Arnold Kling  

My Bizarre Monetary Theory, Continued

Robin's Atoms... Stringham's A Winner... Again...

Before I respond to Bryan's comments, let me first recommend the two Scott Sumner papers that he mentions.

This paper develops a theory that cultural values drive economic policy, which in turn drives economic outcomes. It is an interesting complement to (or substitute for?) Bryan's view that economic education can drive economic policy.

For the issue at hand, monetary theory, Sumner writes,

Almost everything we have learned from recent research in monetary history, theory and policy, all points to the Fed as being the cause of the crash of late 2008. More specifically, an extremely tight monetary policy in the US (and perhaps Europe and Japan) seems to have sharply depressed nominal spending after July 2008. And yet it is difficult to find any economist who believes this. More surprisingly, few economists are even aware that their views conflict with the standard model, circa 2009. Even if I am wrong, and the standard view of the crisis is correct, there is an important message in this paper. The standard view of the current recession conflicts with some of the most important concepts that we currently teach our students in monetary economics courses.

He is saying that the textbooks all tell an MV = PY story for macroeconomic fluctuations. Nobody seems to be telling an MV = PY story for this crisis. What's going on?

I do not tell an MV = PY story for any macroeconomic fluctuations (other than very high inflation). Instead, I tell a Recalculation story. Thus, I do not have a consistency problem. My problem is to defend my bizarre monetary theory. As Bryan puts it,

If Arnold's position were that raising nominal GDP wouldn't help real GDP, I could understand it. But he's actually questioning the ability of central banks to affect nominal variables. Even stranger, he's using hyperinflations - airtight proof of central banks' near-infinite power over nominal variables - to prove that this power is illusory.

I do not think that central banks control nominal GDP any more tightly during a hyperinflation than at other times. That is, in normal times, if nominal GDP wants to grow at a 5 percent annual rate next quarter, the central bank cannot hit a 10 percent target. In hyperinflation, if nominal GDP wants to grow at a 10,000 percent annual rate next quarter, the central bank cannot hit a 10,005 percent target.

Like most economists, I view real GDP in the long run as determined by real factors, such as the supply of factors of production, the state of technology, and the nature of economic and cultural institutions. However, I view average prices in monetary units as reflecting habits. The government can change people's habitual price behavior only by making significant, long-lasting changes in the amount of deficit that it finances by printing money. On the other hand, changes in money-printing that are modest and short-term have essentially no effect.

Think of monetary policy as being like currency intervention. It seems as though it is very difficult for a government to maintain a currency at a value that the market views as unrealistic. Similarly, it is very difficult for the government to maintain an interest rate at a level that the market views as unrealistic.

Another way to express my view is that there is a probability distribution for nominal GDP growth. By printing money much faster starting today and persisting for several years, the government can raise both the mean and the variance of the distribution of nominal GDP growth many years from now. However, in the short run, both the mean and the variance are determined by things that have happened in the past, including past monetary policy but also including Recalculations and other factors that affect real GDP as well as past habits of price-setting.

Sumner's view of the causal ordering of the current recession is that the Fed cut M, which reduced nominal GDP, which reduced real GDP. My view is that the Recalculation reduced real GDP, which also slowed the rate of price increase. The standard view is that there was a huge increase in the demand for M, which lowered nominal GDP. An interesting question is whether it is possible to use data to evaluate these differing explanations.

Comments and Sharing

COMMENTS (15 to date)
Bill Woolsey writes:

Sumner doesnt say that the Fed "cut M." He says that the Fed changed a policy that was changing M to offset changes in V. V dropped and the Fed increased M by less than V dropped, so that MV dropped. And so Py dropped. As the market saw that the Fed had changed its policy, V dropped way more, and M again failed to increase enough to offset that very large drop in V, so that MV dropped a lot. And so, Py dropped alot, and it remains low to this day.

Anyway, the short version is that M increased, but not soon enough, and, in the end, not enough.

He agrees that there is some kind of "recalculation" and that this resulted in lower y. If MV had remained on target, then P would have risen and y would have fallen.

