Do Keynes and Ferguson agree on fiscal policy?

I noticed a recent post by Paul Krugman that expressed his usual exasperation at the theoretical incompetence of those he disagrees with.   Here is a quotation from the post:

. . . further confirmation that we’re living in a Dark Age of macroeconomics, in which hard-won knowledge has simply been forgotten. What’s the evidence? Niall Ferguson “explaining” that fiscal expansion will actually be contractionary, because it will drive up interest rates. At least that’s what I think he said; there were so many flourishes that it’s hard to tell. But in any case, this is really sad: John Hicks knew far more about this in 1937 than people who think they’re sophisticates know now.

This reminded me of something that Keynes said in 1933, the year he worked out the model for the General Theory:

“It is at this stage that a certain dilemma exists; since it may be true, for psychological reasons, that a temporary reduction in loan-expenditure plays a necessary part in effecting the transition to a lower long-term rate of interest.  Since, however, the whole object of the policy is to promote loan-expenditure, we must obviously be careful not to continue its temporary curtailment a day longer than we need.”

When school is out I need to do more research on this quotation.  I seem to recall that Keynes was making a policy ineffectiveness argument for fiscal policy, which was quite close to my interpretation of the famous liquidity trap of 1932, when the Fed’s OMOs seemed to fail.  I don’t have the book with me (Means to Prosperity, or Collecting Writings, Vol. 9, pp. 353-54), if anyone does perhaps they can tell me if the context is international pressure on the currency (devaluation fears) pushing up interest rates.  If so, it would not create any big inconsistency with the IS-LM framework underlying the General Theory, but it would blow Keynes discussion of the 1932 U.S. OMOs right out of the water.  Conversely if this is a closed economy context, then Keynes was hopelessly confused about his own model—as confused as Krugman says Ferguson is.

I was intrigued by Krugman’s “Dark Age of Macro” link, and noticed it linked to a scathing series of comments on Fama and Cochrane.  I suppose I should be the last one to throw stones here, as I am pretty negative on the current state on macroeconomics, however I would observe that just because a famous economist doesn’t accept the Keynesian model, doesn’t mean they are a dolt.  In any case, here is what Krugman said:

There has been a tendency, on the part of other economists, to try to provide cover “” to claim that Fama and Cochrane said something more sophisticated than they did. But if you read the original essays, there’s no ambiguity “” it’s pure Say’s Law, pure “Treasury view”, in each case. Here’s Fama:

“The problem is simple: bailouts and stimulus plans are funded by issuing more government debt. (The money must come from somewhere!) The added debt absorbs savings that would otherwise go to private investment. In the end, despite the existence of idle resources, bailouts and stimulus plans do not add to current resources in use. They just move resources from one use to another.”

And here’s Cochrane:

“First, if money is not going to be printed, it has to come from somewhere. If the government borrows a dollar from you, that is a dollar that you do not spend, or that you do not lend to a company to spend on new investment. Every dollar of increased government spending must correspond to one less dollar of private spending. Jobs created by stimulus spending are offset by jobs lost from the decline in private spending. We can build roads instead of factories, but fiscal stimulus can’t help us to build more of both.1 This is just accounting, and does not need a complex argument about “crowding out.

Second, investment is “spending” every bit as much as consumption. Fiscal stimulus advocates want money spent on consumption, not saved. They evaluate past stimulus programs by whether people who got stimulus money spent it on consumption goods rather save it. But the economy overall does not care if you buy a car, or if you lend money to a company that buys a forklift.”

There’s no ambiguity in either case: both Fama and Cochrane are asserting that desired savings are automatically converted into investment spending, and that any government borrowing must come at the expense of investment “” period.

What’s so mind-boggling about this is that it commits one of the most basic fallacies in economics “” interpreting an accounting identity as a behavioral relationship. Yes, savings have to equal investment, but that’s not something that mystically takes place, it’s because any discrepancy between desired savings and desired investment causes something to happen that brings the two in line.

