Arnold Kling  

Taking Krugman's Side

Morning Commentary... Me on Obama and Brown...

Goodness knows, I don't want to. I still think he owes me a correction--and in fact an apology--in his column for putting in quotation marks an inaccurate, misleading, and distorted version of something I said.

But here is Tyler Cowen continuing to endorse Scott Sumner's monetarism.

I should add that I don't think anyone (I'm not talking about Krugman here) has responded to the claim that quantitative easing, and other monetary reforms, could substitute for a very expensive fiscal stimulus.

I will attempt to provide the sort of response that Tyler is looking for, because I think it can bring out some interesting points about macro theory.

Start with a Tobin model, in which there are three assets--money, bonds, and stocks, and stock prices drive investment. Tobin puts together a substitution matrix in which it is plausible that an increase in the public's holding of money and a decrease in its holding of bonds will serve to raise the demand for stocks, causing stock prices to go up and investment to increase.

Krugman is saying that monetary policy will not work in the current environment. He is saying that when the Fed buys bonds now, it is exchanging what in the public's mind are very close substitutes, so that the effect on other asset prices is minimal. I think that is true.

In fact, I think that it is almost always true that monetary policy is ineffective, because it is almost always true that money and other assets are close substitutes. In Can Greenspan Steer?, I pointed out that long-term interest rates were falling (this was August of 2002) even though Fed policy was to hold rates steady.

Think of an expanded Tobin model, with hundreds of assets and a gigantic substitution matrix. The Fed ordinarily makes short-term repurchase loans to bond dealers. The repo loans use near-term Treasuries as collateral. Out of all the interest rates in the system, the Fed is basically working on the short-term repo rate for holders of riskless collateral. There are plenty of reasons why open market operations might leave other rates unaffected, and those other rates are more important determinants of spending by businesses and consumers.

The only time when I think that monetary policy matters is when there is runaway inflation. When we have runaway inflation, then money is not a good substitute for other assets. Money is depreciating, while hard goods are appreciating and financial assets have nominal interest rates that incorporate an inflation premium. In that environment, with highly imperfect substitutability between money and other assets, a slowdown in money growth will help to end the inflation.

However, in normal times without runaway inflation, money and short-term assets are close substitutes. That makes monetary policy pretty ineffective.

It is true that the Fed could use its balance sheet to purchase assets that are not close substitutes for money. It could buy securities backed by mortgages or commercial loans. That would have real effects. But that would be equivalent to an expenditure that subsidizes mortgage lending or commercial lending. Whether you call it fiscal or monetary policy may come down to a matter of semantics. To see the similarity, consider that the Fed might buy state and local bonds to lower the cost of financing shovel-ready road projects.

Sumner and Cowen talk about the possibility of announcing an inflation target of, say four percent. If announcements of policy targets can change behavior, then that certainly is a cheap way to fight a recession. However, I don't see how the strip-mall business in Peoria is supposed to hire a worker or take out a loan based on the inflation target announced in Washington.

Comments and Sharing

COMMENTS (19 to date)
Tyler Cowen writes:

Of course I favor the latter case, buying the higher-yield assets. Yes it is a subsidy but keep in mind you're not trying to make the subsidy "work" in any traditional sense. You're just trying to get some AD out there (to be noticed in Peoria).

The real comparison is with fiscal stimulus and Congress investing $800 billion or whatever, with limited bang for the buck. Don't you think it beats that for the most part?

I will definitely agree that the AD shortfall is a secondary rather than primary cause of the crisis, and in this regard quantitative easing (and stimulus) can only do so much good. They can't cure the basic problems.

fundamentalist writes:

This is an argument commonly used against the Austrian business cycle theory because it is a monetary theory. However, there are a couple of problems with it. If money and short-term assets are near substitutes, why does anyone trade them? The answer is because they're not exact substitutes. It's hard to spend treasuries. You need cash. Treasuries appreciate, sometimes, whereas money always loses value.

