April 21, 2013
Reinhart-Rogoff data problems
The methods and conclusions of an influential paper by Carmen Reinhart and Kenneth Rogoff published in 2010 have recently been challenged by Thomas Herndon, Michael Ash, and Robert Pollin. Here I comment on both the details and broader significance of the dispute.
Let me begin by stating what I perceive to be the core policy question: are high sovereign debt loads something we should worry about, and if so, why?
The main reason that I personally am concerned arises from the fact that, for any level of the interest rate, a higher debt load means that the government will permanently need to spend more money just to pay the interest on the debt. This is not a matter for arcane debate, but rather is a consequence of the most basic arithmetic. At the moment, the interest rate on U.S. government debt is extremely low, so that despite our high debt load, the government's net interest cost is currently quite reasonable. However, most projections call for interest rates to rise over the next few years, and the most recent assessment by the CBO notes that consensus interest-rate forecasts and existing fiscal legislation imply that within a few years, the U.S. interest cost will be bigger than the entire defense budget, and bigger than all of non-defense discretionary spending.
Higher taxes will be needed to make those higher interest payments, and the adverse economic consequences associated with distortionary taxes are well understood. Moreover, as the debt load gets larger, the government's creditors usually start to require a higher interest rate, which makes the interest burden of the debt even bigger. This creates a potential adverse feedback loop that can lead (and in many unfortunate historical cases, has led) to a major funding crisis. That higher debt loads are associated with higher interest rates has been found by many different researchers using many different data sets and methodologies. For example, Hibiki Ichiue and Yuhei Shimizu (2013) found these effects in a panel of 10 advanced economies over 1990-2010, of which Germany is the only eurozone member. Vincent Reinhart and Brian Sack (2000) found them in the experience of the G7 before establishment of the euro (1981-2000). Thomas Laubach (2009) noted that higher CBO projections of 5-year-ahead debt levels for the U.S. are associated with higher 5-year-ahead forward rates over 1976-2006. And Greenlaw, Hamilton, Hooper and Mishkin (2013) found a relation between debt levels and borrowing costs in the last decade's data from 20 advanced economies. There is also evidence of nonlinearity in this relation, with debt levels mattering more as they get bigger and as the country's current-account deficit grows; see for example Ardagna (2004), Baldacci and Kumar (2010), and Greenlaw, Hamilton, Hooper and Mishkin (2013).
I mention all this to highlight that none of the above issues or evidence has anything at all to do with the data, questions, or conclusions in the paper by Reinhart and Rogoff (2010) that has recently come into some controversy. Instead, what Reinhart and Rogoff studied in their paper was whether higher debt loads are correlated with lower growth rates of real GDP.
The specific evidence reported in Reinhart-Rogoff (2010) came from three different data sets. First, they followed 20 individual countries for up to two centuries, calculating the average growth rate for that country in the years when debt levels exceeded 90% of GDP and average growth rates for years with other debt levels. They reported these separately for each individual country (see Table 1 of Reinhart and Rogoff (2010)), and found that growth rates were slower when debt levels were higher. Second, they combined data from a panel of 20 different emerging economies over 1970-2009 and found that growth rates were slower when debt levels were higher. The recent critique by Herndon, Ash, and Pollin (2013) did not discuss either of these first two claims. Instead, their critique concerns Reinhart and Rogoff's analysis of a third data set, a panel of 20 advanced economies over the last half-century.
Let me first jump to the bottom line, and then review the details. Reinhart and Rogoff originally claimed that for this last of the three data sets, real growth rates were around 4% for low debt levels, 3% for moderate debt levels, and -0.1% or +1.6% for debt levels above 90%, with the latter difference depending on whether the aggregation was done using the mean or the median. Herndon, Ash, and Pollin claim that when the numbers are correctly tabulated, real growth rates are around 4% for low debt levels, 3% for moderate debt levels, and 2.2% for debt levels above 90%, with the latter inference based solely on the mean. The table below reproduces Reinhart and Rogoff's summary of the differences in the bottom-line claims.
RR AER (2010)
|0 to 30||4.1||4.2||4.2||NA|
|30 to 60||2.8||3.9||3.1||NA|
|60 to 90||2.8||2.9||3.2||NA|
|RR AER (2010) (Table 1)|
|0 to 30||3.7||3.9||NA||NA|
|30 to 60||3.0||3.1||NA||NA|
|60 to 90||3.4||2.8||NA||NA|
|RRR JEP (2012),|
Now let's take a look at the details by which Herndon, Ash, and Pollin come to their numbers. First, they found a dumb error in Reinhart and Rogoff's spreadsheet-- Reinhart and Rogoff left the first 5 countries in the alphabet (Australia, Austria, Belgium, Canada, and Denmark) out of the set of cells selected for averaging. This is a numbskull error, but it turns out it would only have changed the estimate they reported by a few tenths of a percent.
The major differences come from a difference of opinion about how one should summarize the mean for these data. For example, the U.S. spent 4 years in this sample with debt levels above 90% of GDP, while Greece spent 19 years. How should we combine these two sets of observations?
One view one could take is that the expected growth rate when a country has a high debt level is a single number across all countries, that is, you expect the real growth rate for Greece when its debt is 90% to be exactly the same number as the real growth rate expected for the U.S. when its debt is 90%. If you further believed that the variance of Greek growth around this mean is the same as the variance of U.S. growth around this mean, then the correct thing to do would be to act as if you have 19 observations on the number of interest from Greece and 4 observations from the U.S., and take a simple average of those 23 numbers. In other words, you should base most of your inference on the data from Greece, because that is where you have the most observations. This is the approach that Herndon, Ash, and Pollin insist is the correct one to use.
Another view you could take is that the expected growth rates for the U.S. and Greece would be different even if the two countries had the same debt levels. From that perspective, there is a different expected growth rate for each particular country when it gets to the 90% debt level, and our goal is to estimate what that number is for a typical country. That view seems to underlie the method chosen by Reinhart and Rogoff, which was to estimate an average growth rate when debt is greater than 90% for the U.S., a separate average growth rate when debt is greater than 90% for Greece, and then take the average of those averages across different countries.
One could go a step further and spell out a complete statistical model of the view just espoused, for which the optimal way of combining different observations would weight the Greek average more heavily than the U.S. average (because the Greek average is estimated with greater precision), but not 19/4 times as heavily as Herndon and coauthors want (because the Greek average is estimating something different from the U.S. average). The optimal statistical estimate from that perspective would be somewhere in between the Reinhart-Rogoff number and the Herndon-Ash-Pollin number.
In any case, as seen in the table above, whichever number you used, you would still conclude that higher debt loads are associated with slower growth in the postwar advanced economy data set, just as they were in the postwar emerging economy data set, just as they were in the centuries-long individual country data sets, and as also was found to be the case in separate analyses of yet other data sets by Cecchetti, Mohanty and Zampolli (2011), Checherita and Rother (2010), and the IMF (2012), among others.
A quite separate and in my mind much more legitimate question is whether this correlation can be given a causal interpretation-- is it high debt levels that cause slower growth, or slow growth that causes debt to accumulate? One can (and should) be persuaded that the correlation is real, but still be in doubt as to what it signifies.
But none of these concerns change the basic reality that is summarized in the first graph above.
Posted by James Hamilton at April 21, 2013 06:50 AMdigg this | reddit
Higher taxes will be needed to make those higher interest payments
Higher GDP means the taxes needed to make those interest payments will be a smaller share of GDP.
whether this correlation can be given a causal interpretation
See http://www.nextnewdeal.net/rortybomb/guest-post-reinhartrogoff-and-growth-time-debt It seems slow growth leads to higher debt.
