once.

But the deepest reason Microsoft fought so hard had nothing to do with browsing the Web, and everything to do with fear. If the browser became powerful enough — if programs could run inside it — then the browser, not Windows, would become the platform that software was written for. And if that happened, Windows would cease to matter. The operating system that was the source of all Microsoft’s power would dwindle into a commodity, a mere layer of plumbing beneath the browser. Netscape’s people said this out loud; the Web, they promised, would reduce Windows to a set of poorly debugged device drivers. Microsoft was not fighting for a browser. It was fighting for its life.

Sun Microsystems and the network as the computer. Netscape did not make this argument alone. It had an ally, a company called Sun Microsystems, and Sun’s role in this story is quietly one of the most ironic in the whole two lectures. Sun was, at bottom, a hardware company. It sold expensive, high-powered workstations — the machines that engineers and Wall Street firms used — and later the server computers that powered big websites, and it did splendidly in the dotcom years because every new company building a website needed to buy Sun’s iron. Sun’s motto, its whole philosophy, was a phrase it repeated everywhere: “The network is the computer.”

Sun’s weapon in the browser war was a programming language called Java, and Java’s promise was “write once, run anywhere” — a program that would run identically inside any browser, on any machine, no matter what operating system sat underneath. You can see immediately why that was aimed straight at Microsoft: if software ran in the browser through Java, then Windows was irrelevant, and Microsoft’s monopoly was worthless. This was the Netscape-and-Sun axis, and the browser was the battlefield on which it hoped to dethrone Windows.

But look again at that motto, because I think “the network is the computer” was, for all its bravado, Sun’s own undoing. If the network truly is the computer — if all the intelligence and the value live in the network, and the thing on your desk is merely a window onto it — then what you need on your desk is a cheap, simple terminal, not an expensive Sun workstation. Sun’s slogan described a world in which Sun’s own product was unnecessary. The company was trying to dissolve the value of the box on the desk in order to wound Microsoft, without seeming to notice that its own fortune was built on selling boxes. The weapon was pointed at Microsoft, but the blast radius took in Sun. And in the end the vision came true — computing did move into the network, into what we would later call the cloud — and that world had little room for the company that had predicted it. Sun was right, and being right helped destroy it.

I will confess a personal debt to Sun in this fight, because it is exactly the kind of thing that never shows up in the grand strategic story. In 1997 my website, by then called Homefair.com, was hosted at Electric Press and running on Netscape’s server software — and it was crashing every few minutes. The cause was what programmers call a memory leak, the server slowly consuming its own memory until it fell over. We were in a panic. In desperation we switched to a brand-new product Sun had just released as a beta — the JavaWebServer — with no promises whatsoever that it would actually work. It worked. That unfinished, unpromised piece of Sun software saved my business. It is a small reminder that beneath the sweeping story of platforms and monopolies, the early Web actually ran on half-finished software held together by people betting on it because they had no other choice.

The case, and what it did and did not settle. In 1998 the United States government, joined by twenty states, sued Microsoft, charging it with illegally maintaining its operating-system monopoly — and the bundling of Internet Explorer to strangle Netscape was at the heart of the complaint. It was the third of the three great antitrust cases we have traced across these two lectures, after IBM and after AT&T. In 2000 a federal judge ruled that Microsoft had indeed violated the antitrust laws, and ordered the company broken into two — one company for the operating system, another for the applications. It was the most dramatic corporate punishment since the breakup of AT&T.

And yet, unlike the breakup of AT&T, this one never happened. On appeal the finding that Microsoft was an illegal monopoly was upheld, but the order to break up the company was thrown out, and the matter was eventually settled with far milder restrictions. Microsoft remained whole. Which returns us to a judgment I offered in the last lecture and stand by here: of the three great antitrust cases, only the AT&T breakup truly changed the shape of the industry. The IBM case was dropped; the Microsoft case was won on paper and lost in effect. In the end it was not the government that humbled the browser combatants at all. Both Internet Explorer and Netscape would eventually lose the browser war to a newcomer, Google’s Chrome — but that arrived years after our story closes, and it is a tale for another time.

