The phrase “lies, damned lies, and statistics” is used almost exclusively in public policy debates. Putting numbers on an issue gives the speaker an air of authority and precision but often comes at the expense of fuller context. Indeed, outright dishonesty is far less common than self-deception and the universal human tendency to seek out evidence that confirms one’s biases.

Consider the stunning statistic that 8 in 10 Americans live paycheck to paycheck. It may not be false, but it hardly captures anything meaningful about the precarity of America’s working class. Instead, it turns out to be a rather literal measure of the fraction of workers who spend their entire paycheck, even if it’s going to college and retirement savings. That is, someone who earns six figures and has a fat 401(k) would still be said to be living “paycheck to paycheck” under this measure simply because they have a high-consumption lifestyle. I doubt this is what the progressive crusader who retweets this statistic uncritically has in mind.

Of course, conservative policy wonks are no less vulnerable to confirmation bias in the opposite direction. Take Angela Rachidi of the American Enterprise Institute’s poverty program. Over the holidays, Rachidi published a blog post titled “Poverty rate reduction of the expanded Child Tax Credit underwhelms.” The post drew on the November 2021 release of the “real time,” monthly estimate of the Official Poverty Measure (OPM) produced by the University of Chicago and the University of Notre Dame’s Lab for Economic Opportunities: 

Notably, the Chicago/Notre Dame poverty rate follows the official poverty measure, meaning it does not include in-kind government benefits, such as food assistance. But it should reflect the monthly CTC payments — making the last point above worth emphasizing [emphasis added]. Advocates of an expanded Child Tax Credit projected earlier this year that it would reduce child poverty by between 39 percent and 42 percent, and cut deep child poverty in half. The monthly child poverty rate fluctuates during this time according to the chart above, but we do not see a steep and constant decline in the magnitude advocates suggested.

As the co-author on one of the studies she links to, I confess this would appear to be pretty damning. The expansion of the Child Tax Credit increased the post-tax income of poor families by thousands of dollars, and yet its impact on poverty seems nearly invisible in the data! How can that be?

Measurement problems

The problem is that Rachidi is incorrect to claim the OPM “should reflect the monthly CTC payments.” While the OPM does count some sources of transfer income, like Social Security and unemployment compensation, it does not include refundable tax credits like the CTC or Earned Income Tax Credit. The Supplemental Poverty Measure (SPM), in contrast, aims to quantify the U.S. poverty rate post-tax and transfer, and thus does include refundable tax credits in its measure of household resources. As it turns out, the analogous, monthly estimate of the SPM produced by the Center on Poverty and Social Policy at Columbia University finds the CTC alone raised 3.8 million children from poverty in November 2021, a reduction in child poverty of approximately 30 percent!

As one commentator on Twitter put it,

Now, 30 percent is less than 39 percent but is still higher than the modest impact Rachidi initially suggested. The main reason the poverty impact has fallen short of initial estimates is that outreach efforts to help eligible families sign up for the CTC are still ongoing. While sign-ups have increased significantly since payments began in July, provisional IRS data suggests a few million children have yet to be reached. These tend to be the poorest families, so increasing program participation remains essential to maximizing the CTC’s poverty impact in the long term.

Our study of the CTC’s poverty impact was done using 2019 data and made no attempt to model the effects of the pandemic, much less the subsequent relief packages. For example, the latest Census data shows pandemic-related unemployment insurance (UI) benefits reduced the OPM poverty rate by 1.4 percentage points. A lower baseline poverty rate could have somewhat reduced the CTC’s impact on SPM poverty relative to estimates based on data that predate the UI expansion. Nevertheless, we stand by our estimate as a reasonable measure of the CTC’s poverty impact in a so-called “normal” year — arguably the more relevant statistic for evaluating whether the reform should be made permanent.

Unfortunately, this isn’t the first time Rachidi has confused AEI’s readers about the evidence around child benefits. This just happens to be a particularly egregious case — a blatant factual error at the height of a controversial legislative debate — and one which has yet to be corrected. While it’s far from general knowledge, the difference between the Official and Supplemental poverty measures ought to be basic knowledge for any U.S. poverty expert. One just hopes that Rachidi did not mislead on purpose and will be more attentive going forward.


Photo Credit: iStock