Posted by: foosion at April 21, 2013 07:03 AM
Thank you! Your summary seems clearer to me than the original articles. It also seems very clear that high debt spells real future problems for the USA. The causation-correlation issue of GDP growth and debt may be a distraction from the real problem, the level of debt as you mention. Certainly at the micro-level, corporations have less liquidity, less solvency and lower flexibility as debt levels climb compared to equity.
Posted by: AS at April 21, 2013 07:40 AM
Why is this claim correct: "Higher taxes will be needed to make those higher interest payments,..."
As I understand things the Fed these days is paying off Uncle Sam's debt about as fast as the Treasury is borrowing more. The Fed doesn't use tax money to pay off the Treasury Department's debts.
If we consider the Treasury to be Uncle Sam's right pocket and the Fed to be his left pocket, we have the Treasury borrowing money, which it puts in Uncle Sam's right pocket. Then the Fed pays back the person who loaned the Treasury that money. The Fed puts the T-bill in Uncle Sam's left pocket. If for some reason you chose to count that left pocket T-bill as part of US Debt Held By the Public, then you will fret about how the Treasury Department, Uncle Sam's right pocket, will pay it back. But the T-bill is already paid back. It was not paid back with tax dollars or even with dollars obtained by borrowing more money.
The situation, it seems to me, is different for the state of California, which has no central bank, or for the nations of the Eurozone.
For some reason of arcane accounting, the Fed arranges for Uncle Sam to take money from his right pocket and put it in his left pocket to pay interest on the left pocket T-bills. Then at the end of the year, the Fed takes this interest money from Uncle Sam's left pocket and puts it in his right pocket. But the money owed to the person who loaned the Treasury some money has already in that person's bank account. She has been paid back by the Fed without tax dollars.
Also, the Treasury doesn't have to collect taxes to pay off maturing bonds. It may just issue new bonds, using that money to pay maturing bonds.
Thank you for your thoughtful answers to earlier questions I have asked.
Posted by: Left Coast Bernard at April 21, 2013 08:04 AM
"Higher taxes will be needed to make those higher interest payments"
Assumes revenue can't increase through growth, which is absurd.
Posted by: cahuenga at April 21, 2013 09:26 AM
You left out the other major data cherry-picking error by R&R, and the confusing of correlation and causality. Intentionally misleading your readers, or just more Reinhart and Rogoff style numbskull errors?
From Konczal: "...they exclude Australia (1946-1950), New Zealand (1946-1949), and Canada (1946-1950). This has consequences, as these countries have high-debt and solid growth. Canada had debt-to-GDP over 90 percent during this period and 3 percent growth. New Zealand had a debt/GDP over 90 percent from 1946-1951. If you use the average growth rate across all those years it is 2.58 percent. If you only use the last year, as Reinhart-Rogoff does, it has a growth rate of -7.6 percent."
You look at this link, and it's clear the dropoff in growth was correlation, not causality. This is really the main point, and it's a point that marks the authors as psuedo-scientists and not scholars.
With Respect, -TV
Posted by: Thorstein Veblen at April 21, 2013 09:50 AM
The problematic concern of RR is that even if you accept their assumptions and data, it still doesn't justify their prescription. Is growth slower due to the burden of debt, or is it slower due to the burden of reducing that debt? The former speaks for austerity, the latter indicts it. Whatever you believe, this does not justify austerity and certainly not currently.
Posted by: Lord at April 21, 2013 10:19 AM
Isn't the real issue whether or not one believes that short run mass unemployment trumps or is trumped by concerns about long run debt issues?
What made RR so powerful was the bright line claim that the magic number of 90 % must never be breached -- that, as George Allen told us, the future is now. Come to think of it, that didn't work out too well for the Redskins did it?
Posted by: malcolm at April 21, 2013 11:35 AM
AS The causation-correlation issue of GDP growth and debt may be a distraction from the real problem, the level of debt as you mention.
I don't think that's an entirely fair summation of JDH's post. For example, as JDH says, "...A quite separate and in my mind much more legitimate question is whether this correlation can be given a causal interpretation-- is it high debt levels that cause slower growth, or slow growth that causes debt to accumulate?
The causation argument is very much central to the discussion. Now it is true that the 2010 R-R paper was properly cautious about pushing the argument that a 90% debt/GDP threshold was a critical and causal tipping point; but in their WSJ op-eds and various media events their claims were a lot more aggressive and certain about the magic power of 90%. So I don't really think they can hide behind the nuances of a non-peer reviewed AER paper. BTW, the R-R experience is a good example of what happens when academics try to "sex up" their findings by carrying their findings over to the WSJ op-ed page. The WSJ op-ed page is where otherwise good economic analysis goes to die.
One thing that hasn't gotten enough attention is the large difference between the mean and median at the 90+% threshold. There's clearly a hell of a lot of skew in the data, and if you look at the HAP charts you can see how everything just goes to scatter after 90%. So how should we interpret that scatter? As evidence of greater risk beyond 90%? Or as evidence that we're probably missing important variables?
JDH the adverse economic consequences associated with distortionary taxes are well understood.
I think this goes too far. We don't really even understand growth theory all that well, so it's a bit of a stretch to say that we "well" understand the distortionary effects of taxes. My problem is with the adverb "well." Yes, there are deadweight distortionary effects of taxes, but taxes can also correct market distortions. The most you're entitled to say is that ceteris paribus certain kinds of taxes can create distortionary effects and reduce welfare; however, our real world ability to observe, measure and evaluate the distortionary effects is very limited and the empirical data is inconclusive.
Here's what I regard as the two stupidest sentences in the R-R paper:
Our topic would seem to be a timely one.
Public debt has been soaring in the wake of the
recent global financial maelstrom, especially in
the epicenter countries.
Timely? Huh? Publishing their paper in 2006 would have been "timely" because it would have addressed the need for fiscal tightening at a time when fiscally contractionary policies would have been appropriate. Their paper is distinguished by it's godawful bad timing because it gave clueless policymakers the intellectual cover to make stupid policy mistakes.
Posted by: 2slugbaits at April 21, 2013 12:42 PM
Thanks, Jim, this is one where I agree with you wholeheartedly. I'm sure a lot of advocates of fiscal and monetary stimulus will loathe you for it.
We should acknowledge though that in developed countries we are in an unusual situation where economies aren't reacting in the usual ways to fiscal and monetary policies. This changes the debate, at least for developed countries.
In a nutshell, the principle that counsels in favor of fiscal and monetary prudence is that lacking one or both always leads to high interest costs or high inflation, or both.
That has held true for most of history, and it still holds true in emerging markets. But it hasn't been holding true recently in developing countries. Japan has had loose fiscal policy coupled with loose monetary policy (at least occasionally, depending how you define "loose") for a very long time now, and still has no inflation and relatively low interest costs.
Keep in mind actual interest costs must be calculated on net debt, not gross debt, and need to take into account that when sovereign debt is held by the central bank, the only interest actually paid by the government is any IOR on the corresponding bank reserves (the government pays interest to the CB, and the CB then refunds that interest net of any IOR to the government). The actual cost of interest to the Japanese budget is much less than what is being paid on gross debt.
Likewise in the US, there is only a bit more than $10bn of generally shortish-term federal debt held outside the Fed and federal government, with a very low average interest rate, plus $1.8bn of generally longish-term federal debt held by the Fed, on which the government effectively pays only 0.25% interest.
