Friday, September 19, 2014

What Have We Learned Since 2009?

In September 2009, Paul Krugman wrote "How Did Economists Get It So Wrong? The key points in that magazine article were:

1. Economists did not see the financial crisis coming, therefore there is something wrong with economics.
2. It's not hard to see what is wrong with economics - it got seriously sidetracked, roughly in 1970.
3. What to do? According to Krugman:
First, they have to face up to the inconvenient reality that financial markets fall far short of perfection, that they are subject to extraordinary delusions and the madness of crowds. Second, they have to admit — and this will be very hard for the people who giggled and whispered over Keynes — that Keynesian economics remains the best framework we have for making sense of recessions and depressions. Third, they’ll have to do their best to incorporate the realities of finance into macroeconomics.
So, what's changed since 2009, at least in Krugman's mind? Not much, apparently. Krugman's Monday NYT column was titled "How to Get it Wrong." Now it's not just the economists who are stupid/misguided/nefarious - it's the policymakers as well. Maybe this is good news if you are an economist. If you are an economist and a policymaker, not so good, I'm afraid.

Not one to shy away from hyperbole and invective, Krugman tells us about an "enormous intellectual failure" in economics, and the "sins of economists, who far too often let partisanship or personal self-aggrandizement trump their professionalism." Economists are clearly a group of bad people, and if you follow the links in his piece, you can figure out who some of the bad guys are. Bob Lucas, for example, who played a starring role in Krugman's 2009 piece. Apparently Lucas was one of those people who imagined an "idealized vision of capitalism, in which individuals are always rational and markets always function perfectly."

So, let's start with that. Part of Krugman's game is to convince you of the moral inferiority of the people he disagrees with, which serves to excuse his name calling. That's what he says in this post:
And if you look at the uncivil remarks by people like, well, me, you’ll find that they are similarly aimed at people arguing in bad faith ... what I’m doing is going after bad-faith economics — economics that keeps trotting out claims that have already been discredited.
The problem of course, is that "discredited" is in the eye of the beholder - Krugman's eye basically. If I follow those rules, I can "discredit" someone, then abuse them all I want.

So, what do the "discredited" economists do? Apparently, those bad people think that "individuals are always rational and markets always function perfectly." First, on rationality, I don't think Krugman has a published paper with irrational economic agents in it (someone please point one out to me, if this is incorrect). So, if we follow what Krugman does, rather than what he says, that might actually conform to my defense of rationality in economics:
In my view, irrationality is the great cop-out, and simply represents a failure of imagination. Rationality is so weak a requirement that the set of potential explanations for a particular phenomenon that incorporate rationality is boundless. If the phenomenon can be described, and we can find some regularity in it, then it can also be described as the outcome of rational behaviour. Behaviour looks random only when one does not have a theory to make sense of it, and explaining it as the result of rational behaviour is literally what we mean by ‘making sense of what we are seeing’.
Second, modern macroeconomics (which Krugman detests) is replete with frictions - it's hardly about markets functioning perfectly. For example, Evil Ed Prescott did some work, with Finn Kydland, on models in which markets work really well. RBC models are basically special cases of Arrow-Debreu. The whole idea is to see how much you can explain without having to resort to frictions. But Kydland and Prescott received the Nobel Prize in 2004, in part for their work on time consistency. The basic idea is that the ability of a policymaker to commit is critical. Even a benevolent policymaker could choose bad policies if there is nothing to stop him or her from breaking promises. So, that's a world that doesn't work perfectly, and the time consistency idea has been highly influential in policy circles and in macroeconomic modeling. Indeed, it's very important in New Keynesian economics. That's an example of specific research by someone in Krugman's bad-person camp, that doesn't fit the profile. More generally, if you look through this year's SED program, you're going to find a lot of frictions - more frictions than you can shake a stick at. And this is from the society founded by the Minnesota macros, basically.

