Stories

The problem with academic macroeconomists is that they’re terrible at telling stories. We say recessions are caused by real and nominal shocks and people snore. And monetary policy is too boring to be the best aggregate demand policy.

“You mean to tell me that buying and selling some debt instruments is the key to solving all of our problems. What, no knight in shining armor riding down from Capital Hill to save us? No brilliant grand plans from genius economic advisors in the White House?”

When I taught the Great Depression this summer, you could see the disappointment on my students’ face when I told them that bad monetary policy was the primary cause of its depth and breadth. They wanted epic stories about Wall Street versus Main Street, evil cabals of foreigners or bumbling Presidents. A student dropped the course when he realized the monetary policy explanation didn’t involve a conspiracy of bankers but just the Fed’s legitimate misunderstanding of the role of money in the economy. He told me “this class isn’t teaching me real economic history”.

Arnold Kling says recessions are caused by Great Recalculations, credit cycles and monetary fluctuations. He talks of great showdowns like “folk-Minsky-ism” v. “folk-Keynesian-ism” or academics v. policy makers. He tells good stories.

As far as I can tell, though, his stories map completely onto what he calls the scholarly consensus in monetary economics. Recessions, including this last one, are caused by real and nominal shocks.

47 Responses to “Stories”

  • swong says:

    But bankers and traders do conspire. It’s just millions of tiny conflicting conspiracies instead of one giant one.

  • gabe says:

    Prescott tells plenty of tall tales….

    I don’t see anything in the Great Recalculation that’s in the mainstream, as it’s more of an Austrian theory. I just don’t see big sectoral shifts towards sectors like housing require no increases in unemployment while shifts out of housing requires major unemployment. The US economy is very flexible and can respond to sectoral shifts easily. The other Kling stuff looks pretty good.

    I don’t think that people like monetary policy is “boring”. It’s almost always better than fiscal policy, just not now and I think most people agree on that. Sumner has a lot of good ideas, but I still think that if monetary policy was going to work, it would have by now. You can only hide behind the fed paying interest on reserves so much, especially because there’s no way to see how effective monetary policy would be absent the fed paying interest on reserves. Clamoring for Bernanke to create more excess reserves won’t do much.

    Since we want to test whether the fiscal stimuls works, how can we test if monetary policy is working now?

  • pushmedia1 says:

    swong, I think the student was thinking of smoke filled rooms, etc.

    gabe, why do you suppose the consensus is that money policy works? Evidence of such perhaps (e.g. the Monetary History, the experience of the Great Depression, the Volker disinflation, etc)? There is limited evidence that fiscal policy works. WWII is about it and the evidence on that is pretty thin. Even in theory, to get fiscal policy to work you have to assume “this time it’s different”. But I’m not really interested in that fight (again… again).

    The Great Recalculation is just a asynchronous technology shock in a two sector economy with adjustment costs thrown in. Also, problems with “systematic risk” in the finance sector can be seen as a real shock. The interesting thing about that sort of real shock is that policy may not be ineffective. The Fed, as lender of last resort, might be able to mitigate some of the shock.

    One take on the failure of the Fed last Fall was that they took their eye off the money supply to act as lender of last resort. The money supply contracted spectacularly relative to velocity while they were busy putting up TARPs to protect the banks. While they were busy fighting the real shock a monetary shock snuck past ‘em.

  • Stories are relevant only for those concerned with doing something more than advance one’s (and hopefully others’) understanding.

    Even for topics for which our understanding is better than for macro issues, the rights answers are not what people (in power?) want to hear anyway.

  • gabe says:

    “The Great Recalculation is just a asynchronous technology shock in a two sector economy with adjustment costs thrown in.” So is the idea that the positive shocks in good times mean low unemployment, but negative productivity shocks cause unemployment during the bust? Do the adjustment costs increase during the bust? I don’t think I really get it. Also, even if this is reasonable, I don’t see it being a mainstream theory.

    Also, is there any evidence that exogeneous productivity shocks drive this recession? Housing bubbles that lead to financial crises are not “exogeneous”, that’s a very endogeneous cause. Many money shocks are not exogeneous, they’re driven by the fed, etc. Naturally exogeneous shocks are much easier to deal with, but that might not explain everything.

    When did the Fed take it’s eye off of money supply (what time period)? Just look at this graph: http://research.stlouisfed.org/fred2/series/BOGAMBNS?cid=124 Is this not the right graph to look at? What do you mean by money supply, M0, M1, M2?