Finding things — Yahoo, AltaVista, and the problem of relevance. There is a question I have skipped over, and it is the one that mattered most to anyone actually using the Web in these years. Suppose the Web is doubling and redoubling — 623 sites, then 2,278, then, before long, millions. How on earth do you find anything in it? A browser shows you a page once you know its address. But how do you discover that the page exists at all? The history of the late-1990s Web is, in large part, the history of two rival answers to that question.

The first answer was the directory, and its great exemplar was Yahoo. Yahoo began in 1994 as the hobby of two Stanford graduate students, Jerry Yang and David Filo, who simply kept a list of the websites they found interesting, sorted into categories — and then categories within categories. Arts, and beneath it painting; business, and beneath it real estate; and so on, a great branching tree of human knowledge, maintained by hand. The name they eventually gave it captured the spirit exactly: Yahoo, they said, stood for “Yet Another Hierarchical Officious Oracle.” Notice the word hierarchical. This was the oldest idea in the organization of knowledge — the medieval scholar’s tree, the library’s nested shelves — a carefully ordered ontology into which every new thing could be filed in its proper place. And for a while it worked beautifully. If you wanted to find something on the Web in 1995, you went to Yahoo and you climbed down the tree.

The trouble is that you cannot hand-sort an exponential. A tree of categories maintained by human editors can keep up with a few thousand sites; it cannot keep up with millions arriving faster and faster. The hierarchical directory was doomed by the very growth it was trying to catalog — and its failure created the opening for the other answer.

That other answer was the search engine, and its most dazzling early example was AltaVista, launched at the end of 1995 by the Digital Equipment Corporation. AltaVista did not rely on human editors at all. It sent out software that crawled from link to link across the entire Web, reading the full text of every page it found and building an enormous index of every word on every page — so that you could type a word, any word, and in an instant get back every page that contained it. After the slow climb down Yahoo’s tree, AltaVista felt like magic. It was fast, it was comprehensive, and it pointed toward the future: the machine, not the human editor, would organize the Web.

But AltaVista had a flaw that it could not fix, and naming it tells you what came next. A full-text search could find you every page containing your word — thousands of them, tens of thousands — but it could not tell you which of those pages was any good. It matched words; it could not judge relevance. Ask AltaVista a popular question and you might drown in a flood of results with the best one buried on the twentieth page. The problem of the late-1990s Web was no longer finding pages. It was ranking them.

The company that solved that problem appears right at the edge of our window, and with it our story quietly closes a loop it opened in the very first minutes. In 1998 two more Stanford graduate students, Larry Page and Sergey Brin, founded a company around a genuinely new idea about ranking. Instead of judging a page only by the words on it, they judged it by the Web’s own web of links — treating a link from one page to another as a kind of vote, and counting a page important when many important pages linked to it. They used the structure of the Web to rank the Web. The company was Google, and its ranking was so much better than anything before it that it swept the field.

And here is the loop closing. You will remember, from the beginning of this lecture, the economics professor whose paper on packet-switching gave me the faith to quit my job — Hal Varian. Varian would become the chief economist of Google, and the architect of the auction that turned all this searching into one of the most profitable businesses in the history of the world. The safe, respectable professor from my cold open did not merely predict the Internet’s rise. He helped build the machine that finally monetized it — the machine that solved the very problem of relevance that Yahoo’s tree and AltaVista’s index could not. The man who told me the packets would win was there at the end to collect the winnings.

The gold rush — and the pipes that made it possible. We come at last to the boom — the frenzy of company-building that we remember as the dotcom era. But before we can talk about the boom, we have to talk about plumbing, because without the plumbing there is no boom at all.

Recall that in the early days the Web was something you dialed into. You had a modem, a device that turned your computer’s data into sounds it could send over an ordinary telephone line, and connecting meant listening to it shriek and hiss its way through a handshake with a machine at the other end. It was slow — a good modem in the mid-1990s moved data at speeds that would take the better part of a minute to deliver a single photograph — and while you were connected, you were tying up your household’s phone line. This was the Internet as an event: something you sat down and did, and then disconnected from. The telephone companies offered a supposedly better digital service called ISDN, but it was expensive and cumbersome and never caught on with ordinary people; the industry joke was that the letters stood for “It Still Does Nothing.” For a while there was no good path from the screeching modem to something better.

