Thursday, June 25, 2009

Saving Glut Smackdown

The Saving Glut theory of the buildup of global economic imbalances and its application to the current economic crisis has been a popular story ever since it was introduced by Ben Bernanke. Menzie Chinn, however, has dealt a serious critique to this view that probably will be followed by others as time goes on. His view is that the Saving Glut (1) should be put to rest as an idea, (2) is mostly a mirage of the data, and (3) did not cause the current economic crisis. I agree with most of what Chinn says in this critique. I would note, however, that some of the key problems with the Saving Glut theory occur because the role of U.S. monetary policy is not properly accounted for in the analysis. Here are the problems:

(1) The Saving Glut theory has an underlying theme of inevitability. The implicit message is that the U.S. was destined to be a profligate spender because of the huge CA surpluses in Asian and oil-exporting countries. Really? Were U.S. policymakers truly constrained by the whims of foreign savers?

A key reason why this inevitability view is suspect is that it ignores a key fact: the Fed is a global monetary hegemon. It holds the world's main reserve currency and many emerging markets are formally or informally pegged to dollar. Thus, its monetary policy is exported across the globe. This means that the other two monetary powers, the ECB and Japan, are mindful of U.S. monetary policy lest their currencies becomes too expensive relative to the dollar and all the other currencies pegged to the dollar. As as result, the Fed's monetary policy gets exported to some degree to Japan and the Euro area as well. From this perspective it is easy to understand how the Fed could have created a global liquidity glut in the early-to-mid 2000s since its policy rate was negative in real terms and below the growth rate of productivity (i.e. the fed funds rate was below the natural rate).

Given the Fed's role as a monetary hegemon the inevitability theme underlying the saving glut view begins to look absurd. Moreover, the Fed's superpower status raises an interesting question: what would have happened to global liquidity had the Fed run a tighter monetary policy in the early-to-mid 2000s? There would have been less need for the dollar bloc countries to buy up dollars and, in turn, fewer dollar-denominated assets. As a result, less savings would have flowed from the dollar block countries to the United States. In short, some of the saving glut would have disappeared.

What is interesting is that many advocates of the saving glut view who argued the U.S. had to run a current account deficit in the early-to-mid 2000s to accommodate the current account surpluses elsewhere in the world later argued in the middle of 2008 that loose U.S. monetary policy was being exported abroad creating too much stimulus in the dollar bloc countries. In other words, these observers had somehow gone from a world where the Fed is a slave to the dollar block countries to a world where the dollar block countries are a slave to the Fed. For example, here is Martin Wolf who argued early on the inevitability of U.S. current account deficits but then had this to say in June 2008:
To simplify, Ben Bernanke is running the monetary policy of the People’s Bank of China. But the policy appropriate to the US is wildly inappropriate for China and indeed almost all the other countries tied together in the informal dollar zone or, as some economists call it, “Bretton Woods II”.

Thus, not only have the imbalances proved hugely destabilising in the past, but they are going to prove even more destabilising now that the US bubble has burst. When most emerging economies need much tighter monetary policy, they are forced to loosen still further.
To be fair, Martin Wolf has since come to acknowledge the U.S. monetary policy played a role. But the point is clear: if the Fed was a monetary hegemon in 2008 then it was also one in the early-to-mid 2000s. It could have tightened policy then and prevented some of the saving glut.

(2) Long-term interest rates were going down across the globe. The saving glut, however, was regionally based in Asian and oil-exporting countries and mostly went to a regionally-based saving deficit area, the United States. How, then, could a regional saving glut cause global long-term rates to decline? (This is why the saving glut explanation for the interest rate conundrum in 2005 is far from satisfactory.) An easier explanation is that Fed's low interest rates in the early-to-mid 2000s were exported across the global economy as described in (1) and transmitted to long-term rates via the expectation hypothesis of the term structure of interest rates.

(3) If the huge CA surpluses in Asian and oil-exporting countries did, in fact, lead to the lowering of long-term rates in the U.S., which in turn fueled the housing boom, why did long-term rates start rising in 2006? How is that the saving glut could fuel low rates in the early-to-mid 2000s but not thereafter? See the figure below (click on figure to enlarge):

(4) Finally, close to 40% of mortgages issued at the height of the housing boom were either subprime or Alt-A. Unlike traditional long-term, fixed-rate mortgages these other type of mortgages had financing charges tied to short-term interest rates . The Fed controls short-term interest rates. The saving glut story typically focuses on the long-term rates. As Larry White notes, the Fed's policies clearly were the big factor here.
To be clear, I do believe this crisis was more than just poor choices made by U.S. policymakers. The securitization of finance, underestimating aggregate risk, the lowering of lending standards, rating agencies failing, aggressive lending tactics, and poor choices made by lenders all contributed to the current economic crisis. However, the Fed's monetary policy choices in the early-to-mid 2000s was in my view key to making these other developments more distortionary and its role helps shed light on the problems with Saving Glut view.