    The Fed only controls the base directly, the rest of the money supply is determined mostly by private agents. The Fed buys assets for reserves and hopes that the banks lend them out and reduce nominal interest rates. If M2 shrinks because demand for reserves increases and the Fed increases the base rapidly, then that’s hardly the Fed’s fault.

  • ssendam says:

    Stories are how we understand causal mechanisms in the real world. Suppose you got the flu and ran a fever – you could say you experienced a “health shock”, which is perfectly consistent with a model where “health shocks” arrive randomly. But if you want to understand what caused it, how you might prevent it or cure it, you have to dig down in to the actual mechanisms. Then you might say you were exposed to a virus which replicated in your body, causing an immune system reaction – but what is that but a story? If we want to explain what caused the recession, why can’t economists trace out the causal chain of events in terms of the things we all do and experience every day – in terms of decisions made to borrow or lend, spend or invest, etc…

    As for the money supply: strangely enough, once upon a time everyone thought the banks _were_ the money supply (which is why every undergrad macro textbook has that whole “inverse of reserve ratio = money multiplier” thing). By keeping the banks going, the Fed _was_ preventing monetary contraction. Somehow we got to the point where the banks disappeared from the theory and the Fed alone determines the money supply…

  • pushmedia1 says:

    ssendam, its at the level of “exposed to a virus” where shocks enter macro models so by your story macro models are telling a story. But I agree “a good story” should be one of the selection criterion for picking models, if only because policy makers insist on it.

    It may be that keeping the banks going prevented a further contraction in the money supply, but there was a pretty severe one in any case.

    gabe, look at what I wrote… I said money supply *relative* to velocity. Read Sumner. He posted a graph the other day showing M-something along with velocity. I understand you believe monetary policy is out of ammunition. I disagree.

    And everything is endogenous so saying {x | x in the set of everything} is endogenous doesn’t really help. The big bang caused the crisis and that was as exogenous as you can get.

    The point is there’s something true about last Fall that wasn’t true about the other 9 months of the recession up to September. There was a depression within a recession. What caused it? Was it Lehman? Maybe, but I don’t think we should accept that narrative just because its the ones the newspapers wrote about.

    Also, suppose you took a hike up the back side of half dome. Once you get to the top you just keep walking off the face. You won’t go splat because gravity isn’t asynchronous?

  • Gabe wrote: “Sumner has a lot of good ideas, but I still think that if monetary policy was going to work, it would have by now.”

    I would have worked by now IF the policy was expansionary. Sumner’s argument is that monetary policy is/was tight! If so, of course we would not have seen it.

  • gabe says:

    Push- I don’t understand what you mean by money supply relative to velocity. I try to read sumner regularly, but I don’t recall that post. Is it “Good monetarism, bad monetarism”?. In any case I’d love to read it.

    M1 velocity is m1/m0, so if you increase m0 without increasing m1, then it just shows that there’s a liquidity trap. The Fed is expanding the base while broader aggregates (money supply) stay relatively constant. This is the main problem with monetarists, which Sumner seems to recognize but still assume, that velocity is stable. It’s not, just look at FRED data m1 velocity. It plummets. Without stable velocity, the Federal reserves control of the money supply gets much more tenuous.

    If Banks get more reserves (part of M0) and then just put the money back at the fed, m1 will be unchanged and velocity will plummet. If anything that just shows evidence of a liquidty trap. Normally pushmedia is very data based, but I don’t see much data here. What could the Fed have done, just increase reserves more?

    Everday- Just look at the graph I posted. Are you saying that’s not an expansionary policy? Sumner’s thesis rests on 1 major assumpitons. 1) Monetary policy works always and everywhere. This means that is there is a decrease in AD/NGDP, then it must be that Federal reserve policy is contractionary. However, if there is a liquidity trap, which we can see beginning in September 2009 with the big increases in excess reserves, then the Fed can do all it wants, and AD/NGDP will shrink. Sumner doesn’t believe in a liquidity trap, but he still needs to deal with this critique.

    From September 2008 to November 2009 Base goes from 900 billion to 2000 billion, more than doubling. Is that not expansionary? Shouldn’t NGDP have doubled too? It’s been over a year, and NGDP is barely changed from Sept. 2008.

  • pushmedia1 says:

    By “critique”, do you mean explain the doubling of reserves if there’s no liquidity trap?

    Interest on reserves, no?

    Regarding velocity: suppose NGDP was your measure of monetary tightness. If velocity goes down 50% but money only increases 40% then that’s contractionary policy. I can’t find that graph… I guess I imagined it.

