Reinhart-Rogoff One More Time: Why the 90 Percent Never Should Have Been Taken Seriously

Print
Saturday, 11 May 2013 10:09

As a general rule economists are not very good at economics. This is why almost none of them were able to recognize the $8 trillion housing bubble that sank the economy. (No, this isn't bragging, it only took simple arithmetic and basic logic.) Most economists are unable to conceptualize anything that someone with more standing in the profession did not already write about.

This is the only reason that the Reinhart-Rogoff 90 percent debt-to-GDP threshold was ever taken seriously to begin with. The point that I have tried to make in the past, apparently with little success, is that debt is an arbitrary number. It is not something that is relatively fixed, like the age composition of the population or the supply of land.

The country's debt is something that can and often is easily altered through simple steps. In this way the debt-to-GDP ratio can be thought of as something like the color of a house. Suppose Reinhart and Rogoff told us that people who lived in blue houses had 40 percent less income than people who lived in houses painted other colors. Presumably people would be skeptical of the results, but if their finding was really true, then we would probably want to encourage people in blue colored houses to paint them a different color.

In effect, Reinhart and Rogoff were making the same sort of claim about debt and GDP. Let me try to explain this in a way that even an economist can understand it.  

I have often pointed out that the value of long-term debt fluctuates with the interest rate. I didn't think this is a secret, but apparently few economists have followed what happens to bond prices when interest rates change. The point is that the value of our debt will plummet if interest rates rise, as the Congressional Budget Office and other forecasters expect. This means that we could buy back long-term debt issued today at interest rates of less than 2.0 percent for discounts of 30-40 percent. This would sharply reduce our debt-to-GDP ratio at zero cost.

Yes, this is really stupid, but if you believed the Reinhart-Rogoff 90 percent debt cliff, then you believe that we can sharply raise growth rates by buying back long-term bonds at a discount. It's logic folks, it's not a debatable point -- think it through until you understand it. 

The other obvious point is that we can reduce debt by selling off assets. We may not want to sell assets, but if we thought there was some huge growth premium from say, going from a debt-to-GDP ratio of 100 percent to a debt-to-GDP ratio of 80 percent, then we would be foolish not to offload some assets. These assets take a wide variety of forms, including the power to tax.

If that sounds strange, let's pick one that has often been discussed: carbon permits. Suppose we decide to issue carbon permits in a quantity where we expect they will lead to a fee equal to 1.5 percent of GDP from carbon emissions. If we sold off permits to emit carbon over the next two decades, then we could probably raise an amount that is close to 20 percent of GDP.

There are good reasons why we might want to do something like this in any case in order to slow global warming, but if we believed the Reinhart-Rogoff 90 percent curse, then this sale of permits would add more than 1 percentage point to annual growth, making us more than 20 percent richer by 2033. Of course countries that already taxed carbon, like those in the European Union, would not have this option, but Reinhart-Rogoff would put them in the exact same situation.

There are many types of asset sales that could quickly reduce debt-to-GDP ratios in big ways. Some countries have far more ability to make such sales than others because they hold far more assets. Again, if anyone believed the Reinhart-Rogoff story, then we should be looking to such sales, even if they might not otherwise be good policy.

In short, the whole idea of a 90 percent curse is just plain silly and everyone who knows any economics should have recognized this fact immediately. The biggest scandal of the Reinhart-Rogoff affair was not that they initially refused to share their data or their silly spreadsheet errors. The biggest scandal is that something so obviously absurd was ever taken seriously in the first place. 

 

Addendum:

Since the issue of interest rates and bond prices has been raised in comments and e-mails, I'll quickly clarify the point. Bonds carry a face value, meaning the amount that will be paid off when they reach maturity. This is what gets entered in our debt figure. However bonds also carry a market price, which fluctuates inversely with interest rates. The longer the term of the bond, the more its price will vary with interest rates. 

If interest rates rise, as just about everyone expects over the next 3-5 years, then the market price of the bonds we have issued in the current low interest rate environment will fall sharply. Since we count our debt at the face value of the bonds, not their market price, we could take advantage of the drop in bond prices to buy up trillions of dollars of bonds at sharp discounts to their face value.

The question is why would we do this, we would still pay the same interest? The answer is that the policy would make no sense for exactly this reason.

However, if we accept the Reinhart-Rogoff 90 percent curse, then reducing our debt in this way could make a great deal of sense. Suppose we can buy back debt with a face value of 60 percent of GDP at two-thirds its face value, or 40 percent of GDP. In our debt accounting we would have reduced our debt-to-GDP ratio by 20 percentage points. If this gets us below the 90 percent threshold then suddenly we can have normal growth again.

This is one of the reasons why a moment's thought should have allowed economists to realize that the Reinhart-Rogoff thesis made no sense. They were telling us that people who live in blue houses earn 40 percent less money. Serious economists should have recognized that this was absurd, even before they were able to get their hands on the spreadsheet.