Of course, all of this is more properly explained in terms of growth rates. A large decrease in V required a very large increase in M so that MV would continue on its previous 5% growth path. Recalculation may reduce the growth rate of y, perhaps even to negative growth, resulting in more inflation, a higher growth rate in P. Becaue the Fed failed to increase M enough, MV decreased and so did Py. This caused y to drop even more than what was necessary for any plausible recalculation. potential y is below its prevoius growth path because of reallocation of resources. y is below potential y because of sticky prices in the face of a drop in MV.

Kling's theory is that recalucation reduces y and V together, and that no increase in M can raise MV so that Py remains remains on target. Efforts to increase M just cause even further decreases in V.

While Sumner believes that if the Fed hadn't paid interest on reserves (and so decreased the money multiplier) then a smaller increase in B (base money) would have been adequate to raise M enough to offset the decrease in V and keep Py growing on target. Again, the reduction in the growth of y (perhaps to negative levels) due to needed resource reallocations, would have resulted in higher inflation.

On the other hand, nothing in his view depends on the notion that smaller increases in B were needed. If the Fed didn't pay interest on reserves and even charged a penalty for excess reserves, and that still required even larger increase in base money to offset any remaining decrease in the money multiplier and the decrease in V so that Py grew on target, so be it.

If some process exists by which open market operations further lower V (which I think is very plausible during this period) then Sumner's approach is to increase B and M even more. Increase base money whatever amount needed so that M rises whatever amount needed to offset whatever drop in V occurs to keep Py on target. If the Fed runs out of T-bills to buy, they must buy longer term government bonds. If they buy up the national debt, they have to buy foreign goverment bonds or private securities. While not entirely worry free, please get back to me after all the T-bills have been purchased. Don't tell me that buying all the T-bills won't work and we don't want to buy anything else, so don't even try.

I think Kling's macro theory is very wrong. The parts that are correct are nothing new. Everyone knows about structural unemployment, and who can possibly not see that potential income is impacted? But the notion that drops in potential income cause velocity to drop just the right amount so that the price level remains on its previous trajectory seems implausible. And that an increase in base money, no matter how large, will always result in a 100% offset by a further decrease in velocity is implausible. That isn't new, really. It _is_ the liquidity trap exactly.

So, we have supply-side factors that exist, but are confused with nominal expenditure. (And some odd unwillingness to see that Sumner and other monetary economists recognize those and say.. yes structural unemployment, we know.) And then there is the liquidty trap.

I will accept that prices are sticky, but the notion that nominal expenditures depend on habit, but at the same time, they drop due to some kind of recalculation, is an obvious contradiction. If they did depend on habit, then nominal expenditures wouldn't change due to this recalculation.

If Kling were to say that there is a shortage of gasoline, but price increases would do no good, because quanties supplied and demanded of gasoline depend on habit, I would be no more shocked by his statements.

I would note, that I prefer to describe all of this using the quantity of money and the demand to hold it. I think "velocity" is generally not a desirable concept. It just confuses matters. Sumner actually prefers "k" which is the reciprocal of V, and the demand for real money balances relative to real income.

fundamentalist writes:

M doesn't affect P with a constant coefficient. My reading of the German hyperinflatin shows that the coefficient of M with respect to P is considerably less than 1 during depressions, reaches 1 during recoveries and exceeds one during bubbles. That's one reason the feds have a very difficult time fine tuning the economy. In addition, the lags are quite long.

fundamentalist writes:

Kling hasn't come around to the Austrian business cycle, yet, but Austrians would say that the monetary pumping that Sumner lusts after caused the misallocation of resources that made Kling's recalculation necessary.

Sumner sees no dangers whatsoever in monetary pumping. As with all monetarists, it does nothing but raise prices across the board. Where's the harm in that?

Even without the misallocation of capital that monetary pumping causes, the amount of pumping the feds would have to engage in to increase ngdp when gdp is falling would be enormous. With the long lags involved, we would be facing hyperinflation in a few years.

Even Hayek eventually came around to endorsing efforts to reduce the secondary deflation that appears in a bust, but that is a long ways from Sumner's view that monetary pumping can cure everything.

