Few popular terms irritate me as much as “deleveraging”.
Yes, the concept is important. In fact, it’s probably central to understanding why we’re in such a rut. But almost everyone talking about it fails to understand why it matters, and why it’s intimately related to monetary policy.
Sure, most people know the basic idea: during the crisis, consumers and businesses experienced an enormous hit to net worth, and now they want to improve their balance sheets. To do so, they spend less—but since lower spending means lower income somewhere else in the economy, in the aggregate balance sheets barely improve at all. The economy is depressed as it gradually returns to the correct level of leverage, and we experience the “long, painful” process of deleveraging. (In the words of, well, every blogger and amateur econ pundit in the world.)
Great story. Too bad it ignores everything else we know about macroeconomics.
After all, why should the desire to spend less and save more hurt the economy? If I want to save, I hand my money over to someone who wants to borrow or invest it. Net saving is channeled into productive investment. If consumers want to save more, we’ll see lower consumption but an investment boom—hardly a disaster for the economy.
Sure, you say, but maybe no one wants to invest this money. Won’t an increase in savings then mean that the economy crashes? After all, the money isn’t being spent on consumption, and it has nowhere else to go.
Although this sounds plausible, it doesn’t really make sense. At the micro level, economists don’t worry about weak demand causing a supply glut: instead, they correctly say that prices will adjust to clear the market. The same is true for macro as well. There’s a price—the real interest rate—that determines willingness to save and invest. Like all prices, this price has a market-clearing level: at a sufficiently low real interest rate, the supply and demand for savings will equate, and consumers’ desire to save will translate into an investment boom, not an economic downturn.
The hitch is that the “market” doesn’t quite control interest rates: the Fed does. It controls both nominal interest rates (by setting them) and real interest rates (by shaping inflation expectations). Worse, nominal interest rates can’t go below zero. And if a real interest rate of -X% (0% nominal minus X% inflation) isn’t enough to clear the market and channel money from savers to investment, we’ll see a downturn after all.
So yes, deleveraging can be very bad for the economy. But this is only because monetary policy doesn’t adjust enough to match the market.
In failing to understand this core logic, most commentary about “deleveraging” is rather bizarre. At some level, it’s the same cluelessness that we once saw from central planners: they’d trip over themselves in the complexity of fixing a shortage in one market or a glut in another, never quite realizing that the price mechanism would do their work for them. Right now, historically low inflation expectations and below-potential output are prima facie evidence that real interest rates are too high. That’s what every macro model tells us is associated with contractionary policy by the Fed. Yet we see pundits lost in all kinds of complicated, small-bore proposals to stimulate the economy—when the fundamental, overriding dilemma is getting the price (in this case, the interest rate) right.
This isn’t to say that non-monetary proposals should be abandoned entirely. Monetary policy doesn’t stop working at the zero lower bound, but it is a lot harder. Even an ideal Fed would probably find itself constrained. This means that we should be open to other policies that affect demand—possibly via government spending, transfers, or tax incentives. But once we recognize that the fundamental problem is monetary, the issues become much clearer. The Fed’s failure to use all the tools at its disposal—in particular, its failure to make a conditional commitment like the one proposed by Chicago Fed President Charles Evans—is by far the most serious failure of economic policy today.
Note: High on the list of people who do understand deleveraging are Gauti Eggertsson and Paul Krugman. It’s even obvious from the title of their paper: “deleveraging” comes right before “the liquidity trap”, i.e. the zero lower bound making it difficult for the Fed to properly accommodate the effects of deleveraging.
Matt Rognlie 
Good post.
Not all nominal interest rates are zero. It is just the interest rates that central banks like to manipulate are very near zero.
The zero bound is really a negative bound equal to the cost of storing hand-to-hand currency.
Also, your account here fails to take into account the feed back of a central bank so tied to manipulating short term interest rates and inflation rates that it is unable to smoothly shift to some other approach, maybe long term interest rates or the quantity of base money, when the problem is a large downward shift in the growth path of nominal expenditure.
Scott Sumner’s key point is that central bank incompence can reduce the equilibrium real interest rate on short and safe assets. A shift to central bank competence can reverse the problem.
I agree that if the central bank doesn’t have much of a record publicly and credibly signaling the future trajectory of policy, it may be harder to signal effectively when simple short-run targeting no longer works and longer-term commitments become necessary to make any headway. (Though I think many of the Fed’s problems in this regard are completely self-inflicted: even given its lack of experience commuting a clear policy target, it could easily bring expectations for inflation and output growth well above where they are now.)
I also agree that skepticism about the central bank’s ability to manage demand properly may make the problem worse. The mechanism I have in mind is that uncertainty on both the consumer and business side increases the desire to save (especially in liquid instruments), pushing down the natural real interest rate and expanding the contractionary gap with the actual interest rate. This sounds similar to the feedback effect that you and Scott emphasize.
