Arnold Kling  

This Should be Good

PRINT
Four Year From Now Plans... Violence, Anarchy, and the Sta...

A Debate Between Scott Sumner and John Cochrane. Sumner thinks that the Fed can hit a nominal GDP target if it wants to. When I heard Cochrane a while back, he was talking about the high rate of substitutability of financial assets, which might make monetary policy fairly ineffective. As you know, I tend to be more sympathetic with the latter view. If my schedule were more normal, I would have watched the video by now.

UPDATE: Scott has a blog post that sums up the debate.

I don't think that the Fed can fine tune an inflation rate. By the same token, if it sets a price level target, it cannot fine tune the date at which it will achieve that target.

I think that inflation expectations move very slowly. Only in a regime of high inflation expectations does money have imperfect substitutability with other assets.


Comments and Sharing


CATEGORIES: Monetary Policy



COMMENTS (19 to date)
Dan writes:

How would the substitutability of financial assets affect monetary policy?

Arnold Kling writes:

Dan,
The Fed operates by trading one type of asset for another. For example, it can trade money for bonds.

To see what happens if these assets are nearly perfect substitutes, imagine that the Fed was trading dimes for nickels at the rate of one dime for two nickels. Using this approach, even if the Fed conducted a massive open market operation in which it bought nickels for dimes, it is unlikely to have much effect. People see two nickels and a dime as equivalent, so they do not change their behavior when they find themselves with many more dimes and many fewer nickels.

Mitch Oliver writes:

Interesting. That view seems to imply that assets act as privately issued currencies in financial markets.

fundamentalist writes:

Sumner is a one-trick pony.

He assumes that economics works just like hydraulics: you pull a lever and something moves. But it's not. It's about human action. You don't need any data to tell you that the Feds can't control ngdp at will. Recall Greenspan's complaint about pushing on a string. Anyone who has been alive more than 20 years has seen the feds try to goose ngdp during a depression; it's impossible.

When the economy starts to recover and people want to borrow again, then the fed's amazing powers kick in and ngdp jumps. Unfortunately, Sumner suffers regular bouts of post hoc ergo propter hoc and thinks that the feds caused the jump when in reality they're just hitching a ride on the market's natural ability to recover from a depression.

Richard A. writes:

The Fed controls both the size of the monetary base and the M1 money multiplier (MM). They drove down MM when they began paying interest on reserves. I am beginning to suspect that they did this deliberately in order to allow themselves to massively expand the monetary base in order to use the monetary base as a slush fund.

Scott Sumner writes:

fundamentalist, I have made exactly the opposite argument from the one that you attribute to me. I said that my hunch is that the so-called "QE program" adopted in March (which wasn't QE at all), probably had relatively little to do with the recovery. If I had to guess, I would say the rapid recovery in Asia is probably the biggest factor in the stock market turnaround. And of course the real economy in the US is not recovering at all, which means we need . . . (here's where my "one trick pony" kicks in) more monetary stimulus.

Richard, there is no question that they reduced the mm deliberately. The goal (according to James Hamilton) was to inject liquidity without creating inflation. I'm not sure why they were worried about inflation.

BTW, It's hard (for me) to present complex ideas in a debate, so I added a Cochrane post to my blog in case anyone is interested in what I was trying to say.

http://blogsandwikis.bentley.edu/themoneyillusion/?p=1868

Bill Woolsey writes:

Fundamentalist:

At one time, the Federal Reserve set a discount rate and then created money based on the amount that banks chose to borrow from them. If banks didn't want to borrow, then no money was created.

For the last 50 years, the Federal Reserve has instead used open market operations. They buy finanical assets with money created out of air. When they buy from someone other than a bank, then the amount of money that households have increaces.

"Pushing on a string" does not apply to modern monetary policy.

You apparently have confused money and credit. That is, monetary policy involves increasing lending. In fact, it works increasing the amount of money people actually have beyond the amount they want to hold, and generating more spending.