As for MV=PY, Hayek wrote that the worst mistake an economist can make is to ignore it; the second is to take it too literally.

David Beckworth writes:


See this write up on how MV=PY can be useful in thinking about the Crisis. Scott Sumner and George both liked it.

Bill Woolsey writes:


P = (V/y) M

So, you found out that V/y is not constant. Very good.

Those of us who favor targetting aggregate nominal expenditure are happy to let the market determine the mix between P and y in the Py product. For those who favor targetting the quantity of money as a means to stabilizing nominal expenditure, then showing the V doesn't change much is important.

Those of us who favor changing M whatever amount is needed to offset any change in V (including any generated by the procedure used to change M) are more interested in whether every change in M is fully offset by a change in V. That is the same thing as the liquidity trap.

The changes in M needed to offset V and maintain nominal expenditure can only lead to malinvestment if entrepreneurs foolishly take the relative prices during the crises as being permanent. And I did say foolishly.

Greg Ransom writes:

Hayek endorsed efforts to reduce the secondary deflation in stuff written in 1931, and published in early 1932.

What people need to come around to is that there understanding of Hayek's intellectual history is mostly myth.

fundamentalist wrote:

"Even Hayek eventually came around to endorsing efforts to reduce the secondary deflation that appears in a bust"

fundamentalist writes:

Bill: "...if entrepreneurs foolishly take the relative prices during the crises as being permanent..."

I'm not sure what relative prices you refer to, but why would that be foolish if the fed makes it clear that under no circumstances will it allow nominal prices to fall?

Anyway, the misallocation of capital happens because of artificially low interest rates. Even if the fed uses the open market to buy securities, that money goes into bank deposits, raises loanable reserves and thereby reduces interest rates. Low interest rates encourage excessive investment in capital goods, such as houses and autos. Now if the current crisis exists because of previous over investment in houses and autos, how will increasing that over investment even more improve things?

Bill: "So, you found out that V/y is not constant. Very good."

If you know it's not constant, why insist that the feds can increase ngdp at will? The fact that it isn't constant, and is below 1 during a depression, makes monetary pumping such a bad idea. Say that V/y is .5 during the depression. Then the feds would have to pump in twice as much money to get the same price increase as it does during an expansion. But the ratio is probably even lower during a depression. So you end up with the feds having to flood the country with money. But there is a long lag between monetary pumping and price increases which makes the practice even more dangerous.

How long do you think the lag between policy and price changes is? I have seen econometric analysis suggesting it's about 4.5 years to max effect. Even if it's half that, say 2 years, by the time the fed's monetary pumping increases ngdp much, the crisis will be over. That's why fed policy, historically, is pro-cyclical.

Lee Kelly writes:


Credit expansion contributed to the boom and bust, but it may nevertheless be an appropriate response in the aftermath. Although this might seem like administering more of the poison, the poison was never credit expansion per se, but rather the mismatch of interest rates with consumers' willingness to defer consumption.

After a bust, demand to hold money can suddenly spike, and thereby defer consumption. The problem is then a mismatch in the opposite direction -- too much saving and not enough credit.

Elvin writes:

I don't think we have an adequate explanation of what happened in 2008 worked out yet. Pressures were building on a daily basis in the credit markets and culminated in October: The auction-rate market collapsed in January, and the muni-insurance market was rocked as well. Bear occurred in March. FNMA and FHLMC had special legislation passed in July allowing the Treasury to seize them (Bill Gross essentially pronounced them toast at that time). A hedge fund was collapsing every week. In the background, you had SIVs, asset-backed commercial paper, and collatrized bond obligations built on mortgage-backed securities with unknown valuations (how low would housing go?). Suddenly, with Lehman, AIG, the GSEs, the cash market froze. It was as if the markets suddenly realized that, despite the efforts of the Fed and government efforts, there was to be no smooth landing or soft recession, that this was real deal, a humdinger of a recession with tremendous financial implications. It's a really big recalculation.