Matt,
Great analysis. I have really enjoyed your last two posts. This one I especially enjoyed since I have been making a similar case. A great example of this is that FDR’s monetary experiment in 1933-1936–which was signaled as price level target–created a robust recovery despite ongoing deleveraging through 1935.
This is indeed an excellent example.
My only qualm in these discussions is that it’s difficult to separate (1) inflation’s effect in bringing down real interest rates and stimulating spending investment directly and (2) inflation’s effect in whittling down the real value of debt burdens and speeding the process of deleveraging. I think most people find it intuitive that sufficiently dramatic inflation (like the double-digit rate of FDR’s first couple years) can accomplish (2), but for whatever reason they have a hard time understanding (1).
Both can be important, of course, but (1) is in some sense a more fundamental mechanism—the additional effect of (2) matters mainly because sometimes it’s not realistic for monetary policy to bring down real rates enough to make up for crippled demand.
Matt,
The experience of economies with persistent negative real rates is not a good one. The list of distortions created by these rates is long: bloated inventories; margin volatility; bloated public sector; capital flight; reduced access to term financing; etc. In many cases these conditions lead to shrinking AS and an increased tendency towards high and variable inflation.
In the Nineties, post-Cardozo Plan, I had the opportunity to tour a formerly state-owned steel plant in Brazil (Cosipa). The rusting, giant hulk produced at around 30% of capacity due to years of underinvestment. This plant was an object lesson in what happens when a country escapes “deleveraging”, but produces something worse.
First, I want to emphasize that the market real interest rate (e.g. the rate we’d get if we had a Walrasian auctioneer finding the equilibrium rate instead of the Fed) right now is almost surely negative, and below its current value. In fact, this is unavoidable if you believe in the “deleveraging” story of economic malaise. If consumers’ desire to save and rebuild their net worth is bringing down the economy even at the current real interest rate, that’s has to be because the real interest rate is not yet at its equilibrium value: it’s too high.
In this light, “bloated inventories” are not a distortion: they’re the efficient response to a very low cost of capital. (After all, if we have spare production capacity now but aren’t using it because no one wants to consume stuff, why not ramp up production and store it for later, when they will want to consume?) On your other points:
1. I’m not completely sure what “margin volatility” is or why it would be related to low real interest rates (as opposed to, say, volatile real interest rates).
2. I’ll make the same point about a “bloated public sector” as with “bloated inventories”: (temporarily) larger public sector expenditures are generally an efficient response to low market real interest rates. I suppose that if you believe government expenditure is currently at a horrifically wasteful level, then low real interest rates could exacerbate the problem, and that would override all other considerations. Keep in mind, however, that holding the real interest rate fixed, the cost of government spending is quite possibly close to 0 (e.g. the multiplier is 1): the only way that it can crowd out private spending in the aggregate is by raising real interest rates, but if they’ve been pushed as low as possible by the Fed (in an attempt to reach the true market rate) this is no longer an option.
3. If you believe in the deleveraging story or indeed many other stories of deficient demand, “capital flight” is exactly what we need right now: a smaller capital account surplus means a smaller current account deficit, which all else equal means higher GDP. This works even if we think about the mechanisms a little more explicitly (rather than reasoning via accounting identities, which I prefer to avoid): lower real interest rates mean a weaker dollar, which increases exports. If we are demand-constrained, it is hard to avoid the conclusion that a weaker dollar is massively welfare-enhancing. (This is, of course, different from the usual situation where a weaker dollar is probably bad for welfare).
4. I don’t see why there should be “reduced access to term financing”. This, again, seems more like a consequence of volatile real interest rates than low real interest rates. I suppose you could say that a more proactive Fed policy would cause exactly this kind of volatility, but I don’t think that’s quite right—if the Fed lowers real interest rates precisely when the economy is doing terribly, then banks making term loans will get a windfall (as their cost of financing is much lower compared to the fixed rate they are receiving) exactly when they are likely to suffer some credit losses. If anything, this should make banks more willing to offer term financing, because it is a useful risk hedge.
In general, I think you are conflating the consequences of irregular, volatile monetary policy—where inflation and real interest rates undergo swings inconceivable in the modern US—with the consequences of a much more systematic policy of bringing down real interest rates when the market suggests they are far too high. I worry about many things, but turning into Brazil isn’t one of them.
Matt,
Eggertsson is right: to impact behavior, the Fed must act irresponsibly. The “we are not Brazil” argument starts to erode at that point. How does the Fed control inflation (real rates) once it sparks an inflationary dynamic? By promising to contain an overshoot. However, this lands us back in the realm of responsibility, which does not change behavior.
So the basic problem I have with your argument is that you seem to want it both ways: moderation to contain volatility; irresponsibility to scare up credit demand. Either policy will be ineffective (in the first instance) or produce volatility (in the second).