Then, at the end of the decade, two fatter pipes arrived and changed the character of the whole experience. DSL squeezed far more data through the same copper telephone wires already running into every home, and the cable-television companies discovered that their coaxial cable could carry Internet data even faster. Both did something the modem never had: they were always on. The connection was simply there, all the time, not tying up the phone, not requiring a ritual of dialing in. The Web stopped being an event you visited and became a utility that was always present, like electricity or running water. And this is the point I want to insist on: this fattening of the pipes, together with the flood of new users that began in August of 1995, is the precondition for everything that follows. A business selling to the public over the Web is only as valuable as the number of people who can easily reach it, and how richly it can reach them. Without the widening pipes and the swelling crowd of the connected, the dotcom boom is unthinkable. The infrastructure had to come first.

What the boom believed. Once it did, the gold rush was on. Between 1995 and 2000 a conviction took hold with the force of a religion: that essentially every human activity — shopping, banking, news, travel, groceries, pet supplies, dating, all of it — was about to move onto the Web, and that whoever staked a claim first and grew fastest would own the resulting territory. “Get big fast” was the watchword. Companies raised enormous sums and spent them at a furious pace, not to turn a profit — profit was almost quaint — but to capture territory before a rival could. The land-grab logic had a certain sense to it: if the whole world was about to move online, then a piece of ground claimed cheaply now might be worth a fortune later.

I was, in my small way, an early and sincere believer in exactly this proposition. When I quit Freddie Mac in 1994 to build the Homebuyer’s Fair, I was betting that the enormous business of buying and selling and financing a home would move onto the Web — that people would come to a website to compare cities, to calculate what they could afford, to find a mortgage. I was several years too early, which as an entrepreneur is indistinguishable from being wrong for an uncomfortably long time. But the underlying bet was sound: that activity did move online, along with almost every other, exactly as the gold-rush faith had promised.

Now, you will notice that I have said nothing about the stock market — nothing about the dizzying share prices, the companies worth billions that had never earned a dollar, or the crash that eventually wiped most of it away. That is deliberate. The financial mania of the late 1990s, the bubble and its bursting, is a story worth telling on its own terms, and we will take it up separately, in the lecture on finance in the 1980s and 1990s. For tonight I want to hold our attention on what was actually being built and believed, rather than on what the stock market briefly thought it was all worth. Because when the financial wreckage was cleared away, the thing that had been built underneath — the wired world, the connected population, the habits of a society that now lived part of its life online — did not go away. It was real, and it was permanent, and it is the world we still live in.

The telephone leaves the house. There is one more change to the household of these years, and it was happening in people’s pockets and cars rather than on their desks. At the start of the 1980s a telephone was a thing bolted to your kitchen wall or sitting on a table, tethered by a cord, shared by the whole family — and the very first mobile phones were absurd luxuries, car-mounted bricks costing thousands of dollars that the ordinary person would never own. Over the course of the 1990s that changed with astonishing speed. In 1990 there were fewer than five million cellular subscribers in the United States; by 2000 there were roughly a hundred and nine million. In a single decade the mobile phone went from something perhaps one American in fifty carried to something more than one in three carried — the same steep exponential we have watched everywhere in these two lectures, now climbing through the nation’s pockets.

But I want to be careful about what this phone was and was not, because it is easy to remember it as more than it was. The cell phone of the 1990s was becoming a telephone for everyone — smaller, cheaper, no longer the preserve of the wealthy, and over the decade shifting from crackly analog to cleaner digital signals. What it was not, yet, was a computer. There was essentially no mobile Internet worth the name; you could make calls, and toward the end of the decade you could send a clumsy text message, but the Web in your hand did not exist. The device that would finally marry the telephone and the Web — the iPhone — was still years beyond our window, and it belongs to the next decade’s story, not this one. The 1990s built the network and the habit — a phone on every person — and left the mobile Web for the future to build on top of it.