A new complaint of Krugman's is that the academic economic journals are ignoring good work:
...starting in the 1980s it became harder and harder to publish anything questioning these idealized models in major journals.
By "idealized models," I think he means Arrow-Debreu. That's quite a charge - apparently the economics profession is so ill that scientific progress is grinding to a halt. So, what's the evidence for the charge? Krugman quotes Ken Rogoff, who has a gripe:
There are more than a few of us in my generation of international economists who still bear the scars of not being able to publish sticky price papers during the years of new neoclassical repression.
Yes, and poor Rogoff was so repressed that he carried his scars from UW-Madison, to Berkeley, to Princeton, and now must reside, scarred, in that intellectual backwater, Harvard University. Complaints about academic journals are universal. There is no shortage of whining about the nasty Keynesians, the nasty theorists, the nasty new old non-neoclassical whoevers who reject one's papers. Being misunderstood comes with the territory in economics, I'm afraid. Best have a thick skin if you want to survive. If the best Krugman can come up with is a complaint from 13 years ago about what happened 30 years ago, this isn't worth much.

If there is an important change in Krugman's approach from 2009 to today, it's in what he expects from the profession. In 2009, he wanted people "to admit that Keynesian economics remains the best framework we have for making sense of recessions and depressions." Presumably "Keynesian economics" includes New Keynesian economics. Further, Krugman was willing to recognize, at that time, that there was interesting work going on outside of what we might call Keynesian thought:
One line of work, pioneered by none other than Ben Bernanke working with Mark Gertler of New York University, emphasized the way the lack of sufficient collateral can hinder the ability of businesses to raise funds and pursue investment opportunities. A related line of work, largely established by my Princeton colleague Nobuhiro Kiyotaki and John Moore of the London School of Economics, argued that prices of assets such as real estate can suffer self-reinforcing plunges that in turn depress the economy as a whole.
That seemed promising. At least Krugman was aware of some of the work, preceding the financial crisis, which dealt with financial frictions. A lot of that work continues, and there is much more of it now, propelled of course by the financial crisis experience. Some of that recent work was presented at a conference that I organized, with Mark Gertler (see above) at the St. Louis Fed late last year. The people at that conference were from both sides of the fence that Krugman imagines divides the macroeconomics profession. The papers were very interesting, and they went some way toward helping us understand the causes and consequences of the financial crisis. Progress!

The 2014 Krugman has lowered his sights considerably.
...the world would be in much better shape than it is if real-world policy had reflected the lessons of Econ 101.
This isn't a call for refurbishing macroeconomic thought with irrationality and frictions. What Krugman wants is a reversion to the technology of 1937. Why? From this blog post:
Why is output so low and jobs so scarce? The simple answer is inadequate demand — and every piece of evidence we have is consistent with that answer...there is a deep desire on the part of people who want to sound serious to believe that big problems must have deep roots, and require many hours of solemn deliberation by bipartisan panels.
So, now I'm confused. Krugman says macroeconomics went wrong after 1970, and one of the primary culprits was Lucas, who apparently told us that economies work perfectly, and don't need help from policymakers. Krugman also links to this 2003 paper by Lucas from which this quote is extracted:
Macroeconomics was born as a distinct field in the 1940’s, as a part of the intellectual response to the Great Depression. The term then referred to the body of knowledge and expertise that we hoped would prevent the recurrence of that economic disaster. My thesis in this lecture is that macroeconomics in this original sense has succeeded: Its central problem of depression prevention has been solved, for all practical purposes, and has in fact been solved for many decades.
This quote is intended as an example to show us how out to lunch mainstream economics was before the financial crisis. Obviously, Krugman is telling us, these guys didn't have a clue what was in store. But, in the conclusion to his paper, Lucas also tells us:
If business cycles were simply efficient responses of quantities and prices to unpredictable shifts in technology and preferences, there would be no need for distinct stabilization or demand management policies and certainly no point to such legislation as the Employment Act of 1946. If, on the other hand, rigidities of some kind prevent the economy from reacting efficiently to nominal or real shocks, or both, there is a need to design suitable policies and to assess their performance. In my opinion, this is the case: I think the stability of monetary aggregates and nominal spending in the postwar United States is a major reason for the stability of aggregate production and consumption during these years, relative to the experience of the interwar period and the contemporary experience of other economies. If so, this stability must be seen in part as an achievement of the economists, Keynesian and monetarist, who guided economic policy over these years.
This sounds nothing like the Lucas that Krugman is telling us about. In fact, he seems to have a lot in common with Krugman. Seems the 2003 Lucas thought that important inefficiencies exist in the absence of government intervention, and that "stabilization or demand management policies" can and did fix these problems. Lucas also tells us that the Depression problem has been fixed. We don't have to worry about it anymore - because stabilization and demand management are at the ready. Krugman says our current problems, as he sees them, can be fixed by standard "Econ 101" stabilization and demand management. Apparently Lucas was a dummy for thinking the depression problem had been solved by stabilization and demand managment. You idiot, our modern depression problem needs to be solved by stabilization and demand management.

