Economics Links, 2/16/2026
Greg Ip on income shares; Peter Coy on the Bitcoin story; Testing for linearity; Jordan McGillis on middle-class anxiety
For the WSJ, Greg Ip writes,
The shift to capital from labor has actually been under way for more than 40 years. Labor received 58% of the total proceeds of economic output, as measured by gross domestic income (conceptually similar to GDP), in 1980. By the third quarter of last year that had plummeted to 51.4%. Profits’ share, meanwhile, rose from 7% to 11.7%.
The Buffet ratio is total stock market capitalization divided by GDP. Call it P/Y. We can think of it as
P/Y = (P/E) x (E/Y)
where E is earnings. Ip is telling us that E/Y has gone up about 70 percent since 1980 (from 7% to 11.7%). But P/Y has gone up about 500 percent since then. Most of the gain in stock prices has come not from higher earnings but from investors willing to pay more for a given amount of earnings.
Bitcoin players borrow money to buy perps. So the bets can be enormous— five, 10, or sometimes even 50 times the money they put up—and far more than the total universe of actual Bitcoins in existence.
That means traders can make a mint if they bet right but can lose their shirts if they’re wrong. And because the Bitcoin exchanges, like Binance, need to keep the price of the perp close to the price of actual Bitcoin, it is constantly balancing payments between the “longs” who are betting on the price to go up, and the “shorts” who are betting on it to go down.
And this is where things get dodgy.
When prices fall, traders who have used perps to bet that the price will go up are forced to post more collateral to guarantee repayment
So you get a cascading effect. A “perp” is a perpetual futures contract. To me, it sounds like a synthetic bitcoin.
If nonlinearity matters so much, why hasn’t the profession moved away from linear models more aggressively?
If this is a problem, it reflects laziness or stupidity. I was taught that you order your sample according to the independent variable X and divide it in half. Estimate the relationship on the bottom half and on the top half. If they are different, you need to allow for nonlinearity.
Nobody should be caught making the mistake of incorrectly imposing a linear relationship on data when a simple test can show that it is wrong to do so.
In the WSJ, Jordan McGillis writes,
the median income in 1975 was $15,000. The 80th percentile for married couples with children was $22,600. Put another way, the 80th-percentile family earned 51% more than the median family.
Since then, the distance between the median and the 80th percentile has grown. In 2000, the median married couple with children earned $59,000 while the 80th percentile family earned $99,000, meaning the 80th-percentile family was earning 68% more.
By last year, the gap had grown even more substantial, with the median married couple with children earning $130,000 and the 80th-percentile family, also a married couple with children, earning $242,000—85% more.
Americans notoriously saw themselves in the middle class if they were anywhere between the 5th and 95th percentile of income or wealth. But that is no longer tenable. If the “middle class lifestyle” is what you see at the 80th percentile, then people at the 50th percentile feel deprived.
substacks referenced above:
@
"If the “middle class lifestyle” is what you see at the 80th percentile, then people at the 50th percentile feel deprived."
The problem is that regional differences in the cost of living have also widened. A ton.
Sure, the 80th percentile couple makes 85% more, but they work and live in an area where housing and other costs are also 85% more. So the actually lived lifestyles are closer that the nominal income makes it seem, and so the "most people are middle class" perception is still kind of valid.
Even putting aside housing, ordinary retail costs in the Big Winner Cities have recently exploded vs exurbs or other regions, and there are countless hilarious sticker-shock stories on Reddit for the DMV area along the lines of, "I just paid $18 for a muffin and I'm pissed!" The urban inflation is real!
The nation is just way too big a unit of analysis, and I'm guessing the results would be more compressed considering a particular city or metro area. At this point it makes nearly as much sense to compare national average 50% vs 80% as it does to lump in Canada, Mexico, and everything north of the Panama Canal to compare aggregate North American 50% vs 80%.
I get nervous about inequality stats like the one you quote, because it sounds like someone is itching to fight but is teeing off at the wrong targets. Here are a couple of stories I tell about all of that.
Mathematically, new technology tends to widen inequality because it will help everyone, but it will disproportionally help those who are more industrious. The rich get richer, and the poor also get richer, just not as much. That's because technology isn't an air drop of food or some other resource that's directly valuable. It's an air drop of a new kind of screw driver. Everyone benefits a little, but some people are going to do more or less with that screwdriver.
One way to think about it, from a societal point of view, is that some people play XBox and some people build XBox. Most people would not enjoy those jobs building XBox. It means giving up a lot of life's normal daily joys in order to take part in large corporate situations, make zillions of small decisions, have a lot of failures along the way from things that didn't work, and having to deal with office politics. Nobody is doing that unless they get paid for it. In fact, the better life gets at the bottom, the higher the premium people are going to charge to leave their early retirement and go slave away in one of these corporate situations.
Despite this, everyone gets XBoxes. Now repeat that thought experiment for phones, Amazon, cars, hot water, lighting, pools, you name it. The more these things are available to the general population, the more you have to pay someone to get them to step away from enjoying it and go do what is needed for the next great thing.
The story is familiar and comfortable to me, and I feel like people who focus on inequality are often missing some part of all of this. The gap is wider than ever for the best and worst slalom skiier, but why on earth is that a bad thing?