The other problem with the argument is that money is fungible. A company may be holding cash to pay short-term expenses and not be able to invest it in long-term projects. But with low short-term rates and willingness by the banks to lend, companies can borrow for short-term needs which frees up funds for long-term investment. In effect, the short term borrowing makes long-term investment possible.

Alex writes:

And what happens if we increase both M and B, wouldn't that drive the prices of stocks up in your model?

Bill Woolsey writes:

Kling's theory is that when the Federal Reserve buys up assets like T-bills, those who were holding the T-bills choose to instead hold more money. The increase in the quantity of money is matched by an increase in the demand for money. The increase in the quantity of money is matched by a decrease in velocity.

That part of base money is paying higher interest than T-bills certainly exacerbates that effect. Sumner called for stopping this policy and instead looking to charging banks for holding reserves.

Sumner's view (and mine) is that if purchasing T-bills with newly created money leads to a matching decrease in velocity/increase in the demand for money, such that this effect is complete, then the Federal Reserve will first buy up all the T-bills and, by assumption, this will have no impact on nominal income. If, as seems more likely, T-bills are imperfect substitutes for money, then there will be only a small expansionary effect. And then the Fed goes on from there with progressively longer term and more risky assets.

The Fed should buy the best assets it can in whatever quantity needed to raise the quantity of money so that any partially matching drop in velocity is overwhelmed and the initial drop in velocity is reversed.

Sumner believes that this will work better if the Fed publicly annoucements that it will do this. If this requires that the Fed take some capital losses from selling assets that it bought at high prices or else credit losses because loans cannot be collected, then it will take the loss.

Kling is claiming that this is equivalent to an expenditure that subsidizes commercial lending or mortgages.

So, raising taxes and using the money to subsidize commercial lending or mortgage lending would have the same effect?

OF course, if zero interest T-bills are the same as treasury currency (Greenbacks,) then issuing T-bills and buying risky assets and the like is exactly what the Fed would be doing. It is monetary policy operated by the Treasury. In that scenario, if the demand for treasury bills falls so that they begin the get positive yeilds, the Treasury sells off the assets, pays off the treasury bills, until it is back to where it was before the "monetary expansion."

This is not the government spending on a bunch of goverment projects and paying whatever interest will be needed to finance them in the future. (Well, perhaps the goverment plans to sell off the roads and bridges.)

Anyway, I believe that the first best situation is negative yields on low risk short term assets like T-bills, FDIC insured deposits, and reserve deposits at the Fed. But, as long as we want to have zero interest goverment issued hand to hand currency, we are left with quantitative easing--the Fed buying up first all the T-bills, and then, if necessary, other kinds of assets.

AI uderstand that financial economists use T-bills and curreny as a sort of riskless benchmark and then head off to the races with other financial assets. But, you know, there are some other issues. For example, is it really possible that a frozen quantity of base money would be consistent in permanent modest inflation? The real quantity of base money could just fall to zero and prices would just keep on rising?

If monetary policy is ineffective (base money and T-bills are about the same in terms of risk) why couldn't the price level be 10% of its current level, or 10 times higher? What provides the nominal anchor in this "model?"

Maybe the risk characteristic of money isn't its "essential" characteristic.

Steve Roth writes:

I truly don't understand. Cash and bonds are inequivalent by the rate of real return on the bonds, no?

So the Fed's monetary lever is made of rubber, and sometimes it's stiffer than others. In the early eighties when Volcker eased, of course, it was very stiff indeed, and the economy started surging back in a matter of months.

Today, it's very, very flaccid. In normal times it is (significantly?) less so.

It's just a matter of degree, no? And the curve may not be straight--it may get very flat (read: flaccid) below, say, 5% inflation. You'd say 10%, perhaps?

But saying that cash and t bills are equivalent in normal times just ain't so, is it? Just that they're equivalent *enough* that monetary policy is weak. But weaker than fiscal policy? At what rate of inflation or real return is that true?

