Reinhart and Rogoff Revisited

2013 May 12
by SJ Leeds

Has it been long enough that you miss me?  I didn’t think so.


First, I received so many emails when I said that I was going to stop my weekly writing – I wasn’t able to respond to any of them.  But, I read them all and I really appreciate all of your comments.


Second, here’s a link to a pdf file that contains the resumes of three of my MBA students.  They all were members of The MBA Investment Fund, a $17 million investment fund that I oversee with Keith Brown.  They are all willing to move anywhere for the right opportunity in asset management (sell side or buy side).  Don’t hesitate to email them or me if you’re interested.


Now, on to an issue that a lot of people have written to me about…Reinhart and Rogoff’s paper and their mistakes.  Here’s a quick summary of the situation and my thoughts about it.



Reinhart and Rogoff (“RR”) are two Harvard professors who have written extensively about our current situation (high debt ratios, financial crises, etc.).  Probably their most influential paper was “Growth in a Time of Debt.”  The paper’s primary finding was that when debt-to-GDP exceeded 90%, the median country experienced a slowdown in growth of 1% and the average was a 4% slowdown.  (In other words, the middle country experienced slower growth, but there were some real disasters that really dragged the average down.)


Remember, a 1% slowdown in growth might not sound like much, but it’s huge!  Imagine that you take the U.S. growth rate from 3% down to 2%.  As a simple example, if we have 2% productivity growth, that means that we don’t need any more employees.  (In other words, our current workers produce 2% more and that’s how much we grew.)  In reality, our productivity is lower than 2%, but our growth isn’t that much higher than productivity.  So, it takes a long time to return to full employment.


A month ago, a doctoral student (and some faculty) at UMass – Amherst wrote a paper arguing that the RR paper contained errors and, as a result, had some incorrect conclusions.  The new paper is titled, “Does High Public Debt Consistently Stifle Economic Growth?  A Critique of Reinhart and Rogoff”.  The authors are Thomas Herndon, Michael Ash and Robert Pollin.


The new paper argues (among other things) that:


1. There was a spreadsheet error in the RR paper;


2.  RR selectively excluded some data (countries with high debt and growth that didn’t slow much); and


3. The weighting system that RR used was not justified.


As a result of these issues, the new paper argues that the average growth rate does not drop nearly as much as RR had said and that there is little discernable difference in growth at a 90% debt-to-GDP ratio.


Many people now argue that the RR paper was the basis of the austerity movement and that the paper has now been disproved.  Reinhart and Rogoff have admitted that there was a spreadsheet error but argue that their main finding was based on the median country (not the average country) and that their findings still hold.  Let me tell you my thoughts…


My Thoughts On This Issue

1.  The idea that there is one magical debt-to-GDP ratio (e.g., 90%) that will cause a country’s growth to slow is far too simplistic.  I think most of us knew this.  Certainly, in my policy class, I always spoke about this and I spoke about this in presentations I’ve given.


All countries are different and the amount of debt (relative to GDP) that they can handle differs.  Does a country issue the debt in its own currency or in a foreign currency?  How much of the debt is held by its own citizens as opposed to foreigners (who may flee)?  Is the country’s currency a reserve currency?  Does the country have growth opportunities?  These are just a few examples of other factors that also matter.


Relying on one single factor (like debt-to-GDP) is like saying that the home team will always win in a sporting event.  The home-field advantage is one factor (and it’s very important), but other issues are also important.


2. I’m always going to be wary of an empirical study that would have significantly different results if a few data points weren’t included.  (The new paper argues that RR left out some data points from New Zealand, Australia and Canada.)  Rather than relying on empirical findings (and arguing about whether New Zealand’s data should be included in these calculations), you should really be thinking about intuition.  Here’s my intuition:


If debt-to-GDP = 100%, that means debt is the same as GDP.  If interest rates return to their 30-year average (prior to the Fed’s zero interest rate policy), that would put rates around 5.7%.  That would mean our interest would equal 5.7% of GDP each year. Our tax revenue has averaged 18% of GDP for the past 60+ years.  So, if interest is eating up approximately 1/3 of our tax revenue, we’re not in a sustainable position.


3. The real problem (that much of the historical data misses because this problem didn’t exist in many of the past years) is that we have huge unfunded liabilities.  If they’re not changed, we will always be running a deficit.  If we’re running a deficit (and the “primary deficit” that we all discuss does not even include our interest expense!), when you add in our interest, we will have huge problems.  As our baby boomers start to retire in force, we’ll see this get worse.


4. To me, these numbers are pretty simple and pretty obvious.  But, just as simple and obvious is that we can’t just cut everyone’s Social Security and Medicare and think that it won’t affect anything.  First, we need to give people time to plan (and change the amount they save – and this will impact consumer spending).  Second, we need to recognize that this will impact retirement for millions of people.


The Social Security Administration says that among Social Security beneficiaries, 53% of married couples and 74% of unmarried persons receive 50% or more of the income from Social Security.  Among elderly Social Security beneficiaries, 23% of married couples and about 46% of unmarried persons rely on Social Security for 90% or more of their income.


5. Any of these changes will impact our growth.  Obviously, if we increase taxes (or the amount of income subject to Social Security payroll taxes), this will also impact our growth.  Add in the political issue (of cutting Social Security or raising taxes on an already-progressive program) and I’m going to have to take issue with anyone who says that Social Security is easy to solve.


6. Even this new paper, which criticizes RR, comes to a similar conclusion: that growth is impacted by debt.  This new paper shows that the average growth for countries with debt-to-GDP below 30% is 4.2% — just like RR.  With a ratio at 90% – 120%, the average growth is 2.4% (significantly higher than RR).  But, move above 120% and the average growth rate is 1.6%.  This might be more optimistic than RR, but this is still an ugly scenario.


7. I’ve read other research that came to similar finding about a slowdown in growth when the debt-to-GDP ratio exceeded 90%.  I’m curious as to whether these papers also had errors.


8. Reinhart and Rogoff had been seen as the heroes of the conservative fiscal movement.  As a fiscal conservative and social liberal, I would argue that their errors have done more harm than their research did good.  People are now (mistakenly in my view) using their errors to argue that the debt-to-GDP ratio doesn’t matter.  I believe it matters, we just don’t know when.  But, we never did know when and the biggest mistake was when people thought that there was a specific number that created a fiscal landmine.


Talk to you sometime in the future…


Have a great week.

If you enjoy this blog, please forward it to others who may be interested.

If you want to receive my occasional emails, here’s how:


1. click on this link  (or type into your browser)
2. toward the top right corner is a place to click on for email service — click and enter your email address
3. you will receive an email which will require you to click on a link to confirm that you want to be on the list

IMPORTANT: if you don’t receive the email in step 3 or you don’t click on the link, you won’t be on the list.  Sometimes, people who use corporate emails get blocked (it’s probably 50% of the time).  So if you don’t get the email, you know you need to use a personal email.



Comments are closed.