A Proposed Petition

I can’t wait for spring break to start next week so I can focus on all your comments.  Today I’d like to give you some idea about how I approach this issue, what I hope to accomplish, and then end up with a petition.  I’m thinking of sending it to James Hamilton first, as he’s been very good on this issue.  I also hope that Tyler Cowen notices, as he has been incredibly supportive.  But in the end, I’ll only make headway if a number of prominent economists throw their support behind this idea.

If you want a little bit of inflation you need the supply of base money to grow a bit faster than the real demand for base money.  And if you want to make sure that it is “a little” inflation, you need mechanisms in place to avoid overshooting.

Let’s start with the demand for base money, which everyone seems to overlook in this endless argument about liquidity traps and quantitative easing.  I have proposed an interest penalty on excess reserves.  But I have also freely admitted that I am not an expert on the intricacies of Fed policies.  So what might be some hidden glitches in my proposal?

1.  Maybe the Fed is reluctant to charge interest on reserves because they are concerned it would impact bank profits.  Then why not continue to pay a positive interest rate on required reserves which is large enough so that the average bank suffers no loss in profits?  The whole point is to control behavior at the margin, to limit the amount of excess reserves that banks demand.

2.  Maybe the Fed wanted to flood the banking system with massive amounts of reserves for liquidity reasons.  While I doubt this does much good, lets assume it does.  Then create three classes of reserves as follows:  Suppose for liquidity reasons the Fed wants 800B in reserves in the system, but required reserves total only about 80B.  Then tell banks they have three classes of reserves.  Required reserves (of 80B) which earn positive interest.  Supplemental reserves (of 720B) which are interest free, and excess reserves which are charged an interest penalty.  In this case reserve demand is still reduced at the margin, and thus monetary policy now has traction as the equality between bank reserve and T-bill yields is broken.  New monetary base injections will go into cash held by the public, where real demand has been far, far, more stable than for bank reserves.  If the banking system gradually improves, the Fed could withdraw reserves and simultaneously reduce the supplemental reserve levels.

3.  Suppose the Fed says that the massive excess reserves are not distributed equally.  Some very sound banks don’t need much excess reserves at all, whereas some shaky banks need more than 10 times the required reserve levels.  Then “grandfather in” current reserve levels when you set up my supplemental reserve proposal.

The point isn’t that I have all the answers; rather what this thought experiment shows is that it ought to be possible for monetary economists who are much more knowledgeable than I am to construct a reserve system that reduces the now bloated demand, or at least controls that demand at the margin, giving monetary policy traction.

Update (March 4);  Jeffrey Hummel of San Jose State has replied to another post with some useful information.  He says that interest is only paid on deposits at the Fed, not vault cash.  I don’t know why they do it that way, as I believe VC counts toward reserve requirements, and thus the Fed should have data.  But it’s not a big deal.  Just pick a VC level that’s plenty big for any reasonable bank’s needs, and say interest only is charged on vault cash above that level.  (Maybe no more than 3% or 5% or total assets.)  The exact number doesn’t matter, you just want to prevent banks from diverting hundreds of billions of reserves into VC to avoid the tax.  And again, some commenters seem to think my proposal would hurt bank profits, but it would be easy to construct it in a way that it doesn’t (the average vs marginal distintion is key.)  Jeff has some good explanations of the interest on reserves  here and here.  The more issues I see raised, the more confident I am that it would work.

There are lots of economists that know a lot more about monetary policy than I do; Mishkin, Woodford, Svensson, McCallum, Taylor, Hamilton, and Hall are just a few of them.  I know that Hamilton has been very aggressive on these issues, and Taylor recently wrote a strong paper on the subject, but it seems the profession as a whole is just drifting on the question of whether monetary policy can be used to boost AD, at least the public face of the profession (I don’t know what’s going on behind the scenes.)  Many of these prominent economists have (in the past) forcefully argued that monetary policy can be highly effective even when rates are at zero.  Why don’t we see more economists forcefully advocating the adoption of unconventional methods to boost prices?  Oh, and I forgot one more prominent economist who as an academic stated that liquidity traps don’t make monetary policy powerless:  Ben Bernanke.

So here’s my proposed petition:

We, the undersigned economists are concerned that inflation and nominal growth are likely to fall short of the Fed’s policy goals under the current stance of monetary and fiscal policy.  We believe that monetary policy can do more than it has, and is preferable to fiscal stimulus because of the risks posed by large fiscal deficits.