This is what happens when credit bubbles are burst. One way to look at is with a very expansive definition of M (cash + assets). If M drops enough, then P and Y can drop as well, even if v is declining (or k is increasing).

fundamentalist writes:

Lee: "The problem is then a mismatch in the opposite direction -- too much saving and not enough credit."

But credit expansion causes over production in capital goods. So you think greater over production in capital goods will cause consumers to spend more?

Saving isn't the problem. People are saving now because they lost a lot of wealth in the boom and bust. The problem, as Kling points out, is recalculation. Over production caused by loose credit needs to be liquidated and people working in the industries with over production, housing, autos and finance, need to find other lines of work.

And consumer spending isn't the problem. That's old stale Keynesian thinking. The problem is over production.

Lee Kelly writes:


The issue is not merely "overproduction" of capital goods, but also a change in the composition of capital goods produced. I do not believe in monetary expanstion to promote greater "overproduction," but aid in the recomposition of the capital goods structure. This recalculation problem is exacerbated by deflation (which adds "noise" into the price signal, to quote Horwitz), so the purpose of increasing the money supply after the bust is to achieve monetary equilibrium. Although this would stabilise the price level, the purpose is not to promote consumption or temporarily prop up an unsustaniable capital structure, but to allow relative price adjustments to occur with the least friction possible.

All this would happen automatically in a free market for money.

winterspeak writes:

FUNDAMENTALIST: I love it! The US has plenty of houses (real wealth) but not enough (nominal) money to pay for them.

Your solution: destroy the houses!

fundamentalist writes:

Lee: "I do not believe in monetary expanstion to promote greater "overproduction," but aid in the recomposition of the capital goods structure."

The major problem is overproduction. The entire capital structure is too large and has to shrink in order to establish the proper ratio between consumer demand and production.

Lee: "This recalculation problem is exacerbated by deflation..."

No, deflation aids the recalculation because the only way to clear the market that is clogged with too many goods is to lower the prices.

winterspeak: "The US has plenty of houses (real wealth) but not enough (nominal) money to pay for them. Your solution: destroy the houses!"

I never suggested destroying anything. That would be stupid. But it's not true that we don't have enough money to pay for houses. There is no such thing as not enough money. Any amount of money will make markets clear if prices adjust. What we have is less credit than we had before because no one wants to borrow. When housing prices fall enough, the market will clear. There is no need to destroy anything.

fundamentalist writes:

Elvin: "This is what happens when credit bubbles are burst. One way to look at is with a very expansive definition of M (cash + assets). If M drops enough, then P and Y can drop as well, even if v is declining (or k is increasing)."

Good point. Still another way to look at it is to let assets be in P. We had a lot of asset price inflation in housing and equities during the past decade. I think velocity is just a catch-all to hold all of the many reasons why people won't borrow or spend. When M collapsed because credit collapsed (people quit borrowing), V probably collapsed, too, which made both P and Y fall.

But we need to know why M collapsed. Was it just a random event? Or could it be because shortages in the economy that appeared in the boom caused some businesses to fail?

It's all about expectations. When business people borrow and expand their businesses, they expect that others have similar ideas, so that production and consumption of others will dovetail with his. For example, a producer of bread making equipment expects that bakers will increase production and buy his equipment. Bakers must expect consumers to buy more bread before he will buy more equipment. Producers require complementary producers in order for their plans to work, and all producers require complementary plans on the part of consumers. Depressions happen when the plans of one group get out of whack with those of another, especially when producers' plans get ahead of plans by consumers. Our banking system promotes such a disconnect between plans when it expands credit too rapidly. The governor of the rate of expansion is savings by consumers. But the banking system can create "savings" out of thin air that work exactly like real savings. The fabricated savings cause overproduction in the capital goods markets.

Lee Kelly writes:


The problem is relative prices.

If price level adjustments up or down are caused by monetary disequilibrium, then "noise" will be added into the price signal. This can be alleviated by expansion or contraction of the money supply. The market can still clear goods from malinvestment lower prices, while spending shifts into investments genuinely backed by real savings.

If it wasn't for the Federal Reserve, this would happen in a free market for money and banking.

Comments for this entry have been closed
Return to top