September 2022

Sun Mon Tue Wed Thu Fri Sat
        1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30  
Blog powered by Typepad

« The Brilliance and Limits of Hayek | Main | New Horwitz on the Web »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Could you clarify what administrative concerns you have? I think I probably have an idea, but you left that a little unclear. Friedman clearly had these on the radar and came to a different conclusion. I certainly agree with you that they always need to be on the radar. I'm just a little unclear on which administrative costs you're concerned about.

I also find this interesting becuase most Austrians seem to oppose this sort of thing from an analytic perspective ("it simply won't work") rather than from a political economy perspective ("it might work in theory but it won't work in practice").


You might have heard about Buchanan's distinction between politics and policy, which has been emphasized by an economist I admire. I think it's pretty clear that you are right at the level of politics. We want institutions that don't force you to decide whether QE2 is a good or bad idea. And there is no prospect of consistently getting monetary policy right in a world with central banks, representative democracy, and a regulatory state. In some sense "the" answer is to stop playing this game. In the meantime, however, we have a central bank and we cannot escape the necessity of monetary policy. In that impossible position, what is the best policy? I would not be so foolish as to offer an answer of my own. But I don't quite see where you are offering a clear argument to trump QE2. I'm not sure there is a clear argument demonstrating the superiority of whatever might happen in fact to be the optimal policy.

Indeed, let me put that as a question to you and anyone else reading this thread: Given the complexity of the situation, the difficulty of the state sending a reliable signal that it will stick to any policy adopted, and the partial breakdown of ordinary market coordinating mechanisms, why should we think the optimal monetary policy is decidable?

I think Milton Friedman would have favored Bernanke's policies:

"The answer is straightforward: The Bank of Japan can buy government bonds on the open market, paying for them with either currency or deposits at the Bank of Japan, what economists call high-powered money. Most of the proceeds will end up in commercial banks, adding to their reserves and enabling them to expand their liabilities by loans and open market purchases. But whether they do so or not, the money supply will increase.

There is no limit to the extent to which the Bank of Japan can increase the money supply if it wishes to do so. Higher monetary growth will have the same effect as always. After a year or so, the economy will expand more rapidly; output will grow, and after another delay, inflation will increase moderately. A return to the conditions of the late 1980s would rejuvenate Japan and help shore up the rest of Asia."

Now, my opinion is that this is quite bad economics, anyway, this would be off topic.

Prof. Koppl: I was trying to understand your (and Rosser's) decidability stuff, but without reading again my computer science textbook I'd probably fail to.

Anyway, my intuition is: we live within the world and we need to make a decision, i.e., either QE2, or not QE2, or any other alternative. What is the consequence of not being able to compute the right answer?

"We" economists (well, I'm not) know some monetary economics (very little) and can analyze some of the consequences of each alternative, and some of the relevant trade-offs.

Among these consequences, I'd put that to perpetuate market-wide distortions, perverse incentives and distorted "knowledge creation" processes is not a solution. This doesn't imply that in the short run some alternatives will fare better. But it implies that in the long run we are actually going nowhere, whereas more principled alternatives (i.e. let the banks pay for their mistakes) would probably solve the problem within a few years.

The problem is no more complex than desocialization: we have a mess in the financial markets because of continual interventions, and everything is distorted. We only have fuzzy ideas about how the world would look like without these interventions, and in the short run everything's possible except the Garden of Eden. But we still can know that the status quo is not feasible, and that after a transition period, desocialization can be successful (if no Hoovers or Mussolinis get the upperhand in the meanwhile).

Isn't it enough to decide?


First of all, please start calling yourself an economist.

Second, "enough to decide" what? I think you sketched an argument for "desocialization," which is politics, not policy. My argument was aimed at policy, not politics. So I think we need to contemplate the same question if we're going to come to a meeting of the minds.

I think most readers of this blog (excluding, perhaps, some beloved members of our loyal opposition) would agree that an economist should not pretend to know what the price of plywood "should" be. Although "the market" can hammer out a good price if it is more or less free to do so, for the observing economist the "right" price of plywood is undecidable in some sense. We can, however, prescribe rules for the plywood market such as well-defined property rights and keeping your promises. In fact, thanks to Vernon Smith and his group, we can now often do some fairly detailed designing of individual markets. We can design markets, but we cannot decide prices. (Qualification anon.)

As I said, I think the above comment is pretty evident for "us" on this blog. Well, I'm just sayin' that deciding the "right" monetary policy is analogous to deciding the right price of plywood. You are saying that designing a good monetary regime is like designing a good plywood market. I think we can design a good monetary constitution, but I doubt whether optimal monetary policy is decidable in the current context.

QUALIFICATION: There are limits on our ability to design market institutions. I don't think those limits affect the substance of the point I wished to make to Pietro, so I waited for this qualification to acknowledge our limits at the constitutional level. It's like act vs. rule vs, indirect utilitarianism.

This is very clear, thanks for the answer.

I recently discovered the works of Vernon Smith, his book "Rationality in economics" is really great stuff, the most interesting book I read since I discovered Hayek and Mises.

I advocate more quantitative easing because I think there is a shortage of money. Roosevelt's New Deal was the political consequence of an unresolved shortage of money. That is, the political consequences of not resolving a shortage of money are far scarier than the administrative difficulties of quantitative easing. If the Fed had prevented the significant fall in nominal income in 2008, then I think Obama's stimulus would never have happened, (or at least it would have been considerably smaller).

Of course, we're forced to choose the lesser of two evils here. Public policy must settle for doing the least harm possible.


My question is this: Would Hayek's Logic Lead Us to QE2?

Hayek, as Larry White has detailed, advocated stabilizing MV. It follows that the proper policy would favor a nominal income target. Specifically, a zero inflation nominal income target would imply that the central bank should aim for nominal income to be equal to the real rate of growth. If nominal income is below that level, doesn't that require more monetary stimulus? Does Hayek's logic lead us to advocate more monetary stimulus on that point? I would point out that Hayek later regretted not advocating this type of policy during the Depression.


You wrote, "I'm just sayin' that deciding the "right" monetary policy is analogous to deciding the right price of plywood. You are saying that designing a good monetary regime is like designing a good plywood market. I think we can design a good monetary constitution, but I doubt whether optimal monetary policy is decidable in the current context."

I am in complete agreement with you in regards to economic planning. However, given that the Fed is in place, what should they do? I see too many people essentially arguing that the knowledge problem precludes action. So my question is whether that means all action. Should the Fed do nothing? The absence of an optimal policy does not necessarily preclude Fed action. Should the Fed stand idly by in the face of monetary disequilibrium?


Where do you get the evidence for this story? I mean both the 1930s and the 2008ff period.

It is this narrative that has me concerned --- I am concerned about bigger narratives no doubt, but this one bothers me because it is promoted by those on our side. I actually had a colleague tell me that he would tolerate 10% inflation if that would have prevented the bailouts and the stimulus.

I literally do not know what this means. I have said from day #1 of this debate that there are two mandatory readings --- Hayek, Tiger by the Tail, and Buchanan and Wagner, Democracy in Deficit. I don't think Monetary History is the right book to be reading at the moment. I am, btw, with Anna Schwartz on this --- Bernanke is fighting the wrong battle --- this isn't the 1930s.

I am willing to be persuaded, I haven't been by Scott Sumner, David Beckworth, or George Selgin; let alone by Tyler and Bryan. I am not blind to their argument, I just don't buy it.

If anything the crisis has made me appreciate the teachings of my undergraduate teacher Hans Sennholz as well as Murray Rothbard on the importance of bright line rules and restraints in monetary policy/system. Sennholz's Age of Inflation would be valuably read. And seriously intrigued by Buchanan's efforts to constrain the fiscal state.

In book 5 of the Wealth of Nations, Adam Smith warned of the juggling tricks that governments always engage in --- deficits, debt, and debasement --- is the natural proclivity of government. In what sense does QE and fiscal stimulus cater to that proclivity or cut against it?

For what it's worth, this gathering of monetarist luminaries I'm at right now thought this WSJ article was terrible and relied too much on Beckworth's paper. In particular, they argued that it ignored Friedman's long-standing commitment to a stable growth in the money supply. To hear it from the horse's mouth, here's Meltzer in the WSJ a couple of weeks ago:

I think QE2 would be a big mistake and I think MF would think so too.

