Tuesday, April 24, 2012

Mitt Math, and the Truth on Women’s Employment

Romney and his campaign claimed: “The Obama administration has brought hard times to American women. Under President Obama, more women have struggled to find work than at any other time in recorded history. Women account for 92.3% of the jobs lost under Obama.”

How have women actually fared in the Obama years? In reality, both men and women have done significantly better under Obama than under Bush’s last two years, the recession years. However, the men’s recovery has been stronger than the women’s recovery.

What Romney did, by deceitful manipulation of statistics and dishonest interpretation of the results, is to characterize a slower recovery for women as an historical hit on women, a time when “more women have struggled to find work than at any other time in recorded history.” This is flat out wrong, as the employment losses and unemployment increases were worse under the Bush recession. And after the effects of the Obama stimulus kicked in full force, from January 2010, there has been net job growth for both men and women.

How did Romney get his 92%, and what is the real story on women in the labor market? First, to understand Romney’s deception, a little story will help.

Once I asked for a quarter from Romney to feed a parking meter, and showed him I only had a penny in change. He gave me gave me two more pennies, and walked away laughing. Then he went around the country saying that he had tripled my money using his financial genius, and that I was a typical liberal with colossal ingratitude.

Ok, that didn’t happen, but the story illustrates a point: when you compare small numbers using ratios or percentages, the results can be totally misleading. Romney started with honest data, but then did dishonest math. The net change in employment in Obama’s months (1/09 - 3/12), compared to the total number of people who were employed last month, is a -.56% net loss. The losses of women alone were -.51%. The ratio of these small percentages, or of the two before & after numbers, is about 92%.

That method of calculation turns a ½% reduction in women’s employment into an apparent massive hit on women. But the reality is that there was only a small loss, especially compared to loss in women’s employment under the last two years of Bush, which was 4%. And, as I said, if you start after the stimulus kicked in strongly in January 2010, you get a net increase in jobs under Obama.

Now if I were using Mitt Math, I would say that Obama has been eight times better for women than Bush. However, I wanted to get an honest picture of the economic impact of the Obama administration on the women’s labor market.

So I talked to the friendly public servants at the Bureau of Labor Statistics. The first thing I learned is that the Employment numbers, which Romney used—or rather misused—do not by themselves give a full picture of what is going on in the labor market. People enter the labor market from school and they leave for retirement. Women often leave to stay at home to raise their small children, and return when their children are bigger. These and other reasons besides unemployment bring people out of the labor market and back in, and this non-unemployment group is as large as those who come into and leave work because of being laid off or fired.

A second problem with the employment numbers is that they are ‘snapshot’ numbers. So if you look at the difference between employment numbers at earlier and later dates—as Romney did—the result doesn’t tell you what happened in between. Further, you can’t just add monthly figures because many people stay employed and unemployed from month to month. Simply adding months would involve double counting.

Fortunately, there are available numbers on the flows in an out of the labor market, from the Current Population Survey. And the BLS does keep track of detailed flows: employed to unemployed, employed to not in labor market, employed to employed; unemployed to employed, not in labor market to employed, and so on. With these detailed numbers, it is legitimate to show average flows of women’s and men’s employment and unemployment over months. That gives you an honest picture of the impact over an extended period. For an example of sound methodology in looking at the flows, the BLS folks directed me to their report on why unemployment rose in the first part of the recession.

Following their methodology, I compared what happened in the last two recession years of Bush, and under Obama. I also compared experience under Obama after the stimulus took hold, in January of 2010. The following tables are based on men’s and women’s seasonally-adjusted numbers from the Current Population Survey, research series, from the BLS. (The flow numbers are published here for the first time.) Note that the total unemployment numbers, widely reported each month, include also those who remain unemployed from month to month, and this is the largest group. That is why total unemployment numbers are much higher than the those entering and leaving unemployment.

[click graphic to enlarge.] So what does an honest reading of the indicators show? First of all the Bush recession years were far worse for women than the Obama years. However, the Obama recovery has been stronger for men than women. During the first Obama year, women were still losing jobs, while men were already net regaining them. From Jan ’10, when according to the CBO the stimulus had its strongest impact, the Obama years are net positive for both men and women. However the recovery is still significantly weaker for women.

As the spokesman for the Bureau of Labor Statistics noted in a Politifact article on the 92% number , this pattern of men being hit first and recovering earlier has been typical. Here Obama tried to funnel federal money to the localities to address the increased unemployment of workers in local government—which accounts for a lot of the women’s job losses—but this effort was blocked by Republicans in Congress.

Even with recovery, the flow numbers reveal that the Obama years have had more “churning” in the labor market compared to the Bush years, with more people entering and leaving unemployment every month. The net post 2010 is positive for both women and men, but with more people changing jobs. This may be an indicator of the kind of “creative destruction” rightly lauded by fans of the free market as part of the transition to a more productive economy. However, it is understandable that it leaves people feeling insecure, especially with the frustratingly slow decline in total unemployment.

Does Romney’s complicated deception with numbers have any significance? Does "Mitt Math" matter? I think it does, because it goes to the heart of Romney’s case for him being a better president than Obama. Romney says he understands business and finance, and so can act more effectively to help the economy. But knowledge of finance can be used in irresponsible and self-serving ways, as well as a good and productive ways. We all are acutely aware that it was manipulation of financial numbers by financiers that led us into the housing bubble and burst, and worst recession since the great depression. When financier Romney starts being dishonest and self-serving about statistics, it raises the specter of a Romney presidency doing to us again what dishonest and self-serving financiers did to us in the last decade.


  1. Excellent analysis with a sobering conclusion. The frightening aspect is that the public's choices are made on the basis of snapshot memes. What does this mean for responsible democracy?

  2. Thanks Bruce, I think responsible democracy is dependent on the electorate not rewarding dishonest campaigning. We shall see...