Debt and Growth

Is the public debt a drag on economic growth? Economist Salim Furth reviews the evidence here and finds cause for alarm.

My own instincts are substantially less alarmist, but it should be noted that unlike me, Furth (and those he quotes) have spent substantial time thinking hard about this question.

Click here to comment or read others’ comments.

Share/Save

34 Responses to “Debt and Growth”


  1. 1 1 123

    “Two economists at the International Monetary Fund studied the link between debt and growth across countries and found that higher debt is associated with lower future growth, especially when debt passes the 90 percent threshold”
    Are there any studies that show that higher employment is associated with lower future growth, especially when the unemployment passes below the 5 percent threshold? Seems that the war on debt is based on pure emotions, as you could use the same arguments to fight employment. Fortunately “employment” is a positive word, while debt is not.

  2. 2 2 Ken B

    It’s odd phrasing Steve: “unlike me [they] have spent [] more time thinking about this than I”. I wonder if there is any issue on which you have spent more time thinking than you.

  3. 3 3 Steve Landsburg

    Ken B (#2): I’m fixing this!

  4. 4 4 Chops

    You’re working on thinking harder than yourself, or you’re merely fixing the wording in this post? I hope the former.

  5. 5 5 KS

    Milton Friedman has this phenomenal example of the pitfalls of regression analysis. I might be a bit off here, but I believe it runs along the lines of a car driving down a slope, such that the inclination of the slope is perfectly negatively correlated with the force applied to the gas pedal, so the car’s speed is always constant. A regression analysis would show, of course, that neither the slope of the mountain OR the force you apply to the gas pedal is independently correlated with the car’s velocity.

    The point is: you need a conceptual if not validated understanding of the mechanism of action between two variables before you can interpret any regression analysis between them.

    So how do these economists propose higher debt levels causatively hinder economic growth?

  6. 6 6 KS

    Btw, I think I need to reiterate that Krugman has stated over and over that US debt is a problem in the medium-to-long term. However, he maintains the current problem is unemployment and NOT debt. He also states deficit spending right now can reduce unemployment and improve the long-term debt crisis.

    It pay seem paradoxical that the current solution to a short-term problem is completely opposite to the solution to a long-term problem. It’s not. Think of an obese diabetic patient who presents to the ER with severe malnutrition from diabetic ketoacidosis.

  7. 7 7 Steve Landsburg

    KS:

    So how do these economists propose higher debt levels causatively hinder economic growth?

    Did you look at the papers before asking?

  8. 8 8 KS

    Briefly. Here’s what I got from the IMF working paper:

    “A key issue relates to the extent to which large public
    debts are likely to have an adverse effect on capital accumulation, as well as productivity, and reduce economic growth. This can occur through a variety of channels including higher long-term interest rates, possibly higher future distortionary taxation, higher inflation, greater uncertainty and vulnerability to crises.”

    Basically a litany of any possible way, and also cites the highly-nebulous term “uncertainty”. From what I understand, two of these — interest rates and inflation — are objective and verifiable, and the fact that none of these are very high right now should alter the pre-test probability. Haven’t had a chance to read the whole paper, so not sure if they address this issue later.

  9. 9 9 Cos

    Questions that immediately come to mind:

    1) Do these results apply when considering only countries whose own currencies are global reserve currencies? (UK and US – so may be hard to answer)

    2) What’s the tradeoff between the expected sacrifice in future growth from a particular amount of debt vs. the expected increase in short term growth from the spending that sets it up? Highly relevant because a lot of today’s politics seem to boil down to “if we have this debt it’ll drag down future growth so let’s harshly slash current growth in order to avoid having this debt”.

  10. 10 10 Ken B

    @8
    Litany is just your pejorative term for their list of ways debt can influence capital formation. Capital formation impairment is their mechanism.

    The relevance of current conditions to this list is unclear. Wasn’t your point about having a theoretical link? Couldn’t those conditions have held in those earlier examples they analyze?

    You are just special pleading.

  11. 11 11 KS

    @10–

    “Wasn’t your point about having a theoretical link? Couldn’t those conditions have held in those earlier examples they analyze?”

    You start with a theoretical link. But you need to update your pre-test probability (ie, understanding of your theoretical link) based on available evidence. Using the result’s of this paper during a contemporary debate while ignoring the fact that available evidence goes against your theoretical link is bad science. Also known as good economics.

