I disagreed with this Washington Post editorial so much I had to comment on it. Here are the opening paragraphs:
The resilience of the American worker is one of the most underreported stories of the 2020s. From red tape to import taxes, successive governments have erected barriers to success. Yet America’s workers have persevered and figured out ways to prosper.
A new American Abundance Index illustrates this. The project from Human Progress, an arm of the Cato Institute, reveals the steady rise of the average worker’s purchasing power. The premise of the index is simple: how many hours do you need to work, compared to the month or year before, to be able to afford the “basket of goods,” which is a standard set of household items and services that comprise the Consumer Price Index used to calculate inflation.
The “time price” is how many hours of work it takes to purchase the basket of goods. The “abundance” is how much of the basket one hour of work can buy. The story told by the index is a very good one: since recordkeeping began, “abundance” for average private sector workers comes out to a net increase of 13.8 percent.
Why, oh why, do people who do not understand statistics rely so heavily on it? Let me offer a hint: whenever you read “average”, ask yourself whether whatever is being measured occurs in a normal distribution. If it does not, the average will be grossly misleading. In this case “abundance” is propaganda not analysis. The index quietly assumes that gains anywhere in the income distribution count equally as ‘abundance’ for everyone.
In the United States incomes do not occur in a normal distribution; they are heavily loaded to the right of the distribution, i.e. a relatively small percentage of people earn a considerable proportion of income. You may argue over whether that is good, bad, or indifferent. I see it as merely a fact. Because income growth in the top percentile is orders of magnitude larger than the rest, even small changes there dominate mean-based measures.
I took the trouble of recalculating the Abundance Index using the same sources as the original source used but omitting the top 1% of income earners. Some interesting things became apparent.
The topmost graph illustrates the change in abundance from March 2006 to the present. As you can see for the lower 99% of Americans abundance was materially flat; it wasn’t until 2020 that it began to increase for us. The original index showed a 14% increase over the period; the revised one reflect a 5% increase over the period. In other words, yes, we have become better off but not a lot better off.
The graph on the bottom right illustrates why that may be so. The top 1% of income earners captured 17.6% of income in 2006; now they capture 20.9% of income. That fact alone discredits the index as a measure of general worker well-being. The moral of the story is don’t use unadjusted means for skewed distributions when making welfare claims.
But the topmost graphic also illustrates something even more interesting. What happened in 2020 that improved the measured abundance of those in the lower 99%? The answer, obviously, is COVID-19 or, more correctly, the pandemic and the federal government response to it. Three things happened:
- There was $4.6 trillion in government spending which prevented collapse and gave workers leverage.
- There was a labor shortage caused by 1.2 million deaths from COVID, retirements, and workforce exits.
- Workers, enabled by enhanced unemployment benefits, refused wages they deemed exploitative.
None of these conditions were desirable; they were merely revealing.
That was anomalous; those conditions were not sustainable. In other words it took a crisis to produce abundance for most Americans.
I think you have the right of it here. Using only the mean often gives a false picture. In my ideal world we always also look at median and even mode.
While I have other pet peeves a current one bugging me is the dates people choose to conduct their studies. Cowen recently cited a study looking at some economic effects in the US from 1939-1963 vs 1963-present. The war years and the years shortly thereafter are quite a bit different than later years. We had essentially full employment in the war years and in the proximate years after the war many/most of the cities and industrial bases of the first world were decimated so we had little effective competition.
Steve
I have similar peeves. For example, I really wish somebody would produce a credible explanation for why the DJIA was essentially flat for 30 years but since 2006 has risen exponentially. They need to explain both sides of that story not just the last 10 years.