If we were living in an emerging market country we wouldn't be having this debate. It would be obvious that higher debts means higher costs means crowding out and lower growth, and the only way to get out of high debts would be austerity or inflation, which we would know from experience are both about equally unpleasant.
But as long as developed countries are able to issue currency and have their central banks purchase much or even all of their net issuance of public debt without generating much inflation, there is no current financial pressure on developed country governments to restrain debt. Calling for debt restraint is something that us more economically conservative folks do because we don't believe this condition will last and we think extrication from high debts would be uglier than current macroeconomic and market conditions would suggest.
Posted by: Tom at April 21, 2013 01:34 PM
The argument doesn't bother me. I expect people would make versions of "cut now" no matter what. Andrew Mellon back in the 1929 didn't know crap about regression and politicians today just use findings to justify what they already intend to say.
But I am troubled by two things.
1. You skip over, as a comment or two has noted, that they excluded years of data. The problem is they've said this wasn't intentional and wasn't a mistake, that the data was too new to them to be included in time. OK. That creates a huge problem for a person of integrity: how in the name of heaven do you justify taking one year of New Zealand when you've admitted you had other years? You can see from the numbers they contradict. So you didn't have enough time to vet the earlier years. How does that justify including the one year? A person of integrity would not include any of those years until they could all be reviewed, especially since their method heavily weighted that 1 included year.
2. It's been 3 years. I'll repeat that: it's been 3 years. Isn't 3 years enough time to look at the omitted years? Don't you have some obligation to complete the work? Don't you especially have an obligation to be honest when you're continuing to give speeches and presentations, including to US senators, in which you use these findings? What are we supposed to believe, that they never bothered to look at the data they say was too new for inclusion or that they looked and decided not to say anything because they were getting so much attention?
The economics profession is trying to find ways to justify results and to skip over what should be treated as a massive failure of personal and professional integrity.
Posted by: jonathan at April 21, 2013 01:54 PM
An elegant quarrel, if successful one could even lose track of the essence. that is:
The ability of the sovereign risks to service the principal repayment « This time is different » rejuvenated memories, when reminding that past same public debts profiles triggered defaults and debts moratoria.
Since the long time series are involved, one still expect to be secured when it comes to homogeneous GDPs comparisons and the GDP deflators to be scrutinized with the same attention. Countries do not share the same elasticity to export, comparing Public debts to GPD growth relationship between, resources based mineral countries, trading countries or industrial based countries, may not drive to the same results. One may hardly understand how a weighing among them can provide for coherent result.
Pr Hamilton « Why I'm more worried than Paul Krugman about the U.S. debt burden » assumes in a subsequent paper Greenlaw, Hamilton, Hooper and Mishkin (2013) interest rates stability to be corner stone of the empirical findings. Assuming that interest rates are independent variables of the debts profile is a given for very few countries.
« 1-percentage-point increase in U.S. debt-to-GDP was associated with a 3-to-4-basis-point increase in the 5-year-forward 10-year rate. «
Have we ever read that a cross relation between interest rates and derivatives was 0 before accepting above relationship?. The question addresses specifically the nature, the duration, the amounts, the proprietorship of the derivatives and their stabilisation effects on interest rates, when the above paper assumes interest rates for granted under natural supply and demand. It is unsure that all countries have the ability to carry among their various domestic banks 660 Trillions USD derivatives
Whilst the list of comparative and unfair advantages when it comes to influence the interest rates pattern and therefore the GDP growth in relation to Public debts and GDPs growth could be made much longer.
Reading the referred documents comes as a disturbing revelation of a uni-dimensional world tropism.
Posted by: ppcm at April 21, 2013 01:55 PM
"...for any level of the interest rate, a higher debt load means that the government will permanently need to spend more money just to pay the interest on the debt. This is not a matter for arcane debate, but rather is a consequence of the most basic arithmetic. "
My understanding of research published (separately) by authors such Greg Mankiw and Tom Sargent is that the United States has, on average, faced effective financing costs on its outstanding debt that were consistently lower than its economic rate of growth (and that this difference persists even if one includes the 1930s in the sample period.) In that sense, the current period (of very low financing costs but slow growth) is not so much of an anomaly; interest costs should eventually be expected to rise, as should growth rates.
I'm therefore uncertain how to interpret the above statements of Prof. Hamilton's. In saying that the govt. "will permanently need to spend more money" to pay interest on the debt, does he mean to imply that the govt. debt cannot be rolled over? (Whether true or not, that would not be an implication of basic arithmatic.)
"That higher debt loads are associated with higher interest rates has been found by many different researchers using many different data sets and methodologies."
I'm also uncertain how to interpret these results in the case of the US. I would have thought that those higher debt loads are typically measured relative to GDP. If that is the key factor, are we right in thinking that the danger upon which you Americans should focus is the debt-to-GDP ratio rather than the defict per se? Regarding the cost of finance, I'm also curious whether Prof. Hamilton believes that markets distinguish between debt issued in a country's domestic currency and that issued in a foreign currency when it comes to the interest-rate & debt relationship.
Posted by: Simon van Norden at April 21, 2013 02:18 PM
Tom Calling for debt restraint is something that us more economically conservative folks do because we don't believe this condition will last...
"Debt restraint"? What a peculiar phrase. You make it sound like running up debt is some sinful pleasure and the role of policymakers is that of good Puritan fathers admonishing the community to purge itself of indulgent spending. As though we're running up sovereign debt today because we just don't have enough self-discipline. This sounds like a morality play posing as economics.
Conservatives would have a lot more credibility on debt concerns if we had been hearing these lamentations back in the mid-naughties when it might have made a difference. And at a time when fiscal contraction would have been good macroeconomic policy.
We need fiscal stimulus today because the economy is stuck at something that is beginning to look a lot like a low output, low growth equilibrium with a lot of unemployed resources. The right question isn't whether or not we'd be better off with full employment and low debt versus full employment and high debt. That's a false choice because we aren't at full employment. The relevant question is whether we want to bring down unemployment and have high debt or would we rather have persistently high unemployment and high debt. The reason we have high debt/GDP levels today is because we made some very bad tax cut and spending decisions 10 years ago. After the Great Recession it was a foregone conclusion that sovereign debt was going to increase. What we've learned from the Euro (and Wisconsin) experience is that austerity in the wake of a financial recession only increases the debt burden.
Posted by: 2slugbaits at April 21, 2013 03:04 PM
Dr. Hamilton, thank you for your thoughtful efforts to use this blog as a teaching and learning tool. I have no expertise in macroeconomics, other than as a well-educated market participant who has paid attention to the macro debates. However, I have significant experience in corporate debt management, which I think has something to say on this issue.
The Debt/GDP measurement is analogous to a debt/EBITDA or debt/sales ratio in corporate finance. (The latter is a closer analogy, but is less useful in practice, so it's almost never used.) In my world, a ratio like this would not be applied indiscriminately across all companies, because different industries are fundamentally different in the source of their revenue, the cost of producing that revenue, the capital investment needed to create and sustain it, and the predictability of the revenue stream.
Let me give an example. The classic case of leveraged finance in North America is the cable TV industry. Throughout the 1980s and 90s, this industry carried debt levels that most investment grade companies considered suicidal. However, cable TV companies, after a high initial capital investment, were able to generate a stable diversified stream of free cash flow protected by locally regulated monopolies. During the LBO boom, corporate managers embraced debt as a way to improve the efficiency of their organizations. The debt level banks and investors considered reasonable -- and the interest rate premium charged -- was fine-tuned based on the economic fundamentals at the firm and industry level. Size, diversification of revenues, flexibility of costs, exposure to commodity prices, access to the public capital markets, competitive position, growth opportunities, regulatory and technological risks… all these were taken into consideration in determining the maximum sustainable debt in a particular circumstance. And the supporting financial models were put into an excel spreadsheet triple checked for accuracy.