  12. 12 12 Ken B

    @KS
    Do you even read your own comments? You criticized what you asserted was a lack of a theoretical causal link, although the link was explicitly stated.

    I have carefully applied Bayes’s Theorem to my priors about you based on your latest howler, and find it makes no change at all.

  13. 13 13 KS

    “You criticized what you asserted was a lack of a theoretical causal link, although the link was explicitly stated.”

    Okay, I am going to assume your job does not fall within the domain of science by your reasoning.

    Say I am a medical researcher trying to investigate whether aspirin reduces the risk of heart attack (it does of course, but bare with me). I propose a longitudinal cohort study to compare the heart attack incidence between patients who do and do not take daily aspirin. Before I do, I want to “guess” the pre-test probability by asserting a theoretical link between how aspirin may reduce the risk of heart attacks.

    Here’s an example of a good theoretical link: aspirin reduces the tendency of platelets to clot, and clots are what reduce blood flow and cause heart attacks.

    Here’s a BAD example of a theoretical link: aspirin may reduce heart attacks by decreasing clots, increasing blood flow, increasing perfusion time, increasing collateral circulation, decreasing blood viscosity, vasodilating coronary arteries, etc…

    You can’t just list EVERY SINGLE POSSIBLE WAY YOU CAN THINK OF and assert that as a theoretical link. Do I need to repeat this to make it more clear?

    Good, now here’s the paper:

    “This can occur through a variety of channels including higher long-term interest rates, possibly higher future distortionary taxation, higher inflation, greater uncertainty and vulnerability to crises.”

    What I see here is a list of every possible way they can think of. Let’s run through these one by one:

    1. Higher long-term interest rates. Great! This is verifiable! How high are interest rates right now?

    2. Possible higher future distortionary taxation. Wow, this isn’t measurable.

    3. Higher inflation. Great! This is verifiable! How high is inflation now?

    4. Greater uncertainty. Wow, this isn’t measurable either.

    5. Vulnerability to crises. Wow, this also isn’t measurable.

    So of the 5 proposed mechanisms, only 2 can actually be verified empirically. And for BOTH of those, the current evidence argues against them. And this doesn’t change your Bayesian priors? Are you aware of what a Bayesian prior even is?

  14. 14 14 Mike H

    Re #3, but the original made for a much nicer logic puzzle!

  15. 15 15 Harold

    No. 5 illustrates one problem with corrrealations. Friedmans thermostat story posits a car on a hilly road. The driver pushes the pedal down when going up hill in order to keep the speed the same. He lets off the gas going downhill to keep the speed the same. The observer finds zero correlation between gas pedal and speed and zero correlation between hills and speed. There would be a strong correlation between hills and gas pedal position. This is a warning about correlations, but not quite the same as the one we are faced with here.

    In fact, given only the data on slope and pedal position, we could not tell if the accelerator and slope were connected with speed, or if the pedal were a slope indicator, or if the pedal caused the car to tilt. It is quite possible the vehicle is being towed, and the pedal has no effect on speed. In the world, we have far more data, which allows us to reject most of these alternatives. Not least, we have data about how th ecar performa when on the flat with differrent pedal positions.

    In the case of the national debt, we have a strong correlation, but we do not know the mechanism, so we cannot say whether there is direct causation or co-correlation with some other, underlying cause. Is the debt level an accelerator, or merely a level indicator? The articles suggest possible mechanisms, so there is no reason to reject the direct causation hypothesis. It is also possible that any co-correlating factor would aslo be affected by our attempts to address the direct causal link.

    Perhaps debt does not lower growth, it is lack of confidence. We may be able to increase confidence by lowering debt, but we could get the same effect in other ways – thus there is not a direct link. However, if we have seen that lowering debt has the effect we want, even without a direct link it may be a good idea to pursue it.

    Also many medical correlations are acted upon without a cause being known – mainly in harm avoidance. Side effects will prevent a drug being used even if the mechanism is not known.

  16. 16 16 Ken B

    @13
    Are you aware we are not debating the truth of the claims in the articles?

    You like to cast aspersions KS, call names, and make assumptions about others. Better you should stick to the point.

  17. 17 17 Ken B

    KS: you need a link? Did these guys think of that? Huh, huh?
    Steve: Did you read the papers?
    KS: No, but here’s a quote. It’s a litany, there is no link identified.
    Ken B: They cite impairment of capital formation.
    KS: Idiot! That’s not a link. Economists are fools!
    Ken B: Oh, you meant a link you agree with!
    KS: Unscientific fool!

