According to Goldman Sachs, we’ll soon need bigger wallets.
I hope someone can telling me what I am missing, because the debate over quantitative easing seems to be taking on an increasingly surreal quality. A recent Goldman Sachs study concludes that it would take another $7 to $11 trillion dollars of asset purchases by the Fed to achieve the necessary 6 percentage point reduction in the real interest rate implied by the Taylor Rule formula. Lucas Critique anyone?
Let’s for the moment assume that the Fed came to their senses and decided to pursue an aggressive policy of quantitative easing (QE.) Does anyone seriously believe that they would continue paying interest on reserves at rates higher than what banks can earn on T-bills? Yet any estimates that GS came up with would almost certainly have been based on empirical data collected under the interest on reserves program—a program explicitly set up to avoid any QE effects at all! (I.e. the program was aimed at keeping the fed funds rate from falling below its target.) In fact, why would a Fed serious about QE allow any sort of massive reserve hoarding at all? My idea of a penalty interest rate on excess reserves has been endlessly discussed here, I’ve thrown it at a lot of famous economists and gotten total silence, so I’m going to assume unless someone tells me otherwise that it will work. I will assume that containing demand for excess reserves is a very solvable problem.
So what about the $7 to $11 trillion in new base money, where will it go? Let’s call it $9 trillion for simplicity. There is only one place it can go—cash held by the public. Does anyone stop to think about how many Federal Reserve Notes that is? We are talking about $30,000 dollars for every man, woman and child in the U.S. Yes, some cash is held overseas, but at least half, and probably much more is held in the U.S. (based on cash to GDP ratios in countries (like Canada) without massive foreign holdings of their currency.) The U.S. doesn’t even print notes with denominations above $100. If we exclude kids, we are talking about nearly $30,000 per capita, even excluding foreign holdings. We’ll need bigger wallets.
OK, there is one good argument against my cry of exasperation. The policy may depress T-bill yields slightly below zero—just enough to make safety deposit boxes profitable. But at some point we need to ask ourselves whether an argument is plausible. Can one actually envision something like this happening, especially to this extent? The total supply of T-bills would be exhausted long before we reached $9 trillion, indeed before we reached $2 trillion. (I don’t know exactly how many are currently held by the public, does anyone else know?) In 1933, despite the fact that nominal rates were near zero and much of our banking system was shutdown, real demand for base money peaked at a level only about 50% above pre-Depression levels. Today, with FDIC, we have only one of the two motives causing higher currency demand in 1933—low interest rates.
To make the sort of argument GS is making, you’d have to make a number of heroic assumptions. (Here I should emphasize that I have only seen the linked summary, so I don’t know the specific assumptions they used, all I know is that I would almost certainly disagree with them.) It seems to me that in addition to the assumption that bank reserves stayed high, or indeed increased even further, one would also have to assume that the policy didn’t have much credibility, that is, that even after several trillion in currency injections the relationship between currency and expected inflation would remain linear. Does it seem plausible that the relationship would be even close to linear once the Fed starts dramatically increasing the base? Especially if accompanied by an explicit nominal target and a commitment to level targeting?
In his blog, Mr. Krugman discussed the GS finding, and used it as evidence to support his skepticism about QE. Since I was a bit sarcastic in my last post on Krugman, I want to emphasize that just because someone links to an article, doesn’t mean one accepts all its conclusions. Indeed Krugman did not have a link, and (like me) may not have read the actual report. Instead, I’d like to make a broader point about the press and the economics profession in general. I simply don’t see a lot of discussion about the most effective way of insuring that any QE doesn’t simply lead to more hoarding. Someone correct me if I am wrong, but I would have thought that ideas like a penalty rate on excess reserves might have created more of a buzz by now. After all, this blog has received some attention in recent weeks. And even if my ideas don’t pan out, where are the other proposals? Other than James Hamilton’s blog, and Nick Rowe’s blog, I simply haven’t seen much discussion of possible solutions to the liquidity problem. But please let me know if other blogs you run across are coming up with creative ideas that I have missed.
