Did Last Year’s Debt Ceiling Debate Hurt the Economy?

Cool economists Justin Wolfers and Betsey Stevenson had an article in Bloomberg on Monday arguing that another debt-ceiling battle could “sink” the economy based on how costly the last one was.

Justin and Betsey believe the government needs to spend more money now to help the economy (after all, interest rates are insanely low right now), so they naturally oppose the Republican’s hypocritical efforts to restrain the debt ceiling. I believe this makes them biased to look for evidence that fights over the debt ceiling are bad for the economy. Meanwhile, I do not think the government should spend more money (after all, interest rates can go up rather quickly,) so I am biased to be skeptical of their evidence, which I believe is weak, cherry-picked, and contradictory.

Let’s look at the evidence they claim:

1. Record level of news articles about business uncertainty

2. A severe drop in the consumer confidence index

3. A roughly 50% drop in job growth

4. The Treasury downgrade of US debt

I do not understand what the number of “newspaper articles” has to do with anything. That may be some kind of indirect indicating proxy, but it is not empirical evidence of the economy being bad. Maybe the media reflects or influences the animal spirits running through the economy, but if it means something, it should be correlated with stronger, more direct data. I think it looks pretty weak to present it as one of their strongest pieces of evidence.

Consumer confidence data is no more direct. It may suggest that people are working less or spending less money, but as far as I understand it is not evidence of those things. Furthermore, the drop in the confidence data itself is not very convincing.

They claim it begins to fall “right around May 11, when Boehner first announced he would not support increasing the debt limit,” but the jagged drop from May to July is similar to a drop a couple months earlier, and is actually smaller than the total drop from the previous peak to the valley just before the shaded portion. How can you say one line is evidence of a costly debt ceiling battle when a half-dozen similar lines have no explanation?

Additionally, it has been two weeks since Boehner fired the first shots for the upcoming debate. But by Justin’s own admission, Gallup’s confidence numbers “remain strong.” Can you confidently assign a narrative to the beginning of the last fall when repeated conditions are not giving the same output?

Now the “freefall” plummet at the end of the shaded portion is strong enough that I can believe it has a narrative, and it correlates strongly with the height of the debt ceiling deal and subsequent downgrade. But where is the evidence that this drop in confidence hurt the economy?

Maybe it’s in the jobs growth data. There is a noticeable difference in the shaded debt-ceiling battle area compared to the areas before and after. But even here I am skeptical that we can confidently apply a convenient narrative to the data.

First, May is the worst month in the graph, before the debate really got going. If its bad numbers have an explanation other than the debt ceiling, how do we know that the next months also don’t?

Secondly, it contradicts the confidence numbers we just looked at. While these numbers were low, consumer confidence was as high as it had been the previous months. When confidence plummeted in August and remained low over the next few months, the job numbers went back to the numbers they had been at earlier. When both sets of data were low and high over different parts of the same time period, how can we confidently claim the low parts of each can be blamed on the debate?

Is it possible that they are just cherry-picking portions of data that fit their hypothesis? Why don’t they look at job openings, which had no drop during the debate? Or retail sales? (If you squint, there’s a tiny slope change around the time period in question, but I could blame it on a spike in gas prices as easily as someone else could blame it on the debate. Besides, it picks up again just as consumer confidence plummets.)

Not only do my cherry-picked data sets fail to correlate with the debate-caused disaster of Justin and Betsey’s cherry-picked data sets, but they’re actually more direct reflections of economic activity than their newspaper articles or confidence levels. How do we know all of this isn’t just noise in oceans of data?

Let’s close with the “permanent scar” of Standard & Poor’s downgrade of U.S. Treasuries. They say “a worse credit rating will probably raise the interest rate on our national debt.” It’s a reasonable claim, but there’s a reason they don’t look at the actual data of the actual debt being debated here; it doesn’t fit the story at all.

Interest rates are supposed to rise since there is less confidence that the US will repay its debts. The first arrow roughly marks the beginning of last year’s debt debate, and the second arrow roughly marks the end of it. Not only did interest rates drop over the doubtful period, but they continued to plummet after the deal and the downgrade, and are now touching record lows.

That’s why they can only say it will “probably” rise in the future. Is there any doubt that if interest rates on Treasury debt had risen, it would be confidently blamed on the debt ceiling debate? But it doesn’t just “not correlate”; it moves in the completely wrong direction!

“But that’s not fair,” you might say. “Treasury rates are low because Europe is in such bad shape that the current flight to safety is overriding any fears about US debt standoffs.” That may be right, but it’s precisely my point! How you can ignore all the data that contradicts your theory because it has “other factors” while confidently asserting that all the data that fits your theory has only the factors you decide it does?

Despite looking at weak indirect economic data, cherry-picking parts of the direct economic data, and completely ignoring the data about the US debt itself, they conclude that “the data tell us that a debt-ceiling standoff is an act of economic sabotage.”

Now I don’t know as much about economics as these two, and I will gladly recant if I have made any errors, but that sounds like hyperbole to me. I could sift through the data and confidently claim there is no evidence that last year’s standoff hurt the economy. At best we can all claim that the data is messy and noisy and we cannot confidently claim anything one way or the other. Unless, of course, we’re trying to reverse engineer the data to fit our narrative fallacy…