I think the problems we now face are much more connected to regime uncertainty than to a simple shortage of money. Over at Thinkmarkets we were talking about this recently.

Ed Dolan mentioned the current unanswered questions:
"(1) Will the Fed try QE2? The FOMC is clearly not unanimous. (2) What effect will QE2 have? There is unusually diverse opinion in the economics profession about the transmission mechanism for this type of operation. (3) Is QE2 part of an inflation targeting regime, or something apart from it? If inflation targeting, an IT regime that includes rebasing, or not?"

I think this is the problem. It's not that normal people need money balances to hold against uncertainty, the problem is that investors need money balances to hold against uncertainty.


I don't think you're getting the structure of my argument, which is my fault. I am assuming that there is an optimal monetary policy. Indeed, I don't really understand how it could be that no policy is optimal. At least one policy must be optimal, even if such "optimality" should be pretty lousy. So I am *not* assuming the "absence of an optimal policy." Nor am I arguing for inaction. I don't know what it would even mean to have the "Fed do nothing." I don't know what "preclude Fed action" could mean. Whatever it might mean for the Fed to "do nothing," that would be a monetary policy too. I am questioning whether the optimal policy, whatever it might be, is decidable. If I asked you to tell me which problems a Turing machine can solve, you would have to say "I can't decide." I would not be responding well if I upbraided you for "standing idly by" when an important mathematical problem is at issue.

Lee Kelly:

you touch many very important topics, in line with the article by Friedman I previosly linked. But I disagree on three points.

"I advocate more quantitative easing because I think there is a shortage of money"

Money doesn't look to be in high demand: they keep it at the Fed at a ridicolous rate of 0.25%, and no one appears to want it. I would expect high profits in producing something that is in high demand, but there are no profits at present in producing money.

What I see no is that we have too much debt. We have too much leverage, which means that a couple of percent point of a negative shock has wrecked the whole financial system. We have too much maturity mismatch. We have too many bricklayers, masons and too many of those tall metal structures used to make buildings whose name I actually don't remember. And, of course, we have regime uncertainty.

If dearth of money were the problem, we wouldn't be here: we would see banks making huge profits creating new deposits, we would see huge profits in reducing the demand for money by improving the efficiency of payment systems. In other words, the market would increase spontaneously both the monetary multiplier and money velocity. People would pay to get money from banks, but no one's actually doing that.

"If the Fed had prevented the significant fall in nominal income in 2008"

The Fed tried it, but it failed.

Total monetary spending is the product of the monetary base and an endogenous factor, which we may call velocity (but defined in M0 terms), which depends on many microeconomic conditions: it depends on the balance sheet of banks, on the average profitability of investment, on risk aversion, on risk perception, and on policies.

The problem with all these factors is that during the boom, and even more so during the bust, they tend to reduce the "wide sense velocity" of the monetary base. The unsustainable boom produces risk, cuts returns, worsens the banks' balance sheets, etc.

As far as the Fed has ammunitions, it can cut interest rates and try to avoid the slump. If it touches the zero level, however, it can't do anything. For a while it can monetize longer term assets, or private assets (if banks are not eager to do it by themselves through interest arbitrage). But in the end the only think it can do is printing money and giving it for free, hoping someone will want it. This would create a stagflation, not a recovery, however.

We could take several trillions of taxpayers' money and give it to the banks as equity. This would solve the problem with the banks' balance sheets, but it wouldn't solve the problem of having too many bricklayers, and the problem of having too little price spreads (low returns on investments).

"Roosevelt's New Deal was the political consequence of an unresolved shortage of money"

I think that thanks to Cole and Ohanian we have a fully convincing argument to explain the Great Depression: it was about cartels and unions, not shortage of money. Shortage of money was only the proximate cause that caused the problem, because the crisis required a cut in nominal wages and Hoover and Roosevelt prevented it to happen.

I'm sorry I didn't realize my comment was so long.

Current: you are right about regime uncertainty, but who's holding money? Banks would pay to destroy their supplies of it, if they could. :-)

4th in a row: I must thank my friend Leonardo for the velocity argument. It's not my merit, or guilt. Now I can shut up with a cleaner conscience.

You are in the middle of a dark room holding a loaded gun. A group of schoolchildren are standing in the dark against one wall. On the other wall a tiger with night vision is crouched ready to pounce on the schoolchildren and eat them. You don't know which wall has the children and which wall has the tiger. What is the optimal shooting policy?

I personally agree with Meltzer's positions. I think people expect too much out of monetary policy right now and de-emphasise too much the structural problems. Given the huge bank reserves, public deficits and public debts and the moderate recovery in GDP growth, but the persistence of an unemployment rate of 9%, there must be something about higher or expected higher tax rates and costly new regulation adversely impacting supply and the whole uncertainty argument holding back th return to normal levels of economic activity.

Horwitz hits the nail on the head by reference to Meltzer.

> Banks would pay to destroy their supplies of it, if
> they could. :-)

I don't think they would for many American banks it's a very important asset currently.

I think that you can say that there are two different types of problems connected with expectations of the future price level. Firstly, there are the shorter-term problems of how output varies and what the central bank do to keep things in line with their policy. Second, there is the "regime" in which we live in, that's about whether the government or central bank *have* a clear policy for the future. I think the current problems are connected with the second issue.

Banks could use reserves that they have to make more loans. However, in many cases that would require making relatively long term contracts in terms of money. However, we've seen that the Fed have done badly both before the crisis and after. They didn't predict it, many think they caused it and their initial responses didn't work. So, are we still in the "Taylor rule" regime of monetary policy? Maybe not. Which leads to the question: what will happen in the future? If banks can't predict that well then they can only lend in situations that are very low risk and high return. If banks offer "variable rate" loans they simply push that same problem onto potential borrower.

If the state monetary authorities can't plausibly promise some future regime that the market can understand then that makes things very uncertain. And that makes low yield investments such as reserves and T-bills attractive.

Roger Koppl, do you think this is what's going on?

Sure seems right to me, Current.

What does Anna Schwartz think about QE2? Whatever she says is what Friedman would have said.


I don't have time for a longer response right now, but I want to clarify a couple of points.

1) I am on your side. The cycle of deficits, debt, and debasement concerns me. I just don't believe quantitative easing threatens deficits, debt, or debasement in the present circumstances. I actually think that had quantitative easing been pursued more aggressively in the first place, the recent ratcheting up of deficits and debt by both Bush and Obama would have been less severe.

2) I also agree that 'bright line rules and restraints' are appropriate for monetary policy. For me, something like a 2-3% level target for some measure of nominal income would be, given the second best world of central banking, better than any alternative. It could almost be administered by a computer.

I don't think we disagree half as much as it seems.

Pete, you mentioned the importance of bright line rules and restraints in monetary policy/system.

Before the crisis, the rule was that the price index more or less followed 2% growth path, this rule was quite credible, and long term private sector debt contracts were relying on this rule. The current crisis was caused by erratic monetary policy. In late 2008, markets expected 20% downward deviation of price level from the previous target over the next ten years. At this moment, there is a huge diversity of expectations, as there is no coherent monetary policy to speak of.

A gold standard analogy would be this - one day government says over the next ten years the dollar will be revalued against gold by 20%. Two years later government says that revaluation is cancelled, but a gold standard will be replaced by a dual mandate approach. Imagine the chaos. An excellent breeding ground for socialism.

The crisis will only end when the vague generalities of dual mandate are replaced by a bright line monetary policy framework that is supported by sound long term fiscal outlook.

QE2 doesn't look so bad in terms of the "administrative side" test, Bernanke can piggyback on markets, as duration risk premia of long term government bonds will continue to be strongly influenced by markets. In Japan 10 year nominal bond yield sent a clear inflationary signal in 2003-04, note that this happened well before the actual inflationary pressures reemerged in 2006.


I don't really agree. Scott Sumner has convinced me that the demand for money is a very forward looking thing. Granted, the demand for money of normal people isn't so dependent on expectations, but the demand for money of investors is. Once the future became unclear investors and lenders were in deep problems. They would have been relatively OK if they had had a good idea of the future path of prices, whatever it that future path looked like. But, the Bernanke fed didn't give them that, rather it see-sawed for a while. As Metzler says in his article quoted above the Fed briefly threatened inflation and then backed down. Under this situation the demand for money will rise. The problem here is that that financial demand for money is much greater than that from those who don't deal in the financial markets. MV=PT is much greater than MV=PQ, and if MV=PT rises then it doesn't matter what MV=PQ does.