For it not to work, the amount of money people want to hold must rise to match the increase in the quanity of money. Kling's dimes and nickels story is an example of who this might happen.

When the Fed buys assets, it does create more money without banks choosing to borrow from the Fed or lend to anyone. However, the Fed decreases the quantity of other financial assets on the market. If people are willing to hold money in place of the assets purchased by the Fed can be corrected by a shift of demand to money. The increase in teh amount of money people want to hold matches the increase in the quantity.

Sumner (and my) view is that one the Fed has purchased all the assets it can, which would certainly include all governmnet bonds, and still there is no excess supply of money or increase in spending, then perhaps I will believe these theories.

Gee, base money has doubled and nominal income has fallen. Doubled! Obviously, open market operations don't increase nominal spending. Well, there is no reason to think that it must only double. Perhaps it must triple, quandrouple, or quintiple.

An argument from Greenspan's authority isn't valid.

Artturi Björk writes:

Fundamentalist: Would you agree that if the Fed bought something which is a compliment to money rather than a substitute that it's inflation targeting (or nominal GDP targeting) would work?

If so, then wouldn't it be fair to say that the less of an substitute the stuff that the Fed buys the less it has to buy this stuff to get a certain market reaction, but that getting a nominal GDP target is deffinately not out of the reach of Feds powers.

To me Sumners argument about his policy resulting in a commodity standard, which is independent from price shocks to the said commodities is very compelling and I'm really surprised at how much criticism hes getting from the austrians as to my view this is exactly what their writing calls out for.

fundamentalist writes:

Bill: ""Pushing on a string" does not apply to modern monetary policy."

Greenspan ain't exactly ancient history.

Bill: "In fact, it works increasing the amount of money people actually have beyond the amount they want to hold, and generating more spending."

But people don't always spend the extra money they have on goods/services. They can also use it to buy assets, such as government bonds, which is exactly what they have done. Or they can invest in the stock market, in commodities or other assets.

The idea that the feds can pump money into the economy at will and people receiving the new money have absolutely no other choice but to buy goods is silly! People have other options and they're exercising those options. When people are willing to accept 2-3% interest on government debt while inflation is 1-2%, what can the feds do to persuade people to exchange that debt for cash? Not much!

Bill: "When the Fed buys assets, it does create more money without banks choosing to borrow from the Fed or lend to anyone."

Who is selling assets to the feds these days? Banks! Not citizens. The banks are increasing their reserves in the hope that someone will start borrowing money but so far it ain't happening. The fed can't sell anything to the public because the public wants gov debt more than anything.

Bill: "Well, there is no reason to think that it must only double. Perhaps it must triple, quandrouple, or quintiple."

Or maybe it doesn't work in a depression. I have no doubt that the feds will increase the money supply by a factor of 5 or 10 once the recovery starts. And you and Sumner will credit the feds for goosing the economy, but you will have cause and effect backwards. As the economy recovers, people will feel more like spending and they will take the cash the feds are pushing. Then, as the recovery turns into a boom, the feds will continue to increase the money supply, and ngdp until very high levels of inflation appear. For anyone awake during the past 40 years this has been the typical pattern.

Artturi: "getting a nominal GDP target is deffinately not out of the reach of Feds powers."

I'm not arguing that the fed is always powerless. Like Austrians, I hold to the quantity theory of money, just not the rigid version that monetarists cling to. The feds can increase the money supply most of the time; they just can't do it during a depression such as the one we're in.

The correlation between fed policy and the money supply or prices is tenuous. In fact, one of the fed banks just came out with a study the claim shows that there is no relationship to price inflation and the money supply. I think the study was severely flawed, but it does highlight the fact that the relationship isn't iron clad.

In order for the feds to sell anyone anything, they have to find willing buyers and those are scarce during a depression. Economists like Sumner and Woolsey who see prices take off after years of fed attempts at pumping up the money supply are guilty of post hoc ergo propter hoc thinking.

fundamentalist writes:

PS, the fact that econometric studies of the relationship between the money supply and prices is tenuous at best is an indication that fed policies don't work during a depression. We know they work quite well at other times. It's the aggregation of the data that destroys the correlation. If modelers would use a dummy variable indicating recesssion and recoveries, they would find that the fed is very powerful during a recovery or boom period, and powerless during a depression.