Krugman is right, in a sense. The solutions to economic policy problems can indeed be simple - once you know the answer. It's determining the answer that is hard. If the solution to our policy problems was all in Econ 101, we wouldn't hire economists with PhDs to do economic policy. We wouldn't spend all the time we do hashing over policy problems, or arguing with Paul Krugman about why it's not as easy as he claims. Further, to say that the solution is in Econ 101 is actually an insult to Econ 101 students. Econ 101 students, if they are taught properly, know that the inefficiencies resulting from price and wage rigidity, if they exist, don't last forever, and if they lasted 6 years after a financial crisis occurred, we would be very surprised. Thus, a smart Econ 101 student might question why conventional demand management is going to solve whatever problem exists. Well-trained Econ 101 students also understand that, if we want to solve an economic problem, it will help a lot to understand the cause of the problem. If as Krugman tells us "textbook economics...didn't predict the crisis," or indeed tell us anything about what a crisis is, or what would cause it, what's it going to say about how to fix things up once the crisis hits, or six years after it went away?

Monday, September 8, 2014

Theories of Inflation and the European Predicament

How do macroeconomists think about inflation? To get a grip on this, it's useful to dig into the history of economic thought. Long ago, in the mid-1970s, when I had no idea what formal economics was about (and, you might say, nothing much has changed), I had heard about "wage-price spirals." A web search will give you plenty of descriptions of what we might charitably call the "wage-price spiral theory" of inflation. Here's one that's as good as any:
When an economy is operating at near full employment and people have money to spend, demand for goods and services increases. To meet the demand, companies expand their businesses and hire more workers. However, at near full employment, most workers already have jobs. So companies have to lure workers with higher wages, which, of course, increases the companies' costs, explains the website Biz/ed. The workers then push for higher wages to meet the higher prices and expected price hikes, which increases company costs again. Theoretically, this continues in an inflationary spiral until a loaf of bread costs the proverbial wheelbarrow full of cash.
You'll notice that the behavior of governments and central banks doesn't enter into the story. Apparently a wage-price spiral is a self-fulfilling dynamic process which could start under any conditions and continue forever - unless something is done about it. In the 1970s, some policymakers thought the solution to a perceieved wage-price spiral problem was (naturally) wage-price controls. For example, Richard Nixon tried a 90-day wage-price freeze, and the government of Canada imposed wage-price controls for a much longer period of time. Such controls were advocated by economists like John Kenneth Galbraith, who perhaps wasn't taken so seriously by most academics, but even mainstream macroeconomists such as James Tobin could sometimes find "incomes policies" attractive.

In current academic circles, there isn't much talk about wage-price spirals, though of course some ideas never die. Perhaps the primary achievement of the Old Monetarists - principally Milton Friedman - was to convince people that inflation control is the job of central banks, and that wage-price controls only produce inefficiencies (though experience with those policies was pretty convincing as well). So, if inflation control is the job of the central bank, and we think there is something wrong with the inflation rate, we know who to blame. Of course, Friedman seems to have failed to give central banks useful instructions for implementing inflation control. He argued that there is a direct link between money growth and long-run inflation, that monetary disturbances are a primary determinant of fluctuations in real GDP, and that "fine-tuning" is inappropriate. So, Friedman reasoned, the appropriate monetary policy is:
...that the monetary authority go all the way in avoiding such swings by adopting publicly the policy of achieving a steady rate of growth in a specified monetary total. The precise rate of growth, like the precise monetary total, is less important than the adoption of some stated and known rate. I myself have argued for a rate that would on the average achieve rough stability in the level of prices of final products, which I have estimated would call for something like a 3 to 5 per cent per year rate of growth in currency plus all commercial bank deposits or a slightly lower rate of growth in currency plus demand deposits only. But it would be better to have a fixed rate that would on the average produce moderate inflation or moderate deflation, provided it was steady, than to suffer the wide and erratic perturbations we have experienced.
Like wage and price controls, targeting of monetary aggregates was tried (in the 1970s and 1980s, primarily) and it failed. The relationships among monetary aggregates, inflation, and real economic activity proved to be highly unstable, causing Old Monetarist prescriptions to work poorly. That doesn't stop the proponents of Old Monetarism -a dwindling breed- from trying. We can still find work involving searches for the elusive ideal monetary aggregate or searches for the elusive stable money demand function.