  • gabe says:

    push-interest on reserves will definitely make excess reserves higher, but does it explain everything? There were excess reserves during the Great Depression as well (The amount of required reserves was raised in 1937 to “address” these excess reserves which worsened the monetary contraction). But there was no interest paid on reserves during the Great Depression. Japan also experienced significant excess reserves in the late 1990s/early 2000s while not paying interest on reserves. So I hardly think paying interest on reserves is sufficient to create a situation where banks hold a lot of excess reserves. Some other factor is causing a liquiditiy trap. (Bad Loans, uncertainty, lack of demand for loans, etc.)

    Regarding velocity: NGDP is not controlled by the Fed, only reserves are. Talking about money in the abstract obscures a lot of the money creation mechanism. NGDP can only be your measure of monetary tightness if you’re not in a liquidity trap. If M1 (or the broadest money supply corresponding to NGDP,PY=MV=M_broadest) will not changed, then the Fed can expand base as much as they want and the velocity will plummet accordingly. (which is exactly what is happening).

    V=M1/M0, so if M1 fixed, then velocity will plummet as M0 increases. I don’t understand why it’s assumed the Fed controls NGDP, it’s clear that unless they can force banks to loan, then any policy they do will seem contractionary. If you’d like the Fed to increase reserves by 2 trillion dollars, that’s fine, but that will mean close to 2 trillion dollars in excess reserves. I’m not seeing a mechanism or the Fed to increase NGDP. Just saying the Fed should do more isn’t informative. Perhaps if the Fed had targeted the forecast things would be different, but I fail to see many concrete actions the Fed could take today that would stimulate the economy. Announcing an higher inflation target would be nice, but I don’t know how credible it would be, given that the Fed is already talking about an exit strategy. This is a terrible signal to send, as it lowers expectations of the Fed’s commitment to monetary expansion.

  • pushmedia1 says:

    Were we in a liquidity trap during the great depression? If a liquidity trap means money policy is ineffective, then the answer is “no we weren’t”. If a liquidity trap doesn’t mean money policy is ineffective, then who cares… i.e. the great depression is not the evidence for ineffectiveness of money policy you were looking for.

    Monetary policy affects NGDP because it controls expectations of inflation. If the Fed announced that it will print enough money to buy every office building in America starting next year, would NGDP increase today?

    The only possible criticism that I can think of is that the Fed’s announcement would be incredible.

  • gabe says:

    During the Great Depression there was a liquidity trap in the sense that open market operations and changes in base would not have been effective. The repegging of the dollar to gold in 1933-1934 is a credible way out a liquidity trap, because it raises prices of importable goods and lowers real interest rates. Also, to defend the new depreciated gold:dollar rate, in the long run PPP holds so price levels must rise. See Svensson 1992 JEP Optimal escape from a liquidity trap. Obviously there are ways out of a liquidity trap or else the USA would have never recovered, but it wasn’t through a traditional monetary policy of increasing bank reserves.

    “If the Fed announced that it will print enough money to buy every office building in America starting next year, would NGDP increase today?” How much is every office building worth? The Fed purchased 1 trillion dollars worth of assets from Sept 2008 to today. If these purchases lead to purely excess reserves, then why would this affect the price level? I’ll be bold- no amount of asset purchases in September 2008 would have been enough to avoid a serious recession like we saw from 2008-2009.

    Basically the question to be asked about the Fed’s credibility is, even if the Fed really wants a high rate of inflation in the future, if it can’t affect monetary aggregates or NGDP almost at all as we’ve seen over the past year, they why would the Fed’s commitment be credible in the future? It’s less a question of the Federal Reserve’s commitment to a goal as their ability to achieve that goal.

  • Gabe,

    Recall the equation of exchange:

    mBV = Py

    where m is the money multiplier, B is base money, and Py is nominal income.

    Do velocity and the money multiplier not matter?! I have shown on my blog that the money multiplier fell from 1.6 to .93 during the period in which the base doubled. That means that for the money supply (M1) to remain CONSTANT, the base would have to increase by over 70%. On top of that velocity declined drastically. Thus, even if the Fed increased the based to perfectly offset the declines in velocity and the money multiplier, such a policy would only be sufficient to keep nominal income constant.

    See my recent post:

    http://everydayecon.wordpress.com/2009/12/02/can-money-be-tight-if-the-fed-does-nothing/

  • pushmedia1 says:

    Oooo…. oooo…. EE has the graph I was talking about in the post he links to.