Take the case of a promise to finance deficits. Credit demand gets pulled along by public sector borrowing, and the economy shifts resources to the public sector. Confidence hinges on the ability to finance the PSBR without constantly accelerating inflation. If the Fed stops to contain that inflation, real rates rise and so does the PSBR — out year deficit projections blow out, and so do real term rates. The Fed steps in to repress the yield curve, and inflation reaccelerates. Do we have to be “Brazil” to have this dynamic? No, we just need: 1) deficits of 10% of gdp; 2) debt of 100% of gdp; 3) reliance on either negative real rates or strong growth to make deficit projections palatable to markets. I would argue we are close to all three conditions being met.
Suppose that medium-term inflation expectations were currently at a level that the Fed judges consistent with its mandate of price stability (generally, 2%). Then it’s true that any further steps to improve demand via future policy would involve levels of inflation that are “irresponsible” next to the 2% target. Right now, however, inflation expectations are well below 2%, meaning that at the margin there is no dilemma here at all. Assuming that the Fed’s goal is actually 2%, for the moment we don’t face the “committing to be irresponsible” issue emphasized by people like Krugman and Eggertsson”: in fact, we’d be committing to be responsible.
But suppose that inflation expectations were at 2%. Then such policy would require inflation to go above trend for a little while. This almost certainly wouldn’t take very long—at the very worst, 2 years of 4% inflation at 0% nominal rates rather than 2% inflation would be an enormous boost. While this is “irresponsible” compared to the 2% target, according to any reasonable estimate the welfare cost would be quite minimal next to the cost of high unemployment. You seem to be worried we’d be letting some kind of uncontrollable inflation genie out of the bottle. I just don’t think this is true: if the Fed clearly says “we’re going to tolerate 4% inflation for the next 2 years and then return to enforcing a 2% target”, markets will believe it.
Matt said: “But this is *only* because monetary policy doesn’t adjust enough to match the market.”
Only? Really?
Ummm,.. no, I don’t think so… maybe if debts were denominated in real, rather than nominal, terms (even then I’d have to think about it some more)? see Steve Keen over at DebtWatch (e.g., http://www.debtdeflation.com/blogs/research/)
I’m not familiar with everything that Steve Keen has ever written, but his discussion of deleveraging seems to emphasize “aggregate demand”—e.g. here. My point is this: it is absolutely incoherent to talk about a concept like “aggregate demand” without including real interest rates. Demand depends on real interest rates; given a particular level of productive capacity, there is a real interest rate at which supply will equal demand.
(Incidentally, this is implicit in the old intro-econ AS-AD diagram, which in the modern era should really be reformulated in terms of interest rates rather than prices: the effect on interest rates is the principal justification for why the AD curve slopes downward. The logic is (higher prices) + (constant money supply) -> (higher real interest rates) -> (lower demand). These days the central bank sets interest rates directly rather than a constant money supply, so we can skip the first step and view AD as a relation entirely between interest rates and demand. You cannot have a remotely sensible concept of AD otherwise.)
Pingback: Economist's View: links for 2011-10-05
But Matt, is it possible for an open economy to delever, in aggregate, while also running a current account deficit? Seems to me that the answer is no.
If you agree with my answer then the next question is whether or not it’s possilbe for the US to close its current account deficit? Seems to me the answer is again no.
just to elaborate, the point I’m getting at is that if the country is going to persist in running a trade deficit then it is necessarily the case that someone (public or private) is spending on credit. Thus the idea that a sufficiently low real interest rate will maintain full employment *and* accomodate a reduction in aggregate leverage is not true.
The reduction in aggregate leverage itself is impossible without a trade surplus and in the presence of a persistent trade deficit any attempt at aggregate reductions in debt will only lead to less income, no matter what the real interest rate. It is a classic paradox of thrift.
Now, that is not to say that reducing the real interest rate doesn’t work to maintain full employment but the mechanism is quite different to what you seem to have in mind. It can only work by making agents choose to maintain, or even increase, their debt to income ratio.
It is certainly possible in theory for an open economy to delever in aggregate while continuing to run a current account deficit, if it swaps debt for equity financing. I don’t even think it needs to pay off any debt to do this; it just needs to increase its wealth by doing equity-financed investment.
But supposing this doesn’t happen, I still don’t see how the releveraging mechanism is different from what Matt seems to have in mind. His point is that when the price (i.e. the interest rate) is set correctly, the market clears. Maybe it clears by making people satisfied with the amount of leverage they have, rather than by allowing them to deleverage. Maybe you think that leverage is unhealthy, but that’s a different discussion. It’s still wrong to blame the depression on deleveraging.
Andy, yes I agree that financing investment by issuing equity to foreigners would allow this. You’re right.
