What’s the point of fiscal stimulus and QE?

If you were to rely on the press, you’d think that the purpose of fiscal stimulus was to boost real output, to “get us out of the recession,” and the purpose of quantitative easing (QE) is to boost inflation, to “get us out of the liquidity trap.”  I’m sure someone will tell me where I am wrong, but this seems like utter nonsense to me.  I had always assumed that the point of both monetary and fiscal stimulus was to boost AD.  And that in the short run both policies raise both prices and output.  And that the way nominal GDP growth is partitioned depends on the slope of the SRAS curve, not on what caused AD to increase.  Did I misunderstand my macroeconomics courses?

I suppose you could argue that fiscal policy also has supply-side effects, but in the current case those are almost certainly negative.  There are no supply-side tax cuts in the recent bill (indeed implicit MTRs rise), and even if more spending on health care, energy research and education boost long run productivity, they do nothing to shift the SRAS to the right.  It would make more sense to ask whether the fiscal stimulus will succeed in boosting inflation, and whether the QE will succeed in boosting real output.

What set me off was a recent piece in the Washington Post by Johnson and Kwak.  Despite my exasperated tone, I am not trying to pick on these two fine economists.  Their views are widely held, and I am certainly in the minority—a minority of one on some issues.  All I can do is provide my subjective reaction to the following quotations:

After the double-digit inflation of the 1970s and early 1980s, why would anyone want to create inflation? Households and companies alike are trying to “deleverage,” or pay down their debts. But deflation makes it harder to pay down debts, because debts are fixed in dollars and those dollars are becoming worth more and more. Moderate inflation in the neighborhood of 4 percent, by contrast, makes it easier for borrowers to manage their debt loads, and stimulates the economy.

First let’s be very clear about one thing, the Fed can create inflation only by boosting AD, not reducing AS.  So Johnson and Kwak are essentially asking “why would anyone want to boost AD?”  My response is; why would anyone even ask a question like this?  When restated in terms of AD, the answer is simple, because NGDP is now falling at a 7% annual rate and is expected to be far below the Fed’s implicit target for many, many years.  Just imagine a discussion of fiscal policy along similar lines:  “Hmmm, I wonder why the fiscal authorities see a need to stimulate the economy?  Aren’t they worried that NGDP growth is likely to be excessively rapid going forward?  Don’t we need to slow down AD growth?”   I’m going to go out on a limb and assume that no one has been asking that sort of question recently.    But they are asking analogous questions about monetary policy.  (BTW, their answer to the question is perfectly fine.)  They then ask:

But is the United States really a normal advanced economy anymore? We seem to have taken on some features of so-called emerging markets, including a bloated (and contracting) financial sector, overly indebted consumers, and firms that are trying hard to save cash by investing less. In emerging markets there is no meaningful idea of “slack;” there can be high inflation even when the economy is contracting or when growth is considerably lower than in the recent past.

First of all this is a perfect illustration of why it makes more sense to talk in terms of NGDP growth, not inflation.  Monetary and fiscal expansion will raise NGDP growth in all sorts of countries, under all sorts of conditions.  That is the Fed’s policy goal.  In addition, the Fed has made it clear that they want low inflation, not high inflation, a point acknowledged elsewhere by Johnson and Kwak.  We’ve had positive inflation in every recession of my lifetime (except the last 6 months), so why would we expect anything different from a monetary expansion today?  In the next paragraph they assert:

If the United States is indeed behaving more like an emerging market, inflation is far easier to manufacture. People quickly become dubious of the value of money and shift into goods and foreign currencies more readily. Large budget deficits also directly raise inflation expectations. This would help Bernanke avoid deflation, but there is a great danger that unstable inflation expectations will become self-fulfilling. We do not want to become more like Argentina in 2001-2002 or Russia in 1998, when currencies collapsed and inflation soared.

What great danger?  Why are inflation expectations at record lows in the long term bond market, if there is a great danger of inflation?  And if those expectations started to rise, couldn’t the Fed just pull the money back out of circulation to keep inflation expectations in check?  (By the way, I’d much  rather be like Argentina after inflation “soared,” than Argentina before inflation soared–although I would prefer to avoid both options.)

At the other extreme is Tim Duy, who worries that we won’t be able to get any inflation at all.  I’ll use Mark Thoma’s post, as it nicely explains all the competing views.  (Be aware of the Russian doll problem, lots of quotations within quotations within quotations.)

The key paragraph in Johnson and Kwak that I [Tim Duy] take issue with is:

“Then in March, the Fed said that it will begin buying long-term Treasury bonds on the open market, hoping to push down long-term interest rates (by increasing the amount of money available for long-term lending) and thereby stimulate borrowing. The implication is that the Fed will finance these purchases by creating money. Not only that, but Bernanke wants us to know exactly what the Fed is doing; he hopes to push up our expectations of future inflation, so that wages and prices will nudge upwards, not downwards.”

The implicit assumption is that the Fed is expanding the money supply via a policy of quantitative easing with the explicit goal of raising inflation expectations.  First off, as Bernanke said once again today, he does not describe policy as quantitative easing:

“In pursuing our strategy, which I have called “credit easing,” we have also taken care to design our programs so that they can be unwound as markets and the economy revive. In particular, these activities must not constrain the exercise of monetary policy as needed to meet our congressional mandate to foster maximum sustainable employment and stable prices.”

Pay close attention to Bernanke’s insistence that the Fed’s liquidity programs are intended to be unwound.  If policymakers truly intend a policy of quantitative easing to boost inflation expectations, these are exactly the wrong words to say.  Any successful policy of quantitative easing would depend upon a credible commitment to a permanent increase in the money supply.  Bernanke is making the opposite commitment – a commitment to contract the money supply in the future.  Is this any way to boost inflation expectations?

So Tim Duy draws the opposite conclusions from Johnson and Kwak, the danger is no inflation at all.  Obviously in terms of likely outcomes, I agree with Duy.  But I also think that in this particular exchange Johnson and Kwak’s interpretation of Fed intentions is just as plausible, perhaps more so.