I think you make an excellent (Steve-Waldman-worthy) point: that buying other (higher-yield) bonds at this point is quite similar to fiscal stimulus. Certainly, monetary policy and tools are so weak at this point that they look downright fiscal.

Which is why, sadly, we need so gosh darned much of it--monetary, fiscal, whatever you want to call it.

The real question from all of this: what is the Fed equivalent of Viagra?

fundamentalist writes:

Keep in mind that banks borrow short-term to lend long-term. That's one way that Fed policy effects long term rates.

But the situation isn't mechanical. The Fed can't guarantee a certain amount of money creation via a specific interest rate. Banks must be willing to lend and businesses must be willing to borrow. When neither happens, the fed is pushing on a string, as Greenspan used to say.

Kirk writes:

Steve Roth said: "The real question from all of this: what is the Fed equivalent of Viagra?"

Debt relief?

We're in an environment where consumers are recovering from a 0% savings rate, the loss of housing equity, and a 50% stock market loss. Quantitative easing could make things worse, not better. Pay me 5% on my deposits and I'll spend more money in the real economy. Pay me 1%, like the Fed is now, and I'm reluctant to spend.

travis writes:

You mock an inflation target, but what if the inflation target is realized? That would make a lot of people prefer to borrow money at cheap rates.

Barkley Rosser writes:

The Fed has an inflation target, 2%, although they have not made too much of publicly propagandizing it. As it is, they have engaged in a quantitative easing the likes of which I do not think we have ever seen in the US, and so far all that has happened is that we got a barely positive CPI increase in January after several months of deflation in late 2008.

As it is, in the face of the reality that we have observed negative nominal interest rates, the liquidity trap is not as firm as many think. But Sumner's solution has already been more than tried and has already failed.

travis writes:

@Barkley: The Fed hasn't tried quantitative easing that hasn't been pretty much sterilized. It started to pay interest on reserves so that it could by debt by printing money and the money wouldn't enter the economy.

Jeff writes:

As has been pointed out by numerous economists dozens of times, Kling is just wrong. Monetary policy is ineffective? As compared to what, exactly?

Sumner points out quite reasonably that if the Fed stops paying interest on reserves, they become a less close substitute for Treasury bills. And the Fed can stimulate aggregate demand by first buying up all the Tbills, then all the longer-maturity Tbonds. Once the entire national debt has been monetized, it can start buying shopping malls, real estate, and pet rocks.

"But wait!" you say, "That will cause hyperinflation!" Well, sure, but you've just conceded Sumner's point.

What Sumner wants is a nominal GDP level target. If expected nominal GDP is less than the target, the Fed buys stuff until that's no longer true. If expected nominal GDP is higher than the target, the Fed sells stuff until that's no longer true. There is no doubt that this could work, although there are a few details about how to measure expected nominal GDP and what the time horizon should be. But those are just details, and they can be adjusted for once the program is in operation.

I'd like to see Scott's program adopted, but I doubt it will be. Why would the Fed want to give up the enormous power it has been accumulating and play by a set of rules? What's in it for them?

Barkley Rosser writes:


One of the more bizarre anomalies going on, among many, is that while the interest rate on reserves is 1%, somehow the fed funds rate and on some T-bills is less than that. Many would argue that this rate should set a floor. I would suggest you explain that before you start handing out the argument you make here.

Jeff writes:


The interest rate paid on reserves is only 0.25 percent, not 1.00 percent. See here for details.

