Harry Truman And The Inflation-Adjusted Heist of The Century
Making a fool of the taxpayers and our inflation statistics
If you’ve got even a passing interest in Presidential history, I cannot highly recommend enough this article from New York Magazine on the post-presidency of Harry Truman [NY Mag]. Truman is among the most iconic pulled-himself-up-by-the-bootstrappers in American history, second only to Mr. Log Cabin himself, Abraham Lincoln. What is less known to the general public but more familiar to scholars is his post-Presidency, where he fell into (relative) poverty and threw himself on the mercy of the Congress.
A Forbes article on the Former Presidents Act [Forbes] tells the conventional story:
These were lean years for Harry and his wife, Bess. Indeed, the former president was living principally on his military pension; while President Truman got nothing from the taxpayers, Col. Truman was entitled to a monthly pension of $112.56, thanks to his service during World War I.
…But Truman considered [corporate work] unseemly. "I could never lend myself to any transaction, however respectable," he later wrote, "that would commercialize on the prestige and dignity of the office of the presidency."
…
Truman’s early struggle to make ends meet after leaving the White House was finally getting some attention on Capitol Hill. The ex-president had been tight-lipped about his financial troubles, but he had acknowledged that he could use a secretary to help answer mail. Eventually, Congress responded with the Former Presidents Act of 1958. The law created a $25,000 annual pension for ex-presidents, while also providing for administrative support. Both Truman and Hoover accepted the pension. Although Hoover didn't need it, he was loath to embarrass Truman.
The New York magazine story, based on Truman’s recently released financials, turns this narrative on its head:
Harry Truman was a very rich man on the day he left the White House, and he became a good deal richer in the five and a half years between that day and the passage of the FPA. Moreover, Truman departed from the White House with so much money because he apparently misappropriated what in today’s terms would be millions of dollars from the United States government.
It’s a great story and I’d highly recommend it - but also one that can sometimes be disorienting to read, due to the much smaller amounts of money compared to the present day. The author has frequent caveats of “adjusted for inflation”, but at various points finds this inadequate for getting the point across:
Truman’s estimate of his net worth when he departed the presidency was approximately $650,000. Simply adjusted for inflation, that $650,000 is equivalent to $6.6 million today. But this latter figure fails to capture the fact that America is a vastly wealthier nation now than it was then. Consider that in 1953, the 99th percentile of household net worth was approximately $125,000. Meanwhile, in 2020, the comparable figure was about $11.1 million. In other words, to be as rich today relative to other Americans as Truman was in 1953, a person would have to have a net worth of around $58 million."
This analogizing-by-percentiles is such an unusual way to discuss income and net worth in historical terms that it seems like the author is trying to slant the data. It’s making Truman “look” richer than it is - why not go with the simple and more objective adjustment for inflation? But the more you think about it, the more you realize that either description is a very important choice and one that encodes very strong assumptions about how we think about historical data.
“Just Adjust For Inflation”
What is inflation? A common definition is “the decline of purchasing power over time”, generally measured as “the increase of an average price level of a basket of selected goods and services in an economy over a period of time” [Investopedia]. This basket approach allows people to easily compare “apples to apples” by comparing the prices of commonly purchased goods. If you want to know the difference of a dollar in 2021 vs. 1921, you just need to calculate the price of a basket of common goods purchased in both periods, like an iPhone 12 or a pair of draft oxen.
Wait, that didn’t sound right.
A key issue is that consumption baskets tend to change over time.1 This won’t matter a ton if you’re comparing pretty short time periods, but even so it can rear its head. Recent inflation figures have been very high by recent standards, but when decomposed into constituent parts it’s very obvious that the inflation numbers are being driven above all by the surging price of used cars [Twitter]. And as an obvious result, Americans are buying fewer used cars [KBB]. So is the weighting of used cars in calculating inflation “right” or “wrong”? It seems like the actual-but-unsatisfying answer is “neither - different measurement choices will give very different answers”.
The problem gets so much more extreme when thinking about comparing the value of a dollar over very long time periods such as between the 1950s and today. When Truman stepped down in 1953, there was simply no price at which he could buy an iPhone, or broadband internet, or unleaded gasoline [Wikipedia]. While you can certainly calculate an inflation adjustment over these periods, whether using a standardized consumption basket or a chained method, it starts to become quite difficult to really qualitatively describe what you’re measuring.
And this is where what social scientists call “scope conditions” really start to bite. The “scope conditions” of any mathematical relationship or formula are, effectively, “what are the pre-conditions under which this relationship makes sense?”. A poignant example comes from V.O. Key’s 1949 masterpiece, Southern Politics [Amazon]. Key mentions early on that the most African-American areas of the South were the most Democratic (and most anti-civil rights). I was initially confused until remembering what went without saying to Key: African-Americans could not vote. The (white) voters in those areas were the most rabid proponents of Jim Crow. Those same areas remain very Democratic today, and so the same calculation would have the same results, but the underlying relationship runs in almost the opposite direction. Claiming that the same demographic predictors of Democratic vote share has persisted for a century is both straightforwardly true and obviously false.
Scope conditions rear their ugly head when thinking about inflation over very long time periods and economic regimes. It is pretty straightforward to calculate an inflation adjustment over very similar periods, where you generally see slow but steady declines in purchasing power. But as you start to stretch that time period longer and longer, covering more and more different economic conditions, that same calculation translates less and less readily to real conditions. It starts to feel more and more like a philosophical residual, something which falls out of the calculation but has no meaning of its own.
This is a long-winded way of saying: we may never truly know how much taxpayer money was stolen by Harry Truman, criminal mastermind and ol’ country boy from Independence Missouri. Not in any way that’s satisfying.
True politics nerds might remember this debate in the context of an Obama-era debate over cutting Social Security benefits by changing inflation-indexing to “chained CPI” which adjusts for shifting consumption patterns.