The Fed clearly doesn’t intend the reader to interpret “stable prices” literally, rather they are referring to the Fed’s inflation target, which is more like 2% inflation.  So you have one side warning of high inflation if the monetary base remains at the current high levels, and others saying we won’t get any inflation at all if the money is pulled out of circulation later.  In a sense both sides are right.  So why doesn’t the Fed clearly say that it plans to leave just enough money in circulation to keep long run price levels 2% or 3% (per year) higher than current price levels?  Isn’t that what the Fed wants?

By the way, here is another case where NGDP growth makes far more sense than inflation.  Most liquidity trap models talk about reducing real interest rates through expected inflation.  But if you look at the logic of these models, it is clearly NGDP growth that matters, not inflation.  Don’t believe me?  Then consider a scenario where the Fed generates 5% NGDP growth expectations, but negative 1% inflation expectations. Do we stay in a liquidity trap?  Obviously not, with 6% real growth expectations the Wicksellian natural rate (in real terms) will rise well above 1%, so you don’t have to worry about that problem.  And even if you stayed stuck in a liquidity trap, would that be so bad with 6% real growth?

Why doesn’t the Fed simply communicate its inflation target so people aren’t playing this zero inflation/hyperinflation guessing game?  They don’t have to tell us exactly how much money they will later pull out of circulation; just tell us that it will be enough to produce X percent inflation.  Perhaps the problem is that Congressmen like Barney Frank get upset when the Fed hints that it might ignore its “dual mandate” and focus solely on an inflation target.  I have a suggestion for the Fed, it’s a nominal target that gives equal weight to inflation and real growth.  Anyone want to guess what I’m thinking?

While I am in a grouchy mood let me get one more thing off my chest.  Consider the following quotation from Johnson and Kwak:

If he succeeds in restarting growth while avoiding high inflation, Bernanke may well become the most revered economist in modern history. But for the moment, he is operating in uncharted territory.

First let me reiterate that I don’t blame just Bernanke for the crisis, I blame the Fed and all the macroeconomists who refused to strongly criticize the stance of Fed policy in October 2008, when growth forecasts fell far below any reasonable policy objective—in other words, just about everyone.  But Bernanke was chairman of the Fed at that time, and they did refuse to cut rates to zero.  Admittedly they did inject a lot of reserves, but then they paid interest on those reserves at rates higher than alternative assets, insuring that almost all the extra reserves would be hoarded.  And this was done when financial markets and private forecasters were (correctly) forecasting falling NGDP.  So when I see people talking about the potential heroism of Bernanke, I can’t help thinking of the fireman who started fires so that he’d have a chance to become a hero putting them out.  Yes, this is almost certainly a bad analogy, Bernanke is both well-intentioned, and a much better than average fireman.  But if it is a bad analogy, then it’s a pretty sad comment on the state of our profession.

Update 4/6/09, Despite my disclaimer before the first J&K quotation, a commenter mentioned that my tone was not very polite.  That was certainly not my intention, but I can see how people might read it that way.  So let me be a bit clearer about this point.  Since early October I have been extremely frustrated with about 99% of macroeconomists; who I blame for the worldwide recession/depression.  I think they have almost everything backwards.  I believe they are looking at totally meaningless “indicators” of monetary policy like interest rates and the money supply, and ignoring the only variable that matters, the expected growth in the central bank’s target variable.  So in the unlikely event that J&K come across this post, just assume you are surrounded by 99% of other macroeconomists, including all the greatest minds in the field, and consider me an eccentric economist at a small school taking potshots from the sidelines.  I was just venting my frustration with the entire tone of the ongoing debate over macro policy, and picked your piece at random.

I’d like to thank the invaluable Dilip for the links used here.  I should probably be paying him for his assistance.  But I won’t, unless someone will pay me for doing this blog.  With a bunch of new links, I will probably not post for a few days, returning later in the week with a piece on international monetary issues.



42 Responses to “What’s the point of fiscal stimulus and QE?”

  1. Gravatar of Clifford Nelson Clifford Nelson
    5. April 2009 at 16:41

    Obviously the concern is not that AD will increase but further contract due to higher rates that result from the QE. Also, the distinction between calling the bond purchases QE or not (based upon the Feds stated intention to watch inflation and sell if and when necessary) is largely illusory because if the Fed is committed to using monetary policy to stimulate growth, and we enter a period of stagflation how does the Fed follow through?

    Do you think there are any limits on untraditional monetary policy (after we reach zero) and, if so, where is that point? Is there a point where the costs
    and risks should prohibit its use?

  2. Gravatar of Dilip Dilip
    5. April 2009 at 16:50


    BTW, Scott, I think its Tim _D_uy (not _G_uy)

  3. Gravatar of ssumner ssumner
    5. April 2009 at 16:51

    Clifford, Thanks for the comment. I don’t follow the “obviously” point. If QE leads to higher rates, it would be through inflationary expectations (the Fisher effect.) But in that case AD would not contract, it would increase explosively. Why? Because the base has already doubled (but is being hoarded.) If the Fisher effect kicked in the base would not be hoarded, AD would explode if the Fed didn’t pull money out of circulation.

    There are no practical limits on untraditional monetary policy. The Fed could buy up the entire world. But that would be unwise, as it would cause hyperinflation. They really don’t need any more money in circulation, they need to get the hugely bloated excess reserves circulating with a penalty interest rate on excess reserves. Any effective policy will bounce us off the zero bound very quickly, the key is getting an effectively policy. Zero interest rates are a sign of failure. Rather than promising zero rates for years to come, the Fed should target 5% NGDP growth expectations, with the hope that rates will quickly rise above zero in the Treasury market.

  4. Gravatar of ssumner ssumner
    5. April 2009 at 16:55

    Thanks Dilip, I really should be paying you. That’s the second silly mistake of mine that you caught.

  5. Gravatar of Clifford Nelson Clifford Nelson
    5. April 2009 at 17:41

    Thanks for the reply, the “obviously” is in response to your response to Simon and James. “First let’s be very clear about one thing, the Fed can create inflation only by boosting AD, not reducing AS. So Johnson and Kwak are essentially asking “why would anyone want to boost AD?” My response is; why would anyone even ask a question like this?”

    As far as the Fisher effect, all I can say is that I wish I saw money being hoard, what I see is a lot debt and insolvency. So … I am not so sure that the inflation target or QE will produce demand… especially with the unemployment situation.