Goal: Increase inflation expectations to about 3%, or nominal GDP growth expectations to roughly 5%.

Policy can be more effective in three areas:

1.  Reduce the demand for base money:

A.  Charge an interest rate penalty on excess reserves.  If combined with positive interest payments on the level of reserves that the Fed considers desirable for liquidity purposes, such a system need not reduce bank profits.

B.  Set an explicit CPI or NGDP target, commit to do whatever it takes to hit that target, and promise to make up for any shortfall or overshooting (“level targeting.”)

2.  Increase the supply of base money:

A.  The preceding proposal for reducing the demand for excess reserves may give monetary policy “traction,” allowing the Fed to boost cash holdings of the public through ordinary open market operations in short term government debt.

B.  If not, commit to buying however much medium, and then long term debt is required.  Fed officials should  try to purchase assets that are not likely to fall sharply in value when inflation expectations return to normal.

3.  Monitor inflation expectations closely:

The Fed should rely on a mix of internal forecasts and market indicators.  No market indicator is perfect, but nominal/indexed bond yield spreads provide some indication of inflation expectations, and should help the Fed avoid sharply overshooting its inflation target.  Other markets can provide some indication of real growth expectations, if the Fed chooses to target nominal GDP.


We are not proposing a cookie-cutter approach that relies on a mechanical formula.  The Fed itself is best able to decide the difficult problems of how to limit excess demand for bank reserves, and what assets to purchase.  Difficult decisions must be made where trade-offs are involved.  For instance, buying indexed debt might reduce the Fed’s risk of future capital losses as inflation expectations recover, but also may reduce the information content of the yield spread.  Despite the very real risks of unconventional monetary policy, we believe that an aggressive multifaceted initiative could quickly turn expectations around, making the actual size of asset purchases required quite manageable.

We would also point to the very real benefits of faster nominal growth, which go far beyond the traditional social welfare benefits of reduced unemployment:

1.  Faster nominal growth would reduce the budget deficit, which is expected to reach worrisome levels.

2.  Faster nominal growth would reduce debt defaults, and this would make it both easier and less costly to rescue the banking system.

During crises it is sometimes the case that an aggressively bold move can achieve results where traditional measures seemed ineffective.  In 1932 traditional open market operations failed to have much impact on confidence in the financial markets or the economy.  Then output and prices began rising rapidly almost immediately after Roosevelt’s dollar devaluation policy was adopted in early 1933.  And this occurred despite near-zero short term rates and in an economy where much of the banking system was shutdown.  Although fiscal stimulus and bank rescue operations may be able to help, nominal growth and inflation are fundamentally monetary problems.  And although FDR’s dollar devaluation policy would not be appropriate during this worldwide slump, we believe than an equally vigorous and unconventional set of steps could stimulate the financial markets and the broader economy relatively quickly.



22 Responses to “A Proposed Petition”

  1. Gravatar of Aaron Jackson Aaron Jackson
    3. March 2009 at 06:36

    Well thought out proposal. I do have one suggestion however. Under point 1B, I would suggest adding something about the use of an explicit target being temporary (at least for now) until it is clear that expectations have recovered and stabilized. After this point, the Fed will re-evaluate the use of an explicit target.

    As you know, having an explicit target, although in use by many banks around the world, is a somewhat contentious policy issue with the Fed. Bernanke has been known to be an advocate of (inflation) targeting, however there are others who are opposed to explicit targets. Having the target temporary I think to some extent would get around this issue (although i’m willing to bet that there would still be resistance to the idea).

    I agree and completely support your idea. It is clear that something new needs to be done with policy, as the Fed has been sitting on the sidelines waiting for its facilities and the fiscal stimulus to gain traction.

    The worst part about it is that I doubt the Fed will take such a proposal seriously. I think they would view it as unnecessary, particularly given their overly optimistic projections going forward. I laughed, and then cried (well, not literally) when the Fed’s unemployment projections came out last week: 2009 estimate of about 8.7%, with a high estimate of about 9.2%. I think somewhere in the low to mid 9% range is a reasonable average estimate, not a high outlier.

    On the other hand, a story just posted in the WSJ quoted the following from Bernanke, so maybe there is hope:

    “we are better off moving aggressively today to solve our economic problems,” the Fed chief said.