On the other point though, what I think Matt had in mind was the idea that a low enough real rate induces the group of creditors to increase their spending enough to offset the reduced spending of the debtors. This is what basically how the Eggertson/Krugman model works.
This is a mechanism that allows the aggregate leverage to fall while maintaining full employment which seemed to be what Matt had in mind, it is different from simply making people choose not to delver after all.
Sorry that I’m so late to reply.
As you’ve both suggested, there are plenty of mechanisms through which lower real interest rates can boost spending. These include, for instance:
1. Creditors (e.g. households with positive net worth) save less and spend more. This is consistent with a decrease in leverage. (Though not the “net leverage” of the country, if anyone cares about such a concept—that has to come through either a reduction in the current-account deficit, as Adam’s first comment suggests, or a swap of debt for equity external financing, following Andy’s suggestion. At least for a country like the U.S. with safe and liquid access to international capital markets, I think leverage at the household level is much more important than the nation’s net debt/equity position, so I’m not too concerned with this concept.)
2. Debtors decide that deleveraging isn’t worth it and borrow more/pay down less.
3. Corporations decide to either borrow more to invest (causing an increase in leverage) or invest more out of retained earnings, rather than hoarding them in cash equivalents or distributing them to equityholders.
4. Low real interest rates cause a decline in the dollar, which leads to an increase in exports and decrease in imports. This implies both a spending increase and a reduction in leverage.
I think all these mechanisms are completely plausible, and I’m fairly agnostic about which one would have the biggest role quantitatively. That strikes me as ultimately an empirical question, one that we don’t have nearly enough data to answer. Generally I’m agnostic about which mechanisms are better or worse; I mainly want to nudge down the real interest rate through appropriate monetary policy and let the market decide.
In particular, I don’t think that excess leverage is a terribly pressing problem except for its impact on demand in the current zero-lower-bound recession (which won’t last forever), so I don’t mind if (2) happens and the “necessary deleveraging” that pundits like to cite doesn’t happen.
Good post.
If the Walrasian auctioneer cleared all markets *taking expectations of future real income as given* (Hicksian temporary equilibrium), then *maybe* the equilibrium real rate would be negative (on some assets, and leaving aside your valid point in reply to Bill about the demand for liquid assets in uncertain times). But if people knew that the Auctioneer would also be clearing markets in future, expected future real income would be higher, so current consumption and investment demand would be higher, and the current equilibrium real interest rate would be higher too.
If good Fed communications could raise expected future real income (as well as raise expected inflation) that would work. (Scott Sumner adds these two effects together by talking about expected future NGDP).
Nick, this statement “But if people knew that the Auctioneer would also be clearing markets in future, expected future real income would be higher,” is not necessarily true.
An economy with a sufficiently large fall in the future dependency ratio, for example, could well have lower future real incomes even under the assumption that full employment is always maintained.
Adam: Agreed. What I meant is that I think it is true for the US right now. And I meant that people in the US currently expect that the economy will be in recession next year, and that next year’s income would be higher if a Walrasian auctioneer got them out of that recession.
Nick, agreed it’s not true of the US right now but it is true of China right now.
My whole point is that China has this problem and is exporting the effect via their Forex peg and capital controls.
But if people knew that the Auctioneer would also be clearing markets in future, expected future real income would be higher, so current consumption and investment demand would be higher, and the current equilibrium real interest rate would be higher too.
In New Keynesian language, I think this becomes the following: if a negative output gap* is expected in future periods, then the real interest rate necessary to eliminate the output gap this period is much lower than it would need to be if there was no output gap (or positive output gap) in future periods. This is undoubtedly true.
That said, I think it is possible that even if the output gap was expected to disappear tomorrow, the equilibrium real interest rate would be negative. This can be true for fundamental reasons like an expected fall in the dependency ratio, as Adam suggests, but I think that increases in financial frictions and the liquidity premium, along with “animal spirits” that shift savings preferences, may be much more important in practice. Alternatively, if the output gap is not expected to disappear tomorrow, there is also a large demand for precautionary savings that drives down the equilibrium real interest rate—this is a different channel (one that’s not explicitly modeled in many places) by which expectations of the future can affect today.
* where by “negative” I mean “output is less than potential”.
Pingback: Balance sheets and reality | Matt Rognlie
Wait, what?
“why should the desire to spend less and save more hurt the economy?”
Isn’t this covered under “everything we know about macroeconomics”? http://en.wikipedia.org/wiki/Paradox_of_thrift
“At the micro level, economists don’t worry about weak demand causing a supply glut: instead, they correctly say that prices will adjust to clear the market. The same is true for macro as well. There’s a price–the real interest rate–that determines willingness to save and invest. Like all prices, this price has a market-clearing level: at a sufficiently low real interest rate, the supply and demand for savings will equate, and consumers’ desire to save will translate into an investment boom, not an economic downturn.”