  18. 18 18 iceman

    “An Indian-born economist once explained…”if you are a good economist…you are reborn as a physicist…[if not] you are reborn as a sociologist”…the speaker was talking about…the sheer difficulty of the subject. Economics is *harder* than physics; luckily it is not quite as hard as sociology…part of the answer has to do with complexity. The economy cannot be put in a box. Physics does very well at explaining simple, contained systems…it has a much harder time when trying to cope with the complexities if nature in the wild…another reason economics is hard is that…it involves human beings, who do not behave in simple, mechanical ways…in the social sciences, one cannot perform controlled experiments: evidence is always historical, and history is complicated enough that its lessons are seldom unambiguous. [Also] social science studies people; and since we think we know ourselves, we all tend to think that we already know the answers.”

    - Paul Krugman from Peddling Prosperity

  19. 19 19 Keshav Srinivasan

    Steve, off topic but I think you may be interested in responding to this post by Matt Yglesias, which discusses whether restaurants will pass on their Obamacare costs to their customers:

    http://www.slate.com/blogs/moneybox/2013/03/11/obamacare_and_five_guys_burgers_won_t_get_more_expensive.html

  20. 20 20 KS

    @ Ken B–

    “Ken B: They cite impairment of capital formation”

    So, please tell me how high debt impairs capital formation? That would be a link.

    A litany of every possible way you can think of is NOT a link. I feel I need to repeat this for you.

    Once again, how does high government debt impair capital formation?

    And finally, I feel I need to restate Paul Krugman’s position. The US has a medium to long-term debt issue. The US has a current unemployment crisis. Deficit spending NOW can alleviate the latter and actually improve the former.

    But how is it possible that the solution to a short term problem is completely opposite the solution to a long-term problem? I feel people have a hard time just “stomaching” this (ie, Ken B). It’s really not all that implausible. When I was in medical school, my student loans (ie, debt) increased heavily my first 7 semester. The best solution was actually to INCREASE my debt even more (for an 8th semester), then graduate, then pay off my loans later. It was not to start paying off my debt after 7 semesters.

    Or, an obese patient with diabetes presents to the ER with ketoacidosis (severe malnutrition). The ACUTE treatment is fluids and glucose (among others). The CHRONIC treatment is weight loss reduction. These two oppose each other. The only paradox is in the simplistic heads of people who dichotomize everything.

  21. 21 21 Steve Landsburg

    KS:

    So, please tell me how high debt impairs capital formation? That would be a link.

    Presumably, the authors did not review this material because they were writing for an audience that could be presumed to be familiar with the standard arguments. The fact that you are apparently unfamiliar with these arguments is not a failing of the paper.

  22. 22 22 Twofer

    The whole thing feels a bit disengenuous. Krugman reads an article about why Krugman is wrong, and finds out R&R is the stated reason, so Krugman reminds everyone about the the correlation/causation problems with R&R.

    Furth writes an article, claiming that Krugman is “ignoring most of the literature,” while ignoring much of what Krugman has actually written. Furth points to several articles, one of which is from the IMF. While the articles don’t quite agree with each other they sort of say the same kinds of things. From the 2010 IMF Study:

    “on average, a 10 percentage point increase in the initial debt-to-GDP ratio is associated with a slowdown in annual real per capita GDP growth of around 0.2 percentage points per year,
    with the impact being smaller (around 0.15) in advanced economies. There is some evidence of nonlinearity, with only high (above 90 percent of GDP) levels of debt having a significant negative effect on growth. This adverse effect largely reflects a slowdown in labor productivity growth, mainly due to reduced investment and slower growth of the capital
    stock per worker.”

    Could be true I suppose. The US during WWII with debt/GDP ratio at 120%, and current productivity at a high run contrary to this finding, but let’s take it as true anyway.

    How does that reconcile with what the IMF is saying now?

    http://www.imf.org/external/pubs/cat/longres.aspx?sk=40381.0

    From the 2013 IMF paper:

    “Studies suggest that fiscal multipliers are currently high in many advanced economies. One important implication is that fiscal tightening could raise the debt ratio in the short term, as fiscal gains are partly wiped out by the decline in output. Although this effect is not long-lasting and debt eventually declines, it could be an issue if financial markets focus on the short-term behavior of the debt ratio, or if country authorities engage in repeated rounds of tightening in an effort to get the debt ratio to converge to the official target. We discuss whether these problems could be addressed by setting and monitoring debt targets in cyclically-adjusted terms.”