  • gabe says:

    I see the graph now, that is a good graph. Velocity and the money multiplier do matter very much. If they fall while the Fed is expanding the money base, then monetary policy will not be effective. Look at how base and the multiplier move in opposite directions almost in lock step. Isn’t this what a liquidity trap would look like? M1 as of October 2009 is 1660 billion, while Base is almost 2000 billion. That’s an odd situation to say the least.

    If you define tight money as one where a 5% NGDP target is not met, then the Fed was tight after September 2008, but I fail to see what could have been done to not be “tight”. I guess loosening earlier in 2008 would have helped before things got so bad (or handling Lehman Brothers better as a forced sale rather than a bankrupcy)

  • pushmedia1 says:

    Yeah, at least on that graph a liquidity trap and a Sumneristic take on the crisis are observationally equivalent. You have to look elsewhere to convince yourself that a liquidity trap/ monetary policy ineffectiveness isn’t likely (e.g. a model with forward looking agents and a credible commitment by the Fed to inflate in the future, the Great Depression, etc).

  • pushmedia1 says:

    Is there an equivalent graph for the Great Depression?

  • Gabe,

    One more point. I think that the difference is that people who believe in what is basically old monetarism (not the caricature of monetarism talked about in the current literature) recognize that money demanded should be equal to money supplied. In other words, changes in velocity should be offset by changes in the money supply. Taken one step further, changes in the money multiplier should be offset by changes in the monetary base. Now, of course, offsetting changes in velocity and the money multiplier is very difficult in practice. Thus, an alternative is to stabilize nominal income.

    In other words, stabilizing nominal income is NOT the objective, but rather the GOAL. By stabilizing nominal income the central bank is ensuring that changes in velocity and the money multiplier are being offset by the correct changes in base money — this is the objective of the policy. However, given the informational limitations, the goal of policy is to stabilizing nominal income which enables the central bank to achieve its objective. Finally, if nominal income is not stabilized, then policy is seen as tight because money is either too loose or too tight in terms of equating the money supply with money demand. The 5% number is not special, but rather summarizes the behavior of nominal income growth during the Great Moderation.

    BTW, I just presented a paper at the SEA meetings that showed that the Fed behaved as though they were targeting nominal income growth during the Great Moderation. Further, the Fed’s responsiveness to nominal income growth was stronger during the Great Moderation than the period 1966-1979. This difference can explain why inflation rose significantly during the late 60s and 70s and explain reductions in the volatility of output and inflation observed in the latter period (as evident from simulations).

  • I should really proofread. The sentence beginning “Finally…” should read:

    “Finally, if nominal income is not stabilized, then policy is either too loose or too tight in terms of equating the money supply with money demand.

  • I haven’t uploaded it, but I could email it to you.

  • gabe says:

    EE: This paper seems interesting. This idea of targeting income is, for me, a good policy. But, looking at negative nominal income growth in several periods last year despite huge increases in base, the efficacy of this nominal incomes policy has to be reconsidered in the current crisis in my opinion. A Hicksian liquidity trap would have infinite money demand, so unless there is infinite money supply, policy will always be “tight”.

    BTW, the Bernanke speech today was terrible for expectations of inflation. Almost everything he said was that inflation would remain low for a long time and that the Fed would act before inflation got too high. Terrible.

  • pushmedia1 says:

    gabe, even in the Gali/Blanchard model I posted on inflation has 3 times the weight of unemployment in the optimal policy. Maybe this ratio explains the amount of inflation hawkery in Bernanke’s speeches.

    Also, its very hard to do hypotheticals at infinity. If money demand was infinite then velocity (or the multiplier) is zero, the correct policy response is an infinite money base. Infinity times zero is… ummm… I dunno.

    EE, do you have my email?

  • gabe says:

    push- Bernanke also said that inflation would stay muted, so I don’t understand why he is pushing an exit strategy. I still think that liquidity trappists and Sumerians should agree that expectations of future inflation are good, because you’re lowering real interest rates/improving future NGDP respectively. If inflation (or expected inflation from Tips spreads) starts to climb, then the Fed should start considering an exit. For now, I’m thinking worrying about inflation is premature. Let’s not forget 1937.

  • pushmedia1 says:

    Well, this was the impetus of the “models” post. We have increasing GDP and increasing inflation. Yes, unemployment was increasing until November (we hope its stopped), but its a lagging indicator. These facts suggest current policy is getting us to where we need to be and any additional policy would be inflationary. This sounds strange, I know, because we’re not at full employment. But, again, what is the mechanism causing unemployment to be a lagging indicator?

