We happened to come across some graphs from several years ago by Pavlina Tchernova of Bard College showing the growth in income by expansion for the top 10% and bottom 90% and wondered what an update would look like. We didn’t do exactly what she did, but we did want to give her credit for the general idea.

The striking graphs below tell the story by themselves, using numbers assembled by Thomas Piketty and Emmanuel Saez. In the expansions from the end of the World War II to 1973, the further up the income distribution you went, the lower the income gains. For the first five cycles, the bottom 90% saw an average of unweighted average of the growth for the bottom 90% first five cycles of 17%; for the second five cycles (the current one only through 2018), of 6%. For the top 1%, the figures are 7% and 52%. The first is less than half the bottom 90%’s performance; the second, almost eight times as much.

Within the top 1%, similar patterns prevail. For the 99.0–99.5% set, the gains were 10% and 29% for the two sets of cycles; for the 99.9–99.99% fellowship, the figures are 4% and 64%.

We live in very separate economic universes. Your opinion of that depends heavily on your percentile, but whether it’s sustainable is open to question.

The Federal Reserve’s distributional accounts are a combined product of two Fed projects, the quarterly financial accounts, which we review soon after their publication, and the triennial Survey of Consumer Finances, a detailed look at household income and balance sheets. Often when we do our quarterly reviews of the financial accounts, we lament that the aggregates (which behave pretty much like means) don’t tell us anything about distribution. Middle-class wealth, such as it is, is largely in housing, while high-end wealth is largely in financial assets. The movements in those assets are proxies for a distributional analysis, and now we have the real thing.

We graphed dollar levels, not shares, but a few words about them first. In 2019Q2, the top 1% of the distribution owned 29.0% of all assets, up from 21.2% in 1989; the bottom 50% went from 7.3% to 6.1%. Percentiles 90-99 saw little change in their share, but the next 40% lost 7 points of share, 37.2% to 30.1%. Holdings of stocks and mutual funds also become more concentrated, with the top 1% going from holding 47.8% in 1989Q3 to 52.0% in 2019Q2. The bottom 1% saw their share rise from 1.0% to 2.2%. Between the extremes, the most interesting development was how those between percentiles 50 and 90 rode the 1990s stock boom, going from owning 14% of stock in 1989 to a peak of 19.1% in 2001; they now account for just 11.4%. The top 1% went from holding 23.9% of net worth in 1989 to 32.4% this year; the next 9% were unchanged at 37.0%; the next 40% went from 35.4% to 28.7%; and the bottom half from 3.7% to 1.9%. In 1989, the top 1% had 6.5 times the share of the bottom half; in 2019Q2 it was 17.1 times—though that is down from a 2011 peak of 150 times, when the bottom half, hammered by the housing bust, was barely above water.

On to the graphs, which show residential and nonresidential net worth by fractile (slice of the distribution), first in the aggregate and then by household. (We estimate the household numbers by dividing the aggregate by the fractile’s share of total households in the quarter.) “Nonresidential” subtracts the value of real estate from assets and adds back mortgages from assets. It also excludes consumer durables, which are illiquid and depreciate rapidly, which the Fed counts as a household asset. We also deflate by the PCE index; the Fed presents nominal figures.

Although the 1% gets all the press, when it comes to residential net worth they don’t hold as much as the next 9% or the next 50%, though of course those populations are far more numerous. So, movements in the balance sheets of the 90-99%, what in economic slang is oddly called the upper middle class (as it is in the new book by Saez and Gabriel Zucman) and the next 40%, the better off segments of the “middle class,” matter for consumption. The bottom 50% has close to nothing in the aggregate.

The picture changes when you take out the residential parts: there, the 1% have more than percentiles 50–89, and is within hailing distance of 90–99. The bottom 50% have a little more than nothing; they’re disproportionately renters (though they took a hit on residential wealth in the crash), and might have a small savings or retirement account, but not much else.

Looking at households, the bottom 50% meanders around the zero line; they had a net worth of $15,280 in 2019Q2. The 40% above them had a net worth of just under $160,000. You can’t see it on the graph, but that’s up 42% since 1989, though there was an intervening hit of -37% during the housing collapse. The top 1%, who don’t have most of their wealth in housing, still have a lot: $2.9 million, up 188% from 1989. But again, the 1% star in nonresidential wealth: $25.3 million in 2019Q2, up 214% from thirty years earlier. The next 9% were not begging: $3.2 million. The bottom 50% have a net of $84,189 to their name, up 90% from 30 years ago, well under half the gain clocked by the top 1%.

Again, people are living in very different universes.