My biggest critique of RR, then, is not their amateur spreadsheet error, but the presumption that an aggregate dataset of countries as different as the US and Greece has anything meaningful to say about managing an individual economy responding to a crisis. OK, so we've learned that, in general, too much debt can be harmful, especially in a high interest rate environment. Duh.
To be useful, this analysis has to go much deeper. Is US GDP more resilient and sustainable than that of Greece or Canada or Japan or New Zealand? Does the borrowing country have access to an efficient and liquid capital market? Does it matter how the funds borrowed are actually invested? Do immature, high growth countries carry lower debt because the risk is higher, or do they have higher growth because they have less debt? Can the borrower control its own costs (interest, currency, health care, defense)? Is high debt putting the country in a competitive disadvantage, or is it the result of a widespread downturn that affects all borrowers at the same time?
I don't mean to be disrespectful of your expertise, Jim, but until I see an actual analysis that takes issues like this into consideration in establishing fiscal benchmarks for an economy, I will assume that the markets are the best judge of the appropriate debt level for any individual country.
Posted by: Tom in Wis at April 21, 2013 04:53 PM
Spending can always be reduced by prudent policy if in fact growth slows so I don't buy the argument that slow growth causes debt to rise to the point economies implode. That argument just makes no sense at all unless you have despots at the helm. I've never heard anyone that doesn't try to justify the excess spending by some rhyme or reason...taking it to the point of implosion is nothing more than politics run amok by insane policy makers and make belief economist.
Posted by: Hondo at April 21, 2013 06:16 PM
Let me just add that we're somewhat lucky in the US that Japan has gone so much further down the road we're traveling. We'll get to see what our future likely holds by what happens next in Japan.
If the Japanese central bank manages to create stable ~2% inflation amid some real growth while continuing to buy most or all of the government's net issuance of debt, and thus hold down interest rates on that debt to below the rate of inflation, then I suppose it will be a brave new world and no one in the developed world will worry about public debt anymore. The government borrows all it wants from the CB, the CB issues that much currency to the private sector, the private sector spends some but mostly just accumulates the currency. Interest on debt is cheap (basically equal to IOR on reserves) almost no matter how big public debt grows. In short, the MMT/NGDP-targeter dreams come more or less true, at least in Japan. If so, something similar will almost certainly be tried here.
Or maybe Japan's stimulus will create too much inflation, and Japan will have to choose between letting inflation run loose, or accepting austerity in order to rein in inflation and public debt - and perhaps suffering a debt crisis if it fails. Those are the choices that have been forced upon every emerging market that has attempted such a scale of deficits and monetary stimulus. Perhaps that might prompt us to get busy deciding before our options get worse.
Or perhaps, Japan's monetary stimulus will turn out to be a dud, and despite all the money supply growth, CPI and GDP will flatline. Then we'll be in a not-so-brave new world where both traditional and newfangled theories have failed. Then I'm not sure what we'll do here, but at least we'll have a good idea where we're headed with current policy.
Posted by: Tom at April 21, 2013 06:22 PM
@slugbait - Actually I'm an economically conservative liberal, and I agree with much of what you write about the Bush era and earlier. I suggest you look back on our highest growth period in recent history, the late 1990s. It was a time of nearly no deficits and high private investment, exactly as classical liberal theory suggest go together. They called it "the productivity miracle." Granted it was also a time of credit bubble and misguided deregulation, but no era in our recent history is perfect. I don't agree at all with your analysis, and ultimately I think such thinking will probably be the undoing of much liberal progress. If liberals don't take responsibility for fiscal consolidation, Republicans will.
Posted by: Tom at April 21, 2013 06:31 PM
Very good analysis.
On a slightly different note, here's my Friday interview with Platts Energy Week re: outlook for the Gulf of Mexico.
Posted by: Steven Kopits at April 22, 2013 06:29 AM
Dear Professor Hamilton,
A small comment about your post. It is true that there is a negative correlation between debt and growth in advanced economies (as you say, we don’t know about causality). However, the evidence for threshold effects is less clear. Andrea Presbitero and I have a survey that, among other things, discusses the threshold issue (Menzie mentioned our paper in this post: http://www.econbrowser.com/archives/2013/04/no_time_for_aus.html)
Posted by: ugo panizza at April 22, 2013 07:12 AM
Tom in Wis at April 21, 2013 04:53 PM: says "My biggest critique of RR, then, is not their amateur spreadsheet error, but the presumption that an aggregate dataset of countries as different as the US and Greece has anything meaningful to say about managing an individual economy responding to a crisis." Let me reiterate that the underlying statistical approach adopted by RR takes the view that the expected growth rate for the the U.S. when debt exceeds 90% is a different number from the expected growth rate for Greece when debt exceeds 90%. By contrast, the underlying statistical approach favored by HAP takes the view that the expected growth rate for the U.S. when debt exceeds 90% is the same number as the expected growth rate for Greece when debt exceeds 90%. For HAP to insist that the RR assumption is an "error" is, in my view, absurd.
jonathan at April 21, 2013 01:54 PM: says "It's been 3 years. I'll repeat that: it's been 3 years. Isn't 3 years enough time to look at the omitted years?" Please note that in their paper with Vincent Reinhart published in the Journal of Economic Perspectives in 2012, Carmen Reinhart and Kenneth Rogoff analyzed the full set of observations from Australia 1852–2011, Canada 1871–2011, and New Zealand 1861–2011. As far as scholarship integrity was concerned, wouldn't you have expected HAP to have acknowledged this in their paper? Because HAP did not even cite the JEP paper in their note, you assume that RR never followed up. And as a result of HAP's intentional omission of relevant information, you conclude that there is an issue of integrity of scholarship on the part of RR.
You've been hoodwinked.
Posted by: JDH at April 22, 2013 07:34 AM
Others have ably critiqued other things you said, but I'm concerned about this:
"Instead, what Reinhart and Rogoff studied in their paper was whether higher debt loads are correlated with lower growth rates of real GDP." True. They would have been in for less criticism if they had stuck with "correlated with". But they didn't.
And then you write: "They reported these separately for each individual country (see Table 1 of Reinhart and Rogoff (2010)), and found that growth rates were slower when debt levels were higher. Second, they combined data from a panel of 20 different emerging economies over 1970-2009 and found that growth rates were slower when debt levels were higher."
Do you not see that you are implying causation with your choice of words? This is part of the problem they created. They didn't say in their paper which way the causation went, but they strongly suggested their views in their testimony. And here you are doing basically the same thing.
Please, stick with "correlated with", and if you don't know the causation, it would be reasonable to alternate variables in those two statements. So, "found that slower growth rates correlated with higher debt levels" and "over 1970-2009 and found that higher debt levels correlated with slower growth."
Posted by: KJMClark at April 22, 2013 07:47 AM
Interesting post, Prof. However, is the question really "whether this correlation can be given a causal interpretation-- is it high debt levels that cause slower growth, or slow growth that causes debt to accumulate?" Greece, and more than a few EU countries, are notorious for mis-stating their debt -- just look at how they mis-reported it to become a euro-based economy, and how they aided and abetted banks that helped them disguise loans within financial products they purchased to put an even greater debt burden on their citizens. You can no more trust Greek debt numbers (or any of the other known mis-reporting sovereigns) than you would trust the title received from the guy who sold you the Brooklyn Bridge.