Isn’t this the Efficient Markets Hypothesis in a nutshell, that the price mechanism will always fix things?
Did someone inject you with Hayek’s DNA while we weren’t looking?
The “paradox of thrift” only makes sense conditional on the Fed holding interest rates constant (or, at least, not being sufficiently accommodative). This is pretty much universally acknowledged at this point, by figures including Paul Krugman. The Wikipedia article does not make this sufficiently clear—another sad instance of the fact that all Wikipedia articles on economics are terrible.
Isn’t this the Efficient Markets Hypothesis in a nutshell, that the price mechanism will always fix things?
No, this has absolutely nothing to do with the Efficient Markets Hypothesis. In fact, the key point here is that although there is some “correct price” consistent with equilibrium (i.e. where deleveraging doesn’t lead to a demand-side recession), the market doesn’t reach that price on its own: since the Fed controls the short term interest rate, it needs to take the appropriate action.
In paragraph 6,
“Won’t a decrease in savings then mean that the economy crashes?”
I think “decrease” is probably a typo for “increase.” (What you really mean is an increase in intended savings. This will likely result in a decrease in actual savings via the paradox of thrift, but I don’t think that’s what you meant to emphasize.)
Definitely a typo. It’s fixed now—thanks!
(I also agree with the point that conditional on an insufficiently accommodative monetary policy, the paradox of thrift will result in a decrease in actual savings, but you were also right that this wasn’t my intent: by savings I really meant desired savings.)
David,
“Irresponsibility” is a vague term. The Fed has to promise a sufficiently high future inflation rate that, as it happens, will turn out to be higher than what will be optimal when the time comes. The Fed promises to be irresponsible in the sense of promising to follow a policy that will not be optimal when the time comes. It doesn’t have to promise to exceed that (already “irresponsible”) inflation rate, and it can surely promise to be “responsible” in the sense of containing any overshoot of that “irresponsible” inflation rate.
Andy,
I don’t think the metric is, “higher than optimal”. I think it is, “high enough to induce a change in behavior.” This is a complicated issue. It has to do as much with the shape of the probability distribution for inflation than anything else. Exiting the gold standard was, “irresponsible” because it created a much fatter tail for the inflation side of the distribution. Promising to bring inflation to 4-5% for two years and then return it to 2% barely influences the perception of that distribution. Certainly, I cannot imagine it inducing actors to run out and buy houses.
The effect of expected inflation on behavior is primarily one of inducing hedging against a loss in purchasing power. This hedging behavior, once it begins, acquires its own momentum. It accelerates without further easing; to brake it requires tightening. Tightening in the presence of high unemployment and high projected fiscal deficits is quite difficult. This is how we end up with high and variable inflation.
Your analysis of inflation stabilization focuses heavily on its implications for nominal contracts: variable inflation introduces unnecessary risk into those contracts, and all else equal it should be avoided. I agree. But if you take this as our goal, we should really be interested in stability of price level expectations, not inflation expectations per se.
Consider the following two processes for the price level. Under process (1), the inflation rate is 0% for one half of each year and 4% for the other half; overall inflation for the year is always 2%, though which half of the year will experience inflation is random. This process has fairly extreme “variable inflation” at high frequencies, and no doubt whenever the 4% inflation period rolls around hawks will give the Fed an earful. But by any reasonable standard, this process does an extremely good job at ensuring long-run price stability: you can never be more than a few percent off in your prediction, even decades hence.
Under process (2), the inflation rate is extremely steady over time, but there is some uncertainty about the exact trend rate of inflation. Suppose that the long-term trend rate of inflation will either be 1.8% or 2.2%. Your uncertainty about the price level 20 years from now is then 8%–quite a bit more, even though the “inflation rate” is far less volatile. This is undoubtedly a more difficult environment than (1) for all but the most short-term contracting.
If your main concern is being able to write nominal contracts with confidence, I submit that you should prefer (1), and more generally a policy of price-level targeting rather than inflation-level targeting.
Such a policy is perfectly consistent with my proposal that we temporarily accept slightly above-trend inflation. In fact, price-level stabilization requires this: we’ve experienced below-trend inflation during the recession, and now we need a catch-up period to make sure that the price level stays on track.
I want to also mention that price level stabilization is not the only valid approach to reducing risk in nominal contracts. After all, why should contracts’ payout be stabilized vis-a-vis cost of a certain basket of goods? Why not stabilization versus the average nominal wage? (i.e. relative to what people can afford to pay back?) Or maybe nominal GDP? I’ve discussed this issue in a previous blog post.
