Many people who found that their college statistics course was a burden to be endured might remember “How to Lie With Statistics” by journalist Darrell Huff. Often assigned as a supplemental text, this classic little book was the only intelligible part of the course for many students.
Huff informed us how it is possible to lie with statistics and gave many examples of how quantitative information, including charts, can be misused whether through ignorance or malice.
Lying with economic indicators is a subset of lying with statistics. One can be thankful that it is something academic economists seldom do — at least in scholarly work that their peers will read. It is too easy to be found out. But misuse of economic indicators is much more common when presented to the public, especially in the context of political arguments.
This is a response to issues raised by readers in reaction to two recent columns: one explained the accuracy of economic indicators, such as the unemployment rate and employment count; the other examined pitfalls in relating economic performance to individual government officials.
One common technique with time-series data is to cherry-pick beginning and ending dates. A Ron Paul acolyte advocating a return to the gold standard emailed me, pointing out that the U.S. consumer price index, tabulated well after the fact, stood at 17.0 in 1816 and 17.3 in 1919. Therefore prices had been stable for more than a century.
This was true on the surface, but there were periods of harsh inflation and deflation between these two dates. Prices increased 70 percent during the Civil War. They fell 30 percent in the decade beginning in 1869. In both cases, the price fluctuations took a harsh toll on many people. Similarly, 17 percent inflation from 1833 to 1837 disrupted the economy fully as much as price-level rises did in the 1970s.
A variation on this is to compare indicators from two periods in which the underlying structure of the economy was very different. Another emailer noted that federal spending in 1909 was about 4 percent of gross domestic product and that it had risen to 25 percent by 2009 and thus was on a pernicious, inexorable rise.
The data cited again was correct, but overlooked the fact that the increase was far from continuous. It jumped somewhat during World War I, fell in the 1920s, rose somewhat in the 1930s and soared during World War II, when it exceeded 41 percent. It then fell sharply in the first years of Truman’s administration only to jump again in the Korean War.
However, after that it did grow, but not as fast as many think. It averaged 17.1 percent in the Eisenhower years, 18.1 under Kennedy-Johnson, 19.1 under Nixon-Ford and 19.2 percent in Clinton’s time.
It has averaged 24 percent during the three Obama fiscal years, certainly a post-World War II high. But it was 21.9 percent over the eight Reagan years and hit 22.9 in the recession years of 1982 and 1983. From that to the current 24 percent is not so great a leap.
Similarly, people who lived through the 1970s find it hard to believe that the fastest annual growth of employment from 1946 to the present took place under Gerald Ford, whom some unfairly tar as an economic dunce.
However, the strong employment growth during his tenure and that of Jimmy Carter was not necessarily due to outstanding economic policies. The two benefitted from the fact that people born in the most prolific years of the baby boom hit the labor force during their terms of office.
One way to burnish one’s record or tarnish that of another is to cherry-pick among alternative statistical series.
The Bureau of Labor Statistics compiles several different measures of the number of jobs. Taking the months each was in office, “payroll employment” grew at 3.1 percent per year under Carter and 2.1 percent under Reagan. That one-percentage point edge slips to 0.7 points if you use “civilian employment,” a marginally different series.
(However if you make the correct assumption that it takes some time for a new president’s policies to take effect and lag each of these by 12 months, the edge in both series flips to Reagan’s advantage by 0.8 points.)
Yet another source of confusion is under-appreciation of the effects of compound interest — that is, applying new rates of growth to new figures affected by previous growth rates. A questioner at a talk I gave recently blasted the Federal Reserve for allowing the purchasing power of the dollar to fall by 95 percent in the first century of its existence. (His calculation was correct, the CPI stood at 10.0 in 1914 and was 231.4 as of September, 2012, for a decline in value of 95.7 percent.)
Just a week earlier, I had queried another group about what they considered an acceptable rate of inflation. Few saw a problem as long as it stayed below 2.5 percent per year. Yet that apparently innocuous rate would have reduced purchasing power by 91.6 percent in the same 98 years.
However, if one removes two high-inflation periods, World War II and from 1970 to 1982 when the Fed went nuts under two incompetent chairmen, inflation averaged only 2.3 percent a year over the past century. This may not be low enough to satisfy some gold bugs, but it is not so high as to alarm most people.
In any case, the more relevant question is what our inflation-adjusted incomes are now versus what they would be if we had strictly limited our money to gold and silver over this interval. On this economists may differ, but it is hard to find one who favors the gold standard.
Finally, always beware of anyone mixing numerical apples and oranges.
I recently heard a political liberal assert that “the national debt tripled under Ronald Reagan while it had fallen as a percentage of GDP throughout the 1970s.” Again, this is true, but comparing one change in absolute dollars to another in the proportion of GDP is confusing and invalid.
If she had compared identical indicators such as average annual growth rates of the inflation-adjusted debt, ratios of annual deficits to GDP or that of overall debt to GDP, for the two periods, she would have made roughly the same point more credibly.
Yet, however she did it, she still would have been making a logically questionable post hoc argument about the responsibility of presidents who ignored the roles of Congress and the Federal Reserve, among other factors, in that whole unfortunate period.