I was puzzled by the Cochrane quote, as I recall reading the paper and not noticing anything amiss.  Sure enough, Cochrane does understand the multiplier, but initially (unfortunately only implicitly) assumed a stable demand for money, and later in the paper relaxed that assumption, allowing fiscal expansion to boost AD by reducing money demand (for a given money supply.)

Back to the Ferguson post; Krugman lectures Ferguson like he was a slow student:

This is the IS curve, taught in Econ 101. Now, we usually explain how this curve is derived in a different way: we say that given the interest rate, you can determine investment demand, and then through the multiplier process this determines GDP. What you’re supposed to understand, however, is that the derivation I’ve just given is just a different way of arriving at the same result. It’s just different presentations of the same model.

So what determines the level of GDP, and hence also ties down the interest rate? The answer is that you need to add “liquidity preference”, the supply and demand for money. In the modern world, we often take a shortcut and just assume that the central bank adjusts the money supply so as to achieve a target interest rate, in effect choosing a point on the IS curve.

Which brings us to the current state of affairs. Right now the interest rate that the Fed can choose is essentially zero, but that’s not enough to achieve full employment. As shown above, the interest rate the Fed would like to have is negative.

Note how Krugman says “we often take a shortcut and just assume” monetary policy picks a point on the IS curve.  Well I guess if you make that assumption, then at zero interest rates monetary policy is not very effective.

Krugman has been blogging for a long time and knows there are sharks like me out there trying to trip him up.  So he doesn’t actually say that this “proves” monetary policy is ineffective.  If he did, he would be making the same mistake as Cochrane made—handing a weapon over to the other side.  After a rocky start at the NYT, he is much more careful now.  But what about Krugman’s readers?  Do they assume Krugman is “proving” monetary policy is ineffective in a liquidity trap?

Krugman’s model of the liquidity trap was built to explain the situation in Japan, when interest rates were stuck at zero, banks were weak, and prices kept falling.  He called it an “expectations trap.”  Later on, a number of famous economists came up with foolproof ways out of the Japanese liquidity trap—generally involving currency depreciation.  Although a central bank may not be able to appreciate it’s currency (if it runs out of reserves), there are no technical barriers to currency depreciation (unless they run out of paper and ink); they can make the market by offering to sell yen more cheaply than anyone else.

These technical facts are indisputable, although some might argue that Japan faced international pressure not to depreciate its currency.  In fact, this was not the reason why they failed to do so, but let’s put that question aside for the moment.  Krugman knows that these “foolproof” plans for currency depreciation mean that the IS-LM model cannot truly show monetary policy ineffectiveness.  IS-LM is a sticky price model; it ignores the fact that there may be flexible prices the central bank can target that are correlated with AD and the price level.  Exchange rates are one such flexible price, gold is another, and, you guessed it, NGDP futures are far and away the best flexible price to target.  You know I couldn’t do a post without putting in a plug for my favorite panacea.

BTW, in case you are curious, expansionary monetary policy generally increases the IS curve, that’s why the IS-LM model doesn’t tell us anything interesting about liquidity traps, or monetary policy in general.

Update (5/5/09):  Bill and Adam pointed out that Krugman presumably was using the real interest rate.  I had assumed he meant the nominal rate, so that somewhat affects my analysis.  But I don’t think it changes the thrust of my argument about how IS-LM is misinterpreted.  Robert King pointed out in 1993 that an expansionary monetary policy that is expected to persist (and what other kind matter?) can actually shift up the IS curve, and that real interest rates will generally rise.



26 Responses to “Do Keynes and Ferguson agree on fiscal policy?”

  1. Gravatar of Bill Stepp Bill Stepp
    4. May 2009 at 16:41

    Krugman really is a free luncher isn’t he?
    The Hicks he and DeLong are fond of, the one who penned “Mr. Keynes and the Classics,” 40-odd years later did a mea culpa over the IS-LM parallel bars. Good thing too, because if they had killed one more student, he would have been sued.
    They always ignore this.
    And a question for Scott: as you’re probably aware, DeLong has a view, committed at his blog, that Milton Friedman (via Jacob Viner) somehow saved Keynes’ bacon, which I gather was the root of his 1969 quip that “we are all Keynesians now.” (JFTR, not me.) As a monetarist, can you vouch for the accuracy of this claim? DeLong has claimed that right-wing economists don’t know their own history (or something like that) when they reject Keynesian-style arguments for the stimulis. (But certainly not Austrians!)
    Thanks for any light on this murkey subject.