Scott Sumner writes:

Fundamentalist, I'm not sure why you keep ascribing to me views that I do not hold. In my blog I have endlessly argued that monetary aggregates and prices are often NOT closely correlated. For instance money is currently very tight, despite rapid growth in the money supply.

The Fed never has any trouble finding people to sell cash to in a depression.

I don't think there is a serious economist in the world who thinks that monetary policy is ineffective in a depression. Rather some economists argue that conventional monetary policy is ineffective. But does anyone argue that if the BOJ set a fixed exchange rate of 1,000,000 yen to the dollar that nominal GDP in Japan would not rise?

Arnold argues that inflation expectations change slowly. Whether or not this is true depends entirely on the sort of monetary regime in place. I favor a monetary regime where inflation expectations (or more precisely NGDP growth expectations) do not change at all.

Artturi Björk writes:

fundamentalist:"they just can't do it during a depression such as the one we're in."

How come? Do you mean that people are unwilling to sell any kind of assets to the Fed in a depression, or that all assets are perfect substitutes for cash? How about foreign currencies?

Bill Woolsey writes:

Fundamentalist:

Let me try again. Quoting Greenspan proves nothing. Federal Reserve chairmen aways evade responsiblity. To hear them say it, they never do anything wrong, and anything other than what they are doing won't do any good.

If the quantity of money is greater than the demand to hold money, and people "put" the money in stock or commodity markets, then the quantity of money remains higher than the demand for money. For every buyer there is a seller, and the person who sells the stock or else sells the commodities now have the money.

If people don't spend on final goods and services (which would include stocks,) then your argument needs to be that the prices of those assets rise enough so that the demand for money rises to meet the quantity of money.

(If stock prices were to rise, I would consider that a good sign for recovery of nominal income. If commodities futures rise, this creates a no risk opportunity for buying on the spot market, which rises commodity prices. Commodity prices are part of nominal income and so this will raise nominal income directly. And, of course, rising demand also expands the output of commodities and employment in producing commodities.)

Open market operations can increase the quantity of money. You say, the Fed is now buying assets from banks. Actually, they are buying from whomever will sell, but even if it is only banks that were selling during some period of time, once the banks run out of assets, then when the Fed buys assets it must be from other people, and the quantity of money rises.

(Ending interest payments on reserves, or better yet, charging banks to keep funds on deposit at the Fed, will get banks buying assets too, rather than selling them to the Fed.)

When you say that people will not spend the money..yes, that is saying that they will choose to hold it. That is saying that the demand for money will rise more than the increase in the quantity of money.

The Keynesian argument that you are making is that people will spend the money on existing bonds. Bond prices will rise and interest rates fall. And because interest rates are so low, people will hold the additional money.

It is such a laugh that you claim to be an Austrian (or Austrian influenced.) Your arguments are straight Keynesian. Really, post-Keynesian.

Anyway, the liquidity trap argument is especially plausible if nominal interest rates on the bonds purchased by the Fed is zero.

But the notion that more money will exist than people want to hold, and they won't spend, or that they will spend on commodities or even equities and this will leave nominal income below target, are just foolish.

Sumner (and I) advocate that when the demand to hold money falls as the economy recovers, the Fed should reduce the quantity of money again.

The notion that the quantity of money will rise when the economy recovers... well, it could.

But it is important to keep straight a discussion of what should happen vs. what mistakes you expect will happen.

You are claiming that the Fed cannot increase the quantity of money. Wrong. They can. Your version of "pushing on a string" doesn't account for open market operations.