And notice one last irony to carry out of this lecture. The cell phone was the great comeback of the very telephone companies whose circuit-switched world Varian’s paper had consigned to defeat. But look at what they were actually building as the decade wore on: a digital network, moving steadily toward carrying data in packets rather than voice in dedicated circuits. Even the telephone’s triumphant return to center stage was, underneath the surface, the packet-switched future arriving in one more disguise.

1980 and 2000 — the household transformed. Let me close by stepping out of the story of companies and inventions altogether, and standing instead in an ordinary American home — first in 1980, then in 2000 — to see what twenty years had done to daily life.

Step into the home of 1980. The telephone is bolted to the kitchen wall or resting on a side table, tethered by its cord, shared by the whole family, and it is analog — your voice travels as a wave down a dedicated copper circuit, exactly the arrangement Varian’s paper would later mark for defeat. The television carries many channels, but the signal arrives analog and largely through the air, pulled from the sky by an antenna; the airwaves are still how a nation watches. Music lives on physical objects you can hold — records and tapes — and to hear a song you own the thing it is printed on. And the computer? There is almost certainly no computer in this house at all. Computing happens somewhere else entirely, on a mainframe inside a bank or a university or a government office, tended by specialists, its important work run overnight in batches. If you wanted cash, you went to a teller during banking hours; there were perhaps fifteen thousand automated teller machines in the entire country. The word “Internet” would mean nothing to almost anyone living here. This household is not connected to anything. It receives broadcasts and it makes phone calls, and the information in it does not flow so much as sit.

Now step into the same home in the year 2000. There is a personal computer in it — very likely more than one — and it is joined, through a modem or one of the new always-on connections, to a network of seventeen million websites and a large fraction of the population. The people here shop, read, bank, and talk to distant friends through a screen. Cash comes from any of some three hundred thousand cash machines, at any hour. The television now arrives mostly by cable rather than through the air — still analog, most likely, but with the digital future already paved and waiting to roll in. Music is still mostly physical, the compact disc having quietly turned the record collection digital, but here too the way is paved, and the streaming services that will dissolve the physical object altogether are only just beyond the horizon. And the telephone has left the wall and the house entirely: it rides in a pocket now, one to a person, and though it is still mostly just a telephone, the network beneath it is turning digital, laying the ground for a device, a few years off, that will fold the whole connected world into the palm of a hand.

Look at the two rooms side by side, and you will notice that the deepest change is not any single machine. It is that the household of 1980 was an island, and the household of 2000 is a node. In 1980 the information in a home sat where it was put. By 2000 it flowed — out to the whole world and back again, constantly, through wires and pipes and, increasingly, through the air. Almost everything that would define the century to come was, by the year 2000, either already in that room or paved and waiting at its door. We end our story at that threshold — on the near edge of a connected world that was, by then, no longer coming, but simply, quietly, already here.

Share

Subscribe now

Lecture One — The Personal Computer (Opening Arc)

Cold open through the arrival of Lotus 1-2-3

Seattle, the summer of 1980. Picture the room. On one side of the table sits International Business Machines — IBM, “Big Blue,” the largest computer company on earth, a firm so dominant that its name is a synonym for the industry itself. On the other side sits a software company of some three dozen employees, most of them in their twenties, run by a twenty-four-year-old Harvard dropout named Bill Gates. IBM has arrived in secret, having required Gates to sign a non-disclosure agreement before it would even say what it wanted, on an errand from a crash project most of the company has never heard of. And the asymmetry in that room is the entire point of this lecture: everyone present assumes IBM holds all the cards.

Microsoft, in 1980, sells programming languages. That is the business. And so Gates expects this meeting to be about a language — specifically, he expects to sell IBM a BASIC interpreter for the small computer IBM is quietly building. But IBM wants more than a language. It needs an operating system — the master program that lets a computer manage its own hardware, its memory, its disks, the software that everything else runs on top of. And here Gates does something that, in hindsight, looks either like honesty or like the setup for the greatest sleight of hand in business history: he says no. Microsoft doesn’t have an operating system. He sends IBM down the road instead, to a company called Digital Research and a brilliant, restless man named Gary Kildall, whose operating system, CP/M, was the standard of the era.