    As for the taxing bank reserves, do you have any concerns that merely pushing the banks to lend without previously stimulating demand will only make things worse? In other words what must come first, demand or investment? And once demand returns will there be a need or benefit to taxing the bank reserves?

    As for your comment that the Fed can only create inflation by boosting AD, I think the concern of Simon and James is that inflation could result from factors other than AD, and they gave an example of emerging markets to demonstrate their point.

    As for practical limits, I meant constructive uses and not destructive uses.

    Thanks for your thoughts, Cliff

  6. Gravatar of malavel malavel
    5. April 2009 at 20:12

    Wouldn’t higher inflation lead to higher nominal interest rates on mortgages and thus higher default rates?

  7. Gravatar of Don the libertarian Democrat Don the libertarian Democrat
    5. April 2009 at 21:01

    Here’s a quote from a paper that I like:


    “The problem is, again, why an expansion of the monetary base today should be viewed as a
    commitment to increased money supply in the future. While the liquidity trap lasts and the interest
    rate is zero, the demand for monetary base is perfectly elastic and excess liquidity is easily
    absorbed by the private sector. However, once the liquidity trap is over and the nominal interest
    rate is positive, demand for money will shrink drastically, in most cases requiring a drastic
    reduction of the monetary base. It is difficult to assess how much the monetary base would have to
    be expanded before inflation expectations and inflation take off. Beyond some unknown threshold,
    deflation may be quickly replaced by hyperinflation.”

    Now, I’m not an economist, but, although Sevensson doesn’t like the idea, he does seem to be talking about QE and inflation together. Am I wrong?

  8. Gravatar of DG DG
    5. April 2009 at 21:51

    Why not just let the markets we all profess to have such faith in set interest rates, and stop allowing fallible individuals to distort the market for capital with such increasingly devastating consequences?

  9. Gravatar of Clifford Nelson Clifford Nelson
    6. April 2009 at 00:10

    I guess my concern is that the Fed’s actions in expanding the monetary base to create some inflation (in the hopes of addressing the liquidity preference) will act to drives rates higher without stimulating demand past the point necessary to effect prices since we have a huge output gap.

    Further, given the amount of debt that must be financed (or refinanced) in our economy, the action may ultimately act to reduce demand. For example, if interest rates run equal or higher to price inflation.

  10. Gravatar of Aaron Jackson Aaron Jackson
    6. April 2009 at 06:27

    I was chatting with my father on the phone last night, and he expressed his serious concern (that he presumably picked up from the media) about how hyperinflation is going to occur soon, which would wreck him financially. Being the good macroeconomist son, I had to explain to him essentially what Scott has so eloquently described in this post. I think that this is a nice anecdote to the “zero inflation/hyperinflation guessing game” that Scott talks about.

    Abstracting from the aggregate welfare costs of inflation, most people view inflation in the economy as bad (even if their nominal wages are completely indexed to inflation), so a policy of consciously and purposefully creating inflation, to the general public would be utterly disastrous and unthinkable. The Fed needs to do a better job of educating the public about this. One of the ways the Fed can and often does make its job easier is to allow expectations to do much of the heavy lifting of policy changes. If half the public believes we will be stuck in a deep recession with zero inflation for some time, and the other half believes that hyperinflation will occur within the next year, what kind of prudent policies can be implemented? What kind of outcome would this produce for a given policy? Likely, not very good.

    Creating an anchor such as a NGDP or inflation target (even if only temporary, and by temporary I mean that it would not necessarily be a permanent, long run switch to a formal targeting framework) from which expectations can be tethered would go a long way to solving this problem, and I think would be the first large scale step to the Fed regaining much of the credibility it has seen slip away in the last couple years.

  11. Gravatar of Aaron Jackson Aaron Jackson
    6. April 2009 at 06:42

    To help out Scott, I’ll respond as I think he would. Scott has several papers (one of which is co-authored with myself) advocating a role for the private sector in setting monetary policy. I think in several points in his blog, he has mentioned that if such a framework were in place, many of the “behind the curve” policy mistakes made thus far by the Fed would have been avoided, or at least mitigated. If a framework was set up such that policy responded automatically to market expectations of nominal GDP, for instance, the signals that the Fed received later into 2008 would have had them cutting rates earlier, faster and deeper, which would have perhaps stemmed, or at least mitigated to some extent, much of the liquidity problems in financial markets, as well as secondary decreases in asset prices.

  12. Gravatar of Jon Jon
    6. April 2009 at 07:14

    Johnson and Kwak == Baseline Scenario

  13. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    6. April 2009 at 07:30

    I’d be interested in Scott’s (and other macro guys) thoughts on this:


    from Vernon Smith. Especially interesting is his idea that the CPI was seriously understating the true inflation rate during the housing bubble.

  14. Gravatar of Alex Golubev Alex Golubev
    6. April 2009 at 09:15

    NGDP Targeting sounds a bit intimidating. Is the suggestion that the Fed’s job is to intice people to switch between assets and paper: 1 printing more paper when people are trying to hoard it and 2 removing it from circulation when people get a bit too exuberant with assets? Makes sense. Particularly considering that the financial system is leveraged and the flight FROM assets is way more important than the flight TO assets. How do we reconcile stagflation in the 70’s and are there any theoretical possibilities of it occuring under “NGDP Targeting”?

  15. Gravatar of David Stinson David Stinson
    6. April 2009 at 09:52

    Patrick. Interesting WSJ article. Sounds rather Austrian:

    “The 2001 recession might have ended the bubble, but the Federal Reserve decided to pursue an unusually expansionary monetary policy in order to counteract the downturn. When the Fed increased liquidity, money naturally flowed to the fastest expanding sector.”

    DG, I am totally with you. I would say the answer to your question is that, in short, many macro-economists, despite their protestations to the contrary, do not in fact have faith in the ability of markets to coordinate economic activity. Many macro-economic policies are essentially central planning activities, designed specifically to reverse or offset the effects of maximizing behaviour on the part of individuals or to manipulate individuals’ incentives or behaviour. Often, these policies affect, by design, some of the most basic and fundamental decision-making in a capitalist economy, such as individual consumption/saving decisions and other capital market decisions.