Part of the problem here is that the consistency of the central banks isn't so important in good times. In good times speaking in mysterious pseudo-economic and pseudo-magical ways, as central bankers do, isn't so harmful. No matter what happens in good times the markets can be sure that there has been no change of monetary regime. That's not the same in crisis times though. In a crisis it must be clear to the markets how the government will respond. That's why some time ago Ambrose Evans-Pritchard said something like "The market wants a Churchill, but central banks keep giving them Chamberlains".

Pietro, you are confusing money and credit.

Current, I agree. QE2 should be unnecessary. The Fed failed to signal its intentions for monetary policy in the long run -- reducing the impact of QE1.


Sorry! That last comment was too brief. Let me explain why an excess demand for money reduces the demand for credit.

Suppose there is an excess demand for money, i.e. desired money balances are greater than the actual quantity of money. People with less money balances than desired try and sell goods and services but cannot find enough buyers -- a 'general glut' occurs.

The general surplus of goods and services will eventually be resolved by market forces, but in the short term the cost of earning money increases. Debts that were manageable can become overwhelming, i.e. the *real* burden of debt increases. This suppresses the demand for credit and lowers interest rates. For the same reason, lending becomes much riskier, and so investors want higher interest rates just when borrowers are least able to afford it. The demand for safe and liquid assets increases for want of any alternatives and so low interest rates prevail.

My point here is that low interest rates can be a consequence of tight monetary policy, i.e. an excess demand for money may cause a decrease in demand for credit.


Another thing. Did we have too much leverage? Isn't that a bit like pretending to know what the price of plywood "should" be. In hindsight, it appears obvious that we had too much leverage, but I am not so convinced. If people had known from 2000-08 that the NGDP was about to decline more than any time since the Great Depression, then I suspect they would have been far less leveraged. Perhaps a lot of lenders would still have suffered heavy losses because of genuine malinvestment, but the consequences would not have been so severe had monetary equilibrium been maintained.

Promote your free hosted blog using Froslo. Froslo - Free Dofollow RSS Directory.

Lee Kelly:

"Did we have too much leverage? Isn't that a bit like pretending to know what the price of plywood "should" be."

Did we have too much money demand? Isn't that a bit like pretending to know that the price of plywood "should" be.

More seriously: I don't need to know the price of plywood to know if plywood price is too high with respect to fundamentals if, for instance, there is a price floor due to successful lobbying efforts: this I know a priori.

We know that markets react to moral hazard by taking on too much risk. This, too, we know a priori: leverage will overshoot. What we don't know is whether it has overshot by 0.0001% or by 300% with respect to normal market levels. This is the problem of relevance: it's a posteriori.

I'm making the relevance judgement (as Mises called it) that the excess amount of leverage is closer to 100% than to 1%: that the Fed, in other words, has made a mess with moral hazard, and that we are not here due to animal spirits, but due to a prisoner's dilemma a là Carilli and Dempster.

Maybe I'm wrong: a 50:1 leverage (actually in Italy it's 20-25:1 for banks) is due to market fundamentals: my relevance judgement is that it is not. But of course is not a purely theoretical point, it's my judgement.


I would ask them: "where are you?" and shoot in the other direction. :)

"Pietro, you are confusing money and credit."

No, I'm just agreeing with Meltzer (I think also with Horwitz, but of course I'm disagreeing with Sumner):

"There is no shortage of liquidity in the economy—banks already hold more than $1 trillion of reserves in excess of their legal requirements, and business balance sheets show an unprecedented amount of cash and near-cash assets."

(from the WSJ link by Horwitz above)

What I'm claiming is not that credit and money are the same thing: but that we have a problem with credit caused by past monetary policies, that present-day monetary policies cannot solve. We have structural imbalances, not aggregate demand problems.

I'm not saying that a constant NGDP may not be the best approximation to a decent monetary policy: I'm saying that it is impossible for the Fed to prevent NGDP from falling when returns on investments are too low and banks' balance sheets are in such a mess.

I'm not even saying that monetary problems may not add to the bill of recessions' costs: they are relevant. But actually the real problems are real (too many bricklayers and... cranes) and financial (too much debt), not monetary. Then there is also regime uncertainty and crowding out effects...

Let's assume that we have no problems of regime uncertainty, structure of production and financial overleverage.

What would happen if we didn't have enough money? People would pay to get it. Banks would gain creating it. Firms would profit in economizing it and improving its efficiency of use. If it didn't suffice, alternative credit moneys would be created.

Nothing like this is happening now. I think that Sumner has only a hammer and he is trying to explain all problems by claiming they are nails. But I prefer to use the malinvestment knife, the regime uncertainty screwdriver, the crowding out chainsaw and the monetary hammer together: the world is too complex for a single tool.

Is there something in menu costs or other causes of price rigidities that prevents people from doing it? No, I just need to go to the bank and ask for new money in exchange for my income, or lend at higher than normal rates on the money market, or improve my cash policy. It's all microeconomic, nothing macroeconomic.

But, we you have structural problems, than everything we are observing makes perfectly sense...

Pete, since you list me among those with whom you disagree concerning QE2, perhaps you will enlighten me concerning where it is that I stated my position on that subject. All I can recall is having argued my belief that "general glut" is not an incoherent concept.

I will say, on the topic, that your own position seems to be that the Fed can do right by doing nothing. Let me insist that, so long as there's a Fed, it is necessarily doing something; it is planning the growth of money and credit. There's no way 'round that. So to argue that the Fed should refrain from doing X is to implicitly endorse its doing Y, that is, to favor one central plan for M over another. The position may be correct after all--Y may be the better central plan than X--but it won't do to suggest that Y is better because it is less interventionist than X. One cannot plead that case unless one has a clear laissez-faire counterfactual in mind, and even then it remains to be demonstrated that the LF outcome is in fact the more desirable outcome. Mere appeal to an implicit ideal of an "immaculate" monetary polity doesn't cut it--at least, I hope it doesn't, or Steve and I and others have been wasting a lot of effort trying to figure out how LF money actually works, when we might simply have assumed that it works just swell.


You did confuse money and credit. You said that the .25 short term interest rate means there is plenty of money.

No monetary disequilibrium theorist every says that money demand is "too high." That is rather an implication of those who favor a quantity of money rule but somehow think money expenditures matter. The problem, in their view, would be that the quantity of money is right, (following the rule) but money demand is too high, causing the problem with low money expenditures.

Of course, if one assumes perfectly flexible prices, then money expenditures aren't important. The level of prices and wages is always at the level that keeps real expenditure equal to the productive capacity of the economy. Any decrease in actual output cannot possibly be due to surpluses--real expenditure below productive capacity, and so it must be a decrease in productive capacity.

That is the real business cycle theory/market clearing approach.

Meltzer's WSJ article was awful.

I see hints of--the demand for money is high because no one wants to invest because of bad Democrat policies. Vote Republican, and investment will rise and the demand for money will fall. Some kind of nonexistent quantity of money rule would work.

But Meltzer actually defends the Taylor rule, which is based upon adjusting short term interest rates to target the inflation rate.

Friedman believed correctly that money expenditures (nominal income) depends entirely on monetary factors. In the long run, unemployment and ouput depend on suppply side factors. And so, monetary factors only determine money prices and wages in the long run. However, short run fluctuations in money expenditures lead to undesirable fluctuations in real output in the short run.

He found that the conglomerations of assets that make up M2 was more or less proportional to money expeditures. In other M2 velocity is was pretty much constant. The implication was that if the Fed could vary the monetary base in such a way that the M2 combination of assets stated on a 3% growth path, then money expenditures would stay on that same growth path. The price level would be stable, wages and other incomes would grow about 3%, and fluctuations in supply side factors would result in fluctuations in output and also the price level and inflation as the price level moves.

But it turned out that the M2 measure of the money supply stopped tracking money expenditures. M2 velocity changed.

Before he died, Friedman noticed that like everyone else. There was never any reason why that combination of assets should necessarily track money expenditures. When they stopped, Friedman began to change his mind.

However, he always maintained that money expenditures depend on monetary factors, that interest rates don't tell us much about those factors, and that stable growth in money expenditures are important in both the short run and long run.

P.S. I support QE2.