  1. You will like the latest from the Adam Smith institute...

  2. Great, great stuff. My problem with the savings glut argument has been not only the various points you make, but the essential optionality of the subject matter itself. Those who’ve advocated it typically have been ambiguous about whether it’s a regional or global issue. How can they argue a position if they literally don’t know what they’re talking about? Moreover, a number of people (e.g. Roach) have pointed out that there has never been an outsized level of global saving, ex post. And the ex ante argument is just speculative theorizing, usually based on the observation of interest rates whose cause may arguably be related to a myriad of other factors, including importantly as you point out official levels and expectations of official levels.

    My view of it is that to the degree it’s an issue, it is a regional glut by definition – according to the existence of current account surpluses. And the current account surpluses exist mostly because of the active operational intervention of central banks. China is limited in correcting its surplus to the degree that the PBOC is siphoning dollars from the market as soon as they arrive there. In any event, the idea that a current account surplus is a savings glut is a tautology. It’s only a question of degree.

    And I’m completely suspicious of the argument that the international flows have caused low US rates.

    Apart from that, I’ve never understood the notion that capital inflows cause US current account deficits. How does the deployment of dollars by the PBOC “cause” the net US expenditure flow that is itself the source of the dollars? Really, how plausible is it that the PBOC purchase of a US treasury bond causes more shopping at WalMart? I view the capital inflow as the portfolio allocation that is part of the clearing and settlement mechanism for the current account position.

    As a non-economist, I’ve always found the savings glut argument to be depressing in its logical weaknesses. A real puzzle for me also has been why Bernanke, somebody who I think is very smart, would sponsor it. But my sense is that his experience with the credit crisis is causing him to interpret the totality of capital inflow portfolio clearing somewhat differently.

  3. David, your argument seems to ignore the Fed's statutory obligation to pursue high employment and price stability (for the US alone). The savings glut theory assumes that the Fed does pursue those objectives, where "price stability" has the conventional interpretation as "low and stable inflation rates." Given that assumption, it is difficult to escape the conclusion that the Fed's easy money policy was inevitable.

    In actual fact, the US did experience low and stable inflation rates during most of this decade, so we may conclude that the Fed was following a policy consistent with its price stability objective. A tighter monetary policy would presumably have resulted in lower employment in the US. (At least the Fed would reasonably have assumed so.) But since the Fed was already succeeding in pursuing its price stability objective, a lower level of employment would have implied that the Fed was failing in pursuing its employment objective, since the level of employment would have been below the level actually achieved and therefore below the maximum level consistent with price stability. Thus the statutory objectives required the Fed to follow a policy at least as loose as the one it did in fact follow.

    You can argue that low interest rates were not inevitable because Congress could have repealed the Humphrey-Hawkins act and changed the Fed's statutory objective, but that seems like a bit of a stretch.

    You could also argue that an enlightened Fed would have foreseen the housing bust and therefore kept interest rates higher in order to avoid letting the boom get out of hand. In doing so, it would have chosen to accept a lower employment level temporarily as the cost of avoiding a much lower employment level later on, and this could be seen as ultimately more consistent with its employment objective. However, that argument gets into a level of subtlety beyond anything ever contemplated by proponents of the savings glut theory.

    Moreover, I doubt that we could have avoided the housing bust merely by avoiding the housing boom. We could have avoided the name "housing bust" because it would have been considered a continuation of the already named "tech bust," but I doubt that the average level of employment over time would have been any higher. If anything, I think that the sudden onset of the recent panic induced policymakers to act quickly in undertaking extraordinary remedies, whereas in the absence of a housing boom, we would have descended slowly into depression and boiled like a frog subject to gradually increasing water temperature.

  4. ECB: Thanks for the link. I will bring it along to read on my trip to the WEA meetings.

    JKH: The regional vs. global nature of the saving glut is a tricky issue that I have seen few proponents of the saving glut view carefully address. I agree with you, it is a big problem with the story.

    BTW, has Scott Sumner changed your mind on the importance of the Fed's interest payments on excess reserves?