    I was at school in the late 70’s and early 80’s and took a fair bit of macro and monetary economics in my undergraduate and MA. Not once do I recall a professor saying, “gee, in economics we generally exalt the free market and competitive processes as a means to organize economic activity and establish prices, so how do we reconcile that with, in effect, the profession’s acquiescence to a government monopoly in currency production and their influence of interest rates via money production?” It’s not that there may not be good arguments in favour of central banking, but what I find astonishing in retrospect, particularly given the magnitude of the intervention, is that not only was the existence of the central bank never really questioned, the apparent inconsistency of its existence with the implications of other aspects of economic theory was never even raised.

  16. Gravatar of David Stinson David Stinson
    6. April 2009 at 10:58


    “The Fed needs to do a better job of educating the public about this. One of the ways the Fed can and often does make its job easier is to allow expectations to do much of the heavy lifting of policy changes.”

    In order to create durable expectations, people not only have to be convinced of the policymaker’s intentions, they have to have confidence in the ability or competence of the policymaker to successfully achieve those intentions. The fact that there seems to be some dispute over what the Fed’s intentions are or even what it expects its policies to achieve (inflation vs real effects, QE vs credit easing, etc., etc) suggests that the Fed has not communicated its intentions clearly enough. Secondly, and not to put too fine a point on it, people’s confidence in the Fed’s ability to determine the right policies or to implement them has presumably taken a rather significant hit over the last year or so. People are aware that the Fed’s previous self-confidence was probably not justified, that monetary policy is by no means a precise or surgical process and that when the Fed makes a mistake, it hurts. A lot. Add to that the generally accepted notion that the Fed is operating in uncharted waters and people’s vague understanding that many of these same policies were attempted unsuccessfully in Japan and it starts to look pretty ugly from the perspective of the Fed managing individual expectations.

    One is almost forced to conclude that the Fed right now is a source of uncertainty, rather than a reducer of it.

  17. Gravatar of David Stinson David Stinson
    6. April 2009 at 11:21

    Hi Scott. Another timely and thought-provoking post.

    Isn’t Tim Duy wrong though when he states:

    “Pay close attention to Bernanke’s insistence that the Fed’s liquidity programs are intended to be unwound. If policymakers truly intend a policy of quantitative easing to boost inflation expectations, these are exactly the wrong words to say. Any successful policy of quantitative easing would depend upon a credible commitment to a permanent increase in the money supply.”

    Inflation, and thus rational inflationary expectations, presumably depend not on expansion of the money supply alone but on the gap between money supply (growth) and money demand (growth). Portfolio demand (as opposed to transactions demand) for money is very high at the moment due to uncertainty and volatility in other asset markets. Once a recovery starts (if not before) one might expect portfolio demand for money to fall significantly as individuals begin to move funds back into other assets. Presumably, any prior commitment to a given moderate rate of inflation would require the Fed to withdraw liquidity at that point in order to preserve the required gap between money demand and supply.

    Wouldn’t it?

    There’s also the issue of the excess bank reserves, which may end up getting loaned out as the economy improves and default risk declines, creating additional money at the very time that demand for money is likely to be declining.

  18. Gravatar of johnleemk johnleemk
    6. April 2009 at 12:16


    In response to the first quotation you cite, you seem exasperated with the obvious inanity of the hypothetical question posed. Unfortunately, I think it’s clear that that is a question which too many laymen are asking. MV = PT is an equation which far too few people know of — definitely fewer people than the number who know that e = mc^2 or F = MA. As far as most people are concerned, an increase in M clearly implies an increase in the price level.


    I’m an undergrad who only took an intermediate macro course a semester ago, but the logic behind central banking seemed reasonably well-explained to me: changes in the supply of gold or any other commodity are unpredictable, and rarely commensurate with changes in GDP. While imperfect, a central banking controlling the money supply can try to appropriately target the growth of the money supply.

    I do agree that most econ courses and a lot of econ professors gloss over the bigger systemic issues, but I’m not sure what can be done about this. My solution to this has been to read avidly — there are some pretty good books explaining central banking out there, although I last read one of them a few years ago, so I can’t recall the title. I would also add that reading this blog has been invaluable in enhancing my knowledge of and interest in macroeconomics.

  19. Gravatar of ssumner ssumner
    6. April 2009 at 13:29

    I can’t fully answer everything, but I’ll say something about each comment.

    Clifford, I’m not really pushing banks to “lend” in the normal sense of the word, just get rid of reserves. If they merely bought T-bills with the reserves, I’d be happy. The act of getting rid of money is what causes AD to rise, indeed it is the only thing (other than more money) that causes AD to rise.

    Malavel, No, higher inflation helps borrowers. If their home price goes up, they can sell it at a profit, even if they cannot afford to service the mortgage. With an ARM, that may be less true, but is probably still true to some extent. And if ARMs are truly tied to S-T interest rates, they should be near zero now. I presume they are not near zero, which means they won’t rise one for one with inflation.

    Don, Yes, Svensson thinks QE can raise inflation but he prefers currency depreciation as it is easier to estimate the impact. He may be right in Japan’s case, although as he says it doesn’t apply to the world economy. Svensson favors using internal Fed forecasts, not market forecasts as I prefer, so he worries more about overshooting than I do.

    DG, I agree about interest rates, and as Aaron says later, I prefer NGDP targeting, and letting markets set rates.

    Clifford#2, An increase in M cannot cause rates to rise, without increasing AD, so I don’t follow this argument.

    Aaron, I agree about the target issue. One point about hyperinflation beliefs. I think it is just a bunch of hot air that people spout off because they want to sound cynical. The people who buy Treasury securities predict 0.8% inflation, so that’s “the public” unless I hear otherwise. Next time someone tells me they expect hyperinflation I will ask whether they expect their house to be worth a fortune in the near future. If the public as a whole has that perception, why are house prices falling?

    Jon, Thanks

    Patrick, The CPI did probably understate inflation during the housing boom, but only slightly. NGDP was growing a bit over 6% during that period, and that number is not affected by CPI mismeasurement. The overall economy was very strong, so clearly real growth was much more than 1%. That means true inflation had to be far below 5%, even during the housing boom.

    Alex, In the 1970s real GDP grew at a decent pace, the problem is that NGDP was growing at over 10%. That’s why inflation was so high. (Not because of anything OPEC did.)