So, if inflation control through money growth targeting works poorly, what to do then? Some central banks opted to target inflation directly or, more generally, there was some recognition that the central bank should take nonneutralities of money into account, and Taylor wrote down a simple rule of thumb that would allow for that. The Taylor rule was subsequently enshrined in New Keynesian models, along with an updated version of the Phillips curve.

The Phillips curve appears to be the modern version of the wage-price spiral. Typically, that's how the inflationary process is described for the lay person. For example, here's from a recent blog post in The Economist:
The American economy, we wrote in July, almost certainly has less room to grow than it used to. Estimates of the economy's potential output, or how much it can produce at a given time without serious inflationary pressure building, have been revised down substantially by the Congressional Budget Office and other economists studying the issue.
That was from a piece on potential output in the United States, but what I am interested in is the part about "serious inflationary pressure." Apparently there is more inflationary pressure the lower is the output gap - the difference between "potential" output and actual output. Clearly the writer(s) of this blog post subscribe to a Phillips curve theory of inflation. The Old Monetarists (Friedman) and modern structuralists (Lucas) may have thought they debunked Phillips curve thinking, but it's remarkably persistent. How come?

If you think a stable money demand function is hard to find, try to find a Phillips curve in the data. As with the money demand function though, nothing stops a committed Phillips curve adherent. Whether by finding the right combination of inflation measure and output gap measure, judicious use of Bayesian estimation, or whatever, by hook or by crook a Phillips curve can indeed be uncovered in the data. But, as I outline here, it's hard to make a case that the Phillips curve is helpful for thinking about inflation, its causes, and what to do about it. For example, Phillips curves are not useful in forecasting inflation (see this paper by Atkeson and Ohanian.).

Diehard Phillips curve folks, in extreme states of denial, will insist that the output gap is a latent variable, and thus the existence of low inflation implies that the output gap must be high. Indeed, from the blog post in The Economist, quoted above, if we take potential output to be defined by the behavior of the inflation rate (as seems to be implied by the quote), we should be able to back out a measure of of the output gap from the actual inflation rate. I'm pretty sure that, if you do that exercise, you will come up with nonsense.

But perhaps the Phillips curve - even as complete fiction - has been useful, if for nothing else than to permit agreeement among policymakers. In the recovery phase of a business cycle downturn, supposing the nominal interest rate and the inflation rate are low, an appeal to the Phillips curve can help in obtaining agreement to "tighten," i.e. to raise the policy rate. Even though inflation is low, it can be argued that "inflation pressure" exists, inflation threatens, and tightening should occur before it is too late. Old Monetarists, Old Keynesians, and New Keynesians alike, might find reasons to agree on that. Indeed, I think we can write down models where this would be self-fulfilling, and that type of reasoning - though it may actually be wrong - could yield the right policy decision. The policy decision would be right in the sense that it would avoid the bad equilibrium (bad in the sense of not achieving the central bank's inflation target) that converges to the zero lower bound and low inflation forever - see this paper by Jim Bullard.

So, what does this have to do with Europe? Here's what's happened to the average of inflation rates across countries in the Euro zone:
The inflation rate has fallen dramatically, and is now well below the ECB inflation target of 2%. As well, long bond yields (in this case of 10-year bonds) are down sharply:
Today's 10-year German bond yield is even lower than when I constructed this chart - it's 0.95%, relative to 2.375% for a 10-year U.S. Treasury bond.

The recent European experience might look familiar to someone who lived in Japan, say over the period 1990-1995. Here's CPI inflation in Japan:
And here's the 10-year bond yield for Japan:
There are some differences of course, for example it took longer for the inflation rate and bond yield to fall in Japan, but the experience is roughly similar.