Prof. Rizzo wrote: "What does Anna Schwartz think about QE2? Whatever she says is what Friedman would have said."

I am not so sure. Recall this post by Scott Sumner.

The punch-line: "Were he alive today, Friedman would be horrified by the neo-Austrian views of Anna Schwartz."


It is the monetary regime uncertainty that really matters.

You said that .25 interest rates indicate that there is plenty of money. But those rates would be negative if there was plenty of money.

You are right that there is a deleveraging pressure. It is one of the factors that makes equilibrium short term interest rates negative. Or alternatively, if you prefer it, too big to fail subsidies should be abolished and savings should be distributed in the form of tax cut, this would reduce the value of money to levels consistent with monetary equilibrium and will facilitate deleveraging.


Thanks very much for the link. It is very interesting.

One of the most pernicious myths about monetary policy is that when the Fed "does nothing" we get to see the free market in action. However, the Fed can never "do nothing", because it effectively monopolises the monetary system with its regulatory powers and control of the supply of base money. We do not get to see the free market in action when the Fed "does nothing" but simply a different kind of intervention.


Hayek did not advocate stabilizing MV. White's (mis)interpretation is based on the selective quotations from Prices and Production, in which he omits the crucial parts where Hayek argues that stabilizing MV can never be a maxim of monetary policy, and that the macroeconomic aggregates can never be a part of a "legitimate chain of reasoning" (Hayek) in economics.

Pietro M,

I agree with you that there still are lots of structural problems. But, I think that most of them have been fixed since the beginning of the crisis. That's why I think that regime uncertainty is more important.

Bill Woolsey, Lee Kelly,

I don't think that any quantitative easing will service the demand for money while the Fed are still unclear about the future. I think this is one of the points that Monetarists have made often. What's important now isn't the specific discount rate that the Fed aims at, it's the words coming out of Bernanke's mouth. He must explain believably what future policies will be. Then he must not deviate from following them.


You're right. Hayek's approval of targeting of nominal expenditures was full of caveats and he later retracted it.

Nicolaj & Current

Stabilizing Mv is still advocated in The Constitution of Liberty (1960), Ch. 21 on "The Monetary Framework".

"Still worse, under all modern monetary systems, not only will the supply of money not adjust itself to such changes in demand, but it will tend to change in the opposite direction. [...]These spontaneous fluctuations in the supply of money [here: effective money, i.e. Mv; amv] can be prevented only if somebody has the power to change delibaretly the supply of some generally accepted medium of exchange in the opposite direction."

He then refers to "single national instituions", namely central banks, which he wants to be independent.

In general: stabilizing nominal income levels is what has to be done independent of the institutional structure (one can argue about the targeted growth path of that level; zero growth is Selgin's productivity norm whereas Sumner opts for 5% growth. LF configurations may work well, no doubt. But given the current setting of institutions, central banks have to do it (because they "dominate" the monetary system). You may reject all non-LF approaches. But isn't that the Nirvana fallacy, too. Why not be a bit pragmatic and study second-best solutions?

After my Rothbardian period, my conscious views of appropriate monetary institutions started with Hayek's denationalization of money. I am sure that half-understood versions of David Friedman's views that were expressed in his short Cato Essay, "Gold, Paper, or is there something better" had some influence, and soon after, Greenfield and Yeager's BFH really became my focus.

But, back to Hayek. He thought the most likely scenario was inconvertible privately issued monies, each of which aims to stable its purchasing power. That would result/require monies whose quantities adjust to meet the demand to hold each of them and all of them in aggregate. Further, money expenditures would grow with the productive capacity of the economy. The trend growth of money expenditures would be the trend growth of the productive capacity of the economy, and the trend growth of the price level would be zero.

Now, I support all of that, except I have been persuaded by Selgin that fluctuations of productive capacity around its trend should result in opposite changes in the price level and inflation or deflation and the price level adjusts.

There may have been some point where Hayek advocated a constant quantity of money and allowing money expenditures and the price level to adjust to keep the real quantity of money equal to the demand for money and real expenditures equal to the productive capacity of the economy, but that wasn't the Hayek of the Denationalization.

Price level stabilization requires changing money expenditures with changes in productive capacity. While I think the trend is fine, fluctuating money expenditures with productive capacity is more likely to cause problems than benefits.


"You did confuse money and credit. You said that the .25 short term interest rate means there is plenty of money."

No, I claimed that money is actually so plentiful that banks don't know what to do with these reserves: they keep them parked at the Fed for free.

It was somewhat of a reductio ad absurdum, but I see no evidence whatsoever that scarcity of money is the problem, now. If money were so scarce, banks would know how to use it. I'd look for problems elsewhere, and this implies that a richer micro view of the financial and the real sectors is needed.

"money demand is too high, causing the problem with low money expenditures."

As I said, there is no evidence of this. NGDP doesn't grow because banks do not want to intermediate because of real and financial problems, not because there is an excessive demand for money.

"Friedman believed correctly that money expenditures (nominal income) depends entirely on monetary factors"

But he was wrong for Japan (my first comment contained the link). He seemed to forget that the money transmission mechanism depends on microeconomic conditions such as riskiness, return rates, risk aversion, banks' balance sheets. Once these factors worsen, then MV or M1 or M2 or whatever Mx target different from M0 can become impossible to achieve because banks will not propagate the monetary policy shock toward the rest of the economy. MV and Mx are not unconditional targets, only M0 can be directly controlled, and in certain circumstances MV and Mx cannot be kept from falling.


You still don't understand.

An excess demand for money means that the quantity is less than the demand to hold money.

While my focus wasn't on bank reserves, it is also true that an excess demand for bank reserves is that the quantity of bank reserves is less than the demand to hold them.

Frankly, "if money is so scarce, then banks would figure out something to do with them," is just odd.

The banks are doing something with the reserves. They are holding them. They would rather do that than purchase securities or make loans. You can "blame" whatever the reasons might be that banks don't want to lend or purchase securities if you like, but the point is that given that, the banks do want to hold them.

You claim that money expenditures are low because that banks don't want to intermediate. Well, I will grant that a smaller quantity of base money would be consistent with money expenditures back up to the trend of the bank moderation if banks did want to intermediate more.

But I would also say that if banks didn't want to intermediate at all, that there was no such thing as financial intermediation, that even then, money expenditures could be at whatever target desired.

Given the banks' incentives and contraints, a larger quantity of base money would result in more money expenditures. Even if banks took actions to offset the expansion of base money, a sufficiently large increase in base money could overwhelm any such efforts, and still get money expenditures back to target.

You seem fixated on the more reserves, more bank lending, more total lending, more spending process.

The process is a quantity of money larger than the demand to hold it, and people spending the excess money at least partially of goods and services, more money expenditures.

The higher money expenditures results in either higher prices or more production or some of both. The higher production is higher real income, and raises the demand for money. The higher prices reduces the real quantity of money. The demand to hold money rises to meet the higher quantity of money given a higher level of money expenditures.

The quantity of money is too low relative to the amount that would get people to want to hold the amount they would like to hold with money expenditures at the target and whatever set of expectations and constraints they face?

Nothing in this argument requires that banks lend anyone more money.

Now, if the Fed expands the quantity of money by purchasing government bonds, then the Fed is now lending to the government rather than those who were holding those government bonds. With a budget deficit and rising national debt, on the margin, the Fed is lending more to the government rather than whoever would have bought the bonds.

Because the Fed operates on banking principles, changes in the quantity of money does impact the quantity of credit. But that isn't the point of expanding the quantity of money.

The additional money is in the hands of those who would have bought the government bonds or who already held the government bonds. They can spend the money on whatever they like, but enough money needs to be created so that enough of it is spent on goods and services, so that money expenditures rise to the equilibrium level.

If banks choose to lend funds deposited with them, and those borrowing from banks purchase goods and services, then a smaller increase in base money (fewer purchases of government bonds) are necessary to reach the target.

There's plenty of money? Compared to what?

And when you bring up banks and their lending, you clearly confuse money and credit. I am not saying anything about whether there is too much or too little lending.

I have opinions about those things, and I don't doubt that a committment by the Fed to get money expenditures back to target and to expand the quantity of money what it would take, would result in a lower demand for bank reserves and lower demand for money, so that the Fed would actually need to contract the quantity of money!

