Here’s a simple way to tell if someone (like the Nat’ Restaurant Assoc.) is abusing numbers to mislead.

July 19th, 2019 at 7:44 am

One sure way to tell in someone is making a biased argument is showing up in various statements by those who oppose the proposal to raise the minimum wage. I wanted to be sure to elevate this dastardly ploy, as it’s a tell that someone is trying to win an argument based on their bias, not on the evidence.

Here’s a tweet from the National Restaurant Association, a group that’s honor bound to oppose the minimum wage, and here’s the same ploy from a Texas Republican member of Congress. In both cases, they exclusively characterize the CBO’s job loss estimate from the agency’s recent minimum wage report as “as many as 3.7 million.”

Now, if you’ve been following the debate over the CBO’s findings, you’ve probably heard the number of jobs lost cited as 1.3 million, not 3.7 million. Here’s how the budget agency summarized their results (my bold).

In an average week in 2025, the $15 option would boost the wages of 17 million workers who would otherwise earn less than $15 per hour. Another 10 million workers otherwise earning slightly more than $15 per hour might see their wages rise as well. But 1.3 million other workers would become jobless, according to CBO’s median estimate. There is a two-thirds chance that the change in employment would be between about zero and a decrease of 3.7 million workers. The number of people with annual income below the poverty threshold in 2025 would fall by 1.3 million.

Put aside the benefit-cost argument as to whether the gains to 27 million are worth pursuing given the estimated median loss to 1.3 million (I get into that here). My point here regards the practice of exclusively citing the upper bound.

First, it’s not wrong. The “as many as” phrasing is the correct way to characterize the upper bound. But it is clearly biased. It would be equally correct—and equally biased—to say the CBO found “as few as zero workers would lose their jobs from the increase.”

Economic estimates like this are highly uncertain and CBO gets extra credit for being explicit about the range. The more of that, the better (see, for example, the top figure on page 4 here where CBO gives us the range of possible outcomes for their long-term debt/GDP forecast). But publishing a range clearly offers ripe fruit to cherry pickers.

The moral of this little tale is simple. When someone—and it will almost always be an advocate for a position supported by their funders—uses exclusively the “as many as” frame, without giving the central estimate and the range, you can be sure they’re all about winning the argument and not about seriously considering the evidence. As such, they should be summarily ignored by anyone honestly trying to figure out what’s actually going on.

Print Friendly, PDF & Email

3 comments in reply to "Here’s a simple way to tell if someone (like the Nat’ Restaurant Assoc.) is abusing numbers to mislead."

  1. Chuck Sheketoff says:

    Spot on, Jared.


  2. Sue Hall says:

    Jared, is the median number of jobs lost actually calculated based on 1.3 million discrete positions being eliminated; or rather, is it calculated based on a median reduction in hours of approximately 2.6 billion (about 2000 hours per job per year)? If the latter, isn’t it possible that the loss of those hours might represent, as an example, 5.2 million minimum wage people being able to reduce their weekly hours from 50 to 40 and still coming out ahead, as they would be being paid at a higher rate? (Just trying to understand the math…) Thanks so much!


Leave a Reply

Your email address will not be published.