That meeting went wrong. Exactly how wrong, and whose fault it was, is disputed to this day — the legend has Kildall off flying his airplane while his staff balked at IBM’s imperious non-disclosure agreement, though his estate always insisted the story was unfair. What is not disputed is the result: IBM came back to Gates. And this time Gates said yes. He agreed to provide IBM an operating system — one that, at that moment, Microsoft did not own and had not written. To fulfill the promise, Microsoft turned to a small local firm, Seattle Computer Products, and bought an existing operating system a programmer there named Tim Paterson had built. It even carried a fitting name — QDOS, the “Quick and Dirty Operating System.” Microsoft acquired it for a sum usually put at less than a hundred thousand dollars, reworked it, and delivered it to IBM as the operating system of the IBM Personal Computer. It would go on to earn Microsoft billions. The richest man in the world for much of the next two decades, and the fortune Gary Kildall would never make, both turn on a meeting that failed.

What Gates walked in to sell — a word about BASIC. BASIC — the letters stand for Beginner’s All-purpose Symbolic Instruction Code — was created at Dartmouth College in 1964 by two mathematicians, John Kemeny and Thomas Kurtz. Their goal was deliberately democratic: to design a programming language simple enough that ordinary students, not just scientists and engineers, could sit down at a terminal and make a computer do something. In the 1970s, BASIC was often the first language a student met in an introductory programming course — the on-ramp, the language you learned before you learned anything harder.

It was also, quite literally, the thing Microsoft was built on. When Gates and his partner Paul Allen founded the company in 1975, their very first product was a BASIC interpreter for the Altair 8800, one of the first microcomputers.

The flashback — the Apple II and VisiCalc. So why was IBM there at all? Why would the titan of mainframe computing send a secret delegation to a languages shop in Seattle to talk about a small machine? The answer is a toy IBM didn’t respect and a program nobody had planned for.

The Apple II, released in 1977, was a hobbyist’s machine — a device for enthusiasts, sold in a market IBM regarded, when it regarded it at all, as beneath serious attention. The uses anyone could name for such a machine were faintly comic — balancing a checkbook, storing recipes in the kitchen — and that was rather the point: no one yet knew what a personal computer was for. Then, in 1979, a program called VisiCalc appeared, and it changed the meaning of the machine underneath it. VisiCalc was the first electronic spreadsheet. It let a person lay out numbers in a grid — a budget, a sales forecast, a loan schedule — and then, when one number changed, watch every dependent number recalculate instantly. Anyone who has ever built a model by hand, erasing and re-adding down a column of figures, understood at once what they were seeing. VisiCalc was so useful that people bought a computer in order to run it. It was, in the language that would later become standard, the first “killer app” — the application worth buying the hardware for.

And the hardware they bought it on was the Apple II. Businesses were carrying these hobbyist machines in through the front door — not because anyone in the data-processing department had approved them, but because a manager needed to run a spreadsheet. IBM’s own corporate customers were among them. That is the pressure that set everything in motion. A toy was quietly siphoning serious money away from the company that owned serious computing. And that is what launched Project Chess: IBM’s crash program to build a personal computer, fast.

The pivot — Project Chess and the open machine. To move at a speed the company had never moved at before, IBM’s small Project Chess team broke IBM’s own deepest rules. IBM built things itself, end to end, in-house. The PC team did the opposite. They assembled their machine from off-the-shelf components anyone could buy, and they outsourced the two most important pieces of all: the microprocessor, the chip at the machine’s heart, which they bought from a young company called Intel; and the operating system, which — as we’ve seen — they got from Microsoft. And they made the design open: they published the specifications, so that other companies could build cards and peripherals and software for it.

Speed was one reason for all this. But there was another, and it points ahead to a theme we’ll return to. IBM had, just the year before, emerged from a federal antitrust lawsuit that had ground on for thirteen years. A company that had spent more than a decade defending itself against the charge of illegal monopoly had powerful reasons not to build a closed, wholly-owned, monopolized product. The open architecture was, in part, the caution of a firm with lawyers in the room.