BTW: In a comment I added to Krugman’s post, I mentioned two problems with the view that QE has recently been found to be relatively ineffective. One was the interest on reserves issue discussed above. The other I just noticed recently; according to the St. Louis Fed the base fell from 1770b in early January to 1547b in early March. To me that data looks a lot like accommodation, not QE. The whole point of QE is to put more base money in circulation than the public wants to hold. Again, please correct me if I am misinterpreting the monetary base data.
FDR was able to do it, so did the central bank of Zimbabwe, are we to believe that we cannot find creative ways to make our fiat currency worth less (that’s two words!), without imposing unreasonable risk of capital losses on the Fed? And let’s not forget that while much of the economics profession has already written off monetary policy, within the Fed they are still trying to decide whether to actually engage in QE, beyond what was required for liquidity purposes in the banking crisis.
(Thanks to Dilip for finding the two articles linked here.)
Update: Bill Woolsey pointed out that as written, I seem to endorse GS’s assumption that we need 6% expected inflation. I do not. Instead, I think 6% nominal growth next year, falling to 4.5-5% thereafter would be fine. If we got 6% nominal growth, we would escape the liquidity trap, and probably end up with roughly 4% real growth and 2% inflation. So please understand that I was just accepting the GS inflation number for the sake of argument. My focus was on their estimate for the base increase required to generate what they saw as needed inflation. I didn’t even mention that the Taylor Rule is normally applied in a backward—looking fashion, whereas I think that monetary policy must be forward-looking.
Tags:
11. March 2009 at 14:58
I think you have a right to be frustrated. I was concerned about the Fed paying interest on deposits back when they started but didn’t have anywhere near as sophisticated an argument. Thanks for the posts.
11. March 2009 at 15:18
I thought the reason the Fed is giving banks interest on reserves is simply a back door way to boost their profits and hence capital. It is a convenient way to bail out the banks without the political calculus of bailing out the banks — though I would think it will take far too long.
11. March 2009 at 17:09
libfree, Thanks for the comment.
Frank, Even if you’re right there is no reason to pay interest at the margin. The Fed should decide how big they want bank reserves to be for liquidity reasons (assume its 20% of deposits for the sake of argument.) Then they can pay all the interest they want on the first 20% of deposits held as reserves. After that have a stiff penalty to discourage excess reserve holdings. That effectively caps reserve demand, and still allows a huge subsidy (putting aside the issue of whether that’s the best way to subsidize banks.) Once capped, it gives monetary policy traction as extra base money is injected through OMOs. Banks won’t want to hold it, and it will be pushed out into cash in circulation where it might do some good.
11. March 2009 at 18:55
I do find it very strange that other leading economists that blog have not discussed the idea of penalty interest rates on excess reserves. I am not sufficiently knowledgeable to state an opinion as to whether it would work or not, but it seems very worthy of discussion.
John Hempton seems responsive to challenging ideas and seems to have an audience. If you were so inclined, perhaps you could email him to request his opinion (or other economists with blogs, whose ear you seem to have)?
Thank you for another excellent post. Thanks, also, for your prior response to an earlier comment. Regards, Jonathan.
11. March 2009 at 19:05
I emailed John Hempton as to whether he had an opinion regarding penalty interest rates on excess reserves, so we will see if anything comes of it. Regards, Jonathan.
11. March 2009 at 19:50
[…] as many do, that the penny is currently worthless), but nowhere near the $9 trillion that this Scott Sumner-linked article on an Andrew Tilton study indicates may be necessary. If the Goldman Sachs […]
12. March 2009 at 04:33
Jonathan, Thanks for your thoughtful contributions here. Let me know if John Hempton responds.
12. March 2009 at 06:02
[…] est intéressant de lire la réponse de Scott Sumner qui reproche à cette étude d’aller un peu vite en besogne. Il se plaint que l’un de […]
12. March 2009 at 07:35
You mention Zimbawe, but is Zimbawe increasing M while B is decreasing? No. The Zimbawe treasury is running huge deficits, increasing B. Since that debt cannot be repaid, the Zimbawe central bank is printing money like crazy and buy the extra bonds. What I say is that if you want inflation you have to increase M+B, Changing M for B does not have the inflationary effect you are looking for, it is like changing 5 $20 dollar bills for 1 $100 dollar bill. Two ways to do this: Run a deficit (we are on our way) and monetize the debt, print money and do helicopter drops (not open market operations). In any case we also have to promise not increase future taxes to withdraw the extra money or to repay the extra debt. Now… are you sure this is what you want? Do we really want to become Latin America? The US has a Great Depression once every 80 years, Argentina has one every decade…
12. March 2009 at 07:46
Scott,
I think that many economists believe that the existing monetary base, even counting .25 interest on reserve balances, is enough to expand the money supply and nominal income more than enough for it to return to the growth path of last year. In fact, high inflation is on the way.