    David, There was no “natural” flow into housing. It occurred because business investment had become overextended in the 1998-2000 period. After the tech crash the marginal productivity of more housing was far higher than more business investment.
    I agree that we expect too much of economic planners, that’s why I favor letting the market determine the money supply, not the FOMC. But as long as the Fed insists, I will be giving them suggestions when they get it wrong.

    David#2, I strongly agree with what you say here (although I don’t see it conflicting with Aaron’s point which is more about educating the public about what terms like ‘inflation’ actually mean.) But they have only themselves to blame if they have lost credibility, as you say.

    johnleemk, I agree with your points, but would add that it is gold demand that is the most unstable. This might seem contradictory, but I also think it is quite possible that a laissez-faire gold standard would have outperformed our actual system during the 20th century. But only because our actual system did a pretty poor job. I also think NGDP futures targeting could outperform a laissez-faire gold standard.

  20. Gravatar of David Pearson David Pearson
    6. April 2009 at 14:04


    You seem to underestimate the difficulty of managing inflation expectations. You also create too much “space” between deflationary and inflationary expectations.

    Let me take the second point first. Imagine a low-savings country facing a deflationary shock. The larger the drop in NGDP, the more the Central Bank would need to inject (or the currency devalue) to reverse expectations. Thus, its quite conceivable that inflation expectations under this condition are volatile and symmetric: they discount discrete outcomes of strong deflation or strong inflation. The same occurs at the opposite end of the spectrum, when possible outcomes are continued hyperinflation or the deflationary shock of a convertibility-plan. Bottom line: its hard to point to the mean of expectations embedded in TIPS as proof of their stability.

    Secondly, you seem to overestimate the ability of the Fed to influence expectations. Assume an “L” shaped recovery where private investment demand is constrained by a moribund banking system and excessive legacy debt. At that point, I would argue that the Fed is not concerned with the mean of the expected impacts of stimulus removal, but with the downside “tail risk”. Put simply, there will be a reasonable chance that stimulus removal will accomplish what you claim the 2006-2007 produced. You might argue that the Fed persisted too long in that episode, and that this was “obvious”. Perhaps obvious to you, but not to the Fed: basically you are assuming that the Fed that disappointed you with its actions in 2006-2007 will now know, perfectly, how to inject and remove stimulus, and that this knowledge will enable them to act free of political interference. Good luck with that.

    I am sorry if I am a bit snippy, but you come across somewhat arrogant in your criticism of Johnson. Remember that he has witnessed, first hand, many episodes of failed Central Bank interventions. This is not the realm of academics but of policy. You would do well to flesh out your fine-tuning recommendations with “real world” scenarios, something I think Johnson has attempted to do.

  21. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    6. April 2009 at 15:19

    Eric Falkenstein mentions your blogingheads performance:


    ‘…banks are hoarding cash because they feel they need a bigger cushion. Indeed, the ‘stress tests’ of Geithner are looking at banks and evaluating whether they have sufficient capital to withstand some hit to their portfolios. Lower levels of capital would make them fail this test, and then they would be taken over.

    ‘In 1937 the US implemented a tax on retained earnings, which seemed like a good idea to economically incompetent Roosevelt because he assumed he was just taxing ‘money’, and unused money at that. The idea was, we wanted to get this money ‘working’. Plus, taxes on capital are progressive (good). See here for a 1937 rationale, and note it mentions Henry Ford by name.’

  22. Gravatar of ssumner ssumner
    6. April 2009 at 16:25

    David, You raise some good questions. First let me say that I added a paragraph at the end of my piece to clarify that no disrespect was intended (and I mean that seriously, I’m not just being polite.) Second, I do not believe that inflation expectations are widely dispersed, I think that markets understand that a Bernanke Fed will not allow sharp deflation. I may be naive, but I still have confidence that he would not allow that. It ought to be possible to resolve our debate by looking at the indexed bond market, and comparing newly issued 5 year TIPS, with 5 year old 10-year TIPS. If I am right the yields should not dramatically diverge. Maybe somebody that knows how to gather that information can do so and prove me wrong.

    Back up plan if I am wrong: Blame Bernanke. After all, if there is such a wide dispersion of possible inflation/deflation outcomes, isn’t that partly because they don’t have an inflation (or NGDP) growth trajectory explicitly spelled out. If the Fed really wants 2% inflation, then they should right now promise to target the five year forward CPI at 10% above the current CPI. Their failure to do so can only increase uncertainty. Now you might argue that given that can’t do that (for whatever reason) it is hard for them to manage expectations.

    Another reason I am frustrated is that for twenty three years I have been promoting a plan that would perfectly manage expectations, what you are now saying is difficult. And for 23 years I was brushed off at conferences by people saying “don’t worry, the Fed has expert forecasters, they don’t need help from the markets.” CPI or NGDP futures targeting would always keep expectations right on target. The money supply would adjust to maintain equality between the forecast and the target. Now the Fed screws up and I am being told that it is hard to manage expectations. Well I wish I could go back in time to those earlier seminars (one at the NY Fed) where I got all those bland assurances that all was fine. That’s another reason I am so frustrated. I have had top level macroeconomists tell me my plan will work. No one has ever come up with a good reason why it won’t work. It would have prevented this debacle. And yet discussion of the idea still seems off limits in polite company. OK, if the Fed has a better idea what is it? And why haven’t they unveiled it yet?

    I think you misunderstood my 2006-07 comment (probably my fault.) Lot’s of people think easy money was the root cause of the subprime fiasco. I don’t think it’s that simple. But to sell my NGDP idea it helps to point out that it would have called for tighter money in the 2004-06 period (for what it’s worth), when NGDP grew faster than 5%. Just to be clear, I was not one of those far-sighted people predicting disaster in 2006-07, I just thought money was slightly too easy. My only serious criticism of the Fed (in 26 years!) has been in the last 6 months. How many other macroeconomists have been so kind to the Fed?

    Here’s the real world scenario you ask for. Ease policy until 5 year inflation expectations get into the 1.7% to 2.3% range, based on indexed bonds. It’s not my dream policy, but we lack a NGDP futures market. But it’s much better than what we have. Then set an explicit nominal target growth path, and commit to return to it when you miss. Do it by first getting rid of interest on reserves, if that’s not enough get rid of excess reserves by using a penalty rate, if that’s not enough then do QE (believe me, QE would not be necessary.)