By keeping for itself only the commodity — the metal box — and giving away the two things that would turn out to matter most, the operating system and the microprocessor, IBM handed its future to two suppliers without understanding that it had done so. It thought it was licensing components. It was giving away the industry.

The launch — August 1981. The IBM PC shipped in August 1981, and IBM’s name accomplished exactly what Apple’s could not: it made the personal computer respectable. What had been a hobbyist’s curiosity became, overnight, a legitimate business tool — because IBM sold it, and, as the saying went, “nobody ever got fired for buying IBM.” The machine sold far beyond the company’s own forecasts.

But look closely at what customers received, because the trap has already closed and no one can see it yet. The operating system on that IBM machine was Microsoft’s — sold to IBM as PC-DOS, but with one clause that would prove decisive: Microsoft retained the right to license the very same operating system to anyone else who wanted it. The IBM PC was an open design that other companies could copy, running an operating system its author was free to sell to those same companies. The door to the clones stood wide open. The audience just doesn’t know it yet.

The consequence — Lotus 1-2-3, 1983. We close this opening arc where we began it, with a spreadsheet — because if VisiCalc was the cause that brought IBM to Seattle, Lotus 1-2-3 was the consequence that made the IBM PC permanent. Released in 1983 and written specifically for the IBM PC and its Intel processor, 1-2-3 was faster, richer, and more polished than VisiCalc had ever been. It became the program that businesses bought the IBM PC to run — and in doing so, it killed VisiCalc, the pioneer that had made it all possible. This is the moment the personal computer stops being a novelty and becomes standard equipment on the office desk.

And notice what that means, given everything we’ve just watched IBM give away. The indispensable machine of American business was now an open design that anyone could clone, running an operating system that Microsoft could sell to every one of those cloners, on a processor Intel could sell them too. The success of Lotus 1-2-3 is therefore not just the end of this story. It is the starting gun for the next one — for Compaq, for the clones, and for the rise of the empire we will come to call Wintel.

The clones — Compaq and the birth of Wintel. IBM had published the specifications of its machine, which meant anyone could build the same collection of off-the-shelf parts. But there was one piece that was not off the shelf and not published: a small, essential program burned into a chip called the BIOS, IBM’s own copyrighted code, the layer that let the operating system talk to the hardware. To build a true IBM-compatible machine, a rival had to reproduce what the BIOS did without copying a single line of how IBM did it. In 1982 a new company called Compaq accomplished exactly that, through a technique that has protected imitators ever since: the “clean room.” One team studied IBM’s BIOS and wrote a precise description of everything it did. A second team, walled off from the first and forbidden to see IBM’s code, read only that description and wrote fresh code to match it. The result behaved identically to IBM’s BIOS but legally contained none of it. The compatible machine — the “clone” — was born.

And then the clones came in a flood. Compaq, and soon Dell, Gateway, and dozens of others, sold machines that ran the same software as the IBM PC, often faster and always cheaper. For the customer this was wonderful: prices fell year after year while the machines grew more powerful. But consider who won and who lost. The standard that every one of these machines had to meet was defined by two components: the Intel processor inside and the Microsoft operating system on top. It did not matter whose name was on the box. Every clone sold meant another sale for Intel and another sale for Microsoft. The box itself — the thing IBM had kept for itself — had become a commodity, made by anyone, earning almost nothing. The profits pooled instead around the two suppliers IBM had so casually taken on. We have a name for the empire those two companies built together: Wintel, from Windows and Intel. And here is the full irony of the story that began in that Seattle conference room — IBM built the standard that organized the entire industry, and captured almost none of the money that standard threw off.

Moore’s Law — the engine under everything. Step back now from the personalities and the deals, because none of what follows makes sense without understanding the force that drove all of it. In 1965 an engineer named Gordon Moore, who would go on to co-found Intel, noticed something about the chips his industry was making. The number of transistors that could be packed onto a chip of a given size was doubling at a steady pace — roughly every couple of years. He predicted the doubling would continue. It did — not for a few years, but for decade after decade, long enough to become the most reliable prophecy in the history of technology. It came to be called Moore’s Law.