Quantitative easing? Penalty rates on excess reserves? That will just exacerbate the problem. Doing more now will not impact anything in the near future. What will happen in the Spring and Summer is already determined. Further actiona now is just going to create over expansion next year.
It comes down to the long and variable lags.
Then there are the people who don’t believe that nominal income is ever a problem and that all of the real problems are about problems with sectoral shifts, reduced productivity, more realitic views about what consumption can be maintained, and the like. Worse, they think that any effort to increase nominal demand will interfer with proper adustment to the real shocks.
And, then, there are the liquidity trap people who agree that low nominal income is a problem, and that is where you a getting most of your traction. Frankly, I think many of them favor increased government spending to solve social problems and you cannot let a crisis go to waste.
Of course, doesn’t Barro advocate solving the problem by lowering marginal tax rates? What is that about? Think about it. Hey, I think lower marginal tax rates (and a smaller size of government) are a good idea too.
12. March 2009 at 07:57
http://blog.atimes.net/?p=722
This is why we cannot nationalize the banks – why we won’t. and why we must resist the web based deflationary bank-nationalization lynch mob – such as Krugman, Roubini, Naked Capitalism, and BaseLine.
what to do with the senior debt – or even the preferred or unsecured debt.
1. Guarantee all of it? Congress would never allow it. There would be a populist revolution – lead by the crazed deflationists.
2. or – do a debt cramdown – like the totally mad and raving Buiter wants. In which we get a run on the world wide insurance industry – and all assets go very promptly to zero.
This is why we must have an inflation – a moderate controlled inflation – in a perfect world lead and managed by Volcker – who would replace the child currently running the Treasury.
12. March 2009 at 08:10
Scott:
Don’t try to anticipate all possible objections to charging for excess reserves and come up with ways they can accomplish some kind of foolish goal they might have.
For example, if the Fed is using interest on reserves to subsidize banks, then they should go to jail.
Getting banks to hold more reserves so that they are more liquid is insane.
Teh reality is that the Fed has long wanted to pay banks for reserve balances so that banks would no longer have an incentive to report the true amount of transactions balances. And then, they actually started doing it to allow them to direct lending without increasing the money supply and nominal income. As said, they wanted to increase lending to banks while keeping the Federal funds rate from falling.
Also, you need to take into account the impact of the penalities on reserves to the willingness of banks to pay interest on deposits. Even if they must pay on their required reserves, it will reduce the equilibrium interest rates they pay on trasactions accounts. Maybe below zero.
12. March 2009 at 08:16
Scott,
You are falling into your “inflationist mode.”
You advocate raising the inflation rate so that the real interest rate on T-bills (at a zero nominal interest rate) is -6%?
12. March 2009 at 09:40
Alex, You say we need to increase M+B. Most people who say that (like Krugman) define B as T-bills. What’s your definition? Would you include T-notes, T-bonds, Foreign government bonds, agency debt, Aaa bonds, Baa bonds, stock index funds? Tell me exactly what you put into B, and I’ll still favor monetary policy–advocating the Fed buy the safest assets outside your “B category” and still avoid massive deficits. Of course they might lose some money when they sell off the assets later to prevent high inflation. But wouldn’t you rather have a 50-50 chance of the Fed breaking even, than have the fiscal authorities certain to run up a trillion dollars in future tax liabilities? And BTW, the risk is less than people think because the $9 trillion figure is absurd.
