woody_zimmerman_118_2007My grandpa often used the expression that titles this article. He also referred darkly to “lies, damned lies, and statistics” – although Mark Twain might have said it first. Both waggish lines seem apropos today, as we are routinely subjected to a blizzard of figures and statistics describing myriad public and private activity, including the performance of the national economy. With respect to the latter, however, there is more here than meets the eye. That’s why I’m writing this “primer” for my readers.

The monthly Department of Labor report of the unemployment rate for the country is an important case in point. The basic mathematical expression for this statistic is the following: i.e.,

% Unemployed = 1.0 – [total # employed/total worker population]

One would think that there could be little chance for confusion about this simple expression, but because it is a government-issued statistic, one would naturally be wrong.

For most of the past year, the official unemployment rate has dropped lower, month by month. Originally well above 9%, the rate has gone down 1/10 or 2/10 points each month, until it now stands at 8.1%. This would appear to be good news for the country – not to mention the Obama administration – since each of the first three years of that administration saw unemployment averaging above 9%. But now, just in time for the election, the economy seems to be recovering and people are finding jobs.

The difficulty with this happy scenario is that things are not quite as they seem. In fact, the denominator of the employment-fraction (above) changes from month to month. This is not unexpected, of course, since new workers typically join the workforce each month. Normally, this would make the worker population increase over time, but that is not what’s happening. Instead, that population has been shrinking by hundreds of thousands of workers each month, at an accelerating rate. In April, the shrinkage was 342,000. For March and April, the cumulative decrease was over 500,000.

The official explanation for this anomaly is that the “official” worker population includes only people currently working or looking for work. Unemployed people who have stopped seeking work are deducted from the worker population – i.e., the denominator of the employed fraction. A smaller denominator makes the employed fraction increase, producing a smaller unemployed fraction.

This convention lets Mr. Obama celebrate a lower unemployment rate, while leaving the “man on the street” in the dark about how the numbers were adjusted to produce that result. Since Mr. Obama’s inauguration, the official worker population has shrunk by some 2 million workers because those who stopped looking for work were subtracted out. Calculations suggest that if these workers were still included in the worker population, the true unemployment percentage would stand between 9.5% and 10%. Some economists place the corrected figure between 11% and 15%, if accounting is made for “underemployed” workers – i.e., people who are working jobs significantly below the level they desire.

Media reporting on all this has been depressingly simplistic, with most news organs simply parroting the official unemployment fraction without providing any supporting explanations of the methodology used by the government to calculate the statistic. Only recently, with release of the figures for April, has the news agency Reuters mentioned how 342,000 “vanished” workers – not an increase in employment – has reduced the unemployment measure. A few media organs have picked up that aspect of the story as well.

Since not everyone grasps the mathematical details of the “vanishing worker” scenario, I thought it might be useful to demonstrate how it works on a smaller scale that doesn’t involve numbers in the millions.

Let’s say a company has 1300 workers on its rolls, but only 1170 – or 90% – are working. This means the “idle” rate (1 – 0.90) is 10%. In a later quarter, it is found that only 1157 of the 1300 are on the payroll. This is 89%, so the idle rate (i.e., 1 – % working) is 11%.

After a year, only 1118 are working, apparently giving a working rate of 86% and an idle rate of 14%. However, the company’s accountants find that the number of employees working or wishing to work has now dropped to 1250. This actually raises the employed rate to 89.4% and lowers the idle rate to 10.6%. It is a much happier result to report.

By the next quarter, the number of working employees has dropped by another 50, to 1068. If the number of employees working or wishing to work were still 1250, the employed rate would have dropped to 85.4%, and the idle rate would have risen to 14.6%. (Not a good picture for the annual report to convey to stockholders.) Luckily, the company’s accountants find that only 1150 are now working or wish to work for the company. This makes the employed rate 92.9% and the idle rate 7.1%.

Thus, over the period of time represented, the company’s active payroll has dropped from 1170 to 1068 – a decrease of 8.7% – yet its official idle rate has dropped from 10% to 7.1%. At a casual glance, the latter looks like an improvement. But a deeper look into the diminished payroll and degraded worker-base reveals a more sobering picture. Indeed, another loss of 102 from the payroll, and shrinkage of 150 more from the reportable worker-base would give an employed fraction of 96.6% and an idle fraction of just 3.4%. Splendid, it seems, but not truly so.

If all of the numbers in the example are multiplied by 100,000, the scale approaches the level of our national economy (although not necessarily in every detail of the example). The instructional detail is that shrinkage of the worker population produces a deceptive employment picture.

Both the media – at least those parts that care about accuracy – and the public are beginning to realize that we’re not getting a correct reading on the country’s employment. For the Obama administration, this is politically advantageous at the moment, as the official unemployment statistics make it appear that the economy is gaining in strength – a nice story to be able to tell during campaign season.

The view of many (including Yours Truly) is that 8% is considered a magic unemployment number. No incumbent president since Franklin Roosevelt has been re-elected when unemployment has exceeded 8% at the time of the election. Thus, I expect to see the official figure drift down into the 7-8% range as the election nears, while the finer details are lost in the background noise.

This will be an illusion worthy of Houdini himself, but Americans need to be aware that it “ain’t necessarily so.” Millions of people who are out of work – including 41% of the jobless who are considered long-time unemployed – already know this.