Understand what a doubling every two years actually means, because the human mind is bad at it. It does not mean chips got a little better each year. It means that a computer’s power grew a thousandfold roughly every twenty years, and a millionfold over the span of a career. The machine on an office desk in 2000 was not an improved version of the machine from 1980; it was thousands of times more capable, at a fraction of the price. Every single thing this lecture describes after this point — the graphical screens, the ever-heavier software, the ability to store photographs and then music and then video, the network that could suddenly move all of it — happens because of this one relentless curve. Moore’s Law is not an item on our list of milestones. It is the engine underneath every milestone on the list.

Andy and Bill — the upgrade treadmill. There is a joke from this era that explains the economics of the whole PC industry in six words: “What Andy giveth, Bill taketh away.” Andy is Andy Grove, who ran Intel. Bill is Bill Gates. Every two years, faithful to Moore’s Law, Intel delivered a new processor twice as fast as the last — Andy giveth. And every two years, Microsoft’s software grew heavier, more demanding, hungrier for exactly that new power — Bill taketh away. The new version of Windows, the new version of the office software, needed the new chip to run well, which meant the machine you bought three years ago now felt unbearably slow, which meant you bought a new one, with a new Intel chip and a new copy of Windows inside it. The treadmill never stopped, and it kept both companies growing for twenty years.

And recall where Microsoft’s half of that treadmill came from. The tollbooth Microsoft collected at, on every one of those hundreds of millions of machines, was the operating system — the descendant of QDOS, the “Quick and Dirty Operating System” that Gates had bought from a programmer across town for less than the price of a house. That is the return on the meeting that failed.

Xerox PARC, the graphical interface, and the duel between Gates and Jobs. This had happened once before — a company holding the future in its hands and failing to understand what it held. In the 1970s the Xerox Corporation, whose fortune came from photocopiers, funded a research laboratory in California called the Palo Alto Research Center, PARC for short. What the researchers at PARC invented in a single decade is difficult to believe when you see it listed in one place. They built the graphical user interface: the screen not as a wall of text but as a “desktop,” with overlapping windows, little pictures called icons, and a pointing device called a mouse that let you simply point at what you wanted. They built the laser printer. They built Ethernet, the wiring that lets computers in a building talk to one another. They built, in a machine called the Alto, essentially the first personal computer as we would now recognize one. Xerox, a company that sold copiers, had invented the future of computing — and had almost no idea how to sell it.

In 1979 a young Steve Jobs, whose Apple was flush with success from the Apple II, was given a tour of PARC. He saw the graphical interface — the windows, the icons, the mouse — and he understood in an instant what Xerox never had: that this was how everyone would one day use a computer. Apple’s engineers took the idea and refined it into a product, and in 1984 Apple released the Macintosh: the first computer ordinary people could buy that worked the way computers work today, by pointing and clicking instead of typing arcane commands. It was, and everyone admitted it, the more elegant machine. Microsoft’s answer, a system called Windows that laid a graphical interface over the plain DOS of the IBM world, arrived later and looked clumsier for years. But Windows ran on the Wintel machines — on all of them, the whole cloned flood.

And there is the duel. Jobs and Gates spent two decades as rivals, and the contest between them was, at bottom, an argument about architecture. Jobs believed in control. The Macintosh was a closed machine: Apple made the hardware and Apple made the software, sealed together, and no one else was welcome inside. That control is exactly what made the Mac so elegant and so reliable — but it also made it scarce and expensive, one company’s machine at one company’s prices, and it meant relatively few outside programmers wrote software for it. Gates believed in the opposite. Windows ran on every cloner’s hardware, so the machines were cheap and abundant, and — crucially — the entire world’s programmers wrote their software for the PC, because that was where the customers were. The more software appeared for the PC, the more people bought the PC; the more people bought it, the more software appeared. That self-reinforcing spiral, and not any question of quality, is what decided the contest. The versatile, open, inelegant machine crushed the beautiful, closed one.