Bill, I agree that “many economists” are worried about high inflation, but I think that’s because they look at their old textbooks on long and variable lags, rather than market indicators like indexed/nominal yield spreads. Japan is entering it 15th year of almost continuous deflation of the GDP deflator. Many years ago they rasied their monetary base sharply, then more recently pulled the money back out of circulation. I’d like to know what these economists who expect high inflation have to say about the Japanese case.
I agree with Barro’s idea, but Obama won’t do it (unfortunately.) In contrast, I still do hope to help persuade the Fed to change its views. (Although I won’t be able to unless other economists pick up on the ideas in this blog, and similar blogs by a few others.)
On your second post, I agree that the argument that we need interest on reserves to boost profits, or high levels of reserves for liquidity, are bogus. But I continue to make these points to show people that there are NO GOOD ARGUMENTS that I have heard against my idea. Maybe they are out there, but I haven’t seen them yet.
Finally, I don’t favor 6% inflation. I meant that that was what GS thought was necessary. My point was that their estimate about the money creation necessary to hit that target was farfetched (unless the Fed was really foolish, and did nothing to discourage reserve build-up during their QE program.) I’ll add a footnote.
12. March 2009 at 11:14
You mentioned here and many other places the issues associated with paying interest on excess reserves. Is there a place somewhere I can go to find out the quantity of reserves being kept in excess, drawing interests. I have looked around some and am unsure how to identify this value. Thanks for your help.
rb
12. March 2009 at 12:02
ssmuner,
In B I include all types government bonds. Changing M for B affects the composition of nominal liabilities, this can have some effects on the economy. But if you want to generate inflation you have to increase the total stock of nominal liabilities, not the composition.
12. March 2009 at 12:17
rb, If you google monetary base data, the St Louis Fed should pop up, and they have lots of data. If you cannot find excess reserves, you can get a rough idea from total reserves. Right now roughly 90% of total reserves are excess.
Alex, When the economy is not in a liquidity trap, then swapping base money for bonds definitely creates inflation. The assets are not at all close substitutes during normal times, as banks don’t like to hold excess reserves, and people don’t spend T-bills at stores. Right now your point is debatable. But even Krugman would concede that swapping money for government bonds in a liquidity trap would be inflationary if the policy were credible (i.e. likely to persist.) So it’s way too simple to say monetary policy won’t work because swapping money for government debt doesn’t work. But even if this were so, I would have the Fed do OMOs with triple A bonds to generate a bit of inflation, I would not waste time and money on fiscal policy. And by the way, I do not think that you are correct in assuming that buying long term T-bonds, or Treasury indexed bonds, or foreign government bonds would have no impact on AD. I know of no example in all of world history of a central bank that tried QE of a fiat currency and failed. Japan never tried, we haven’t tried (as I explain above.) Where is the liquidity “trap” that everyone has in their minds?
12. March 2009 at 13:24
ssummner,
In a frictionless economy swapping M and B has no real effects (this is just an application of Modigliani-Miller to the government instead of a firm). Of course the economy is not frictionless (one friction is that the government can borrow at lower rates than the public for example), so swapping M and B can have real effects (magnitudes to be determined). Why? Well think of the intertemporal trade off. Increase M lower B and lower R, consumers and firms want to spend/invest more today and demand goes up this puts some inflationary pressure on the economy. But eventually bond yields go to zero and then holding money dominates holding bonds and you enter a “liquidity trap”. Once you are in this situation the swapping will have no effect but you can then increase both M and B without increasing taxes. This last step is crucial, since if you increase taxes to pay for the debt then people will use the extra nominal resources to pay for the taxes and not to demand goods and services. This should do the trick. Notice that with the helicopter drop money infusion you give money without buying bonds. I mention the case of issuing bonds and then buying them back with newly printed money because the Fed is not allowed to perform helicopter drops.
12. March 2009 at 13:38
Alex,
I money is simply an asset that people hold as a store of weatlh, then perhaps what you say is correct.
But money is the medium of exchange.
The Fed doesn’t have to “sell” money to people who are willing to hold it. They buy things from people who don’t necessarily want to hold the money.
People who want to hold more money don’t need to buy the money from someone willing to sell. They just cut back on their expenditures.
I presume you have lived in a monetary economy and experienced the flow of income through a cash balance. Think about the significance of this.