    Real world fiasco’s like Argentina post-2002 weren’t stumbled into accidentally. We don’t need to worry about that. We need to worry about Japan in the 1990s-2000s, where policy was too tight at the very same time that experts were warning the bloated base would eventually lead to high inflation.

    Patrick, Correct me if I am wrong, but don’t things like T-bills count toward bank capital? If so, then his comment is off the mark–I am not trying to reduce bank capital. If he is worried about bank profits, then have the Fed pay banks a subsidy on T-bill holdings to offset the penalty on excess reserves. I don’t blame him for misunderstanding my intent, one must grossly oversimplify in a short talk. But the issues of capital and excess reserves are completely different. Dealing with one need not affect the other at all. Ironically, the bigger cause of the 1937 recession was the Fed encouraging banks to hold more reserves, by raising the reserve requirement. Does that sound familiar?

  23. Gravatar of Bill Woolsey Bill Woolsey
    6. April 2009 at 17:24

    Capital is the difference between assets and liabilities.
    T-bills are an asset, and so “add” to capital. Commercial loans add to capital as well. Liabilities, like deposits or overnight borrowing, take away from capital.

    However, as a practial matter, banks “add” to capital by earning profit and retaining earnings or else by selling new shares of stock.

    Capital requirements are a fraction of total assets. The assets are weighted according to a formula that is based upon risk. T-bills (and all government bonds) have a zero weight. That means that if a bank changes the composition of its assets from commercial loans to T-bills, the amount of capital the government requires it to hold is reduced. It doesn’t change the amount of capital it actually has. Changing the form of assets leaves capital unchanged. But because government bonds are considered less risky by the regulations, shifting from commercial loans to T-bills reduces the amount of capital a bank is legally requred to have.

    More importanly, if banks switch from holding reserve balances at the Fed to holding government bonds (including T-bills) this has no impact on capital requirementes. It has no impact of the banks’ capital (changing the form of assets) and it has no impact on how much capital the goverment requires a bank to hold.

  24. Gravatar of Jon Jon
    6. April 2009 at 19:32

    “But because government bonds are considered less risky by the regulations, shifting from commercial loans to T-bills reduces the amount of capital a bank is legally requred to have.”

    Or if you’re a European bank, you buy CDS contracts from AIG to render a lower risk-weighting…

    “The assets are weighted according to a formula that is based upon risk. T-bills (and all government bonds) have a zero weight.”

    US Banks are also subject to an absolute leverage ratio of 33:1–this applies regardless of whether the bank holds tbills. European banks were, and are, much more highly leveraged:

    Bank of America $1,715B $146.8B 11.7
    Citigroup $2,187B $113.6B 19.2
    JPMorgan $1,562B $123.2B 12.7
    Wells Fargo $575B $47.6B 12.0

    Deutsche Bank €2,020B €38.5B 52.0
    UBS Fr2,272B Fr42.5B 53.4
    Credit Suisse Fr1,360B Fr59.88B 22.7
    Fortis €871B €34.28B 25.5
    Dexia €604B €16.4B 36.8
    BNP Paribas €1,694B €59.4B 28.5
    Barclays £1,227B £32.5B 37.8
    RBS £1,990B £91.48B 21.7

    Its ironic that just a year ago, the US approach was chided as antiquated. Now its the only metric the market trusts because it cannot be engineered.

    (If you’re wondering how your local bank is doing look for its Regulation F reports).

  25. Gravatar of Jon Jon
    6. April 2009 at 21:07

    Scott writes: “It occurred because business investment had become overextended in the 1998-2000 period. After the tech crash the marginal productivity of more housing was far higher than more business investment.”

    I have to disagree. The apparent yield on housing was higher–there was a bubble–but in an absolute sense the marginal productivity was NOT higher. From my own small sample of associates, the mantra was that the (stock) market rolls but housing grows.

    Scott also writes: “NGDP was growing a bit over 6%… The overall economy was very strong, so clearly real growth was much more than 1%. That means true inflation had to be far below 5%, even during the housing boom.”

    Sorry, I don’t think this logic makes sense, it’s only right if the trade-balance was zero–which certainly was not the case. The large trade imbalance implies that nominal consumption grew much faster than GDP. …The fuel of this growth was inflation. Notionally the numbers work out in support of the 1980 CPI methodology.

  26. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    7. April 2009 at 07:38

    I don’t know enough about banking to be correcting anyone, Scott. However, I don’t see the purpose of changing the form of bank capital from reserves to T-bills. Just what does that accomplish?

    Btw, I don’t know Falkenstein, but I’m probably responsible for him noticing your discussion with Thoma. Falkenstein and I have a mutual friend (to whom I had pointed to your bloggingheads piece). Falkenstein isn’t an academic, but does have a Northwestern Phd.

  27. Gravatar of ssumner ssumner
    7. April 2009 at 16:56

    Bill, Thanks, that’s what I thought.

    Jon, Thanks, I didn’t know those French and German’s favored such wild reckless cowboy capitalism. They should emulate our more cautious system, with it’s much more conservative leverage ratios. seriously, do you know why the European regulators allowed such extreme ratios?

    Jon, I feel more strongly about this inflation issue than almost anything I have said on this blog. The numbers that people like Vernon Smith are suggesting are simply not credible. He argued inflation was actually 6.2 percent in 2004, which is absurd in my view. (And back in 2004 I was also arguing we were understating inflation, while most economists argued the opposite.) He relies on the C-S index, which was wildly unrepresentative of the national scene in 2004. I also don’t understand your trade deficit argument at all. Wasn’t inflation lower among import goods than domestic goods. The only inflation rate I care about is the inflation rate for domestically produced output. That is the GDP deflator, and it is not biased at all by trade deficits. The CPI is slightly affected by trade deficits, but it is of no interest to me. And even if you do care about the CPI, no one has refuted my basic argument that you can’t argue the CPI is grossly understating inflation without arguing that America is getting poorer. And when I see the middle class moving from 3 bedroom ranches to McMansions with 4 baths and granite countertops, I have trouble visualizing a country getting poorer. I know that leftists think that is happening, but I don’t see it. As recently as the 1940s a large fraction of Americans didn’t even have indoor plumbing. When I grew up in the 1960s one bath was the norm.