Scott Sumner writes:

Arnold, Thanks for the comment. I haven't had time to read all the comments, but let me simply address your observation that money and bonds are close substitutes in normal times. This makes no sense to me, and even Krugman would not make that argument. In normal times almost all of the monetary base is cash and required reserves, and the latter is also fairly small and is fixed. That's because in normal times a bank would much rather earn 4% on a T-bill than 0% on cash. They are not close substitutes at all. Now let's look at cash held by the public. If the Fed increased the base sharply in normal times, as a practical matter almost of of the increase would go into cash. By why would that make people want to hold more cash? If the Fed doubled the money supply would you suddenly change your preferences and want to hold twice as much cash in your wallet? I have never understood this argument, but maybe I am missing something. The only thing that would make me want to hold twice as much cash in my wallet is if my nominal expenditure doubled. Sorry if I repeated arguments made by others, I'm rushed today.

Barkley Rosser writes:


OK, so the reserve interest rate has been lowered to .25%. But that is still higher than the fed funds rate. This is still not supposed to be the case, even if the gap is now smaller.

Also, why is paying banks interest reducing the money supply? Offhand it strikes me as being a way to increase it. And, given that it is now down to .25%, I do not see that a move of it back to zero, where it was previously, is going to have all that much effect on anything.

And I would be curious to see Scott Sumner reply to my argument that the Fed has gone way beyond what he is suggesting. Why does not the massive purchases by the Fed of such things as failed European SIVs amount to a massive increase in quantitative loosening well beyond your proposed extra use of inflation-indexed bonds?

Jeff writes:


Three points: First, its true that cutting interest on reserves from 25 basis points to none probably wouldn't make much difference to bank behavior. No doubt that's why Scott Sumner has also called for the Fed to charge banks a penalty for holding reserves, i.e., a negative interest rate. (Probably only on excess reserves.) The whole idea is to get banks to lend out their excess reserves rather than keeping them on deposit at the Fed. Personally, I don't think a penalty rate is such a great idea, but I can see where he's coming from.

Second, the idea of buying indexed rather than nominal long-term Treasuries is something Jim Hamilton proposed to counter one of the objections to Fed purchases of long-term Treasuries. If the Fed purchases "worked", they would increase inflationary expectations and long rates would rise. If the bonds were nominal, this would mean that the Fed would incur embarrassing capital losses when it sold the bonds. If the Fed hesitates to drain reserves once aggregate demand expectations come back on track, you get overshooting into high inflation. This is not an issue if the bonds bought and sold are indexed to inflation, since their prices don't change with expected inflation. (Note that the purchases of risky assets the Fed has already made could lead to a similar problem since their value is also declining.)

Third, I think Scott would argue that the right amount of monetary stimulus is whatever it takes to get the Fed's own (presumably unbiased) forecast of aggregate demand to grow at 5 percent. The fact that the Fed sees slower growth than that ahead shows that they know they're not doing enough to hit their own target. In which case it's not a target at all, is it?

Dezakin writes:

Uh, seriously how can Kling not see the purpose of quantitative easing? I'm not an economist, and it seems obvious to me. Is there something I'm missing here?

I can see why he would think that it wouldn't work or would have unintended consequences, but right now we have a drop in aggregate demand and a deflationary cycle that if it weren't for the huge monetary pump priming done now would be a mirror of the great depression, down to the debt crisis that caused it.

The purpose is to break deflation by printing money. It comes back to bite as high inflation during the recovery with a huge monetary base, sure, but it gets money moving again because no one wants to hoard it and everyone wants to put it into some equity, commodity, or loan.

Jonathan Dean writes:

Dr Kling, Thank you for rising above the lack of politeness and professional courtesy that seems to often accompany these debates in the "blogosphere".

I genuinely don't understand why Dr. Krugman has these lapses of manners. He seems a very nice and mild-mannered man on TV. Perhaps it is just incredibly stressful being a prominent "point person" currently for this raging and critical debate?

Thank you for all you do and please continue to lead these discussions. Thanks too, to Dr. Sumner for reaching out and contributing to the comments here. Regards, Jonathan.

Dezakin writes:

I know very little about Krugman, but if thousands of people I didn't know were calling me an idiot, I'd probably be a bit rash from time to time as well.

Comments for this entry have been closed
Return to top