12. March 2009 at 19:06
Scott,
Do you care where all this money goes? I no you are not an Austrian, but surely there would be a risk of expanding demand in sectors of the economy where it would be unsustainable even in the medium term, leading us back into another bust or downturn.
More broadly, I don’t understand why you fear deflation so much. Isn’t a downward adjustment of prices (and especially a fall in the prices of some assets relative to others) what we need to get out of this mess? Looking at things in a large-scale comparative context, economies in the 20th have been harmed far more by high levels of inflation than deflation (N.B., I do realize that you are looking at the U.S. in the 30s and Japan in the 90s as the your (ominous) parallels).
13. March 2009 at 02:13
Craig,
Why does demand fall?
Not why does demand shift between one good or another, but why does demand fall overall?
What determines interest rates? Not how does monetary policy distort interest rates. What determines interest rates undestorted by monetary policy? Do the factors inolved in the downturn impact any of things things?
How exactly does a lower price level return the economy to equilibrium? Why is the pattern of demands generated by a lower price level result in something especially sustainable.
Fortunately, it is unnecessary to choose between the harms of inflation or defllation (though I do think the deflation of the thirties had horrible consequneces.) Have neither.
13. March 2009 at 03:05
I think QE should atleast seek to address some other inequalities. Give every citizen a credit card with a payment period of 2 years (a moderate idea of by when the crisis might be over).
To prevent moral hazard, let those who use this facility lose their voting rights for a certain period of time. Automatically, the only ones left with voting rights in th next election will be the prudent and policy that will be made will be one that is fiscally prudent and will avoid the temptation to go to the printing press one more time.
13. March 2009 at 05:53
Alex, The point you raise is so important I need to discuss it in a post. But briefly, you misinterpret the implication of the M-M theorem. The assumption is not just a zero interest rate environment now, but from now to the end of time. Nobody believes that, which is why Krugman developed his expectations trap argument, the assumption that the monetary increase was temporary, and would be removed if you exited the liquidity trap. Otherwise a permanent M increase will be expected to be effective when the liquidity trap is over, and will then be expected to raise prices when the liquidity trap is over, and thus will lower real interest rates today. In other words, Krugman is basically using a Ricardian Equivilence-type argument against QE. OK, but does that also apply to fiscal policy? It applies more strongly to fiscal policy, as deficit stimulus MUST be temporary, whereas monetary stimulus CAN be permanent.
Craig, It’s not so much deflation that scares me, it is falling nominal GDP. Why, because I think it leads to socialism, as it gets wrongly interpreted as a failure of capitalism. Have you noticed that even right-wingers like Posner seem to simply assume the current crisis is a failure of capitalism? That it is caused by financial turmoil, not monetary policy. (Actually I don’t know that this is his view, but the title of his forthcoming book strongly suggests it is.) We know that falling NGDP led to statist policies in the U.S. (1930s), Argentina (recently), and too a lesser extent Japan (big spending on pork projects.)
Your worry about unsustainable sectors implicitly assumes that the current crisis of falling NGDP was caused by the housing bubble. But it was not. The housing sector has been falling sharply since mid-2006. For two years real GDP kept growing, as labor redirected from housing construction into exports and services. Only after mid-2008 when monetary policy started allowing NGDP to plummet, did mass unemployment develop. I am not advocating some radical new policy, but the same old 5% NGDP growth that worked so well in the past, until we stopped doing it.
I am very strongly opposed to the sort of high inflation we had in the 1970s, but let me add that deflation did probably 10 times as much damage as inflation in this century. The Great Depression imposed human suffering far beyond anything in the post war period. In Germany the suffering was so great that the population voted for extremist parties out of desperation. BTW: Some on the right have created a myth that hyperinflation led to the rise of extremism in Germany. But in mid-1929, six years after the hyperinflation, the Nazi’s were still a tiny party. 4 years later they were in power.
I basically agree with Bill’s comments, regarding Prakesh, there is just no chance that your plan would be enacted, and I still think that monetary policy is the best way out of the mess.
13. March 2009 at 06:50
Bill: “People who want to hold more money don’t need to buy the money from someone willing to sell. They just cut back on their expenditures.”