These countries also have extraordinarily low tax receipts -- a small percentage of those with tax liabilities actually pays taxes (remember the tax on pools in Greece?). The debt in these countries grows and endures because they are dysfunctional. Not to say the U.S. isn't becoming similarly dysfunctional, as the whole "carried interest" debate illustrates, or the ability of super-high earners to pay lower rates than their secretaries. But that's a whole 'nother story, as they say.
It's impossible to see what historical facts can be drawn from the data of corrupt governments, which never have and never will accurately report the true state of their economies' indebtedness or tax situation.
Posted by: markets.aurelius at April 22, 2013 08:17 AM
Excellent! Thank you for words of reason.
Posted by: Ricardo at April 22, 2013 08:33 AM
“Between public debt/GDP ratios of 38 percent and 117 percent, we cannot reject a null hypothesis that average real GDP growth is 3 percent.”
Herndon et al.
Hardly a cliff at 90%. R&R are defending the indefensible.
Posted by: mobk at April 22, 2013 09:34 AM
mobk at April 22, 2013 09:34 AM: Please provide a specific citation in which RR ever used the word "cliff".
Posted by: JDH at April 22, 2013 09:38 AM
Anyone really think you can run a democracy for long when you just kick people off the boat because your spreadsheet says there might be a danger of inflation somewhere down the road.
Posted by: rd at April 22, 2013 09:47 AM
The best analysis about the weight criteria adopted by RR that I saw is:
"So if I take two baseball batters and one of them has 500 at bats with a .200 BA and the other has 1 at bat with a 1.000 BA, then their combined batting average in .600 !!!"
Yes, I think I got hoodwinked
Posted by: Marco at April 22, 2013 10:21 AM
Professor, how about here???
Posted by: Marco at April 22, 2013 10:25 AM
The one point that seems to be lost in the controversy is that whatever you think of either R&R or the UMass students or whatever you think about correlation vs causation, there is still nothing that shows a nation with a high debt / GDP ratio continuing to run large deficits that miraculously result in the ratio going down growing its way out of the debt / ratio. The multiplier isn't big enough to make that happen.
Posted by: Mark T at April 22, 2013 10:28 AM
Marco at April 22, 2013 10:21 AM: You use the word "combine" rather loosely. Let me suggest being more focused.
Suppose you have a team consisting solely of these two players, and you want to predict what the batting average for the team is going to be. Or more generally, you want to predict the batting average for a typical baseball player. How do you answer that question?
As I explained in detail in the post above, the optimal way to answer that question would be to downweight the data for the batter for whom we only have one observation, because the historical average for batter B is going to be a poorer predictor of his next outcome than the historical average for batter A will be for predicting batter A's next outcome. The correct number for predicting the average outcome for the team would be somewhere in between 0.600 and 0.200, much closer to 0.200 than to 0.600, because for your example the ratio of observations is 500 to 1. But the optimal predictor of the team's future average would be larger than 0.2016.
Marco at April 22, 2013 10:25 AM: I did not find the word "cliff" in the article you linked to-- did you? What I found was the following: "We aren't suggesting there is a bright red line at 90 percent; our results don't imply that 89 percent is a safe debt level, or that 91 percent is necessarily catastrophic." What I further found was that, when they refer to their empirical results, their reference is to the median (which HAP never challenged) rather than the mean (to which HAP are devoted exclusively).
Let me suggest a simple standard for academic integrity, a word commenters above are throwing around with casual venom. If you want to say, "RR claimed x", then fill in "x" with a direct, exact quote. If instead "x" is your own paraphrase, intentionally selected to exaggerate for purposes of attacking your straw man, then please don't describe your exercise as pursuit of academic integrity.
Posted by: JDH at April 22, 2013 11:08 AM
James Hamilton: "Higher taxes will be needed to make those higher interest payments, and the adverse economic consequences associated with distortionary taxes are well understood."
--Why is this presented as unquestionable truth, instead of being acknowledged as the indication from some theoretical models, with econometric verifications that are disputable?
Posted by: Lee A. Arnold at April 22, 2013 12:13 PM
" If you want to say, "RR claimed x", then fill in "x" with a direct, exact quote."
R-R( from the Bloomberg piece):
"Those who remain unconvinced that rising debt levels pose a risk to growth should ask themselves why, historically, levels of debt of more than 90 percent of GDP are relatively rare ..."
Virtually every speaker of English will tell you that the above quote clearly implies (without explicitly saying so) that
a. Causality runs from debt to growth and
b. There is something special about the 90% ratio.
Rogoff (Project Syndicate):
" we find that very high debt levels of 90% of GDP are a long-term secular drag on economic growth"
Reinhart(testimony to Congress):
"If it is not risky to hit the 90 percent threshold, we would expect a higher incidence."
You still maintain R-R never implied anything about causality or a magic 90% debt/GDP ratio?
Posted by: arthur at April 22, 2013 12:38 PM
If you mean that honesty in economics is not amending your prior work but instead doing other work in which your prior omissions or manipulations are subsumed to be made invisible, then I concede. Of course what you point out is that they did in fact review the data but saw no reason to post anything about that anywhere. Not on the website. Not in an addendum. Nowhere. That's a rather low barrier to cross to be considered honest. Or complete.
Posted by: Jonathan at April 22, 2013 12:41 PM
But it's not .600. Assuming that they are both wrong (I'd rather use HAP modelling, but I digress), my problem is that the professor chose not to emphasize the problems in RR, only in HAP (I know people are piling up on RR, but this doesn't change the fact that they are nor right).
And no, he didn't said the exactly word "cliff", I won't argue that...
Posted by: Marco at April 22, 2013 12:51 PM
arthur asks, "you still maintain that R-R never implied anything about causality?" No, I never said "R-R never implied anything about causality." I did not say it because I have never believed that. Why do you make up a position you attribute to me that you then want to argue with?
When I raised doubts about interpreting the correlation causally, I was presenting my views, not the views of RR. They interpret this causally. I am not as persuaded. I am not RR.
And as for a "magic 90% debt/GDP ratio," sorry, none of the quotes you produce seem to me to be remotely valid to describe as a belief in magic, nor do any of them use the word "special" in talking about the value of 90%. They seem to me primarily to be using 90% as an example of a large value.
If you are willing to think about this objectively, I think you will see that the statement "levels of debt of more than 90 percent of GDP are relatively rare" is, in fact, true. It is, in fact, a statement that I could imagine you, or any sensible person, making. Why does the statement become irresponsible when it is made by RR?
Posted by: JDH at April 22, 2013 01:26 PM
Thanks for you update. Interesting and important data.
Posted by: Ricardo at April 22, 2013 01:48 PM
Jonathan at April 22, 2013 12:41 PM: The JEP 2012 was not presented as an amendment, because results are very much consistent with the others. Please look again at the table summarizing the different results from different data in my original post.
The claim that the HAP findings are inconsistent with those of RR (2009) is simply another intentionally misleading claim made by HAP.
And in fact RR have been posting updated versions of their database as it has been expanding. That's precisely where HAP got the data. They then complained, hey, this isn't the same as your 2009 database.
You've been hoodwinked.
Posted by: JDH at April 22, 2013 01:53 PM
Marco at April 22, 2013 12:51 PM: On the contrary, I explained that what I would regard as the appropriate estimate would be somewhere in between that proposed by HAP and RR.
But the reason I am jumping in the way I am is HAP make these strident statements that RR made an error and that HAP is the only acceptable method. Those claims are nonsense and that is what I am trying to correct.