    The alternative, he warned, “could be a prolonged episode of economic stagnation that would not only contribute to further deterioration in the fiscal situation, but would also imply lower output, employment, and incomes for an extended period.”

    story: http://online.wsj.com/article/SB123609244895319071.html?mod=

  2. Gravatar of Bill Woolsey Bill Woolsey
    3. March 2009 at 06:57

    Why price level targeting? Why a policy aimed at raising the CPI to a certain point?

    Given nominal income, a higher price level implies lower output.

    I can think of various ways in which higher expected prices would generate more nominal income. But I can also see ways–increases in reservation prices, that expectations of higher prices would directly result in higher prices and lower real output.

    Why the scatter gun approach? Target nominal income.

    Leaving asside expectations, that is what the Federal Reserve can impact through creating money.

    Further, many people (ordinary ones) understand that “the problem” is too little spending. Have the Fed commit to getting spending up.

    Can’t you see that there are many people who think paying lower prices is a _good_ thing?

    Stick to targeting spending.

    Second, trying to think of all the reasons why the Fed might want to pay interest on reserves and then figuring out all the different ways that they might accomplish that is futile.

    I would note that the “profits” angle, ignores that banks are free to charge people for keeping balances in savings and checking acounts if they want. And, that is part of the goal.

    I doubt that the Fed thinks that all banks need to be liquid makes any sense. Much less the theory that the Fed thinks the weak banks need to be more liquid.

    Why would banks need to be liquid? To be ready for currency runs? Fluctuations in net clearings? What?

    The reality is that the “strong” banks are hoarding reserves which allows the Fed to make loans to the weak banks, without expanding the money supply and aggregate demand “too much.”

    I see no evidence that the Fed has changed its policy. It is still trying to reconstuct the shadow banking system (securitization) so that the demand for money will fall and velocity will rise. But most importantly, so that “real” intermediation can support real production. Having regular banks sell loans and having them packaged and then sold to wall street banks, and funding them with commercial paper and having that held through mutual funds–this is all supposedly essential to the market economy.

    Your approach (and mine) that the Fed should focus solely on keeping nominal income on target, and let “the market” deal with credit markets, and face up to some problems in real intermedation, hampered productivity and stagflation, is not their plan.

    And the result is that nominal income is in free fall. They still haven’t fixed the fiancial sector. And now, as a back up, government spending and debt is skyrocketing.

  3. Gravatar of Phil P Phil P
    3. March 2009 at 07:17

    Scott, I’ve read your previous posts and am trying to better understand your views on how monetary policy works, and specifically, how essential RE is to the effect of monetary policy. The old Friedman monetarist view was that changes in M effect nominal income through the cash balance effect, which I think you agree with, at least in part. This view did not depend on RE, in fact when the RE theorists first came along (Lucas et. al.) they argued that anticipated changes in monetary policy would have no effect. Perhaps you can combine the cash balance effect with RE by saying RE speeds things up as people anticipate the expected effects of government policy. I take it you believe that happened in 1933.

    However, there is one difference between now and 1933. Then it was generally believed that abandoning the gold standard would lead to inflation. Now there is controversy about whether unconventional monetary policy (UMP) can work under conditions of a liquidity trap. Can the Fed announcing an inflation target work if there are widespread doubts as to whether the Fed can effectuate its policy? Consider too the widespread pessimism. Even Bernanke, who believes in UMP says the recession won’t end unless the banking system is fixed, in effect saying UMP isn’t enough by itself which must effect the credibility of any policy the Fed pursues.

    Of course, UMP may work anyway through the cash balance effect, but then the inflation target isn’t essential. One final thought: Can a monetary policy work purely through expectations even if the underlying theory is wrong, if people believe the theory is correct, through a kind of placebo effect, a kind of irrational expectations theory?

  4. Gravatar of Rob Rob
    3. March 2009 at 07:37

    Nice proposal. Quick reaction on the packaging, fwiw.

    “We believe that monetary policy can do more than it has, and is preferable to fiscal stimulus because of the risks posed by large fiscal deficits.”

    Hmmm…. Is it necessary to take on this emotional debate (monetary verus fiscal policy)? It could be a distraction. Isn’t your point “here are things we need to do”? If fiscal policy makes the problem worse, then all the more reason to adopt your proposal, I suppose.