A bit paradoxical, I admit. But the point is that even if people are myopic, a sufficient expansion in the quantity of money will get money expenditures back to target. And it doesn't take any bank lending. And if people see it working, then the decrease in reserve demand and money demand could result in money expenditures returning to target with only modest purchases of government government bonds and expansion of base money. And, its possible that selling bonds and a contraction base money would become necessary, even as money expenditures rise.


I don't think that any quantitative easing will service the demand for money while the Fed is still unclear about the future.

I think quantitative easing will service some of the excess demand for money and 'stimulate' nominal spending. The notion that everyone who the Fed purchases assets from will 'hoard' the money is not plausible. However, unless the Fed can stabilise expectations of future nominal spending (or just inflation), the result of more quantitative easing will be underwhelming. In short, I think more quantitative easing would be better than none in the current environment, but I otherwise agree with your argument.

I would add to bill wooley's post that

#1 the Fed pays interest on reserves (higher than the federal funds rate) and thus helps commercial banks to generate risk-free profits so that they can improve on their balance sheets. This is counterproductive, if nominal spending is the target, as it should be.

#2 today, credit creation or intermediation in general is dominated by market-based financial activities, not by commercial banks, who are however still deeply interwoven in the credit creation process. See Adrian and Shin:

Thus, targeting PY has become very, very complicated. The best way to do so seems to be by the control of (nominal) market expectations. Markets do the rest. This is the only disagreement I have with monetary equilibrium theorists. They don't put enough emphasis on expecations. Manipulating base money is certainly no proper alternative today.

Oh, Mr. Woolsey, of course. Sorry.


The only reason interest on reserves is higher than the federal funds rate is that the Fed does not pay interest on funds held by Fannie and Freddie at the Fed.

Interest on reserves arbitrage with Fannie and Freddie pushes the fed funds rate a bit below the 25bps. However this arbitrage does no harm to monetary equilibrium. In fact, the opposite is true. This arbitrage lowers the opportunity cost of reserves and pushes it a bit closer to the cost that is consistent with the monetary equilibrium.

In support of the points made by Lee Kelly and Bill Woolsey about money and credit, I'd also point out a piece of clear evidence against Pietro's argument. Interest rates are not low.

Five years ago I took out a personal loan at an interest rate of 6.7%. When I looked into rates recently (from the same lender), their offer was 17.9%. Nothing has changed from my side (if anything, my credit rating and income are better now), but the demand for money (and credit) coupled with an economy-wide change in the perceived risk profile of borrowers has clearly resulted in an increase in market interest rates. The reduction in Federal Reserve interest rates no doubt helps to counteract this a little by increasing supply, but not enough to cancel out the shift.

Anecdotal, sure - but this illustration is in line with lending data and with lots of other anecdotes economy-wide.

Separately, a point about Pete's original article:

Please folks, lets emphasize the #1 rule of public policy --- "Do No Harm."

Who says that's a rule of public policy? (Austrians, I know. But anyone else?) It's a rule of Pareto efficiency, and it could be a rule of deontological philosophy, but it's really not a rule of any realistic public policy.

Any ethical public policy must start with "first, do no harm." I reject the distinction between "realistic" and "ethical." It's like saying it's unrealistic to expect women not to be raped -- so we shouldn't try to do anything about it. It is unethical, and therefore, we must try to stop it. The same is true of unethical public policy.

@ the money demand:

interesting. can you please provide a link ... I'm not sure I understand your point ... thx!

This is James D. Hamilton:

"Unfortunately, until the beginning of 2009, the Federal Reserve was doing everything it could to prevent its actions from stimulating the economy in the usual fashion. It was viewing the slowly unfolding credit problems as primarily a crisis in lending, in which the Fed felt it needed to step in as lender of last resort on what ultimately proved to be a massive scale. [3] The Fed wanted to lend extensively, but did not want to see currency held by the public increase. For this reason, it sold off a significant portion of its holdings of T-bills through September 2008, in a paired set of actions, lending with one hand and selling T-bills with the other, that might be described as “sterilizing” the lending operations so as to prevent them from having an effect on the money supply. [4] When the Fed ran out of T-bills to sell, it asked the Treasury to create some more for the Fed to use just for sterilization. More importantly, In October the Fed began paying interest on reserves, in effect borrowing directly from banks, and creating an incentive for banks to hold the newly created deposits as a staggering burgeoning of excess reserves, again preventing its actions from increasing the value of M1. Excess reserves amounted to $833 billion by August 2009, or more than the sum of all the currency issued by the Federal Reserve between its creation in 1913 and 2008."

There is no Englishman who repeats Hamilton's error, as Bank of England started paying interest on reserves in early 2006 with no adverse consequences. More here:

I agree with Arash that the comments I've seen by some monetary equilibrium folks don't take enough account of expectations.

Lee Kelly,

Without management of expectations I think QE2 could be counterproductive. The problem is that the more base exists the more difficult it will be to remove when the time comes. It increases the risk that prices move without being really under Fed control in the future. So, without a clear policy attached to it it does more harm than good.


I know that Hayek advocated flexible money and advocated meeting the demand for money. He did that even in the 30s. But, his advocacy for an NGDP type rule isn't without caveats. I think that in "The Constitution of Liberty" which you mention Hayek says that an NGDP rule wouldn't work well in times of high productivity growth.

I should add... I don't support QE2 and I'd never support any monetary policy named after an ocean liner.

Also, "The Money Demand", I've agreed with everything you've said so far except the bit about interest on reserves. It's important here to distinguish between interest on reserves that are in use and interest on *excess* reserves.

@ money demand.

first of all, you have a great blog.

now, on IOR policy.

#1 the problems with IOR pointed out by hamilton are independent of the fact that the BoE introduced them in 2006 without negative implications. 2006 were regular times; accordingly, there was no question of pushing extremely high excess reserves into circulation, thereby increasing nominal income. I totally agree with the harmless nature of IOR policy in regular times, so I also accept the results by Curdia and Woodford.

#2 you write: "Fed's IOR policy was immediately criticized by James Hamilton and David Beckworth. But their criticism suffers from two shortcomings.

a) First, they picked the wrong target by saying that interest rate paid on reserves is too high. Because the primary tool used by the Fed is fed funds rate, Hamilton and Beckworth should have argued instead that fed funds rate is too high.

b) Second, it is not true that hoarding by banks of massive quantity of excess reserves under IOR regime is somehow more harmful than hoarding of a small quantity of required reserves under previous zero interest rate regime."

ad a) the IOR is a floor rate for the Fed fund rate. This is from the NewYork's FAQ side:

"Without authority to pay interest on reserves, from time to time the Desk has been unable to prevent the federal funds rate from falling to very low levels. With the payment of interest on excess balances, market participants will have little incentive for arranging federal funds transactions at rates below the rate paid on excess. By helping set a floor on market rates in this way, payment of interest on excess balances will enhance the Desk’s ability to keep the federal funds rate around the target for the federal funds rate."

Thus, as Sumner points out (link in your post), in times of near-zero Fed fund rates the above zero IOR is the rate to focus on. In saying that IOR rates are too high, hamilton is saying that short-term market yields are too high, relative to the stance signalled by the official funds rate.

ad b) hamilton point is that a lower IOR would shift the LEVEL of the yield curve downward. You again provide the link. He argues against Altig and Abate that substitution of short-term assets for excess reserves decreases real risk-free yields along the term structure, which implies higher nominal spending. Altig and Abate argue that the only effect would be that the SHAPE of the yield curve alters, becoming steeper, without much effect on aggregate demand.

here the link to the NY Fed FAQ's:

There is nothing wrong with paying interest on reserves. Paying interest on reserves to keep interest rates up when market forces are pushing them down is a mistake. And that is what the Fed said it was doing in 2008. The answer is to drop the interest rate on reserves with market rates. I would peg it 1/2 percent below 4 week treasuries.

So, I believe that the current appropriate value is less than zero.

With the Fed purchasing mortgage backed securities and even long term treasuries, while paying banks for reserve deposits, the Fed is providing a subsidy to the banks.


NY Fed has certainly fooled a lot of people, and the FAQ you linked to is a clear attempt to whitewash the record.