    So, all in all, after reading some of the literature, I’m inclined to agree with Krugman and KS. In the current set of economic conditions austerity will worsen the debt/gdp in the short run, furthering the drag Furth wishes to avoid (if indeed the drag is real, there aren’t that many actual examples). Address the employment issue first.

    BTW: It strikes me that debt/GDP is really a pretty crude measure, and it would be more important to look at the interest payment to GDP ratio. Bond prices and interest rates are inverse to each other, so when interest rates go up, the total amount of debt could be reduced…But I’d need to think about that some more.

  23. 23 23 iceman

    twofer – you do need to think about that last bit more. When interest rates go up the market value of debt goes down, but debt service payments (which are based on par value and coupon rate) do not. That where if you run the #s assuming the govt has to roll over our debt burden at more ‘normal’ rates it gets a little scary.

  24. 24 24 iceman

    KS – of course some might think the analogy is, you keep switching your major and your parents are starting to wonder if you’re actually going to graduate, or if they’re just throwing good money after bad and the quickest route to getting you on your feet is to pull the plug and make you get an apartment and a job. Some short-term pain leading to a quicker recovery. Seems to me the attempt to prevent ‘hysteresis’ (the key to the debt debate) can be circular.

    Sometimes a sophisticated mind also needs to recognize when it is NOT necessary to reconcile two conflicting ideas.

  25. 25 25 Twofer

    @iceman #23,

    I wish I had followed my instincts and deleted my last paragraph. It’s a sideshow to the point I was attempting to make. Bu all in all, I don’t think you and I are saying anything different. You asked me to run the numbers, so here goes:

    Suppose that we issue $100 billion in 30 year bonds at or near the current interest rate of 2.75 percent. Let’s imagine that in 3 years the economy has largely recovered and that long-term interest rates are back at a more normal level; let’s say 6.0 percent for a 30-year bond.

    In this case the bond price would fall by over 40 percent meaning, in principle, that it would be possible for the government to buy up the $100 billion in debt that it issued in 2013 for just $60 billion, shaving points off the debt-gdp ratio in the process. If we had been flirting with the magic 90 percent debt to GDP ratio before the bond purchase, we will have given ourselves a huge amount (potentailly) of leeway by buying up these bonds.

    Of course, this would be silly. The interest burden of the debt would not have changed; the only thing that would have changed is the dollar value of the outstanding debt.

    That’s what makes me think the real concern should be interest payment to GDP, not debt. My understanding is that interest to GDP right now is actually pretty low historically when compared to the post WWII period.

  26. 26 26 iceman

    Agreed when rates are unusually low just looking at the debt arguably overstates the immediate risk, while just looking at interest understates the longer-term risk. For the latter the concern is that rates rise while we’re still borrowing to finance ongoing deficits and roll over maturing debt rather than buying back 30-yr bonds in the open market.

    On your main point, there is certainly a debate going on over whether we are currently in a special time of extraordinarily high multipliers. I would just suggest this involves a much stronger case than reflected in your IMF quote of temporary, partially offsetting effects. It quite literally suggests a free lunch in which spending pays for itself — advocated, naturally, by people who have been the most derisive of analogous ‘extreme’ claims on the supply side.

  27. 27 27 KS

    Dr. Landsburg–

    “Presumably, the authors did not review this material because they were writing for an audience that could be presumed to be familiar with the standard arguments. The fact that you are apparently unfamiliar with these arguments is not a failing of the paper.”

    Now, I haven’t had a chance to peruse the entire paper, but I plan to do so. I add that caveat now.

    However, IF the paper doesn’t review this material — or doesn’t interpret the results of its regression analysis in the context of what the theoretical causal relationship between high government debt and slower economic growth might be — THEN that is a failing of the paper. Furthermore, IF the paper ignores the fact that current evidence (low interest rates, lower-than-anticipated inflation despite massive annual deficits) argues against at least two of the proposed mechanisms, that’s also a failing of the paper.

    I’m no economist; I dabble in this because I find it fascinating, and your blog is one of the few econ I’ve found where the audience is very intellectually-curious and your writing isn’t intended for technical economists.