Matt,
My first point was that price level targeting is not “irresponsible”. It may, in a steady-state condition, work wonderfully. In the current condition, it would arguably be insufficient to change behavior. When a “fat tail” risk exists for deflation, price level targeting won’t work. The policy promises an asymmetric expected profit distribution; that is, a large loss if the policy fails and deflation occurs; a small gain if the policy works. How would actors facing this probability distribution behave? By hedging against deflation, not inflation.
Adam P:
You are mistaken.
Smith owes Jones $100 and borrows $10 from Davis this year
Simth has debt of $110. Jones has no debt. (He has an asset of $100, but we ignore that for these calculations.)
Smith pays Jones back next year, but borrows another $10 from Davis.
Smith’s debt is now $20.
Smith’s debt fell from $110 to $20.
Smith and Jones debt fell from $110 to $20.
Smith and Jones owe Davis $20 more.
If Smith and Jones are the U.S. and Davis is the rest of the world, then total debt in the U.S. has dropped but debt to the rest of the world has increased.
Of course, Matt pointed out that the capital flight means that Smith is borrowing less from Davis, say an additional $5, so today debt drops from $110 to $15. Still a net capital inflow for the U.S., but a reduction in debt.
Bill I’m not mistaken, at least not in the way you think. Your example is simply wrong.
Smith’s income is not exogenously given, he can earn income only by selling to Jones or to Davis.
Now, suppose Davis is willing to lend money to Smith or Jones but doesn’t buy anything from either of them. Smith and Jones each make a different good and trade with each other, Smith’s income is then only what he sells to Jones and vice versa.
Suppose Jones never, ever spends more than his income but did once spend $100 less and he lent that money to Davis so that Davis could buy $100 more from him (he had a very high output year that year and is absolutely committed to keeping the asset until the last year of his life where he will rest while Jones will have to work for nothing – assume Jones will do this, he won’t default).
Now we are at your case where Smith owes Jones $100.
How can Smith pay back the $100 debt, he doesn’t have the money because he borrowed it and spent it. The only way to get $100 more than expenditure is for Jones to spend $100 more buying from Smith than Smith spends buying from Jones. But we assumed Jones doesn’t ever spend more than his income (and Davis never buys anything from Smith)! Under that assumption it’s impossible for Jones to pay back the loan out of his income, he can’t delver. If he tries to save $100, this reduces Jones’ income by $100 and causes Jones to reduce expenditure by $100 which means Smith’s income falls by that same $100.
Now, perhaps a reduction in the real interest rate might induce Jones to spend more than his income, essentially selling the bond back to Smith in return for his output. But if the real rate is fixed, say at zero, then you can’t get him to do this.
Now Smith can pay back Jones by borrowing $100 from Davis but he hasn’t delevered, he still owes $100. Without some period in which one or both of Jones or Davis spending more on Smith’s output than they have income Jones can never delever.
If Jones and Smith are the US and Davis the rest of the world then the US, in aggregate, can’t delever either. Of course in this story we could have just had Jones be the rest of world and Smith be the US. Same thing applies.
I see a couple of potentially reasonable objections to a policy of allowing the real interest rate to drop low enough to produce full employment, but people who make these objections need to be aware that they’re objecting to the way that markets normally function (or at least the way first semester microeconomics instructors tell us they normally function).
The first — and I think it is at core the objections that Adam and David are making — is what one might call the “bumper crop” objection. If there’s a bumper crop of some commodity, one could at least imagine that the destabilizing effects of allowing the price to fall to a market-clearing level would be worse than the surplus condition that develops if there is a binding price floor. For most markets, with most economists, this objection is a non-starter, but when the commodity involved is savings, you get a lot of objections. And they may not be entirely without merit. The lower the real interest rate goes, the more sensitive asset prices become to forecasts of future returns. After having lived through the last 15 years, it doesn’t seem too radical to suggest that the resulting instability could be a big problem, perhaps (though I personally don’t think so) a bigger one than the unemployment that results from a binding floor on the real interest rate.
The second objection is that the equilibrium may not be stable, or there may be no equilibrium at all, at least within the range of what’s reasonable. We like to think we live in a nice world where demand curves slope downward and supply curves slope upward. But perhaps savings are a Giffen good. Or perhaps…I dunno, it’s not too hard to imagine that saving and investment don’t respond to interest rates the way they’re supposed to. I think this is what the MMT people have in mind. Generally, a deleveraging theorist could argue that the market clearing real interest rate is so extremely negative as to be ridiculous, and as a result the theory of market clearing isn’t the best way to think about the problem. Then there is Nick Rowe’s variant of the second objection: the equilibrium in terms of interest rates is unstable, so it’s more useful to think in terms of money.
I see a couple of potentially reasonable objections to a policy of allowing the real interest rate to drop low enough to produce full employment, but people who make these objections need to be aware that they’re objecting to the way that markets normally function (or at least the way first semester microeconomics instructors tell us they normally function).