Expectations matter. Expected opportunity cost of reserves is what matters for the monetary policy. The problem was that the expected opportunity cost of reserves was way above the fed funds rate target in September-October 2008. Erratic behaviour by NY Fed in September-October 2008 has led to frequent interest rate overshooting and undershooting in actual fed funds rate transactions, and financial institutions were very concerned with the risk that the effective fed funds rate will spike in the nearest future. The fear that NY Fed has lost the control of the actual fed funds rate has led to the hoarding of reserves, and IOR was a necessary step to lower the opportunity cost of reserves closer to the fed funds rate target (on the other hand the fed funds rate target was too high, it was a lesser problem).

Bill Woolsey,

I would put it this way:
Interest rate on reserves should be equal to the fed funds rate target. Ignoring market forces when setting the fed funds rate target is a mistake, and fed funds rate target was too high in September-October 2008, and fed funds rate target is too high now.

But the main mistake in September-October 2008 was different. NY Fed has lost control of the expected nominal opportunity cost of reserves, it failed to bring it down to the fed funds rate target.

The reason why the current fed funds rate target and the current rate on reserves is too high has very little to do with the yield on 4 week treasuries.

4 week treasuries have incremental utility as a collateral, the Fed should pay a little bit more to compensate for that.


I agree with all of those who emphasize the importance of expectations. The Fed needs an explicit nominal target. I think a quarter by quarter target for the level of final sales of domestic product is the best way to go. I think the target should be substantially higher than its current value, but then increase at a 3% annual rate afterwards.

Promising to buy long term government bonds at a certain rate is probably better than nothing, but a commitment to a time path for bond purchases is not a good idea. The committment should be to buy whatever amount and type of bonds are needed to get money expenditures to a particular target.

I think a price level target would be better than an inflation target now, but in other scenarios it is much worse. It is a bad rule. (One Fed governor said we should have a price level target now and then go back to inflation targets later. What kind of rule is that?)

Targeting some particular measure of the quantity of money has looked bad for decades now, and this crisis made it look worse. Were overnight repos money? If you don't count them as money, and you target all the other stuff, then the "true" money supply is out of control. And even if you know the true measure of the quantity of money, the notion that the demand for money will grow at a slow, steady rate has no theoretical justification. Maybe it does sometimes, but what if it doesn't?

Base money targets have looked bad for centuries, and they look worse now. Of course, with free banking we would have been under suspension, with the banks paying bonus interest, which would probably be counterproductive in terms of monetary equilibrium (though maybe a good idea in terms of keeping contracts and allowing the banking system to operate at all.) On the other hand, we could at least hope that there would be some kind of expectation that money expenditures would return to equilibrium in the long run. That would help.

Interest rate targets have always been a bad idea. That they worked tolerably well in the Great Moderation was a fluke. But Taylor, Bernanke, and those guys just don't want to give it up.

I just can't express strongly enough what a failure the Taylor rule turned out to be. Or more exactly, targeting short term interest rates in order to get the expected price level to rise at a 2 percent annual rate from wherever it happens to be.

Central banks, including the Fed, have always liked to make slow, deliberate adjustments in short term interest rates. It seems to me that the preference for slow and deliberate suggests a bias to keeping them unchanged. Thank the lord they figured out that they need to raise them faster than expected inflation. If the expected inflation rate goes from a 2% to 5%, then then target for short term rates must go up by more than 3 percentage points. OK, an improvement over the seventies.

But this whole approach needs to go.

I have problems with your language. Perhaps you may clarify the following point:

FFR is a rate at which banks borrow from the Fed, whereas the IOR is a rate at which banks lend to the Fed. This is at the heart of the free-lunch argument: banks borrow from the Fed at FFR and lend it to the Fed at IOR. The difference is risk-free return to bolster balance sheets.

If this is the case, I don't get your opportunity cost reasoning. The most valuable alternative foregone if one invests in excess reserves at IOR seems to be some other kind of short-term investment in securities (see Hamilton). The opportunity cost of FFR is the best borrowing rate foregone, which cannot be the IOR. If you borrow to invest, FFR and IOR are no alternatives but possibly complements and opportunity cost reasoning doesn't apply.

I'm sure the mistake is mine, because I don't know the semantics of your economic peer group. I just don't what my mistake is.

FFR is the interest rate at which depository institutions lend balances to each other overnight.
IOR is the rate banks lend to the Fed. Zero is the rate Fannie and Freddie lend to the Fed. As a result, FFR is between zero and IOR. I see no evidence that this Fannie Freddie arbitrage hurts the economy.

In late September 2008, there was no IOR, but banks were earning convenience/scarcity yield on reserves that was much higher than the fed funds rate target.
In November 2008, banks were earning IOR on reserves, but additional scarcity yield was quite low.

I think about it. Thx

"The Money Demand",

I don't understand your point about Fannie and Freddie arbitrage, could you explain it?

I read about the Bank of England's policy on their website. Before 5 March 2009 the commercial banks had to pre-declare how much reserves they were going to carry over a particular time period. If on average they met their target then they got paid reserves. After 5 Mar 2009 that no longer applies and reserves are paid interest at the bank rate.

Aresh is mistaken concerning my position on optimal MV growth. I favor growth of MV at a rate equal to that of (weighted) factor input growth, and I'm perfectly clear about this in Less than Zero. Absolutely constant MV may be someone's ideal, but it has never been mine.

Sorry Mr. Selgin,

you indeed account for changes in total factor supplies in efficiency terms, and not just for changes in labor productivity as I ascribed to you. Both fall under the label of productivity norm, which I associate with your name. That's why ... The point is that the debate focused on more fundamental issues which can be discussed in a simplified version (just growth of labor in efficiency units).

Let me add that I also don't ascribe the constant MV position to Steven or any other monetary equilibrium theorist. It is just that most of the debates here and some of the publications (not yours) assume constant labor force (n=0) and positive growth in labor productivity. Technological progress seems to be Harrod-neutral, since labor reaps all productivity gains. In such a framework your policy rule boils down to targeting nominal wages. It is quite natural that the policy rule alters in light of more detailed models.

Nonsense! What I meant is that Selgin allows Mv to grow with total factor supplies and not changes in total factor supplies in efficiency terms (and I ascribed to him Mv because L is constant in our discussion here - at least this was my assumption) .... To target growth in efficiency units is no productivity norm anymore.

The use of T-bills for collateral (or the use for reserve balances for settlements) really doesn't matter.

The point is for the interest rate the Fed pays on reserves to float.

My choice of the 4 week T-bill yield is that reserve balances are shorter and more liquid (at least to banks) than T-bills. The Fed is intermediating, borrowing from banks and the public (by issuing reserves and deposits) and lending the proceeds in a variety of fashions, but holding short government bonds can (and should) dominate.) The Fed shouldn't provide intermediation for free, and should pay less than it earns. Since even the short government bonds it buys are longer than the reserve balance it issues, the interest rates it pays should generally be lower than the interest rates it pays. 1/2 percent is arbitrary, really.

But the point is for it to float with market yields.

I can undestand some market forces that would result in the federal funds rate being no lower than the interest rate the fed pays on reserves (why would any bank lend?) and I also was aware that freddy and fanny hold reserve balances and are paid no reserves and for some reason regular lenders on the market so banks can borrow for less than the interest rates they recieve on reserves. And I resume banks don't borrow from fanny and freddy just to hold reserves and collect interest because of risk or transactions costs.

But why is this a problem? Or, with different conditions, the federal funds rate is above the interest rate that the fed pays on reserves?

Do I misunderstand what you mean about the federal funds rate should equal the interest rate on reserves?

Anyway, paying interest on reserves raises reserve demand and exacerbates an excess demand for money. It keeps money expenditures higher than they otherwise would be. This can be offset by a larger quantity of base money--in other words, the Fed holds even more longer and riskier assets.


I agree that QE2 'could be counterproductive'. If banks' demand for base money fell very suddenly, then an excess supply of money could develop. It is possible the Fed would be unable to contract the money supply before prices start rising. But I think this is unlikely. The Fed doesn't want inflation to rise much above 2 percent; in this case the Fed has managed expectations well. That is, expectations of future contraction would counter inflationary tendencies. I understand your point, but I just weigh the costs and benefits differently -- I don't know if either of us is right.

I'm late to this because, along with Steve Horwitz, I spent the weekend with a happy band of monetarists. There was not much sympathy for either the policy of QE2, or David Wessel on what Milton would have said. (N.B. The WSJ piece is not an editorial, but Wessel's column. Wessel is the Fed reporter and must be read carefully.)