    The Gali/Blanchard model says its real wage rigidities. Those will be there whether or not Ben pumps more money into the economy (i.e. additional money will be inflationary and it won’t speed up decreases in unemployment). To justify *increased* stimulus, you have to have a mechanism that causes unemployment to lag AND the mechanism has to be such that the lag is responsive to stimulus.

  • gabe says:

    But is there any reason to think that inflation will be high in the near future? We do have increasing inflation, but it’s still fairly low. Actually, looking back at Bernanke’s comments, they don’t seem as hawkish as when I was listening to them, but still I don’t understand talking about a commitment to lower inflation. Higher inflation is what the doctor ordered until the recovery is well in place.

    The mechanism making unemployment a lagging indicator is not known from what I can tell. This lag is much worse in the past 3 recessions (The current one, early 2000s and early 1990s. All of these are “Great Moderation” recessions, but this may be coincidental) that had unemployment lagging severely. (Jobless recovery). But yes, I agree that it’s odd. I’ve heard some stories about labor hoarding or efficiency cutting labor inputs rather than expanding output, but haven’t seen any good models yet. Brad De long has a post on these issues: http://seekingalpha.com/article/149542-jobless-recovery-fasten-your-seatbelts

    Additional money will be inflationary eh? Not in the last year it hasn’t….

    Why do I need to provide a mechanism for unemployment to lag? It just means that stimulus has to continue until we get to NAIRU (5.2 percent I think). Inflation is still under 2%, so with a Taylor Rule, interest rates should be lowered more (can’t, I know). I’ll discuss the GB paper on the other thread, but I don’t think it’s a benchmark. It’s fine, but it has issues.

  • “Additional money will be inflationary eh? Not in the last year it hasn’t….”

    Only if you measure additional money by the monetary base. What we need isn’t an increase in the supply of money, but an excess supply of money to generate inflation.

  • gabe says:

    “Only if you measure additional money by the monetary base. ” I’m kind of confused.

    The Fed only controls base directly, doesn’t it? Do you just think it’s coincidental that base and multiplier move in almost exactly opposite directions? Is this just coincidental?

    Do you really think that the Fed doubling it’s balance sheet and pumping 1 trillion dollars into the banking system is insufficient to increase inflation? How much is needed? 10 trillion?

    There’s plenty of excess money, it’s excess reserves.

  • Gabe,

    I think that we are talking past one another. Clearly I recognize that the Fed controls the monetary base. However, you seem to think that it is a good indicator of the stance of monetary policy. I share this view so long as velocity and the money multiplier are constant.

    You are correct that there is plenty of excess money and that excess money is sitting in reserves. But why do we have excess reserves? Because the Fed is PAYING banks to hold them.

    For us to be in a liquidity trap money must be a perfect substitute to ALL other assets (see Brunner and Meltzer, JPE 1968). If money, debt, and capital are all distinct and imperfectly substitutable, then changes in money can affect asset prices and output.

  • gabe says:

    I agree that we seem to be talking past each other. The Fed paying interest on reserves is a huge mistake, but do you really think that not paying interest on reserves would make much of a difference? It’s half a percent or less isn’t it? If banks aren’t willing to lend at half a percent, then will they be willing to lend at zero?

    If you look at this graph from the Riksbank, it looks similar to the USA’s. Credit is not expanding much with new reserves, causing the multiplier to fall.

    http://www.riksbank.com/upload/Dokument_riksbank/Kat_publicerat/Rapporter/2009/mpr_3_09oct.pdf (warning pdf)

    Excess reserves began spiking in September with financial crisis. Naturally people could have expected the Fed to begin paying reserves next month (which they did), but it fits with a story of financial panic causing high demand for reserves, which will make monetary policy much less effective.

    Japan and the US in the Great Depression saw significant excess reserves without paying interest on reserves, so it’s obviously not a necessary condition.

    I’ll check out the Brunner and Meltzer paper.

    Here’s a different way to think about a liquidity trap: Normally, bank with extra reserves have increased desire to lend out these reserves to make profits. This increases supply of loanable funds, which lowers interest rates and clears the market by increasing quantity demanded of loans. But with interest rates close to zero, how can the Fed increase bank lending? If investors and credit-based consumers have optimized their purchases given prevailing conditions and interest rates, then why would extra reserves lead to more lending? Due to the zero bound, this price mechanism of interest rate declines can’t work to increase demand for credit.

  • gabe says:

    Sorry, the graph is on pg. 50 of the report, it’s really long.