Gates paid a real price for that openness, and it is worth naming, because it is the mirror image of the Mac’s elegance. To keep Windows running every piece of software that anyone had ever written for the PC, on every strange combination of hardware the clone-makers had ever shipped, Microsoft had to pile compatibility upon compatibility until the system grew vast, tangled, and notoriously prone to crashing. The Mac was polished because Apple controlled everything; Windows was bloated because Microsoft controlled almost nothing but had to support everything anyway. That was the trade, and Gates made it knowingly.

The coda tells you who won. By 1997 Apple, having lost the war, was months from running out of money. And the company that rescued it was Microsoft, which invested a hundred and fifty million dollars to keep Apple alive — taking non-voting shares, extending its commitment to keep making Microsoft Office for the Mac, and settling old legal quarrels in the bargain. Gates had reasons beyond charity: a Microsoft with no competitor at all would look, to the government lawyers already circling, exactly like the illegal monopoly they suspected it of being. But strip away the motives and the picture is stark. The winner of the personal-computer war wrote a check to keep the loser breathing.

The office made modern — the laser printer and the network. While that drama played out, a quieter transformation was turning the personal computer from a thing on a desk into the nervous system of the workplace. Two pieces mattered most, and both, as it happens, were more children of PARC. The first was the laser printer — another invention Xerox’s researchers had built — which, once it became affordable, meant that documents produced on a personal computer could be printed with a crispness that had previously required a professional print shop. The office could now design and produce its own polished paper. The second was the local-area network, and above all Ethernet — PARC once more — the wiring that let the separate computers in an office stop being islands and start being a connected whole, sharing files, sharing that expensive laser printer, sending messages down the hall. By around 1990 the modern office had taken shape: a personal computer on every desk, those computers wired to one another, a shared printer humming in the corner. Every element of the office we would recognize today was now in place but one. The computers could talk to each other within the building. They could not yet talk to the world.

Coda — the antitrust ghost and the threshold. One thread remains to be tied off. That thirteen-year antitrust case against IBM — the one whose long shadow helped push IBM toward the open architecture that cost it the industry — was finally dropped by the government in 1982, abandoned as “without merit.” Sit for a moment with the irony the whole lecture has been building toward. The case never convicted IBM of anything. But the caution it bred, the instinct of a company that had spent thirteen years with its lawyers explaining why it was not a monopoly, was part of what led IBM to give away the operating system and the processor — and so to hand the monopoly it feared being accused of straight to Microsoft and Intel. The government would spend the end of the 1990s prosecuting Microsoft for precisely the dominance that IBM’s antitrust anxiety had helped create. But that is a story for the second lecture.

Because we end this one on a threshold. By the close of the 1980s the personal computer is everywhere, and inside the office building the machines are wired together. Everything is in place for the last and largest transformation: joining not the computers in one building, but every computer on earth, each to all the others. That is where we begin next time.

Share

Subscribe now

From an economic perspective, a health care system is a set of institutions that processes information and provides incentives to consumers and health care providers.

…We ask a lot of the institutions that make up our health care system. These institutions must govern who can supply various medical services. They must channel resources for innovation. They must encourage the use of cost-effective medical procedures while discouraging the use of procedures that are not cost-effective. They must provide adequate and appropriate compensation for health care providers, taking into account their incentives to take advantage of any system, including by committing fraud. They have to take into account political constraints, including the resentment that Americans feel toward any institution that is perceived as restricting insurance coverage of medical services.

In all of these activities, our capabilities are limited by a lack of available information.

When America finally does adopt “Medicare for all,” which I believe is likely, this will put the government in an impossible position. It will have no idea how much to pay providers, because there will be no “usual and customary charges” for comparison. It will have little ability to refuse coverage for procedures, because that will be politically suicidal. It will be a sitting duck for fraud on an even larger scale than is the case now.

As government provision of health care expands, the quality of service of service declines. Private alternatives become an escape valve for the affluent. A goal of equal access to medical care is not achievable.

I think that a better health care system would be one in which individuals pay for the care of non-catastrophic ailments. I can understand why it seems wrong to make someone with a devastating illness requiring expensive treatment to have to pay for his own health care. But you and I sharing the cost of every minor doctor visit, prescription, and X-ray through a third party (insurance or government) is unnecessarily wasteful and expensive.

Share

Subscribe now