  5. Gravatar of W Mitchell W Mitchell
    3. March 2009 at 07:52

    Mr. Sumner

    I comment with trepidation as I am way out of my league on this site, so this will be short. First, you have an amazing blog here – thank you very much for your time and insights. I am learning heaps by reading.

    I think the idea of a penalty on excess reserves is marvelous – one of the smartest things I have heard this entire cycle. I certainly don’t know the legal ramifications of my suggestion, but what about funneling portions of those penalty payments to the FDIC? It seems that Chairman Bair is within a few days of collecting returnable aluminum cans to shore up her funds.

    With humility,

  6. Gravatar of smokedgoldeye smokedgoldeye
    3. March 2009 at 08:58


    You very modestly inform us that there are plenty of economists smarter than you. Perhaps one of them was Murray Rothbard? Perhaps not. Please elucidate on Rothbard’s observations below as provided by CunningDove on Vox Popoli this morning:

    A leader in the easy money policy of late 1929 and 1930 was
    once more the New York Federal Reserve, headed by Governor
    George Harrison. The Federal Reserve, in fact, began the inflationist policy on its own. Inflation would have been greater in 1930 had not the stock market boom collapsed in the spring, and if not for the wave of bank failures in late 1930. (Rothbard, America’s Great Depression p 240.)

    The the money supply increased during the time from June 30, 1921 to June 30, 1929 from $45.3 million to $73.26 milliion. (Rothbard p 92)

    In fact, if you read Rothbard’s work, you might find in chapter 9 that the reason the money supply stayed “almost constant” from 1930 to 1931 was because the Federal Reserve was trying desperately to keep the easy credit flowing.

    During the entire year, 1930, total member bank reserves increased by $116 million. Controlled reserves rose by $209 million; $218 million consisted of an increase in government securities held. Gold stock increased by $309 million, and there was a net increase in member bank reserves of $116 million. Despite this increase in reserves, the total money supply (including all money-substitutes) remained almost constant during the year, falling very slightly from $73.52 billion at the end of 1929 to $73.27 billion at the end of 1930. There would have been a substantial rise were it not for the shaky banks which were forced to contract their operations in view of the general depression. (p 240)

    The truth is that while the Federal Reserve Banks and the Federal Government tried to keep the easy money flowing, the investors & market wanted nothing to do with the easy credit. If you examine what happened in Japan in the 1990’s you will see a very simular pattern. So, if you are planning to prevent the Great Depression II by providing easy credit to a market that does not want credit – All respect to Milton & Anna, but you’re wrong, and we can’t dig our way out of a hole.

  7. Gravatar of David Pearson David Pearson
    3. March 2009 at 09:18


    Speaking for myself, I wouldn’t sign such a petition (don’t worry, I’m not an economist!).

    Why? Because there is no discussion of the cost of inflation targeting, and roadmap for removing the stimulus. A promise to closely follow a leading inflation indicator is just not sufficient.

    Here’s some questions that might help:
    -What happens if the indicator(s) flash red but the output gap is still rising (at an accelerating rate) and the credit markets are fragile? Won’t removal of stimulus threaten to upend the financial markets?
    -What happens to the Fed Chair nomination process if the removal of stimulus begins in 2010, the year of a congressional election and the last year of the FOMC Chairman’s term? Can the inflation targeting policy be consistently implemented in the face of political pressure?
    -How much of an uptick in unemployment would result from stimulus removal, and what unemployment increase would the Fed tolerate before again injecting stimulus? What span of time should elapse between removal and re-injection?
    -Did the 1937 Fed err in removing stimulus? In hindsight, what was the signal that they should have heeded?
    -Under price level targeting, how can one avoid overshooting when leading inflation indicators are flashing red, but the shortfall in the price level has not yet been recovered? Doesn’t that either guarantee overshooting or undershooting?
    -What happens if stimulus removal contributes to crowding out of U.S. Treasury borrowing? What if that crowding out has an out-sized effect on real interest rates (i.e. causes a spike in yields)?

    I’m sure these are just a small subset of questions that one could think of. The issue here, the broader issue, is that monetary stimulus is exceedingly easy to undertake; the rub is always in its removal. So any proposal for stimulus better start with the removal phase and work backwards.

  8. Gravatar of Richard A. Richard A.
    3. March 2009 at 10:09

    Targeting the velocity adjusted money supply (nominal GDP) is superior to targeting inflation. NGDP responds much faster to changes in the money supply than does the inflation rate.
    Targeting the inflation rate directly could lead to overstimulating the economy.