  • Gabe,

    I think that our fundamental disagreement boils down to the monetary transmission mechanism. Forgive me if I am simplifying, but it seems that you believe that the mechanism is the interest rate channel. I view the monetary transmission mechanism as acting through a variety of different channels.

    Thus, the fundamental source of our disagreement is not whether we are in a liquidity trap, but rather as a result of the fact that we have different underlying models from which we are generating our predictions.

    BTW, isn’t Sweden going to start charging a penalty for excess reserves?

  • Gabe,

    Also, the example of the US during the Depression is quite different in the sense that the Fed did not even try to offset changes in the money multiplier caused by excess reserves and a rising currency-to-deposit ratio with changes in the monetary base thereby letting broader measures of the money supply fall by 1/3. I don’t think that this makes your case very well. If you are interested in reading other people who similarly have differing models argue about this with one another, see the exchange between Nelson/Schwartz and Krugman in the JME last year.

  • gabe says:

    Yes, I generally see the interest rate mechanism as the channel. I don’t really see any other mechanism working now, as excess reserves started piling up once interest rates got close to 0. What other mechanisms are you think of, and are they working now? could they?

    It seems like many people have a view that Fed->aggregate demand, but without really specifying the mechanism that gets you there.

    Sweden has been paying a negative interest rate on reserves since July. It would be better if more countries did the same, but at least they’ll be one country to look at to see if their recovery is rapid or not. Bank reserves are still quite high despite this, from what I read in the above paper (though the swedes use different definitions of money than the USA).

    The case of the USA in the Great Depression was simply to show that excess reserves are possible without paying interest on reserves. You said this:

    “But why do we have excess reserves?Because the Fed is PAYING banks to hold them.”

    which doesn’t follow at all because countries experiencing deep recessions experience excess reserves while not holding excess reserves.

    But yes, it’s a much different scenario that currently in that the current Fed is very expansionary.

    Why does the money multiplier seem to fall exactly when money base increases? Do you think this is coincidental? For me, it’s that broader aggregates (M1, PY, etc.) aren’t changing, not that the multiplier and velocity happen to move exactly when the Fed is expanding base.

  • gabe says:

    I finally found how much interest is paid on excess reserves, and it’s a quarter of a percent or 0.25%. It seems for Sumner and EE that this is enough to have excess reserves, while an interest rate on reserves of 0.0000 would cause little to no excess reserves, right? 0.25 This is hardly a leg to stand on.

    http://www.federalreserve.gov/monetarypolicy/reqresbalances.htm

  • Gabe,

    Six things:

    1. Empirical evidence for the interest rate channel is weak. Show me an empirical paper that identifies a statistically significant relationship (with a sizable magnitude) between the real interest rate and fixed capital investment or inventory investment. Allan Meltzer has a paper in a book on the transmission mechanism printed by the Bundesbank (I will find the reference) in which he shows that the real interest rate does not provide a useful guide for explaining monetary policy or the downturn and recovery. (There is a great deal more evidence on why one should doubt the interest rate channel. Perhaps I will write up a post on it soon.)

    2. Given that you believe in the interest rate channel, what is wrong with Svensson’s “Foolproof Way”?

    3. I did not say that banks never hold excess reserves. I also did not say that banks would hold zero excess reserves without the interest payments. What I said is that if we are going to pay interest on excess reserves, we can expect excess reserves to be higher than without interest payments. Do banks not respond to incentives? This is bad policy whether or not a liquidity trap exists.

    4. You seem to think that the money multiplier is falling because the monetary base is increasing. The timing suggests otherwise.

    5. If we are really in a liquidity trap, why not just have Bernanke come out and announce that he is going to start to monetize the debt. After all, money and bonds must be perfect substitutes for a liquidity trap to exist. Thus, one of two things occur: (1) people view this as credible, inflation expectations rise, or (2) people substitute money for bonds and the national debt is eliminated. (I am being facetious — but not entirely.)

    6. Finally, we are ignoring the elephant in the room regarding fiscal policy. It is not sufficient to just “do something” with fiscal policy. The current stimulus package is a giveaway to political interests. What’s more, it does everything wrong according to economic theory. Temporary tax cuts don’t work. Much of the spending is a permanent expansion of government. Whether monetary policy works or not, I wouldn’t mind of some type of fiscal policy so long as it is consistent with what we know from theory and evidence — of course, I also have little faith in politicians to create such policies. If they are going to waste money on policies that don’t work, I prefer we not use fiscal policy at all.

