It has never been advisable to take the statistics thrown out in the heat of political battles other than with a grain of salt. As the old saying goes, “Figures don’t lie, but liars can figure.” There are many ways to slip in the lie. For example, one can introduce variations in the way that common terms are understood, or compare apples and oranges, or simply imply that facts have a significance that lacks any reasonable justification. The battle between the Left and Right in Wisconsin over public unions has generated some interesting examples.
One of the most egregious comes from the left, although the right is hardly without sin in these matters. Specifically, Ezra Klein of Journolist fame is citing a study by the Economic Policy Institute (EPI) that purportedly “proves” that Wisconsin public workers are actually under-compensated compared to their counterparts in the private sector. The basis for his claim is a nice graph included in the study comparing public and private sector compensation as a function of educational attainment. In all these comparisons except at the high school level, the public sector workers seem to be taking a huge hit, amounting to a deficit of anywhere from a quarter to a third compared to the private sector. However, Ezra’s post quotes a couple of paragraphs from the EPI source document citing some caveats regarding this rather striking graph. For example, at the very end of the quote, which appears in somewhat finer print than the bulk of the post, we learn that,
Controlling for a larger range of earnings predictors—including not just education but also age, experience, gender, race, etc., Wisconsin public-sector workers face an annual compensation penalty of 11%. Adjusting for the slightly fewer hours worked per week on average, these public workers still face a compensation penalty of 5% for choosing to work in the public sector.
There is no explanation of why these controls weren’t factored in when the bar graph referred to above, which seems to show that public sector workers make a much greater sacrifice in order to serve the people of Wisconsin, was created. It happens that one can find some possible reasons for the discrepancy if one “Googles” the EPI. It turns out that Ezra somehow forgot to mention that the organization describes itself as “non-partisan but progressive.” For those who happen not to be astute followers of US politics, those who deem themselves “progressives” are rather more likely to be found on the side of the public sector workers than the Republican party in Wisconsin. Ezra also forgot to mention that the source of a big chunk of the EPI’s funding is unions. Perhaps he thought it was too insignificant to mention.
The cost to the state of public pensions is, of course, one of the major bones of contention between Wisconsin governor Walker and the public sector unions. It would, therefore, seem a matter of some importance to calculate this cost with some rigor, and to explicitly document the method used in any document citing that cost. Unfortunately, the EPI source document does not do so. It merely states that,
Retirement benefits account for 8% of state and local government compensation costs compared with 2.5% to 4.9% in the private sector.
It is unfortunate that the details of the method used to arrive at this 8% figure are not described. It seems rather dubious on the face of it. For example, Wisconsin teachers who retire after 30 years service will draw 48% of their top pay in pension for the rest of their lives. It would seem plausible to assume that “top pay” is rather larger than “average pay.” A teacher hired at the age of 25 would reach retirement age at 55. At this age, the average life expectancy for US males is about 25 years, and for females about 28. Any way you figure it, the cost of providing a pension of 48% of top pay for over a quarter of a century dwarfs the 8% figure cited by EPI. Throw in the fact that this figure does not include retiree health and other non-cash benefits, and the discrepancy gapes even wider. On the other hand, the average teacher will likely work for less than the required 30 years. The EPI article does not mention how these and other seemingly salient factors are included in the data. Apparently, its figure is based on the amount of money the state is currently setting aside to fund the pensions, a wildly inaccurate metric for determining what they will eventually actually cost. Given that the organization is anything but an unbiased third party, this would seem to be a rather prominent red flag to anyone tempted to cite them as a source.
In a word, dear reader, to credit statistics thrown out by ideologues is to skate on thin ice. Their main value lies in pointing the way to source material. Should you really be so bold as to seek to isolate a small fragment of something as evanescent as the truth, you will have to endure the tedious task of sifting through a great deal of that source material on your own.