Exactly. There are reasons why we might want to implement a real interest rate that differs from the “flexible-price equilibrum” real interest rate. (Though I can think of more cases where we’d want to err in the other direction—for instance, maybe the equilibrium real interest rate isn’t low enough to overcome asymmetric information failures that prevent job matching and keep unemployment high, and we want to temporarily go below in order to introduce an “unnatural” boom and push down the stock of unemployed.) But this should be recognized for what it effectively is: government intervention in prices that is supposed to serve some kind of social function.
The lower the real interest rate goes, the more sensitive asset prices become to forecasts of future returns. After having lived through the last 15 years, it doesn’t seem too radical to suggest that the resulting instability could be a big problem, perhaps (though I personally don’t think so) a bigger one than the unemployment that results from a binding floor on the real interest rate.
This is an interesting point, but I am a little skeptical about the comparative magnitudes. A 15% decline in expected cumulative real interest rates for the next 3 years would be absolutely huge: it would almost certainly pull us out of the current malaise into healthy growth. Yet it would only increase the sensitivity of asset prices to future returns by… 15%. In general I suspect that the direct effect on current economic conditions absolutely overwhelms this “sensitivity to expectations” effect.
Generally, a deleveraging theorist could argue that the market clearing real interest rate is so extremely negative as to be ridiculous, and as a result the theory of market clearing isn’t the best way to think about the problem.
Responding to this kind of objection is essentially the purpose of my follow-up post: looking at the data, it’s hard to swallow the arguments from Koo & Co. that everyone is so hopelessly overleveraged that the incremental impact of real interest rates is nil.
Matt,
You argue a -5% p.a. real rate would have a large impact on growth. Let’s say in year four the economy grows at 3% with 4% inflation. At that point the “natural” real rate might be 2%. What should the stance of policy be? Can the Fed raise the real rate by seven percentage points without tipping the economy into recession? What if actors perceive in year 1 that in year 4 this would happen? Or what if, instead, they perceive that the Fed will not slam on the brakes in year 4, but instead allow inflation to accelerate?
It is almost always easy to imagine how entering into stimulus might, “work”. The trick is to imagine the exit.
“A 15% decline in expected cumulative real interest rates for the next 3 years…would almost certainly pull us out of the current malaise into healthy growth.”
I’m not so sure. House prices went up way more than 15% during the last business cycle, and that was barely enough to bring us to full employment. And then house prices crashed, because (in a roundabout sort of way) the market decided its estimates of the returns on those houses had been a little bit off. Why should we expect things to be different this time?
Adam P:
You have now tried to figure out everyone’s income too.
Jeez.
I simply said that there could be a decrease in debt in the U.S. even while people in the U.S. borrowed more from foreigners.
You decided that the U.S. creditors never, ever, spend more than their incomes.
Where did that assumption come from?
By the way, I don’t generally think in terms of three person economies. So, the debtor and creditor don’t earn incomes from buying and selling to one another. They may be trading with other people to earn these incomes, and perhaps none of those other folks have debts at all.
There is no doubt that what must happen to reduced debt in the U.S. is that the debtors spend less than than their incomes (paying down their debts,) and the creditors spend more than their income, receiving the debt payments.
Now, if the “creditors” spend on capital goods or even buy equity, then they don’t have to dissave. They can swtich from lending to direct investment or equility investment.
But if we ignore that, then yes, the former creditor must dissave.
Now, if we are in this world, where net worth in aggregate is zero anyway, then we cannot increase net worth by paying down debt. It stays zero.
But that isn’t what is generally meant by deleveraging. It is just paying down debt, and so former debtors consuming less and former creditors consuming more reduces debt.
And your claim that this cannot be done in an open economy if there is a net capital inflow is false.
(and yes, there can be a net capital inflow with direct investment or equity investment too, as Harliss said.)
The simple way to “deleverage” is to swap debt for equity. Firms sell stocks and pay off bonds. Creditors use the funds received from the bonds to buy stocks. If “leverage” were really the problem–fixed.
Bill, in a closed economy *net* debt is zero, thus so is net leverage. I was quite careful to say “aggregate” in my comments on deleveraging. Net debt of a country is its external debt.
Thus the assumption that the net creditors never spend more than their income is just a restatement of the assumption that the debtor country is running a persistent CA deficit.
The story I’ve just told says nothing more than that it is not possible for the Western world to collectively run a CA surplus without the non-western world collectivey running a CA deficit.
And my point is that it is not possible for the Western world to lower their net debt (which equals their collective “external” debt – the net debt that the West owes to the non-west) without being able to run a CA surplus.
Furthermore, switching to equity financing doesn’t solve this problem. The issue is a limit to the amount of output that the West is willing to pledge externally, the form of the claims doesn’t change that limit.
The west wants to reduce the volume of claims they are issuing, that is to reduce its collective CA deficit with the non-West. The non-West is making this impossible.