Some observations.

IN my opinion, Beckworth's paper was a selective reading of Freidman. Friedman of the long and variable lags is absent from his story. That part of his theory was not so far from Hayek.

John Taylor captured Milton's position the closest of all those quoted.

Milton was not an inflation targeter, but favored targeting the money stock (M2). Where did Milton ever advocate discretionary monetary policy to counteract unemployment?

Allan Meltzer and I have been singing the same tune, so of course I liked his WSJ column. Allan favors a rule; the Taylor rule in his analysis has worked; and the Taylor rule is consistent with the Federal Reserve Act (the dual mandate). Many rules that economists concoct are not consistent with the law.

As Steve Horwitz, Richard Ebeling and I have observed here, Brunner/Meltzer always contained Hayekian and UCLA insights (non-Walrasian price theory and radical uncertainty). Monetarism is not a monolith.

Anna Schwartz is just being a good economic historian who understands manias. She has opposed Bernanke's policies and called for him not to be reappointed.

I learned that fear of deflation no longer grips monetarists. (Not that we have had conventionaly measured price deflation.) In that, they are going back to the findings of Freidman and Schwartz (1963: 15). No historical case can be made from US monetary history that deflation and economic growth are incompatible. In his history of the Fed, Meltzer reconfirms that finding.

Bill Woolsey,

During the financial crisis there is a great uncertainty about the growth of broad money supply, and as a consequence, there is a great uncertainty about the demand for monetary base. This means that so called "floor" system is optimal. In this system FOMC (or preferably a market mechanism) sets the fed funds rate target that is expected to result in monetary equilibrium, interest rate on reserves is set equal to fed funds rate target, and the system is saturated with reserves so the fed funds rate drops until it is equal (or at least very close) to the IOR rate. The Fed de facto announced the switch to the floor system on October 13, 2008, and in fact on this day excess demand for money started falling. The Fed should have switched to the floor system just after Lehman, failure to do this has led to the concern that the quantity of monetary base might be too low in the future, this concern has created excess demand for monetary base.

I don't agree that the Fed should hold only short maturity government bonds, as this can severely distort market estimate of sovereign default risk and can lead to a underestimation of the cost of fiscal stimulus. During financial crisis there is a great uncertainty about the growth of broad monetary aggregates, as a result private sector credit assets are cheap, and by purchasing a diversified basket (or index) of short maturity private sector credit instruments, the Fed can earn the safe spread between the IOR rate and private sector credit instrument rate.

You said:
"My choice of the 4 week T-bill yield is that reserve balances are shorter and more liquid (at least to banks) than T-bills."

In fact, 4 week T-bill is much more liquid than reserve balances at the Fed.

The yield on 4 week T-bill should be higher than IOR because of duration risk premium. The yield on 4 week T-bill should be lower than IOR because 4 week T-bill provides excellent liquidity services to a wide spectrum of market participants, many of them are unable or unwilling to hold accounts at the Fed. At present, the second effect clearly dominates, and in equilibrium 4 week T-bill rate is lower than IOR. To fix IOR - 4 week T-bill spread at 50 bps would result in a severe market distortion.

The strange impact Fannie and Frieddie causes to the fed funds rate doesn't really matter, my remarks about them were directed at Arash, and my intent was to prove that we can safely ignore the fact that fed funds rate is lower than interest on reserves.

Perhaps I can explain my support for QE2 in different language.

The Federal Reserve is a central bank. Although the Fed has been liberated from many of the ordinary constraints of ordinary banking, it still operates along similar principles to any other financial intermediary. The notes issued by the Fed may no longer be liabilities, but good monetary policy should treat them as such. An increase in demand for base money is effectively lending to the Fed, and it is the Fed's responsibility to invest those savings by increasing credit in proportion.

The situation is complicated by the absence of of any reserves at the Fed; the principle of adverse clearings that ordinarily would signal when to expand or contract is missing. Instead, the Fed must use crude substitutes like interest rates, inflation, or unemployment. But given present institutional arrangements, the Fed must do the best it can to match savings and investment.

Given all this, I do not believe the Fed is currently fulfilling its obligation to its "liability" holders. QE2 is merely a way to ensure base money demanded equals base money supplied, because nobody but the Fed can satisfy that demand.

Oh, and future expectations of monetary policy are also really really important.

I don't agree with the policy approach of--broad monetary aggregates are uncertain, so a floor on interest rates is optimal. I don't favor any flooding with liquidity.

You are in error that having the Fed pay less on reserves than any target interest rate creates are severe market distortion. It just reduces the quantity of base money the Fed needs to create to be consistent with monetary equilibrium. The micro consequence would be that banks have more transactions costs to clear payments because they economize too much on reserve holdings. Under current conditions, I don't think that is relevant.


After M2 velocity began to change in the eighties, there are a number of quotes from Friedman that suggested that he quit supporting M2 targeting.

The stable, nearly constant, relationship between money expenditures and M2 were always theoretically suspect. Why should that set of assets (including CDs less than 100k and money market mutual fund balances less than 50k) closely track money expenditures?

Quasi-monetarists of today are not that different from the monetarists of the past, but I think all of us have given up fully on the notion that there are numerical constants in the economy, like Y/M2 = k (where k is a constant.)

We all know about new classical and real business cycle theory. This is inconsistent with the monetarism of Friedman.

Friedman's version of monetarism (like that of Yeager) never assumed perfect market clearing so that aggregate money expenditures are irrelevant and real expenditures always equals the productive capacity of the economy. However, many, if not most, free market macoeconomists took that turn and see fluctuations in real output and employment as always being due to changes in productive capacity.

I will grant, however, that the "long and variable lags" of Friedman is the most likely difference between those of us advocating a further expansion in the quantity of base money and Friedman. The current level of base money may be sufficient (and likely excessive) and its impact on future monetary expenditures will appear sooner or later. Maybe.

However, I would make one correction to some of these claims. Friedman advocated M2 targeting and not monetary base targeting. During the Great Depression, for example, M2 fell, but the base rose. My understanding of Friedman's old view (assuming M2 velocity was more or less constant) was that the increase in base money was inadequate. In Friedman's view, (and that of orthodox monetarists) possible large fluctuations in the money multiplier can and should be offset by perhaps heroic changes in the quantity of base money.

If a target for the growth path of M2 were undertaken today, nearly all the change in base money we have observed would be necessary. Further, the "long and variable lags" between open market operations and the level of M2 created by the banking system didn't appear to be a problem for Friedman. In other words, if and when the weekly M2 figures start to show excessive growth in M2, then base money can be decreased as rapidly as needed to offset any increase in the money multiplier.

The only real change--what made many of us into quasi-monetarists--was giving up on any assumption that some measure of the quantity of money could be found that would have a constant ratio with money expenditures. When that is given up, then open market operations to offset changes in velocity as well as changes in the money multiplier is where you are left.

Unless, of course, you quit worrying about aggregate money expenditures and convince yourself that changes in prices and nominal incomes, including wages, are sufficiently rapid and smooth to keep real expenditure equal to productive capacity, so money expenditures don't matter. And then try to explain away all the firms cutting output and employment as being a response to bad sales. And try to explain it all as some change in the productive capacity of the economy--you know, new classical and real business cycle theory.

That Meltzer went over to the Taylor rule is distressing. Targeting short term interest rates based upon the output gap and 2 percent expected inflation from wherever the price level might happen to be? And, there are some supposed "constants," right? The coefficents in the rule.

When you read old monetarists fighting the last war, correctly arguing against targeting real output or unemployment with monetary policy, it is distressing. Who is arguing in favor of that?

I realize that the quasi-monetarist focus on a target for the growth path of nominal expenditures is a minority view. Raising the expected inflation rate and lowering long term nominal rates appears to be more dominant approach. The reason, however, we have such difficulties in overcoming that approach is because of the baleful influence of the Taylor rule.

Anyway, targeting a growth path of money expendtures isn't an inflationary disaster waiting to happen like targeting real output, unemployment or interest rates (all things that play a role in the Taylor rule.)

It is consistent with the "dual" role of the Fed, though I hardly count as a virtue, other than using monetary policy to force prices down in the context of a negative shock to productivity is a really bad idea, and forcing prices back up after a positive shock to productivity is a bad as well.