    And speaking of housing, how can you say the higher marginal benefit was only apparent? Why would society benefit more from an extra mile of fiber optic lines, when we already had 10 times more than needed, as compared to an extra house that someone might actually enjoy living in. And even if the higher rate of return was only apparent, that’s all we have to go on. Decisions should be made on the basis of what is apparent.

    Patrick, I don’t have any complaint about Falkenstein, there is no way anyone could have understood my interest penalty idea solely on the basis of the talk. Only if they read my blog would they be able to understand why I think it is needed. I am surprised that you don’t see the merit of getting banks to hold less reserves. Monetary policy cannot be effective if banks are hoarding reserves. The whole point of an expansionary policy is to raise the money supply and/or reduce its demand. My idea does the later. If banks substitute T-bills for reserves, where do you think the reserves will go? Certainly not back to the Fed.

  28. Gravatar of JimP JimP
    7. April 2009 at 17:46

    Yes – its Japan we have to worry about. And to me that is entirely beyond comprehension. Bernanke studied Japan and yet is giving us another Japanese gradual soul destroying deflation. He surely knows better. And he surely knows exactly what he is doing. He surely knows that paying interest on reserves is deflationary – which has to mean that he consciously desires deflation.


    The only reason I can imagine is that he is worried about the dollar. I can think of no other.

  29. Gravatar of JimP JimP
    7. April 2009 at 17:48

    We are being crucified on a cross of gold – to invent a quote.

  30. Gravatar of Jon Jon
    7. April 2009 at 19:49

    “I also don’t understand your trade deficit argument at all. Wasn’t inflation lower among import goods than domestic goods. The only inflation rate I care about is the inflation rate for domestically produced output.”

    Yes, inflation in import prices was lower…

    So how do you know the the real value of domestic output? Suppose production doubled with a constant-level of labor and capital. Price per-unit of output should fall, but aggregate value should be constant given constant money. Are we more prosperous? Yes, but inflation and nominal growth is zero. So I do not accept your assertion that prosperity implies that GDP growth less inflation must be greater than zero.

    Now suppose instead that the supply of goods doubles because some foreigners send us some goods without getting anything in return. Price per-unit should fall, but it didn’t; the price still rose. Yes we’re more prosperous, but inflation was huge.

  31. Gravatar of Jon Jon
    7. April 2009 at 20:32

    It occurs to me that my first example conflates “price inflation” and “money supply inflation”.

  32. Gravatar of Bill Woolsey Bill Woolsey
    8. April 2009 at 02:21

    Nominal GDP doesn’t include purchases of imported goods.

    Also, personal saving has been positive for all but 2 quarters in the last 10 years. While some households have expanded consumption beyond income through consumer debt, other households have been saving quite a bit, “offsetting” that.

    I think it is impossible to deny that Real GDP was less that reported because residential investment was worth less than it seemed. Hindsite is 20-20. But this sort of thing happens all the time.

  33. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    8. April 2009 at 07:09

    Scott, it seems to me the reserves go to the people from whom the securities are purchased. What they do with the proceeds would be crucial. True, they could spend, but they also could also deposit it right back into a bank. We’d be back to square one.

    Ultimately the banks might just find ways to pass the tax onto their depositors, or create some new savings vehicle that would avoid the excess reserves definition.

  34. Gravatar of Jon Jon
    8. April 2009 at 08:09

    “Nominal GDP doesn’t include purchases of imported goods.”

    Precisely the problem.

  35. Gravatar of Bill Woolsey Bill Woolsey
    8. April 2009 at 14:37


    The assumption is that the “reserves” stay in the banks. “Reserves” are balances that banks hold in deposits at
    the Federal Reserve bank, or else, vault cash.

    Don’ think of the penality fee as a tax on the banks. It is a fee for storing money at the Fed. Why should the Fed provide storage servcices to banks for free?

    If the banks don’t want to pay these fees, then they can stop storing money at the Fed. They can buy government bonds.

    If banks respond in the expected fasion, and purchase government securities, then a checkable deposit is created for the person from whom the bank purchases the securities. The bank of the person who sold the securities now as the reserves. But checkable deposits have increased. This increases required reserves and reduces excess reserves. And so, the indiviual bank that got rid of the reserves by purchasing the securities pays less tax, but because there are more balances in checkable deposits, then the amount of require reserves increases, reducing the amount of excess reserves and the total amount of “tax” that must be paid by the banking system.

    The point of the exercise is to expand the amount of checkable deposits. Of course this money is all in some bank. And all the reserves are held by banks.

    Your argumetn is supposed to be that the people who sold the government bonds to the bank will just now hold the balance in their checkable deposit. If they hold all of it, then the demand for money rises to match the supply because people substitute holding money for the government bonds that are no longer availible for the nonbanking public.

    Scott’s proposal is that the Fed buy enough assets so overwhelm any such effect. Create an excess supply of money.

    To the degree that banks instead continue to hold reserves, and pay the storage fees, and pass it on to depositors–then that is good. This will reduce the demand to hold money and raise velocity. Interest rates on checking accounts, savings accounts, and the like, will be lower, perhaps negative.

    As for the new savings vehicle that would avoid the excess reserves definition, that is backwards. What they have to do is expand the amount of checkable deposits that people hold and so increase the amount of reserves required. That is how they create less excess reserves. And so, they might turn existing savings accounts into checking accounts. or they might refuse to renew CD’s and pay them off with increased balances in checking accounnts.

    Sure, there are many possible responses. But none of these possible responses suggest that it is sensible for the Fed to pay interest on reserves. Or that it isn’t desirable for them to charge storage fees on excess reserves.

  36. Gravatar of ssumner ssumner
    8. April 2009 at 16:06

    JimP, I am puzzled about Bernanke as well. The dollar is actually much stronger than last summer, so I don’t know why they would be so concerned about the exchange rate. I hope to do a post on the exchange rate now.

    Patrick, Bill has a good response. I was just thinking of pushing the reserves out into circulation where it becomes cash. But one can also push it into required reserves, if the amount of deposits rises. The process is counterintuitive, because the individual doesn’t see it happen. We each try to get rid of excess cash balances. But society as a whole cannot get rid of excess cash balances, except through inflation.