    Andy: You raise an interesting point. How binding in practice, though, is the Humphrey-Hawkins Act? Do Fed officials really base their decisions on this legislation or because they genuinely believe in price stability? My impression is the latter and I suspect that as long as the Fed has price stability over the business cycle Congress will not object to occasional deviations.

    Your point still has merit, however, given the genuine belief or conviction the Fed itself has in price stability. So, yes, one could argue it was inevitable in that sense. But this inevitability I still find troubling since it boils down to a price stability orthodoxy that handcuffed the Fed in 2003-20004. I am someone who believes that there is an important distinction to be made between deflationary pressures arising from weak aggregate demand and those arising from gains to aggregate supply. As I read the data, the deflationary pressures in 2003 were of the latter form--there were rapid productivity gains at the time--and therefore, they did not warrant the Fed's accommodative action. In short, the Fed's fixation on price stability did not allow it to think outside the box and allow some potentially mild and benign deflation in 2003.

    In case you are interested, I have an article that explains the AD-AS deflation distinction more thoroughly and how it applies to the early-to-mid 2000s. Also, here and here are a few post where I talk about what the data indicate for the 2003 deflation scare.

  5. David,

    No, my assessment of interest on reserves remains distant from that of Scott Sumner.

    But it’s actually not the polar opposite of Scott’s view.

    I discovered the polar opposite articulated best by Art Laffer on CNBC the other day.

    The Sumner view has been that the Fed should charge interest on reserves in order to stimulate commercial bank balance sheet expansion, which would help restore nominal GDP growth.

    The Laffer view is that the expansion of the monetary base is a disaster in waiting due to inevitable bank balance sheet expansion, overstimulation, and inflation.

    My view is that the payment of interest on reserves is indicative of very wise foresight on the part of the Fed. This is how they can ensure their future ability to tighten monetary policy by raising the Fed funds rate, but without necessarily having to reverse all credit easing by the time such rate tightening commences. The payment of interest on reserves ensures a golden mean of policy flexibility and transition capability.

    This view is also expressed indirectly by Paul McCulley at:

    This should be required reading. It’s by far the most useful description of current Fed policy regarding excess reserves that I’ve seen, apart from what can be gleaned by speeches from Fed officials (which is actually a lot more than they’re given credit for.)

  6. David: But capital markets are not regional; they are global! Borrowing and lending do not stop at national boundaries, or regional boundaries, any more than they stop at state boundaries.

    Under perfect capital mobility, there is one world capital market, and one world natural rate of interest. Even under imperfect capital mobility, there is nothing magic about an increase in savings in one part of the globe lowering the natural rate in another part of the globe.

    I must check out Menzie Chinn's post, then do my own.

  7. Nick:

    Can I take your silence on my monetary superpower claim as an endorsement from you on this point?

  8. David: Nope, at least, not yet. I haven't thought through how one large country (the US) setting interest rates below the natural rate would affect the natural rate in other countries. You might be right, or might not.

    OK, let me do it here and now. The world consists of 2 countries: US (very large) and Canada (very small). Assume perfect capital mobility. US shifts LM right, lowering US nominal interest rate temporarily below the natural rate. This shifts Canada's BP curve down, if we initially assume no expected change in the exchange rate.

    The Bank of Canada needs allow the real exchange rate of the C$ to appreciate to offset the lower interest rate, and keep Canadian inflation on target.

    The real exchange rate of the C$ will therefore be expected to depreciate over time as the US interest rate returns to the natural rate. But US inflation will also be expected to exceed Canadian inflation, which would make the nominal exchange rate of the C$ to appreciate, for a given real exchange rate. So we have two offsetting effects on the expected rate of appreciation or depreciation of the C$.
    Canadian inflation.

    Dunno (yet). My guess is you might be right.

    But it's got nothing to do with what I would call "monetary hegemony". Inflation targeting countries like Canada do have independent monetary policies. It's just that the US is a large country. So a monetary "mistake" in the US might lower the Canadian natural rate of interest temporarily and raise (appreciate the Canadian natural exchange rate temporarily.

    I read Menzie Chinn's post. I don't think Frenkel(?) is saying the savings glut argument is wrong, just that it doesn't apply any more. I totally disagree. Fear of the recession has caused private savings to increase even more. Not sure if I agree with Menzie's own argument either. The argument quoted from the Thai economist in the WSJ was totally wrong. Low consumption causes high net exports, not the other way round.