    Since real GDP growth rate averages about 3% a year a 6% growth rate for nGDP would on average give about a 3% inflation rate.

  9. Gravatar of travis travis
    3. March 2009 at 15:22

    @smokedgoldeye: Do you think that the dollar devaluation vis a vis the price of gold in 1933 had little to do with the subsequent rise in production? A dollar devaluation is essentially a monetary easing.

    @richard: The difference between targeting nGDP or the cpi is an interesting one. I thought that according to the ‘sticky prices’ idea, inflation helped speed wage adjustments that the markets weren’t making on their own. So shouldn’t theoretically the cpi increase before nGDP? As an historical matter, does nGDP increase before the CPI and if so, is that contrary to theory?

  10. Gravatar of Richard A. Richard A.
    3. March 2009 at 18:34

    The monetary equation is M x V = GDPn

    GDPn = P x GDPr

    This equation can be differentiated to give
    dGDPn/GDPn = dP/P + dGDPr/GDPr

    For small percentage changes only, the above equation can be written as
    %ΔGDPn = %ΔP + %ΔGDPr where %Δx = (Δx/x) x 100%

    Our economy has become use to about a 6% annual growth in GDPn. The average growth rate in real GDP is about 3% per year and for the above equation to balance inflation will be 3%. If GDPn growth rate were to suddenly change from 6% to 8% — because of sticky wages and prices GDPr in the short run may increase as much as 5% while inflation stays close to 3%. In the long run, GDPr would move back to a 3% annual growth rate as inflation moves to 5%. This is why I believe targeting GDPn growth rate is superior to targeting the inflation rate.

  11. Gravatar of Jon Jon
    3. March 2009 at 20:03


    The current interest-on-reserves policy does divide reserves into tiers, although the distinction was only used when the policy was first implemented.

  12. Gravatar of ssumner ssumner
    4. March 2009 at 04:10

    Smokedgoldeye, I will eventually do a post on 1929-30, but for now I”ll say that (contrary to Rothbard) the 12 months after October 1929 saw some of the tightest money in world history. First, a short lesson on the gold standard. To some extent prices, money and interest rates were “endogenous”, determined by the automatic forces of the GS. But central banks, especially large central banks, had one policy instrument that they could control–the gold reserve ratio. This was basically the ratio of gold to the monetary base. There was a sort of unwritten understanding that the ratio was to be kept fairly stable–this was known as the “rules of the game”. Most people see that in terms of gold flows–if gold flows in you’re supposed to raise the money supply, and vice versa.

    In practice, gold reserve ratios did fluctuate somewhat. Between December 1926 and October 1929 the world gold ratio rose at a rate of about 2.5% per year–a slightly contractionary policy. Why contractionary? Because a higher ratio means more demand for gold, and more demand for gold means a higher value of gold, and since gold was the medium of account, a higher value of gold means a lower price level. There was slight deflation internationally during this period.
    Then after October 1929 an extraordinary thing happened, the world gold reserve ratio started soaring, rising 9.62% over the next 12 months. This highly contractionary policy (which as far as I know I discovered–but I’m sure I’ll eventually find someone who got there before me), is the smoking gun that triggered the Great Contraction of 1929-33. It may also have played a role in the stock market crash of October 1929.
    What did the major central banks do to trigger this policy—nothing much, but perhaps I should move this later to be a formal post, and talk more there about the details. You might, however, notice a pattern here. Big deflationary mistakes in October 1929, October 1937, October 2008, were errors of omission. Rothbard was wrong.

    David, My policy would boost nominal GDP growth sharply, and thus reduce the budget deficit. I don’t claim it would eliminate the crowding out problem, but it would reduce the deficit. I think some of your worries on policy in general would be taken care of by a forward looking nominal GDP growth target. The Fed needs to consistently set policy so that it is expected to reach that goal, even after recovery begins.

    Richard, I agree, although I thinks the Fed’s implicit target is closer to 5%.

    Jon, You are right. I think my only new concept is the supplemental reserves category.