  2. Gravatar of Adam P Adam P
    4. May 2009 at 23:02

    Scott, about those famous economists way out of the liquidity trap. Credible commitment was not an issue?

    That is, your example of a yen dollar rate of 1000 works even if everyone knows it’s temporary or do they have to commit?

  3. Gravatar of Kevin Donoghue Kevin Donoghue
    4. May 2009 at 23:58

    The Means To Prosperity is available online. It’s an early version, not the Collected Works version, which includes a discussion of the multiplier.

  4. Gravatar of ssumner ssumner
    5. May 2009 at 03:49

    Bill, I am not certain of the DeLong argument, but I know he once argued that new Keynesianism adopted some monetarist ideas. This allowed it to revive after the difficulties posed by events in the 1970s. Maybe this is what Brad meant. With a few exception like David Laidler and Brad DeLong, most left and right wing economists don’t know their own history. Few left-wingers know that old Keynesianism is a gold standard model. Yet right-winger Henry Simons figured that out.

    AdamP, It’s been a while since I read the relevant papers, but I seem to recall the the interest parity theorem made the policy credible. Exchange rates cannot be expected to appreciate by more than the interest rate differential. So the very zero boundary that creates problems for interest rate targeting, also puts an upward bound on the expected rate of appreciation. Push to spot yen to 1000, and the forward yen will fall almost as much.

    Kevin, Thanks. I suppose I’d have to read through to find the exact page, perhaps I’ll wait and look at the collected works.

  5. Gravatar of Bill Woolsey Bill Woolsey
    5. May 2009 at 03:50

    thought the Cochrane and Fama arguments were bad, much for the reason Krugman did. They were abtracting away from monetary disequilibrium. The proper analysis of fiscal stimulus is that in theory government debt could subsittute for money and so create a surplus of money at current levels of nominal income, resulting in an expansion of spending. To the degree production is below capacity, that would result in more income and output. _And_ that in reality, if this doesn’t occur, the quantity of money will be increased to fund the government spending. That is, newly issued bonds will be sold to the central bank, which will pay for them with newly created money.

    Of course, if there is no liquidity trap, then an increase in the quantity of money can generate more spending just as well without producing a bunch of government goods and services and leaving an overhang of government debt and the need for tax financed interest payments. (If the alternative is just to leave production and employment low, that isn’t a “real” cost, but if the alternative is increased output and employment funded by private expenditure, it is.)

    The “right” answer is that fiscal stimulus is an indirect way of correcting monetary disequilibrium. It also impacts the allocation of resources. It is only sensible if you want to change the allocation of resources anyway.

    By the way, IS-LM is in real interest rate/real income space. An increase in expected inflation lowers real interest rates. It shifts the LM curve “down” or the the right. There is no zero lower bound in real interest rate space.

    I think that Krugman and company are making an argument about nominal interest rates with a tool that doesn’t use nominal interest rates.

    Of course, I am more interested in policy options that keep nominal income on a 3% growth path and the expected inflation rate at zero, and so, the nominal zero negative bound is a zero real bound (unless we are already talking about reversing a deflationary error.) But then, as I never tire of saying, not all interest rates are at zero. Just because the Fed can’t carry out its traditional practice of targetting the federal funds rate by trading T-bills (and even repurchase agreements with T-bills) doesn’t mean that it can’t lower nominal interest rates by other means.

  6. Gravatar of ssumner ssumner
    5. May 2009 at 03:52

    BTW, I forgot to thank Dilip for links to two of my three most recent posts. In addition, those interested in policies targeting gold, etc., should look at the Nick Rowe post from a few weeks back on that subject.