Posted by: JDH at April 22, 2013 01:58 PM
"When I raised doubts about interpreting the correlation causally, I was presenting my views, not the views of RR. They interpret this causally. I am not as persuaded. "
In that case, clearly I wronged you. Apologies.
So we can agree then that R-R have been making unwarranted claims about causality.
'And as for a "magic 90% debt/GDP ratio," sorry, none of the quotes you produce seem to me to be remotely valid to describe as a belief in magic, nor do any of them use the word "special" in talking about the value of 90%. They seem to me primarily to be using 90% as an example of a large value.'
Curious, then, how very limited their imagination is, for they are only able to come up with the exact same example of a large number time and time again.
Since you seem to have missed it, I'll repeat the quote from Reinhart:
"If it is not risky to hit the 90 percent threshold, we would expect a higher incidence"
To say that the talk of hitting the 90% threshold is just 'an example of a large value' is -- how shall I put it? -- to lay upon the logical and semantic resources of the English language a heavier burden than they can reasonably be expected to bear.
Posted by: arthur at April 22, 2013 02:29 PM
Tom I don't agree at all with your analysis, and ultimately I think such thinking will probably be the undoing of much liberal progress. If liberals don't take responsibility for fiscal consolidation, Republicans will.
Just to be clear, I not arguing that we won't need fiscal consolidation. My complaint is that fiscal consolidation at a time of high unemployment and a weak economy in which virtually all of the post-2009 GDP gains have gone to the top 5% is just stupid. Give me a full employment economy and I'll go back to being the deficit hawk that I was 8 years ago.
Posted by: 2slugbaits at April 22, 2013 05:38 PM
@slugbaits - Fair enough, but you should understand that high deficits are a very important factor driving disparity.
You're likely familiar with the principle that all financial savings is somebody else's financial liability, the world's net financial savings sums to zero, and a country's net financial savings (or borrowing) is equal to its current account balance. This is often used to derive the truisms that all countries can't improve their CABs at the same time, and any country's private sector can only increase its financial savings if the public sector increases its borrowings or the CAB improves. The latter of those is often badly misinterpreted, but that's another story.
Using that same logic, the business savings of a country plus its household savings plus its public savings equals its current account balance. Another accounting identity: business savings = profits - dividends. Perform some subtractions and sign-switching and that becomes: profits = public borrowing - household savings - current account deficit + dividends.
In other words, public borrowing feeds directly into profits. It is a source of income for consumption that does not come out of the wages and taxes that companies pay. Record high deficits explain why wages are a record low share of national income.
I submit that there is more than enough waste in the federal budget to cut total spending while at the same time introducing a federal re-employment program. I think it's pathetic that liberals defend the current US budget and want it even bigger when deep down the vast majority of liberals know perfectly well that the federal budget is chock full of waste. Economics is about rationalizing the use of resources. Yes, we could and should give work to the unemployed, we're a plenty wealthy country to afford that. The way to do that is to hire them, not to play with macroeconomic parameters that will mostly miss the target and cause all kinds of other problems.
Posted by: Tom at April 22, 2013 07:28 PM
@JDH - Thanks for all your great comments on this thread. You're absolutely right that HAP are far more ideological than R&R.
Posted by: Tom at April 22, 2013 07:29 PM
Arthur and Marco are right: R&R repeatedly implied that 90% was an important number. It is very disappointing to me to find JDH trying to imply otherwise. No, they didn't use the word "cliff", but they did use the word "threshold" or "level"; and as ARthur points out, "hit the 90% threshold"--cannot reasonably be interpreted as you interpret it. And the word "magic" is put in by R&R's critics for obvious reasons: because R&R adduce no theoretical reason that 90% should be significant, yet continually harp on it as if it has some special significance. And when Senators and high European officials and lots of other prominent people were using their research to argue that 90% is some kind of threshold, when the data, not very compelling at any level, show a stronger relationship between debt and gdp at less than 30%, then R&R should have spoken up.
Posted by: EC at April 23, 2013 03:57 AM
Left Coast Bernard at April 21, 2013 08:04 AM: You are entirely correct that the Fed can be a net revenue source for the government through the mechanism you describe. However, there is a limit to how much revenue the Fed can raise for the Treasury which is based on the real value of currency and reserves that the public and banks are willing to hold. This amount does not automatically increase when interest rates go up. On the contrary, when long-term interest rates go up, the real demand for cash and reserves goes down, not up. If the Fed tries to create more cash and reserves than the banks and public are willing to hold, the result is just inflation. I would include inflation as another example of a distorting tax.
Posted by: JDH at April 23, 2013 06:13 AM
Thanks very much for this clear, informative post and for your comments.
Piqued by your post, I read the HAP paper and indeed your point is very clear in table 3 of their paper. Commenters who disagree with you should really take a look at Table 3 before commenting further. Table 3 shows that the "spreadsheet error," "the selective exclusion of data," and the "transcription error" together account for a few tenths of a percentage point and don't change the R&R conclusion in any significant way. What changes it significantly is HAP's alternative averaging scheme.
The HAP paper as well as the FT editorial dishonestly lump together the data issues with their alternative averaging method, implying that R&R are guilty of some kind of chicanery in using a non-standard method. They want the reader to think that the averaging method that R&R employed is some sort of mistake equivalent to a data error. However, R&R's averaging method is much more plausible than HAP's. As you pointed out, the only correction you might make to R&R is to account for the differences in sample size. My guess is that R&R didn't make this adjustment since they were just trying to establish stylized facts.
Data errors unfortunately happen to the best of us but in this case they don't change the result. I'm sure R&R appreciate someone checking and correcting the data. But HAP exploited minor data errors to sneak in an alternative implausible assumption in an effort to both discredit R&R's research and besmirch their reputations, most likely for ideological reasons. It's shameful, really.
Posted by: Rick Stryker at April 23, 2013 07:36 AM
Jim Hamilton: Very nice post. Great table. The eye easily sees the expected 1 percentage point drop in growth when debt rises above 90%. And the growth differential is robust across both the original authors and Herndon-Ash-Pollin. Causality is something that economics needs to probe into and explain, and you go right to that point. I’ve thought long and hard about causality ever since the original Reinhart Rogoff paper. Of course, there is bi-causality. A financial crisis episode results in deficit spending and a concomitant surge in debt. Initially the recessionary growth slump causes debt. There will be a policy response. But the beneficial aspect of the fiscal impulse is quickly spent. The initial positive impulse multiplier gradates and morphs into a longer-run multiplier. No paper estimating the multiplier should receive grade A approval without showing both short and long-run multiplier estimates. For all too long papers deficient in this have misled. Nor does any study that does not include debt on the right hand side of the equation give a true picture of the time-dependent multiplier. This invalidates most studies out of hand, and surely pre-crisis multiplier estimates are now outdated.
The pernicious aspect of the debt burden kicks in slowly over the course of the decade following the crisis. Other than what the normal forces of recovery and the added fiscal stimulus do to raise tax revenues and reduce stabilizer spending, the deficit spawned by the crisis until eliminated lingers to do damage in the out years. It pushes debt still higher, holds it at too-high a level, or slows its exit from the high debt region.
The channels by which debt causes slow growth are numerous, and hardly limited to the higher interest payments on the debt. They work adversely on society’s time preference, on the political process, and through manifold direct economic effects. Any creative macro economist can come up with a long list, though teasing out the magnitude of the causal effects of each channel will be a Herculean task. The important thing is they be ferreted out and brought to light. Reinhart Rogoff’s work has forever changed the face of economics in this regard.
Context is important. Context is always important, and this is little understood. The initial level of the deficit and debt going into the crisis fall under the umbrella of context. The initial state is, with few exceptions, critical to how the process plays out over time.