  13. Gravatar of Carsten Valgreen Carsten Valgreen
    4. March 2009 at 07:10


    Just a couple of perhaps stupid questions:

    1. If excess reserves carry an interest penalty, you will create an interest spread between what banks pay depositors for funds (which cannot be meaningfully below 0% because depositors can hoard 0% physical notes) and what they get for reserves. This will work as an extra tax on banks. Not quite needed right now? The incentive for banks would be to cut deposits as much as increase lending. Why are you so sure that this wouldnt just make banks shrink deposits (through fees for instance) and create unregulated shadow banking. If bank force non-financial to hold less deposits, this wont imply more spending. It will imply more t-bill purchases and money market mutual funds will have a feast.

    2. If you cannot affect the interest rate paid by non-financial entities on funds, then how on earth do you expect to affect their demand for funds or behavior in general?

    3. Your approach seem to assume that banks are hoarding cash and wont lend. This is not obviously the case. The Senior Loan officer suvey is showing tightening standards but not more so than in 1975 and 1981 recession. It actually looks pretty cyclical. Bank stocks are hammered by the market pricing of their assets through FAS 157 accounting rules an an illuquid crazyly priced ABS market. Banks are not the end investors in ABS. It is a mystery to me, why banks paying interest penalty on reserves should help stop this death spiral of ABS pricing bank balance sheets to zero. There is plenty of liquidity around (as Taylor also noted in his paper). What hasnt been around is counterparty risk transparency. Complex balance sheets which the CEOs themselves had no real idea about the struture of, has led to complete mistrust. Taxing bank reserves hardly creates visibility or counterparty faith. Neither is it clear to me, why monetarists think it will create any real activity, as the incentives for all non-financial agents appear unchanged, and credit risks is not lessened by it.

  14. Gravatar of Jon Jon
    4. March 2009 at 07:23

    Scott wrote: “Then after October 1929 an extraordinary thing happened, the world gold reserve ratio started soaring, rising 9.62% over the next 12 months.”

    Sounds like you’re using the gold-cover ratio to measure monetary policy against a fully autonomous gold policy, but to square the circle don’t you need to establish a change in the price-level?

    The weak point of MV=Py is that the equation should really be written \sum_{instruments} m_i v_i = \sum_{goods} p_j y_j. I.e., broad money can decline substantially but the components of broad money have different velocities. I posit that, ordinarily, a decrease in M actually results in an increase in V because the marginal units of M are claims on money, such as bonds and debt, not money itself.

  15. Gravatar of ssumner ssumner
    4. March 2009 at 08:52

    Carsten, They are hoarding incredibly large amounts of cash (which, when in the banking system, are better called reserves.) I never said banks won’t lend, and don’t really care whether they “lend” in the normal sense of the term. If all they do is exchange reserves for T-bill, that would be great. (Technically that’s lending to the gov., but I think not in the ordinary sense of the word.) The extra reserves would flow out into cash held by the public, and boost AD.
    As my proposal shows (with supplimental reserves, it need not hurt bank profits. I could construct the charge in a way where bank profits went up, yet banks had an incentive to limit their ER holdings. It’s the difference between average cost and marginal cost. The average interest payment on reserves can be positive, while the marginal cost is strongly negative.
    Why does it boost AD, if it doesn’t “solve” any of the problems with the banking industry? Because it forces $100s of billions of federal reserve notes out of reserves into the publics’ wallets. That’s more than they want to hold, and they then try to “get rid of them”. Since by assumption, banks don’t want the reserves, the public buys goods services and assets, driving up AD.

    Jon, I am just not very interested in the broader aggregates, as the Fed controls the MB, and the broader aggregates are not reliable indicators of monetary policy. You say M and V ordinarily move in opposite directions. I’m not sure that’s true. I think F&S showed they normally move in the same direction. But the current situation certainly fits your scenario. And I suppose you mean that in modern times, when the Fed is targeting NGDP they tend to move in opposite directions. Then I would agree with you, indeed under NGDP targeting it is almost a tautology.

  16. Gravatar of Jon Jon
    4. March 2009 at 10:35

    “broader aggregates are not reliable indicators of monetary policy.”

    Certainly you are in good company with this remark. See re: http://www.federalreserve.gov/newsevents/speech/Bernanke20061110a.htm

    But that’s a very American position, Trichet made some fairly disparaging remarks at the time in response. The german bundesbank used M2 targeting very successfully–and the ECB has largely continued the tradition.

    “I’m not sure that’s true. I think F&S showed they normally move in the same direction.”
    Sure, at least they do during hyperinflation. That’s why the situation becomes unstable.