While the "long and variable lags" between changes in M2 (or some measure of the quantity of money) and money exenditures may be the biggest gulf between Friedman and the quasi-monetarists who have given up on constant velocity, the biggest divergence between the new Meltzer and orthodox monetarism is blather about interets rates being low and excess reserves--the confusion between money and credit. If the quantity of money is below target, then open market operations should be used to expand base money to offset any change in the money multiplier and get the money supply back up to target. Traditional Fed blather about free reserves, excess reserves, short term interest raets, and the like were properly dismissed as being confused.

Of course, I think the Taylor rule _is_ mostly just giving in to the central banker's preference to focus on short term interest rates. To bad Meltzer has bought into it.

Bill Woolsey,

Our main difference is that you ignore the risk and uncertainty in determining the parameters of monetary policy that are consistent with the monetary equilibrium.

During normal times, it is possible to obtain a reliable estimate of quantity of base money that needs to be created to achieve monetary equilibrium. In this case, the spread between IOR rate and fed funds rate should reflect the difference between the credit risk of commercial banking system and the Fed, and a corridor system is optimal.

During a financial panic, it is hard to obtain a reliable estimate of quantity of base money that needs to be created to achieve monetary equilibrium, and this estimate is subject to significant intraday revisions. Estimation errors are unavoidable, but the harm is asymmetric. If the quantity of base money is overestimated, the harm is minor, however if the quantity of base money is underestimated, a sharp deflationary destabilization of monetary equilibrium results. From September 16 till October 13 2008, the Fed has frequently and severely underestimated the quantity of reserves that need to be created to restore monetary equilibrium. For this reason many central banks have switched to the floor system for the duration of the financial crisis.

The market distortion I mentioned is not related to the issue of corridor vs. floor system. The distortion results if the Fed attempts to perform harmful financial intermediation by issuing relatively illiquid monetary base and mopping up relatively liquid 4 month treasuries. Thankfully, the Fed understands this and has reduced its holdings of treasury securities during the financial crisis. ECB has taken steps to increase the liquidity of its liabilities, for example, term deposits at the ECB are eligible form of collateral at ECB liquidity auctions.

Bill Woolsey's post covers a lot and I will let Allan Meltzer speak for himself. I was mainly concerned that we get Milton Friedman right, because I found the Beckworth paper -- and even more so, the Wessel column -- to be tendentious.

The last written piece by Milton that I can recall made an effort to defend M2 targeting. I asked Anna Schwartz a few years ago what she thought about the reliability of M2. She responded that only time would tell. The St. Louis Fed folks are still trying to hold on to targeting aggregates.

Milton never said V was constant, just less variable than the expenditure multiplier. Deviations from predicted values of V are not predictable and should not be offset by discretionary policy (the long and variable lags argument). I don't see how Woolsey and others can escape that critique.

Milton followed the Bagehot rule for lender of last resort. That was not "discretion" in his mind. In any case, it was not a response to unemployment. I don't think you can find an instance in which he advocated a change in monetary policy in response to unemployment or any other real variable.

There was a general consensus in the group of monetarists at the weekend conference that there is absolutely no theory behind QE2. There is no way to estimate how much to inject because the models being used have no quantities in them. There is only one observation QE1, so the whole enterprise is subject to the Lucas critique in spades.

In short, Bernanke is acting without reason. Draw your own conclusions.


I do not think Bill or anyone else here is in favour of discretionary monetary policy. The 'quasi-monetarists' all prefer a rule targeting some measure of nominal spending. In principle, such a rule is hardly different from a target for the growth path of M2 -- it simply aims at a different aggregate.

@ money demand,

I think I got it. you are right. I was wrong .. and surrender. For those who still disagree, I found the following link quite helpful:

@Lee Kelly,

Again, my principal goal was to get the record on Milton straight. In passing, I wondered how Bill would solve the knowledge problem. I appreciate his posts.


Congrats on your election to the CC in DC, :-).

Would I be right in saying that the problem with the Taylor rule is that it creates a sort of "dual target" situation? On the one hand the central bank are targeting an inflation rate, explicitly or otherwise. They can do so by the Taylor rule. But, when things go wrong, as they have it isn't clear if they will continue with the rule and see what happens or continue targeting inflation (or NGDP) using feedback and ignoring the rule. And that creates uncertainty.

So, in this view the real problem isn't the Feds dual mandate for inflation and employment. It's the conflict between:
1) Targeting a short-run interest rate, and
2) Targeting the long run intended consequences on prices.

Bill, is that what you mean?


The Taylor rule does target real output (the output gap, anyway.)

Quasi-monetarists don't target real variables--money expenditures, nominal GDP, or whatever, are nominal variables.

I suppose some people who favor quantitative easing at this time may favor targeting real output or unemployment. Still, the dominant view seems to be that we should raise the target for the inflation rate (a nominal target) on the view that this will raise output and lower unemployment. That is not the same as targeting higher output and lower unemployment. If the higher inflation rate leaves production or unemployment unchanged, then so be it.

Of course, I certainly hope that increased money expenditures raise output and reduce unemployment, but I don't favor targeting either. If higher money exenditures leave output and unemployment unchanged, then so be it. Apparently, then, the entire decrease in real output would have been due to a decrease in productive capacity.

My solution to the knowledge problem? Don't you know? Index futures convertibility.


Thanks. Hope to see you there.


OK. But why would someone target inflation even if it might leave real output and unemployment unchanged?


Interest rate targeting is just a bad idea.

Mr. Money Demand:

I grant that during a financial crisis the relationship between the money exenditures and the demand for base money calculated in the past will not apply.

But it is equally true that the relationship between short term interest rates and money expenditures that applied in the past will no longer apply during and in the aftermath of a financial crisis.

The notion that using interest on reserves to create a floor for interest rates provides any extra benefit during or in the aftermath of a financial crisis seems wrongheaded to me.


Perhaps I misuderstood Friedman, but my understanding is that he favored having the Fed use open market operations to serve as lender of last resort to the banking system as a whole, rather than freely lending to individual banks at a penalty rate. Why else would Friedman have favored closing the discount window?

M2 targeting means changing the quantity of base money to offset changes in the M2 money multiplier. Trying to make sure that market interest rates rise during this process appears to be a distraction. There is plenty of reason to believe that Friedman was free of any confusion regarding money and credit and was more than willing to let interest rates be determined by market forces.


Yes on lender of last resort. And, of course, on interest-rate targeting.

Friedman's original formulation was the x percent rule: increase M2 at a constant rate estimated to bring about long-run price stability. I can't recall his buying into feedback rules, but am prepared to stand corrected on that.

I haven't looked into it much, but Sumner and Beckworth have found quotations from later years (after M2 velocity appeared to have some permanent shifts) complementing Greenspan on his ability to offset that shift and then later, support for the concept of targeting the market expectation of inflation found by the difference in yields between tips and ordinary bonds.

Bill Woolsey,

Taylor rule is a very dangerous oversimplification that has created a cargo cult. But the real harm was done not by the Taylor rule, but by the failure to supply sufficient quantities of base money during the financial crisis, even though the quantity supplied was frequently revised by the Fed in September-October 2008. Only when the Fed promised to supply unlimited quantity of monetary base and switched to interest rate targeting to ration the base using the floor system, the problems with insufficient quantity of monetary base disappeared (a lesser problem with fed funds rate target that was too high and LOLR interest rate that was too high still remained).

The goal of the floor system is not to prevent short term risk free interest rate from falling (higher short term risk free rates are just a side effect). The goal of the floor system is to prevent short term risky interest rates from skyrocketing.

The interest rate floor in the floor system should be frequently revised, and failure to do this was a secondary cause of Great Recession.

O'Driscoll & Woolsey:

On pages 227-229 of Monetary Mischief Friedman calls for legislation that would have the Fed target expected inflation from the bond market. He notes it would address many of the long and variable lag problems. He also notes that it would make the Fed more accountable since there would be instant feedback on the Fed's performance.

Relax for ten minutes and rinse with cold water.- To remove stains and purify the skin, crushes a banana and mix with five tablespoons of oatmeal powder, add water until thick and use it to massage the face.The kiwi is the basis for masks used in oily skin.- Peel the kiwi fruit, chop and puree, add boiling water and let sit for thirty minutes. Then Add one tablespoon of vodka or white liquor (by alcohol), stir well and let sit again. Apply to clean face for fifteen minutes and then rinse.

The comments to this entry are closed.

Our Books