    Finally, I’d have an easier time believing that increased government spending results in misallocation of resources if the unemployment rate is close to a “natural minimum” (my terminology, probably made up and arguably bogus). I also have an easier time believing that with a higher unemployment rate, we can assume that resources are not optimally allocated (unless you laughably believe most current unemployment is voluntary), and that increased government spending might actually put idle resources to use rather than redistribute them in a suboptimal way. I think Krugman would agree with this, but that’s just my intuition.

  28. 28 28 Twofer

    @iceman #26

    “For the latter the concern is that rates rise while we’re still borrowing to finance ongoing deficits and roll over maturing debt rather than buying back 30-yr bonds in the open market.”

    Well, that’s been the argument for at least 4 years now. Here’s Brian Reidel the leading budget analyst for the Heritage Foundation making this exact point in Feb 2009:

    http://www.nationalreview.com/corner/176615/case-no-stimulus/david-freddoso

    I’m sure he’ll be right some day. Interest rates will be high again. If you know when that will be, you have an excellent opportunity to make a whole lot of money.

    In the meantime, while I’m not good at predicting the future (the market is smarter than me and right now it is calm about interst rates) the CBO is fairly optimistic about the trimming of the deficit in the coming years, so it could still be a number of more years before your scenario comes to pass.

    http://www.cbo.gov/publication/43977

    If we were at full employment, I’m willing to bet you and I would be sharing the exact same concerns.

  29. 29 29 KS

    @ iceman–

    ” you keep switching your major and your parents are starting to wonder if you’re actually going to graduate, or if they’re just throwing good money after bad and the quickest route to getting you on your feet is to pull the plug and make you get an apartment and a job. Some short-term pain leading to a quicker recovery. Seems to me the attempt to prevent ‘hysteresis’ (the key to the debt debate) can be circular”

    Right. Exactly. There are two competing hypotheses for what the current problem is — do we have a deficit crisis (your analogy), or an aggregate demand crisis (my analogy)? I would argue historically low interest rates support the latter.

  30. 30 30 iceman

    We’re talking about risk not an actual prediction; why just looking at the current interest payment may understate the issue. Maybe ask someone who had a teaser mortgage rate. I also agreed just looking at the amount of debt arguably overstates the issue (currently) for the reason you mention.

  31. 31 31 Twofer

    @ Iceman #29

    “We’re talking about risk not an actual prediction; why just looking at the current interest payment may understate the issue.”

    This is a fair point. In theory, the market has already taken this risk into account…in practice we now have pretty good information that the market had no idea of the risk during the bubble (at least not until near the end), so it’s not infallible.

  32. 32 32 iceman

    Twofer – Well a market price indicates the average expectation but not the distribution profile of possible outcomes. The tail risk by definition reflects unexpected events. Ask the guys from Long Term Capital Management. Sometimes markets don’t move until in step function. Or when, to quote Bob Dylan, money doesn’t talk it screams.

    KS – Actually I think even if we’re just talking about aggregate demand absent a debt “crisis”, there’s a legitimate debate over whether govt intervention smooths but also prolongs a recovery below trend growth (e.g. bringing growth forward while creating a debt overhang), whereas normal corrective / clearing forces produce perhaps a sharper ST contraction but a quicker path to trend growth. And this seems very relevant to the issue of hysteresis (which has been a leading source of calls for even more stimulus). Or maybe we accept this tradeoff as generally true, but the Great Recession was sufficiently unique (e.g. credit-induced) and potentially dire it called — and still calls? — for extraordinary measures. The difficulty is in proving negatives, like where would unemployment be without some of the actions we took? Could the 30s have been more like the recession of the early 20s?

  33. 33 33 Ken B
  34. 34 34 Ken B

    A remarkable straw man here
    KS:”But how is it possible that the solution to a short term problem is completely opposite the solution to a long-term problem? I feel people have a hard time just “stomaching” this (ie, Ken B). It’s really not all that implausible.”

    I’m curious if you can cite even a single comment of mine suggesting any such stomaching problem. Like everyone here I understand that even if you live in a valley you might have to walk uphill a bit to get to the sea. This is not some insight vouchsafed only to you and Krugman.

    And again, you persist in confusing debt and spending.

  1. 1 Some Links - Unofficial Network
Comments are currently closed.