Two more things. It will ultimately be understood that sovereign debt – the focus of Reinhart Rogoff’s work – is only part of the story. The private sector debt level must also be taken into account to get the unbiased full picture. Case in point is the 7-year WWII sovereign debt-to-GDP ratio of 105%. War-bond-forced private sector saving during the war greatly cushioned the impact of the sovereign debt on growth in the decade that followed. Exactly the opposite is the case now. The personal saving rate had fallen to nearly zero before the crisis, and private sector debt is still sky high.
These public and private debts set in motion both inflationary and deflationary forces in the wake of a crisis. Private sector debt deleveraging is deflationary. Public sector debt via the nexus of the QEs and need to repay debt obligations in out years is inflationary. In conjunction with private sector deleveraging, the newly embarked upon fiscal austerity in the US implies deflationary forces have now gained the upper hand out across the horizon. But to compare today’s modest deflationary pressure with the serious farm and trade collapse-driven deflationary spiral of the 30s would be like comparing apples and oranges. Context is important.
Posted by: JBH at April 23, 2013 07:46 AM
Studies that show excessive debt are a drag on any economy are actually common but are often ignored by academic and government entities (is there a difference). Governments do not want to be told that expanding they hurt national economies. No one wants to go on a diet and least of all politicians.
Posted by: Ricardo at April 23, 2013 08:05 AM
The fact that interest expense is so high when interest rates are near zero is what scares me.
Posted by: aaron at April 23, 2013 08:34 AM
While I firmly support the concept of selective reasoning, the FED has been subsidizing the Treasury through its "massive" large-scale asset purchases (LSAP).
Taxation mechanisms are infinite and oblique at best. Monetary, income, sales, etc.
Whatever happened to the moral fiber proviede by a monetary system based on honest weights and measures, it seems to me special interest groups have nothing to fear if their bad bets go against them, their losses come directly out of my wallet, selectively.
Okay fine, we can play that game, I'm protesting with my wallet.
Posted by: Chicken at April 23, 2013 09:47 AM
Prof. Hamilton and readers,
Here is an (apparently) knowledgeable and reputable economist who draws a causal link between national debt and growth based upon the body of R&R's work:
[ the European Union’s Commissioner for Economic and Monetary Affairs, Olli] Rehn,...:
"Carmen Reinhart and Kenneth Rogoff, in their wonderful book, “This Time is Different,” have coined the 90-percent rule, or should I say rule of thumb, that countries with public debt exceeding 90 percent of annual economic output grow — tend to grow more slowly. High debt levels can crowd out economic activity and entrepreneurial dynamism, and thus simply hamper growth. This conclusion is particularly relevant at a time when debt levels in Europe are now approaching the 90-percent threshold — it’s currently, on average, around 85 percent — and when the U.S. has already passed this threshold."
It seems to me, an economic amateur, that readers of R&R's works would naturally draw the conclusion that if a nation's debt exceeded about 90% of GDP that nation's government could increase the GDP growth rate by taking immediate steps to cut its debt below the threshold.
Where did Dr. Rehn, a professional economist, get the idea that there was a causal relationship and a threshold? Does he have a problem with reading comprehension in the technical literature?
Posted by: Left Coast Bernard at April 23, 2013 10:01 AM
You wrote: "The eye easily sees the expected 1 percentage point drop in growth when debt rises above 90%. And the growth differential is robust across both the original authors and Herndon-Ash-Pollin."
Your eyes have been fooled. The dropoff in growth after 90% is not statistically significant... And if you look at lags and leads of growth there is hardly any different at all.
When R&R fooled you the first time, shame on them. Now that you've been fooled twice, shame on JBH.
Posted by: Thorstein Veblen at April 23, 2013 10:16 AM
JDH, several years (I think shortly after QE1) ago you estimated that the Fed could maintain policy for about 10 years and remain profitable (IIRC). They've eased since and it's been about five years. Do you have a new estimate?
Posted by: aaron at April 23, 2013 11:38 AM
re: "within a few years, the U.S. interest cost will be bigger"
The GAO long term forecast from December shows in its more realistic "current policy" scenario that interest+entitlements alone will be more than revenue by 2025.
They are using an optimistic assumption for GDP growth, if you plug in more conservative assumptions from one of the Social Security Forecasts as this page:
does, then *this decade* we reach the point where interest+entitlements are more than revenue, they couldn't balance the budget if they spend money on anything else.
Posted by: Politics Debunked at April 23, 2013 11:52 AM
aaron at April 23, 2013 11:38 AM: Yes. My current best estimate is that the Fed will be operating at a sizable loss in 2017-2018 if they try to sell off their MBS.
Posted by: JDH at April 23, 2013 12:32 PM
Left Coast Bernard and others,
Not sure what all the discussion about causality is about. Ken Rogoff has been pretty clear about how he interprets the historical evidence--as 2-way causation. Although recessions certainly increase the debt ratio, his research with Reinhart suggests that the slow growth-high debt periods last far too long to be explained just by recessions, which typically last 1-1.5 years. Thus, the direction of causation can't be just from recession to higher debt. Rogoff believes that causation must also run from the high debt to lower growth for 2 reasons: 1) governments raise taxes to reign in the budget; and 2) investment is lower, crowded out by the high government borrowing.
Rogoff is not defining 90% to be some kind of magic threshold but rather an indicator of a high debt regime based on the historical evidence.
Posted by: Rick Stryker at April 23, 2013 12:50 PM
Rick Stryker Rogoff is not defining 90% to be some kind of magic threshold but rather an indicator of a high debt regime based on the historical evidence.
Nonsense. R-R specifically call it a "threshold." And they also call it "an important marker." And in Senate testimony they referred to it as a "tipping point." And even our host JDH apparently believed there was something special about the 90% threshold because in his recent paper on debt he said:
"Reinhart and Rogoff (2010) and Reinhart, Reinhart and Rogoff (2012) documented that in advanced countries, levels of sovereign debt above 90% of GDP lead to a substantial decline in economic growth..."
Note the definiteness of the 90% figure. Debt above 90% leads to slower economic growth. Not "high debt around 90%." Not "high debt is associated with slower economic growth." But high debt leads to slower economic growth. Nothing fuzzy or wishy-washy about there being an "association". This is a definite statement written in the declarative indicative mood and not the indefinite subjunctive mood. So if even JDH attached particular significance to the 90% figure, then I think it's fair to say that most reasonable non-professional economists would be justified in interpreting the 90% figure as a critical threshold value. BTW, in my world Title 10 of the US Code specifically defines a "threshold" value as one that cancels an acquisition agreement if that threshold is breached. So the normal (and in some contexts legal) understanding of "threshold" is a bright line. In any event, R-R had plenty of opportunities to correct any misunderstandings about their use of the terms "threshold," "important marker" or "tipping point."
R-R have not been clear about the direction of causality. In fact, they have been deliberately obfuscating. For example, several times they tell us that 90+% ratios are relatively rare; but then when we look at their actual data we find no less than 96 historical observations with debt/GDP ratios over 90%. Is that your idea of "rare"? Or to take another example of deliberate confusion. In their NBER working paper version they say that high debt during wartime is exceptional and probably does not lead to lower economic growth. War is considered an understandable and transient reason for high debt and does not reflect fiscal profligacy. Fair enough. They then say that high debt during peacetime is evidence of profligacy. So apparently they have the Reagan/Bush41/Bush43 years in mind. Okay, fair enough. But then in a little noticed footnote they say the following:
"It is important to note that post crises increases in public debt do not necessarily push economies in to the vulnerable 90+ debt/GDP range."