    “And I suppose you mean that in modern times, when the Fed is targeting NGDP they tend to move in opposite directions. ”

    Sure but so can prices and real output. I don’t see how NGDP targeting gives you any influence over a hyperinflation for instance.

  17. Gravatar of Winton Bates Winton Bates
    4. March 2009 at 12:10

    If I understand your earlier posts correctly, targeting the CPI inflation rate would have done nothing to prevent the excessive monetary expansion a few years ago that led to the current mess.

    Rather than encouraging the use of monetary policy to get a short term fix, it seems to me that it would be far preferable to propose that the Fed should create expectations that monetary policy will be used in future to promote a stable rate of growth in nominal GDP.

  18. Gravatar of Carsten Valgreen Carsten Valgreen
    5. March 2009 at 10:19


    How can you “force money into the publics valet”. This is where the monetarist logic sometimes fails me. It (might have) made sense in the 70s and partly in the 80s, when many legal entities where credit rationed by the Fed – but not really now.

    Lets assume Im a household. At interest level 0% I hold cash and some other assets and have some debts. I have my income and spending decision. Also assume that the credit spreads I pay is in equilibrium with the supply of risk/credit in the economy so that they perfectly reflect the expected risks on me in the current macro environment.

    Now you charge banks for reserves. How does that change my behavior if it doesnt change the price of credit the bank charge me, or the interest on my deposits? Why do I “take more money into my valet” which I didnt want before in the first place? And assuming that my motive for spending is that I have “too much money in my pocket and dont know what to do with them” (and I dont know why I took those money in my pocket) appear somewhat irrational to me. If the additional reserves do not change market interest rates (and they are still zero bounded) it really eludes me why it should work. At least it takes a very irrational economic man for it to do so. In general I think the data, including the dynamics the savings/spending equation work I have seen or done my self, show people, if not entirely living the life cycle hypothesis out, then at least coming close. If for instance the marginal propensity to consume out of receiving $1 more equity gain is 2-5 cents. Then why should I be expected to spend more when you try to circulate more money? I have had no gain, I have had no interest rate reduction on liabilities I have had no interest rate increase on deposits. My life cycle problem hasnt changed. Maybe you can affect my inflation expectations by a target, but it appears to me that Bernanke is already trying to do all he can to tell us he has a target without telling Congress he has one.

    Fundamentally the use of the monetarist approach to this must either assume that liquidity constraints exist in the system in the sense that some entities, for some reason are not able to access the current credit market at the equilibrium credit prices given their risk profile. Or that banks or other agents act in a rather irrational way. Which is it?

    Can you refer some recent empirical work showing that the MB has any Granger causing relationship with real or nom. GDP growth (that is any lead to it).


    I live doing independent macro research for asset managers, in part related to the Euro area. The M2 or M3 has NO relationship with GDP growth in Europe over the past 20 years or so. It used to when there were more regulated markets in the 70s – but hasnt for many years. ECB likes to pretend that M2 & M3 makes sense as targets. The papers they have written on it is hardly persuasive. It appears mostly a religion to the ECB because it was to the Bundesbank. However, as opposed to the US data, M1 (especially when deflated by CPI og PPI) clearly has a lead on Euro area growth. This does suggest that Scotts proposal has more merits in the Euro zone than in the US.

  19. Gravatar of ssumner ssumner
    5. March 2009 at 10:26

    I should add a few more comments to my earlier response to David Pearson.
    I will do 1937 soon, hopefully this weekend.
    I agree about the endgame being important, but I believe the only way to resolve that problem is with a policy that is not ad hoc, i.e. that is not time inconsistent. If you target NGDP at 5%, then you will do the same policy in the future as today (except for recovery from possible under- or overshooting.) Thus when you escape the liquidity trap with a 5% nominal growth target guided by market expectations, you continue on with a 5% nominal growth target guided by market expectations. I don’t like policies that say we’ll do X until the crisis is over, and then we’ll do Y. I don’t think they are credible.

    Jon, I agree that the Euros still look at money, but I don’t think it is very useful. I don’t understand your comment about hyperinflation. While hyperinflation is technically possible under NGDP targeting, as a practical matter it is not.

    Winton, I agree NGDP targeting is a good idea.

  20. Gravatar of Links « Phuco Links « Phuco
    9. March 2009 at 09:29

    […] A petition for Monetary Stimulus by Scott […]

  21. Gravatar of Devin Finbarr Devin Finbarr
    9. April 2009 at 19:47


    Bank lending has nothing to do with boosting AD. In a world with no banks, QE would still boost AD.