    Just too much for me to counter all at once. And the savings glut argument is getting to be politically incorrect. You risk get accused of "China bashing"! Maybe a Canadian can get away with it, since we don't have much of a CA deficit or surplus, so don't have a dog in the race.

  9. David: On thinking about it some more.

    It depends on parameter values. Under PPP, or close to PPP, a US monetary policy that set US interest rates below the natural rate would reduce the Canadian natural rate also. (Since real exchange rate changes needed to keep Canadian inflation on target would be tiny, so we can ignore the effect of expected real depreciation on the Canadian BP curve).

    But this result does not depend on the US$ being the reserve currency, or Canada having fixed exchange rates. It just depends on the US being large (which it is).

    It's a straight application of Mundell-Fleming.

    So, basically I agree with you (at least under some parameter values). But for slightly different reasons.

  10. Nick:

    There are a number of studies--I have been compilling them--that show short-term interest rates in the U.S. granger cause short-term interest rates elsewhere. This is not surprising for dollar block countries or small open economies (as you mention), but what is surprising is that these findings hold for the Euro area as well. My interpretation of these studies is the story I give above. Here is one of the studies:

    Given all the thought you are putting into this issue, I am really look foward to reading your upcoming post on this issue.

  11. David: I don't find it any more surprising that US interest rates should cause interest rates in large open economies as well as small open economies. (Though I would be a bit surprised if ECB interest rates did not also Granger cause US interest rates, unless it's just that the ECB is so darned slow to Granger cause anything!)

    I assumed Canada was very small and the US very large only because it made it mentally easier to model, because I could then treat the US as equivalent to the world, a closed economy, with causation flowing one way only.

  12. David, Good post. I have a couple reservations, although I have a fairly open mind on the whole savings glut issue.

    I tend to think the so-called "liquidity effect" is overestimated. I.e we tend to misidentify monetary shocks. I guess this relates to Andy's point about the Fed being fairly successful in stabilizing the price level. To the extent that is true, then interest rates were actually at the Wicksellian natural rate. What looked like "easy money" was actually a weak economy in 2002. I don't want to push that too far, as I think monetary policy did become too easy by 2004, but I just want to emphasize that the assumption embedded in your post--that monetary shocks are easily identified, is not so obvious.

    2. Another way of making the same point is that I think a good case can be made for the proposition that the Fed's low interest rate policy prevented another Great Depression. Here's my argument. The tech bubble was just as big as the stock bubble of 1929. If the Fed had not cut rates so sharply in 2002, why should we have expected the outcome to be any different from the 1930s? Of course if a depression had occurred, there would have been no huge CA deficit and no housing bubble, but I don't think that's what you mean by the hypothesis that easy money caused the housing bubble and CA deficit. I think you mean we could have had a decent recovery, w/o those nasty side effects. I'm not so sure.

    One small aside, I don't disagree with JKH's view that an interest on reserve policy might eventually prove useful in restraining inflation, but that doesn't preclude the interest rate from being set at negative 2% right now. It could be raised when and if the Fed needed to tighten policy without pulling a lot of reserves out all at once.

  13. Scott,

    That’s a very fair point – i.e. charging interest on reserves now doesn’t preclude paying interest later, so that rate increases can begin before the balance sheet is returned to normal size.

    BTW, congratulations on the Riksbank move. (It’s as if you were on their policy committee.)

  14. Thanks JKH, It's weird that Lars Swensson was a name that appeared very frequently on my blog during the first few months. But I had no idea he was on the Riksbank board until this story broke. I think he also knows Bernanke quite well, which is another interesting angle. (They were colleagues at Princeton.)

  15. Nick:

    The article I reference above shows that short-termU.S. rates only respond to changes in the fed funds rate, not to ECB rates changes. Rates in the Euro area an in Great Britain do, however, respond to fed fund rates. There are other papers I could point to that similarly show the Fed being the leader. When 60+ of the world's economy is informally or formally linked to the dollar (Rogoff's number), and hence it monetary policy gets exported across the globe, it should be no surprise that it has this influence even in Europe. Now, yes, those dollar block countries do not have to peg to the dollar. But as long as they do the Fed effectively sets monetary policy for a large part of the world. That makes it a monetary superpower.


    I agree that one can make the case that the Fed's policy was appropriate as late as 2002. However, I see no evidence that can justify its actions in 2003 and 2004. This is a point I made to Brad DeLong recently. As I show in that post by 2003 nominal spending was rapidly growing and productivity growth accelerated.

  16. Oops, that number above should read 60%+ of the world economy informally or formally links to the dollar according to Rogoff.