    Jon, I’m not sure the inflation argument is meaningful unless we have a use for that particular inflation rate. Inflation is not some objective thing, where there is a true rate of inflation, it is rather a human construct, and there is no right answer to the question of what is the true rate of inflation. Some techniques are more useful than others. I am interested in the inflation rate of goods produced here. Thus I use the deflator. Others care about the inflation rate of goods consumed here—and use the CPI.

  37. Gravatar of Jon Jon
    8. April 2009 at 20:06


    I think you missed my point: calculating the GDP deflator is tenuous at best in the presence of a trade imbalance because the real-value of goods produced could be declining. That real-value can be declining because of net foreign supply.

  38. Gravatar of Bill Woolsey Bill Woolsey
    9. April 2009 at 03:00

    Jon writes:

    “Scott also writes: “NGDP was growing a bit over 6%… The overall economy was very strong, so clearly real growth was much more than 1%. That means true inflation had to be far below 5%, even during the housing boom.”

    Sorry, I don’t think this logic makes sense, it’s only right if the trade-balance was zero-which certainly was not the case. The large trade imbalance implies that nominal consumption grew much faster than GDP. …The fuel of this growth was inflation. Notionally the numbers work out in support of the 1980 CPI methodology.”

    A large trade imbalance does not imply that nominal consumption grows much faster (or any faster, or as fast) as GDP.

    The obvious and simple point is that real GDP was overstated because the value of the homes produced during the bubble were overvalued.

    However, if you are interested in tracking the productive capacity of the economy or determining whether or not resources were being fully utilized, then the acutal measures of real GDP are doing their job.

    Poor investments and business losses happen all the time. If there is a period where it turns out that they were extra high, human welfare wasn’t enhanced very effectively. If you are trying to track how well off people are in each period, then this needs to be taken into account.

    But the notion that there was necessarily monetary disequilibrium, that is, excess money balances, during a period because real investment decisions were exceptionally poor, is a mistake.

    Higher prices or output of particular goods raise the demand for money, other things being equal. We can imagine that if the supply of money hadn’t increased, then that particular output increase or price increase might not have occurred. And so, form that highly impractical perspective, we can say that the increase in the money supply allowed for those higher prices or higher production. And, we can imagine, that if the money supply and not increased, then those output increases or price increases would not have occurred. If the output increases turned out to waste resources, then one might imagine that the increase in the money supply caused the waste.

    I find that very implausible.

    I think your notion that a large trade deficit implies that nominal consumption grows much faster than GDP is false. It is rather that consumption could grow faster than GDP without reducing investment or government spending if a trade deficit is possible.

    However, before we take seriously any of these stories, you must keep in mind that personal saving was positive all but 2 quarters in the last 10 years. Private savings stayed strongly positive. Nonresidential investment grew rather than shrank. And, of course, the Bush administration and the Republican Congresses instituted an orgy of deficit financed government spending.

    I support Scott’s call for a return to the 5% nominal GDP growth path. But that is because I think it is a good idea to get through the current crises before instituting a disinflatoin to a 3% growth path for nominal income. As I explained before, I think the notion that a different monetary policy would choke off what in hindsight turn out to be bad investments is implausible.

  39. Gravatar of Steve Roth Steve Roth
    9. April 2009 at 07:43

    >why doesn’t the Fed clearly say that it plans to leave just enough money in circulation to keep long run price levels 2% or 3% (per year) higher than current price levels?

    Realistically, isn’t that where effective expectations currently sit?

    Doesn’t everyone know that’s what the Fed wants to get to? (Regardless of excruciating parsing of particular Fed statements?)

    At some point, we all believe, inflation will kick in. (May the good Lord smile upon us.) If it kicks in excessively, the Fed will tighten. Doesn’t everybody expect that?

    The uncertainty comes from concern that the Fed won’t be able to manage that successfully. But historically, monetary tightening seems almost magical in its efficacy–even at relatively low inflation rates. There’s not technical difficulty.

    So the concern is that the Fed will be constrained from tightening by continuing sluggish AD and NGDP growth–IOW, stagflation.

    That’s where fiscal stimulus can help. It can give the Fed its monetary mojo back.

    Unfortunately fiscal stimulus is a weak lever. Which is why (unfortunately) we need so much of it.

  40. Gravatar of ssumner ssumner
    9. April 2009 at 16:12

    Jon, When you refer to the trade deficit, do you mean the deficit itself, or the increase in the deficit from the previous period? It seems to me that it is the change in the deficit that might matter. I’m not saying that you don’t have something here, but I just find the size of the bias that some people throw around to be implausible. I’m not opposed to someone saying the the actual CPI is one percent higher. But if someone says its three percent higher, then alarm bells go off. The economy is like a big puzzle. If you greatly change the size of one piece, the other pieces don’t fit.

    Bill, I agree with most of what you have to say. Of course if you get into the “human welfare” definition of GDP, then everything is up for grabs. There is not much evidence that Americans are getting happier, so perhaps real GDP doesn’t grow at all in that sense. The more one thinks about the conceptual problems in the CPI, the more I like NGDP targeting–whether 3% or 5%.

    Steve, Inflation expectations are not 2-3%, they are only about 0.8%, which is precisely the problem. I do agree that the Fed will tighten if inflation gets out of control. But I don’t understand your last point–you seem to slide from arguing money may be too easy, to saying we need fiscal policy. If the Fed is “constrained from tightening” (i.e. too easy) we don’t need fiscal policy do we?

  41. Gravatar of Jon Jon
    11. April 2009 at 18:49


    I don’t mean the accumulated deficit, but the delta. While your puzzle analogy is a good one; *I* am not convinced that an 8% (peak) rate is implausible.

    The early 90s methodology which supports 4-9% per annum CPI for the past decade is quite reasonable as it goes.

  42. Gravatar of ssumner ssumner
    13. April 2009 at 11:13

    Jon, The midpoint of that range is way above the average income gain for Americans. I just don’t see average Americans getting poorer, if anything this country seems to be getting richer. I know that sounds vague, but that’s what the inflation debate really mounts to. Indeed it is all it amounts to. NGDP is something definite, it is completely unaffected by the methodology issues you refer to. Beyond that, we are simply making a judgement about how much better (or worse) off we are.

    I have one question, does NGDP include implicit rents on housing? If so, maybe I am wrong about the inflation biases not affecting NGDP, as implicit rents don’t involve market transactions.

Leave a Reply