Harless:
Where would any real interest rate be “ridiculously” low if it were required to coordinate saving and investment?
I agree exactly with your examples of market clearing prices in particular markets.
Why are giffen goods a problem if the market is stable.
And while I can imagine a locally unstable potatoe market, a globally unstable one is absurd. At a high enough price, potatoes are normal again. But it would be sad for the poor potato consumers.
Now, suppose saving is negatively relatived to the real interest rate, but less so that investment. So?
If it is unstable, with saving more negatively related to interest than investment, then that would be bad, but I am pretty sure that saving will be zero at an interest rate of -101%. There is a budget contraint on the savers that is going to put a stop to it.
I am pretty sure that an interest rate of -101% will drive saving to zero. All output is consumed. Now, if there are some investment projects that can be funded for free, I don’t think we will get that far.
I think the normal situation is positive real interest rates. But if people want to produce goods for now, and earn income for now, they need to figure out how to spend those funds now, unless they can figure out an investment project or can find someone else who wants to consume now or has an investment project.
“The economy is depressed as it gradually returns to the correct level of leverage, and we experience the ”long, painful” process of deleveraging. (In the words of, well, every blogger and amateur econ pundit in the world.)”
Isn’t this Richard Koo’s argument?
“If I want to save, I hand my money over to someone who wants to borrow or invest it. Net saving is channeled into productive investment.”
What if the money is used to buy treasuries? That’s not exactly a productive investment, is it?
Pingback: Do Balance Sheets Matter? « Modeled Behavior
Pingback: Is Focusing on Deleveraging a Useless Distraction? | Rortybomb
The fed is just as informational and time constrained as other economic actors. Just because it has the power to set real interest rates at an optimal level doesn’t mean we have any reason to believe that it will do this.
The zero lower bound is important, but the main reason we should be focused on deleveraging and debt in general is that they are integral in how shocks are propagated throughout the economy. Certain structures increase the risk of catastrophic breakdown, (borrowing short and lending long is the classic example of this). During the crisis the fed can act as lend of last resort, but this is obviously very difficult.
After the crisis lots of deleveraging is a drag on the economy for the reason any badly handled bankruptcy destroys value. It is better for both the creditor and the debtor if the debt is quickly wrote off to a level to a level where the creditor can pay back what is owed and is able to get new loans at an affordable rate. Lowering the interest rate can make this process easier, but it is secondary to the actual process of deleveraging.
TLDR version: This analysis needs more Minsky and less Friedman.
I like this post, and think it broadly supports the idea that the Fed should engage in nominal GDP targeting, perhaps rules-based.
As for deleveraging, I am all for it, especially as expedited by inflation.
The most important thing to avoid is an unhealthy yet banal obsession with inflation, and a fetish for zero price increases (if that could even be measured in a world of rapidly evolving goods and services). If the USA ran five percent inflation and five percent real growth for five years, I assure you everyone would be happy campers (except possibly for gold buts and Austrians, and possibly drug dealers who hold lots of cash, or so I see in movies, and maybe a few people in 100 percent bonds), This seems so obvious to me.
The economic policy goals should be economic growth, prosperity, innovation, and commercial and economic freedoms. Price stability comes on the second train.
If you start with the premise that the moral and economic goal is price stability, you will end up like Japan.
Pingback: Incrementally reducing your negative equity is not saving « Synthetic Assets
Pingback: Incrementally reducing your negative equity is not saving — Clearing and Settlement
Pingback: Dificuldade em ler coisas da direita | Blog Pra falar de coisas
“Net saving is channelled into productive investment.”
Deleveraging is not necessarily about net saving and investment – no more than gross saving is about net saving. Gross activity is larger by the amount of dissaving.
“At a sufficiently low real interest rate, the supply and demand for savings will equate, and consumers’ desire to save will translate into an investment boom, not an economic downturn”
That may be true in part, but again consumers’ desire to gross save doesn’t necessarily translate to a desire to net save in the same amount. There’s no direct investment involved in a resold house, but the buyer has saved for the down payment. The seller on the other hand may consume rather than invest. The result is gross saving with zero net saving.
“So yes, deleveraging can be very bad for the economy. But this is only because monetary policy doesn’t adjust enough to match the market.”
Agreed, but not just because negative rates aren’t available to strike a better balance between saving and investment – also because negative rates aren’t available to ease the carrying cost of debt, whatever the situation is regarding investment.
So I think there’s a high correlation between the ideas of “deleveraging” (as currently viewed) and the zero bound.
But specifying the definition of deleveraging is complicated, given the differences and interactions between gross and net saving activity.
Pingback: The Problem’s Not (Monetary) Policy, it’s Politics « thecurrentmoment
Pingback: interfluidity » The negative unnatural rate of interest
Pingback: Wednesday Morning Links | Iacono Research