    You seem to have a very different concept of QE than both Ben Bernanke and Wikipedia.

    What is your vision of the ideal QE regimen? Should the Fed print money to buy bank assets? Or should it just buy treasuries? Does it buy treasuries at market prices or at a premium? Or does it just print money and mail checks? Or throw it out of a helicopter? ( If you’ve already posted an answer to these questions I apologize, I’ve read through a bunch of your archived posts but I don’t think I’ve seen you describe your interpretation of QE).

    Sorry, I don’t see any evidence that inflationary policies cause bubbles.

    Monetary dilution is a necessary but not sufficient ingredient. Other ingredients include: tax incentives, government policies, supply rigidities, mass psychology, cultural memory, and marketing by Wall St. Also, in our modern system new money enters through credit creation, so assets exposed to credit creation are especially prone to bubbles. In the 20’s and the 90’s it was a lot easier to buy stocks with leverage than in the 70’s.

    I don’t know precisely why there was no stock bubble in the 70’s, although I could certainly pose a number of plausible theories. I do know that there were a bunch of other bubbles – gold, real estate, oil, paintings, etc.

    As for the 1920’s, while price inflation was flat, monetary inflation was not ( and my point is all about monetary inflation, if Ford can produce cars for 10% cheaper each year, that does not magically make loose money policy sound). While there is no Fed M2 or MZM stat from the time, there are stats on the growth of bank deposits, savings and loans deposits, savings and loans deposits. Overall, broad money grew at about 6.5% a year from 1922 to 1928 (source, page 91). And total mortgage debt increased by 200% ( I’m am conflating credit creation with broad money supply growth, but in a system of fractional reserve, the two are conflated). In short: the 1920’s was a period of very loose money.

    While the specifics of bubbles depends on many factors, what can also be said is that the period of very low monetary inflation under the British classical gold standard ( 1860 to 1914) had far fewer bubbles than the periods of greatest monetary growth in the U.S. ( 1917 to 1928, and 1970 to 2008 ).

    You can also look at it from the opposite direction. In a system of zero money supply growth, broad indexes like the S&P would be basically flat ( dilution and stock buybacks pretty much offset ). Given that there would be no price appreciation, it would be obvious to everyone that the only way to make money off stocks was through dividend yield. Thus, when dividend yields drop below the interest rate on BAA bonds, it is plainly obvious to everyone that stocks are a bad buy. Not even the greatest Fidelity ad can argue, “Don’t bother with dividends, because stocks always go up in long run”, because the statement would so plainly be false.

    But when you have monetary inflation at a rate of ~8% a year, and a typical bond interest rate of 5% a year (as was the case from 2000-2008), then even a 0% dividend rate can be plausibly a good buy. If everyone has a 401K and allocates 5% of income into equities, and everyone keeps this ratio constant, than stocks will appreciate at 8% a year (all other things being equal). That makes stocks a much better investment than bonds or dollars in a vault, despite paying no dividends. But the problem is that now instead of basing your investment decisions based on predictable dividends, you’re basing investment decisions based on herd behavior. And the herd could easily change its asset allocation from equities to gold or equities to dollars, thus wiping out your investment.

    Bubbles are defined as irrational behavior. Why should inflation cause people to behave irrationally?

    No, bubbles are defined as disequilibrium. The disequilibrium may be rational by market actors in the context of a badly mismanaged currency.

    Here is a great explanation of money, via the blog Interfluidity:

    Money is the bubble that doesn’t need to pop. As long as there is demand for indirect exchange, at least one asset will be stockpiled by hoarders, hence experience demand that is not a consequence of any direct utility, hence be overvalued. As long as the storage cost for this asset is zero and the supply in existence is fixed, you have a perfect Nash equilibrium – using any other asset as a medium of indirect exchange provides no advantage, and runs the risk of buying into a bubble which will subsequently pop as punters revert back to the stable standard.

    When the supply of dollars increases faster than the supply of stocks or gold, it breaks the Nash equilibrium. The rational actor must weigh the risk of buying into a bubble versus the risk of dilution. This creates a very unstable situation.

  22. Gravatar of Devin Finbarr Devin Finbarr
    9. April 2009 at 19:48

    Oops, I posted the above comment to the wrong post, please delete it. Sorry!

Leave a Reply