What is the mandate of the ECB? From the ECB's web site:
The ECB’s main task is to maintain the euro's purchasing power and thus price stability in the euro area.
In case it's not clear to you what that means, there is an explanation here. Apparently price stability means that inflation rates between 0% and 2% are more or less OK, negative inflation rates are not OK, and maybe 1.8% is more OK than 0.8%. Suffice to say, though, that the ECB is changing policy, or about to change policy, on several dimensions, so it appears to think that the inflation you see in the first chart is definitely not OK, or projected to be not OK. Draghi's press conference after the policy change makes it clear that he's worried about a decline in inflation expectations, which you can see in breakeven rates on European government bonds.

So, since it appears the ECB wants to increase the inflation rate in the Euro zone, how does it intend to to it? First, the ECB has reduced its policy rates. The interest rate on the ECB's main refinancing operations was reduced to 0.05%, and the interest rate on deposits at the ECB was reduced to -0.20%. The refinancing rate is important, as the ECB does not typically intervene directly in an overnight market as is the case, for example, in the United States, but by lending to financial institutions, and the key lending facility is "main refinancing." The ECB has also taken the unusual step of charging for the privilege of holding reserves at the ECB, i.e. the ECB currently has a -0.20% lower bound rather than a zero lower bound. People - Miles Kimball among them - who think that relaxing the zero lower bound on the nominal interest rate will solve the world's problems get very excited about this. Second, there is a central bank credit program about to get underway, i.e. TLTROs (targeted long-term refinancing operations), which is central bank lending to European financial institutions with attached incentives to encourage these institutions to lend to the private sector. Third, there are planned asset purchases by the ECB - quantitative easing (QE) - with some of the specifics to be worked out later. Again, Draghi answers some questions about this in his press conference. It seems the asset purchases will take two forms: asset-backed securities and covered bonds. A covered bond is basically a collateralized bond, secured by a specified set of assets on the issuer's balance sheet, rather than being subject to the usual seniority rules in bankruptcy proceedings.

What theory of inflation could we use to think about the ECB's change in policy? Though some deflation-scare stories sound something like old wage-price spriral stories in reverse, let's dismiss that. What about Old Monetarism? It's reported in the press conference that M1 in the Euro area was growing at 5.6% in July, and that wouldn't alarm a quantity theorist, I think. However, belief in the Phillips curve is certainly consistent with what the ECB is doing, or proposing to do. Real economic activity is deemed to be weak, so a Phillips curve adherent might think that has something to do with the low rate of inflation in the Euro zone. Then, if we follow Old or New Keynesian prescriptions, conventional monetary easing - lower nominal interest rates - should increase output and inflation. Unconventional easing (QE and relaxing the zero lower bound) could be added to the Keynesian mix, given the zero-lower-bound problem.

But what if we think more carefully about the key elements of the "easing" program?

Starting with the last policy change first - ECB asset purchases (QE) - we might ask why this might matter. The press conference I think makes clear what I have in mind. Draghi says:
So QE is an outright purchase of assets. To give an example: rather than accepting these assets as collateral for lending, the ECB would outright purchase these assets. That’s QE. It would inject money into the system.
Then later:
Let me also add one thing, because the ABS may sound more, I would say novel, than they are in the ECB policy-making, and indeed, the modality is novel, because we would do outright purchases of ABS, but the ABS have been given as collateral for borrowing from the ECB for at least ten years, so the ECB knows very well how to price and how to treat the ABS that’s accepted, especially since we have, - and this is in a sense another dimension that makes any precise quantification difficult at this point in time - we have narrowly defined our outright purchase programme to simple and transparent ABS.
So, you might ask why it would make a difference if the ECB purchases a given asset outright, vs. extending a loan to a financial institution with the asset posted as collateral. Why would there be a bigger effect - and on what - in the first case relative to the latter?

Next is the TLTRO. Details of this program can be found here and here. This is basically an incentive program for lending by ECB financial institutions, tied to main refinancing operations - a kind of subsidized lending program. There's no particular link to inflation, unless we think there is some mechanism by which more credit - or credit reallocation - matters for inflation.