And what kind of "crises" are they talking about? They were specifically talking about financial crises. And this makes sense. The Great Recession is a "one off" kind of event even rarer than war, so why should this be seen as evidence of fiscal profligacy? Especially since many of the countries with high debt ratios were actually running surpluses before the Great Recession. And here they do offer a causal connection that their Austerion allies seem to have overlooked in their zeal for fiscal contraction:
"Periods of sharp deleveraging have followed periods of lower growth and coincide with higher unemployment....In varying degrees, the private sector (households and firms) in many countries (notably both advanced and emerging Europe) are also unwinding the debt built up during the boom years. Thus, private deleveraging may be another legacy of the financial crisis that may dampen growth in the medium term."
So why didn't R-R "man up" and emphasize this important point in their WSJ and Bloomberg op-ed pieces? (Okay, I'll cut Carmen Reinhart some slack on that one.)
One of the ways in which high debt/GDP ratios is supposed to weaken economic growth is distortions coming from higher inflation. Sounds plausible. But guess what? R-R did not find any relationship in advanced countries between inflation and high debt/GDP ratios.
And here's another embarrassing factual error. The mean and median growth rates shown in Figure 2 (advanced countries 1946-2009) of the abbreviated AER paper do not match the growth rates shown in their longer NBER paper version of the AER paper.
This is just the tip of the iceberg. There are so many errors and inconsistencies in the R-R papers that it isn't funny.
Finally, if R-R recognize the reality of the households and firms deleveraging in the wake of the financial crisis, then how are households and firms supposed to save if the government doesn't create the financial assets needed to absorb that saving? Isn't that an obvious question that deserves an answer from a Harvard economist?
Posted by: 2slugbaits at April 23, 2013 03:40 PM
@debunked and similar
While I'm on the side of fiscal restraint, you need to get something very clear. The cost of interest does not depend mainly on the size of debt, although that is a factor. The cost of interest depends mainly on the spread between interest and inflation.
So long as interest rates can be held below the pace of inflation, as they are now, interest costs will not be an issue. If debt grows interest costs will grow in absolute terms, but the debt will be getting inflated away faster than the interest costs are adding to it. That's what's happening now in the US. Debt stocks in real terms are growing by less than the scale of the primary deficit. Real interest costs are negative.
There are lots of forecasts around, none of them worth much. They mostly look at the absolute interest cost in terms of gross debt multiplied by the expected weighted average interest rate, which has little to do with the real interest cost. First you must use only the net debt. Second you must take into account that the government effectively only pays the IOR rate on debt held by the Fed. Third you must forecast not simply the weighted average of those interest rates but its spread relative to inflation. Fourth you might want to forecast possible Fed capital losses, which effectively add to government interest costs, if you expect the Fed to sell assets as rates rise.
In the end the details of the interest calculations hardly matter. What matters is whether the government and Fed can continue to hold interest rates below the rate of inflation. As long as they can, real interest costs remain negative. That doesn't mean they have achieved Nirvana, unless you think Japan has.
Most countries can't hold their interest rates below the rate of inflation, and thus for them growing debt always bites them back one way or another, either by rising inflation or by rising interest costs. The US, UK and Japan and much of Europe are in a different situation. There are all kinds of problems associated with what they're doing, but interest costs, for now, are not one of them, and won't be as long as this situation lasts.
Posted by: Tom at April 23, 2013 04:21 PM
Tom I submit that there is more than enough waste in the federal budget to cut total spending while at the same time introducing a federal re-employment program. I think it's pathetic that liberals defend the current US budget and want it even bigger when deep down the vast majority of liberals know perfectly well that the federal budget is chock full of waste
First, cutting waste is something you worry about when the economy is operating near capacity. Worrying about waste when the economy is operating well below capacity is silly, and in some cases counter-productive.
Second, there are two kinds of "waste." If by "waste" you mean the kinds of stupid projects pushed by way too many congress critters (and especially Republican congress critters), then I think you have a point. But if by "waste" you mean inefficiency, then I don't think there's any evidence that government agencies are any less efficient than private sector entities. For example, a few weeks ago I completed a large stochastic frontier cost analysis study that compared cost efficiencies of government operated versus contractor operated Army repair sites. For example, repair of equipment at Ft Hood and Ft Lewis is done by government personnel; repair of equipment at Ft Bragg and Ft Campbell is done by contractors. I assumed a gamma inefficiency distribution, but the results were the same whether you assumed a half-normal, truncated normal or whatever. There was no real difference in cost efficiencies between contractor or government entities. So the idea that government entities are inherently less efficient than private sector entities just isn't supported by the data.
Posted by: 2slugbaits at April 23, 2013 04:40 PM
2slugbaits at April 23, 2013 03:40 PM: While you're at it, why not also quote this line from our paper?
We analyze the recent experience of advanced economies using both econometric methods and case studies and conclude that countries with debt above 80% of GDP and persistent current-account deficits are vulnerable to a rapid fiscal deterioration as a result of these tipping-point dynamics.
The model we use is unambiguously continuous, the effect we describe is one that gradually builds, and as we review in detail, the point at which the problems arise can turn out to be very different for different countries. Does the statement I just quoted somehow convey to you that we are suggesting there is something magical about the 80% number, that it is a sharp threshold, a bright red line at 80 percent, so that a country with 79% debt is safe, and one with 81% faces catastrophe?
If so, I must apologize to you and any other readers, because that is not what we intended the use of the specific number 80% to convey. We might just as well have written "countries with high debt levels and persistent current-account deficits are vulnerable to a rapid fiscal deterioration." The reason we used a specific number (80%) was that someone reading this might reasonably ask, "well, how high is high?", and we thought it useful to give a particular number for the point at which we personally would start to have these concerns.
Posted by: JDH at April 24, 2013 07:49 AM
Very nice careful empirical work!
The causality is inevitably complicated. High debts could reflect good cooperation among political agents, or it could reflect poor cooperation among political agents.
Posted by: westslope at April 24, 2013 08:49 AM
JDH If so, I must apologize to you and any other readers, because that is not what we intended the use of the specific number 80% to convey.
No problem there. Your paper was clear enough that you view the 80% thing as more like a light that gradually flashes from yellow to amber to orange to red. Evidently I wasn't clear enough. I used the quote from your paper to illustrate the point that just about everyone...you, me, Paul Krugman, Paul Ryan...everyone, interpreted the R-R 90+% threshold to be pretty much a bright line beyond which a country dares not cross. In fact, that interpretation may be the only thing that Paul Krugman and Paul Ryan agree on. This is what I would regard as the strong interpretation of the R-R 90+% threshold. That may or may not be the interpretation that R-R intended, but it was certainly the dominant interpretation. So dominant that they surely knew this even while living in an ivory tower behind ivy walls. I didn't mean to suggest that you particularly endorsed the definitiveness of that threshold or the causal interpretation that they (at least passively) encouraged.
Posted by: 2slugbaits at April 24, 2013 02:04 PM
2sb, so you're saying that because the government is as bad at contracting as it is at actually doing things, we should continue to do more stuff through the government.
Posted by: aaron at April 25, 2013 05:47 AM
Basic finance will tell you a dollar today is worth more than a dollar tomorrow.
If we have higher taxes in the future, or if we have higher interest payments, what's the NPV on those future cash flows?
Bond prices vary inversely with price. If interest rates double, outstanding debt is worth half as much, but the interest payment remains the same. So rising interest rates will only affect any net new borrowing, not the current stock of debt.
Posted by: Nate O. at April 26, 2013 08:59 AM