Cutting back on expenditures means a fall in the AD which under the assumption of an upward slopping supply curve will give you a fall in output. You avoid this by increasing the money supply. But notice that the increase in the money demand is not driven by a desire to hold more cash and less bonds. It is driven by people wanting to hold more government assets, so again substituting M and B does not help. You need more of both.
ssumner: “Nobody believes that, which is why Krugman developed his expectations trap argument, the assumption that the monetary increase was temporary, and would be removed if you exited the liquidity trap.”
I actually agree with Krugman, you need the increase of money (or M+B I should say) to be permanent. This is why I say that you have to make sure that you do not increase taxes in the future to pay off the debt or retire the additional money supply.
I repeat, I don’t advocate this policies, I just say if you want inflation then it is very easy to do it. But you have to live with the consequences afterwards. There is no such thing as a free lunch. You want inflation, you will get inflation. We know that in the end we cannot play games like “we have to make people believe that there will be inflation but then once they do it we then undo what we did and do not deliver the expected inflation”. Notice that if you can actually do that you would actually plant the seeds for another recession like in 1982. In any case we are talking about a once and for all increase in the nominal liabilities of the government, not an increase in their growth rate. This should not increase the long run inflation rate. My concern is that it is not very easy to convince the market that this is a once an for all thing. Once you open the Pandora’s Box it is very hard to close it.
13. March 2009 at 07:25
Terrific comments in here. Thank you all. I did receive a response from John Hempton (in response to a request for his views on charging interest on excess reserves):-
“We live in a world where governments confiscate banks (WaMu) or dilute the shareholders out of existence if there is a perception of capital inadequacy.
That encourages excess reserves.
Then you want to tax those reserves?
Would be a nice idea if you could agree not to confiscate banks on the other side –
Otherwise it looks like cruel and unusual punishment.”
Hope this helps in some way. Regards, Jonathan.
13. March 2009 at 17:54
Alex, I am not sure you quite get Krugman’s point. He is not saying you need a permanent increase in M+B, he is saying that in a liquidity trap all you need to do is permanently increase M. And I think that what you need is just 2% inflation, you don’t need some high rate of inflation to recover from the recession, 2% would do the job. Then when your out of the recession, keep on with 2% inflation. So there is no credibility problem about whether markets believe you will cut back on inflation later.
Thanks Jonathan, Unfortunately it is hard to convey a complex idea to someone with limited time. If we sat down and explained to Hempton that it could be done without hurting bank profits, it might have helped. But based on his response I doubt it. If he was interested he would have quickly seen that option on his own.
14. March 2009 at 06:03
Hempton’s response was a bit snarky.
It would have been nice if he would have explained why the threat of taking over banks whose regulatory capital fails to meet requirements results in an increase in the demand to hold excess reserves. I presume it has to do with the “risk adjustment” made for risk adjusted capital requirements.
I don’t really care about bank profits, but if banks really feel a need to hold reserve balances at the Federal Reserve and so must actually pay, then they should charge depositors for holding funds in FDIC insured checking and savings accounts. It is hoarding of FDIC insured deposits that is causing the reduction in velocity at least as much as the hoarding of reserves by banks is keeping the quantity of money from rising enough to offet that decline in velocity.
Perhaps banks are holding large amounts of reserves now because of the nature of regulations of capital requirements. If so, then we are seeing how capital requirement regulations are having the unintended consequence of turning past bad investments by banks into monetary disequilibrium and a drop in nominal income.
The managers and stockholders nad junior debt holders of financial institutions that lent into a speculative bubble, and financed their activities at least partly with government guaranteed debt have no room to complain about injustice. Their entire business model is about shifting risk to the taxpayer.
15. March 2009 at 05:54
Bill I agree with you on the fairness issue.
Your point about the interest subsidy indirectly causing hoarding of FDIC insured deposits is a good one. I hadn’t thought of that before.
I would guess that any capital requirement could be met by short term T-bills (someone tell me if I am wrong here.) But because they now earn interest on reserves at a rate higher than on T-bills, banks may prefer to meet these capital requirements with reserves. If so, just another example of how the interest payment on reserves was probably a big mistake.