  • pushmedia1 says:

    gabe, do you mind spelling out by what’s implied in your questions “Why does the money multiplier seem to fall exactly when money base increases? Do you think this is coincidental?”

    You’re suggesting the measures of the multiplier are a statistical artifact, but could you spell it out? I’m wondering if there might be an alternative measure or other evidence that could resolve the issue.

  • gabe says:

    Sorry for the delay,

    1) I will look around re: interest rate channels. I am not knowledgable on the channels- how else does monetary policy get transmitted? I just don’t get how credit increases without lower interest rates. Why else would the quantity demanded for loanable funds increase?

    2) I am not opposed to the “foolproof way”, indeed I think this explains a lot about the recovery from the GD. Many people think the US left the gold standard in 1933, which is true, but the new peg of $35 a troy ounce was set in 1934, to continue until the collapse of Bretton Woods. This seems akin to the “foolproof way” and indeed inflation increased and the economy recovered rapidly.

    3) Fair enough. I think we agree that paying interest on reserves is a mistake. I think we disagree on how much this explains of excess reserves.

    4) I’ll explain more on this with my next post responding to push

    5) This is a completely fair question. I think that this would be helpful, as it would increase inflation expectations, but not too helpful, as most of the increase would show up in excess reserves without affecting the broader economy. To be clear, I think that the prescence of excess reserves is good in that once the economy recovers, then there will be reserves available for new loans and plenty of money supply to raise inflation temporarily. Also, monetizing debt is good for the long term budget concerns. However, if the monetary stimulus will be withdrawn eventually, then these bonds will need to be resold at lower prices when rates are higher, so that means a significant capital loss. So it’s a tradeoff between leaving plenty of reserves available with future capital losses. But obviously enough reserves should be available for the recovery. I am surprised that the Fed has been selling treasuries while buying other assets. If the goal is to increase reserves, then selling bonds shouldn’t be an option.

    6) Agreed that temporary tax cuts don’t work. We will see whether these will be permanent expansions of government- I don’t see that happening for many of the projects- broadband, aid to states, etc. Much of US infrastructure is woefully underfunded, so permanent increases in infrastructure would be beneficial imo. Similarly, funding of basic research is also sub-optimal. Is the stimuls imperfect? Yes, but the stimulus package is still a net positive. However, I think debating the stimulus details is much more fruitful discussion than the long discussion of whether government spending works.

  • gabe says:

    push- I should have been clearer. The M1money mulitplier was around 1.6 for most of 2008 until 9/10/2008 when the multiplier began to plunge, falling under 1 by mid-December. If all new reserves are being held at the Fed as excess reserves, then broader aggregates shouldn’t be affected by this monetary policy. M1 being below 1 is odd, as M1 includes M0, so how could base exceed m1? (I still don’t get this). But If m1 is relatively unaffected by changes in base, then the multiplier should move inversely with the money multiplier, as m0*multiplier=m1.

    I think the issue is the EE believes that Fed is simply offsetting increased money demand, with increased demand for base offset by increases in reserves, leading to no real change, which is “tight” policy. Perhaps I am misunderstanding, but I think this is too coincidental, especially considering the excess reserves, to be anything but a liquidity trap.

    I made a spreadsheet to illustrate. It’s money base-excess reserves, which shows how much of the increase in reserves is not becoming excess reserves. This has gone from ~840 billion before the Sept 2008 to 940 b now, while base has gone from 900 billion to 2000 billions, an increase as 1.1 trillion dollars. This is not that money and bonds are complete substitutes, but pushing on this string isn’t moving it much.

    http://tinyurl.com/ya2rrfx

  • pushmedia1 says:

    I don’t have permission to see the spreadsheet…

  • I can’t access it either. Here is the relevant data:

    Here is excess reserves (from the FRED, in billions of dollars):

    2008-07-01 1.916
    2008-08-01 1.971
    2008-09-01 60.053
    2008-10-01 267.902
    2008-11-01 559.036
    2008-12-01 767.397
    2009-01-01 798.233
    2009-02-01 643.482
    2009-03-01 724.623
    2009-04-01 824.367
    2009-05-01 844.076
    2009-06-01 751.364
    2009-07-01 732.996
    2009-08-01 765.857
    2009-09-01 860.074
    2009-10-01 994.734
    2009-11-01 1077.268

    Here is the monetary base (for beginning and end of period for brevity):

    2008-07-02 863.332
    2009-11-04 2024.393

    So the base increases by roughly $1.2 trillion and excess reserves increase by roughly $1 trillion. Gabe interprets this as a result of the fact that banks are simply willing to hold excess reserves as a near perfect substitute for the bonds that the Fed is taking off their books.