Then you promote the liquidity trap argument, which is that the demand for money will rise to match any increase in the quantity of money. Ths is more plausible. I think it is wrong in a global sense. It may will be that some increases in the quantity of money will be matched by increaes in demand, especially if the Fed is purchasing bonds with zero nominal interest rates. But the open market operations just need to be more agressive. They will have to buy other sorts of assets that don't have zero yields.

Anyway, as I said before, I think you have confused money and credit. When monetary economists write that the quantity of money needs to be higher to get nominal income on target, you read that banks should lend more money so that people will borrow more and spend more.

Well, sorry, but that isn't the argument. And Mises, Hayek, and even Rothbard all know it perfectly well.

Bill Woolsey writes:

Kling:

Sumner argues that the base money should increase enough, whatever amount that is, so that nominal income will return to target. (Actually that its expected value one year from now will be on target.)

Your argument is nothing more than the Fed will have to purchase assets that aren't close substitues for base money. If the Fed starts off buying T-bills, for example, you are saying, that buying all of them will not be enough.

In the extreme, you are saying that the Fed will have to end up buying Baa corporate bonds or something.

So?

Your argument, then, isn't that increases in the quantity of money generated by open market operations cannot increase nominal income to the desired growth path. It is just a claim that the increases in base money must be really large, and that the Fed must buy some assets that are not close subsitutes for money before nominal income will get back to target.

And, perhaps you can argue that this is a bad idea. That having the Fed buy these assets would be more undersirable than just leaving nominal income below target.

But that is the argument you need to make.

Of course, Sumner will respond with the somewhat paradoxical view that if the Fed is committed to do whatever it takes, then the needed increase in base money is much smaller.

The Fed's operating procedure of open market operations with T-bills (and even repurchase agreements) aimed at targeting the Fed funds rate, with the Fed funds rate being adjusted in a discretionary fashion so that prices will be expected to rise from where they are about 2% and no more... has failed.

But the Fed's particular policy framework isn't the same thing as monetary policy.

fundamentalist writes:

Sumner: “I don't think there is a serious economist in the world who thinks that monetary policy is ineffective in a depression.”

Would you count Greenspan as a serious economist? Besides, most economists adhere to the strict, mechanical version of the quantity of money equation. Of course they would believe that monetary policy is always effective at all times. But where is your evidence for it, not your belief? If mainstream economists understood money as Hayek and Mises did they wouldn’t believe such nonsense.

Artturi: “Do you mean that people are unwilling to sell any kind of assets to the Fed in a depression, or that all assets are perfect substitutes for cash?”

None of the above. The Feds can’t sell something to people who don’t want it, just as it’s difficult to sell ice to Eskimos, though not impossible. Today, banks are buying more cash, but they can’t loan it out because few people want to borrow. The Feds could buy gov debt from individuals, but who wants to sell? Interest rates on short term gov debt are close to zero. How is the fed going to persuade them to trade that debt for cash? It would have to offer a higher principle, which would mean a much lower interest.

Bill: “Quoting Greenspan proves nothing.”

Greenspan was at one time considered one of the most successful Fed chairmen in history. His comments on fed impotency doesn’t prove anything, but it lends support.

Bill: “You are claiming that the Fed cannot increase the quantity of money.”

No I am not. I am saying the Feds can’t increase the quantity of money in a depression. You and Sumner believe that the feds’ monetary policy is always strong all of the time. Nothing affects its superpowers, not even krypton. I am making the Austrian argument that the fed is not omnipotent. It occasionally needs Viagra. It does not have the power to control the economy. If it tries to keep a boom going by expanding the money supply it will get hyperinflation, as in Germany in the 1920’s, the third world in the 1970’s and in Zimbabwe today. If it doesn’t try to keep the boom going, a depression will happen. While the depression is on, the feds are powerless. The real economy has to do the heavy lifting. When the depression is over, then the feds can put on their superman tights and take credit for saving the world as they always have

Artturi Björk writes:

fundamentalist:"The Feds can’t sell something to people who don’t want it, just as it’s difficult to sell ice to Eskimos, though not impossible."