Finally, let's look at the ECB's interest rate policy, which I want to spend some time on. There are two parts to this. The first is conventional easing, which in the case of the ECB is a drop in its main refinancing rate - the rate at which it lends to financial institutions. The second is unconventional - a drop in the interest rate on reserves to a lower negative rate. The key worry here is that the ECB becomes trapped in a state with low inflation and low nominal interest rates forever, and can't get out. Note that this is a policy trap, not the "deflationary trap" that some people worry about. This is what Jim Bullard discusses in this paper. It's well known that conventional Taylor rules can have poor properties, and can lead to policy traps of this type. If inflation is low, the central bank lowers the nominal interest rate, which leads to lower inflation in the long run due to the Fisher effect. Ultimately, the economy converges to a steady state at the zero lower bound with inflation lower than what the central bank wants, and it can't get out unless the policy rule changes. Further, note that relaxing the zero lower bound won't help. If the central bank charges negative interest on reserves, this only lowers the inflation rate at the low-inflation steady state.

Could a central bank actually get into a low-inflation policy trap and not figure it out? The Japanese case shows us that central bankers can be very stubborn. In the case of Europe, here's what's been happening to the overnight interest rate and the inflation rate over time:
And here's the same data as a scatter plot:
You can see that there is substantial variation in the real ex post interest rate, but the Fisher relation shows up in the scatter plot, just as it does for the United States. Given this, and the first four charts in this post, it might enter your mind that Europe might be going in the same direction that Japan did 20 or more years ago.

How could we think about monetary policy in this context? Our goal is to determine what it takes to increase the inflation rate in an economy, taking into account short run effects, and the Fisher effect, which will dominate in the long run. One model that captures some of what we might be interested in is a segmented markets model. I'm going to use it because it's simple, and the key implications may not be so different if we were to include other types of short-run monetary non-neutralities. A very simple segmented markets model is in this paper by Alvarez, Lucas, and Weber. The details of what I did are in these notes that I've posted.

In the model, there are many households which each live forever. Each has a fixed endowment of goods each period, and sells those goods in a competitive market for cash. Goods are purchased from other households subject to cash-in-advance. There are two types of households - traders and non-traders. A trader can trade in the bond market each period, and holds a portfolio of money and bonds. Non-traders do not trade in the bond market, and hold only cash. The central bank intervenes by way of open market operations, the result being that an open market purchase of government bonds initially affects only the traders. There is a nonneutrality of money, which comes about because of a distribution effect of monetary policy. An open market purchase will increase the consumption of traders in the short run, and it is the consumption of traders that determines bond prices. Thus, when the open market purchase of bonds occurs, this tends to increase consumption of traders and nominal and real interest rates go down. But standard asset pricing implies that there is a Fisher effect - higher anticipated inflation implies a higher nominal interest rate.

In my notes, I work out the local dynamics of this economy. I don't worry about liquidity traps as I'm interested in what happens when the central bank gets off the zero lower bound. If the central bank experiments with random open market operations, it will observe the nominal interest rate and the inflation rate moving in opposite directions. This is the liquidity effect at work - open market purchases tend to reduce the nominal interest rate and increase the inflation rate. So, the central banker gets the idea that, if he or she wants to control inflation, then to push inflation up (down), he or she should move the nominal interest rate down (up).

But, suppose the nominal interest rate is constant at a low level for a long time, and then increases to a higher level, and stays at that higher level for a long time. All of this is perfectly anticipated. Then, there are many equilibria, all of which converge in the long run to an allocation in which the real interest rate is independent of monetary policy, and the Fisher relation holds. A natural equilibrium to look at is one that starts out in the steady state that would be achieved if the central bank kept the nominal interest rate at the low value forever. Then, in my notes, I show that the equilibrium path of the real interest rate and the inflation rate look like this:
There is no impact effect of the monetary "tightening" on the inflation rate, but the inflation rate subsequently increases over time to the steady state value - in the long run the increase in the inflation rate is equal to the increase in the nominal rate. The real interest rate increases initially, then falls, and in the long run there is no effect on the real rate - the liquidity effect disappears in the long run. But note that the inflation rate never went down.

We could add things to this model - liquidity effects on the real rate from scarce safe assets, real effects of monetary policy on aggregate output, etc., and I don't think the basic story would change. The story is that a sustained increase in the inflation rate is not possible unless the nominal interest rate goes up. Further, note that the real interest rate goes up temporarily in the process, and with a more fully developed model, there may be pain associated with that. Of course, in economics free lunches are always hard to find. If we think that higher inflation in the long run is good for us, it's hard to imagine there wouldn't be some cost to getting there.