    Whether I am correct or Gabe is correct, I still don’t understand why monetary policy cannot be effective. Lars Svensson has his foolproof way:

    (1) an explicit central-bank commitment to a higher future price level

    (2) a concrete action that demonstrates the central bank’s commitment, induces expectations of a higher future price level and jump-starts the economy

    (3) an exit strategy that specifies when and how to get back to normal

  • Gabe,

    BTW, one thing that I forgot to mention on an earlier comment. You wrote, “I finally found how much interest is paid on excess reserves, and it’s a quarter of a percent or 0.25%. It seems for Sumner and EE that this is enough to have excess reserves, while an interest rate on reserves of 0.0000 would cause little to no excess reserves, right? 0.25 This is hardly a leg to stand on.”

    What is the current federal funds rate? The effective federal funds rate as of December 9 is 0.12%. So let’s see, I can loan out my excess reserves to another bank at an overnight rate of 0.12% (and bear some degree of risk) or I can hold the excess reserves (without any risk) and receive 0.25% (TWICE THE RETURN) from the Fed.

  • gabe says:

    EE- yes, I mean that most of the reserves end up as excess reserves. Sorry about the spreadsheet, I’m not good with google docs, but you got the punchline.

    How would monetary policy be effective? Will the next trillion be more effective than the last? I just don’t understand the mechanism.

    I support Svennson’s foolproof way, but how do you commit to future inflation? By increasing reserves? If massive monetary expansion can’t get inflation above 2%, then how can the Fed credibly commit to future inflation?

    I think the temporary currency peg is the best way, as this means that arbitrage will raise the domestic price level of import competing goods, which is a mechanism to increase expectations of inflation. Like I said, the 1934 repeg of the dollar to gold I would cite as a classic example. How do we do this though? Thinking of this, my first guess would be a repeg to gold temporarily, but it is already at a high price, so would this work? (I don’t know). We have substantial oil reserves, so maybe an oil peg would work, but considering that oil price increases tend to lower GDP, it doesn’t seem like it would.

    Do you like Svenson’s foolproof way? If so, how would it be implemented? I.e. if you want more QE, then what’s the mechanism for increasing aggregate demand?

    EE, yes this is true that the Fed funds rate is lower than the interest rate on reserves. This is why it is a mistake. But if the interest rate on reserves is 0%, then how much excess reserves would be held? To put it another way, given that interbank rates are only 0.12%, then banks obviously don’t much like the prospect of loans for anything other than interbank overnight or 3-month T-bills. Why would more reserves change that? Would a Fed funds rate of 0.06% spell recovery? At a certain point 0.0025 and 0.0012 are just an epsilon.

  • “I support Svennson’s foolproof way, but how do you commit to future inflation? By increasing reserves? If massive monetary expansion can’t get inflation above 2%, then how can the Fed credibly commit to future inflation?”

    The Fed could start by announcing an explicit target for the price level. Your issue seems to be credibility. I don’t think credibility is much of an issue when there is an explicit target.

    “Do you like Svenson’s foolproof way? If so, how would it be implemented? I.e. if you want more QE, then what’s the mechanism for increasing aggregate demand?”

    Bernanke floated the idea of buying long term assets. Ed Nelson has a paper on whether this would work with theory and evidence (I am mentioning it, so he obviously agrees with me). I will find the cite.

    Unless you believe that sole transmission channel is the federal funds rate, then conducting open market purchases of longer term assets should have an effect on aggregate demand.

    To understand my position, I think we need to delve into the literature. Ed Nelson has two papers that I think can help you to understand my position:

    “The Future of Monetary Aggregates in Monetary Policy Analysis” JME 2003

    “Direct Effects of Base Money on Aggregate Demand: Theory and Evidence” JME 2002

  • gabe says:

    “The Fed could start by announcing an explicit target for the price level. Your issue seems to be credibility. I don’t think credibility is much of an issue when there is an explicit target.”

    For me, the issue is, if you’re in a liquidity trap, then you can be commited but not creible, as the Fed is ineffective. This is not inconsistent with your views, but it’s definitely a problem for someone like Krugman, who believes Japan was in a liquidity trap but could commit to a higher price level.

    Buying longer term assets has also been proposed by Gagnon, though you’ve probably seen this on Sumner’s blog.

    http://www.piie.com/publications/interstitial.cfm?ResearchID=1451

    I’ll check out the references today, thanks for that