"the Feds could buy gov debt from individuals, but who wants to sell?"

hmmm.... First you are comparing cash to ice as to imply that cash has pretty much no value what so ever and thus can't be exchanged to anything. This is clearly false if you look at our everyday lives.

Then you are saying that the Fed can't do anything because gov. dept is a perfect substitute for cash. What if the Fed bought something else than gov. dept? (like wallets and piggybanks... just kiddin...)

If you agree that the Fed could buy something that is not a perfect substitute then don't you agree with Sumner that reaching a nominal GDP target is in the Feds powers?

Bill Woolsey writes:

Fundamentalist:

You are saying that the Fed can't increase the quantity of money _in a depression_.

You are wrong.

The claim that the Fed cannot buy assets from people because they refuse to sell is false. Bond markets are clearing today. This morning. There aren't shortages where people call their brokers, ask to buy T-bills, and the broker responds.. no one is selling, sorry.

You seem to be confusing, "increasing the quantity of money" with "increase spending." The Fed can increase the quantity of money. But, will that increase in the quantity of money result in more spending? Well, yes, if the quantity of money rises more than the demand to hold it. But, maybe the demand to hold money will rise to match the increase in the quantity of money.

And that is the argument you need to make. In a depression, the Fed can increase the quantity of money using open market operations, but the demand to hold money will just rise to match the increase in the quantity of money. The Fed can increase the quantity of money in a depression, but it cannot no matter how much it increases the quantity of money, it cannot cause spending to rise in a depression. In other words, the liquidity trap argument.

fundamentalist writes:

The ice analogy wasn't meant to indicate that money is worthless, but that people have as much as they want, just as Eskimos have as much ice as they want. You might could sell ice to Eskimos in the summer time, though.

"Then you are saying that the Fed can't do anything because gov. dept is a perfect substitute for cash."

That's not what I meant. Short term gov debt can be a substitute for cash, but if you think banks are going to fail then short term gove debt is better than cash.

The feds can't always buy as much as they want. The supply isn't totally elastic. Suppose the feds decided to buy used cars in order to get them off the market and give people cash to buy new ones. Would the feds be able to buy as many cars as they wanted? If they offered a high enough price. For example, if they offered me a million bucks for my 1991 Ford Probe, of course I would sell it.

In other words, if people are holding on to assets instead of cash, that's because the market rate for those assets is lower than what the holder wants for them. The feds would have to offer more than the market rate in order for them to sell. That's true of gov notes and bonds as well as cars.

But since the people holding those assets had as much cash as they wanted before they sold to the feds, what prevents them from just putting the cash into another asset instead of buying consumer goods? Nothing! If people don't want to buy more consumer goods, the feds can't force them to do it.

The main problem with Sumners economics and mainstream economics in general is the believe that people are robots and the economy works mechanically.

Artturi Björk writes:

fundamentalist:"The ice analogy wasn't meant to indicate that money is worthless, but that people have as much as they want"

So you mean that cash is worthless on the margin? This is also clearly false since people accept cash in all sorts of exchanges even in a depression.

"he feds can't always buy as much as they want."

Ofcourse. No one can. However the Fed can spend as much as it want's just like anyone else. For example if I was to buy potatoes and I wanted to buy 5 kg's, but I didn't have to money for it. Well I can buy a smaller amount and still spend all that money.

"The feds would have to offer more than the market rate in order for them to sell."

Well this depends on how you define the market rate. If the market rate is a rate, where supply and demand are in equilibrium, then when Fed increases the demand the market rate must rise so that the demand is met. So you could say that the Fed has to pay more than te previous market rate, but that's true always regardless if we are in a depression or not.

" just putting the cash into another asset instead of buying consumer goods?"

Where do these assets come from? If they come from somewhere else than the Fed then there must be a seller in the economy who's cash balance has increased. The only way that the increased balance doesn't result in nominal GDP growth is that it is a perfect substitute for the stuff Fed exchanged it for.

Comments for this entry have been closed
Return to top