  7. Gravatar of ssumner ssumner
    5. May 2009 at 03:58

    Bill, Regarding IS-LM, I agree it should use the real interest rate. My question is: do I need to revise my post? I thought Krugman was using the nominal rate (hence liquidity trap), and replied on that basis. Isn’t that right? If not, let me know.

    Cochrane doesn’t look at things as I do, but he clearly says fiscal policy might reduce the real demand for money, and hence expand AD. Obviously for a given money supply, that is the only way fiscal stimulus can boost AD. My point is that Cochrane was not ignorant in the way Krugman implied, he just expressed his ideas in an awkward way.

  8. Gravatar of Bill Stepp Bill Stepp
    5. May 2009 at 03:59

    Few left-wingers know that old Keynesianism is a gold standard model. Yet right-winger Henry Simons figured that out.

    Is DeLong on the list of left-wingers who don’t know the golden fetters of old Keynesianism?

  9. Gravatar of Bill Woolsey Bill Woolsey
    5. May 2009 at 04:22

    He is drawing his supply of saving and demand for investment and his IS curve in “r” space. The convention is that this is the real interest rates. As you mentioned, he is a bit vague about the interest rate business isn’t he.

  10. Gravatar of Adam P Adam P
    5. May 2009 at 04:50

    Scott, what I meant was whether pushing the yen dollar rate to 1000 necessarily has the desired real effects. In the other thread I conceded that yes the BOJ could get that fx rate if they want it.

    However, you might have mistakenly gotten the impression that I was also conceding that devaluing the currency without a credible commitment to keep it there for a sufficiently long time would still break the liquidity trap. It won’t. You still need the credible commitment to a higher future domestic price level to break the liquidity trap and you need to keep the currency devalued for as long a time as that takes.

    The reason is simple. The liquidity trap is a problem where the return to holding money is too high (relative to required real returns). If you devalue yen to 1000/dollar but this is assumed to be temporary then you’ve just increased the future return to holding yen. Thus it gets hoarded even more.

    Here is a link to a fairly recent paper on the subject by Svensson. He says that the exchange rate must be pegged until the domestic price level rises. You still need a credible commitment to future actions to break the trap, exactly as I’ve been saying all along. Futures targeting solves nothing in this respect.

  11. Gravatar of Adam P Adam P
    5. May 2009 at 04:57

    Scott, regarding what Krugman meant by r on his IS curve I’m pretty sure he meant the real rate. I pointed this out on the WCI blog.

  12. Gravatar of Kevin Donoghue Kevin Donoghue
    5. May 2009 at 06:30

    “I suppose I’d have to read through to find the exact page, perhaps I’ll wait and look at the collected works.”

    Click on Chapter III and scroll down to where it says [pg 21] in the left-hand margin. The collected works version is certainly better, in that Keynes had had a few more weeks to think it over, but I don’t think there’s any huge difference.

  13. Gravatar of ssumner ssumner
    5. May 2009 at 09:12

    Bill S., I am not sure. I hope to spend more tiome looking at Brad’s blog after school is out (in three days.)

    Bill W. and Adam, Thanks for the tip. I’d appreciate if you could take a quick look at my short update. IS-LM is not my strong suit (I just know there is something wrong with it.)

    Adam, I see your credibility point. But here’s how I see it. The BOJ can no longer say they are powerless as an excuse. Here (depreciation) is something they can do, and as long as they keep doing it prices will rise. And to some degree prices will rise right away, because some traded goods prices are pretty flexible. This is what happened when we devalued in 1933–prices rose right away, especially the commodity-intensive WPI. But I agree, unless they carry on with the policy, they’ll be right back in a liquidity trap. If they are not expected to carry on with it, rates may stay at zero for expectational reasons, but, and this is important, prices and output can rise anyway. So when I say something like “get out of a liquidity trap”, I really mean boost AD despite being in a liquidity trap, there is as you say no guarantee nominal rates will rise above zero.

    As an aside, in all my reductio ad absurdum points about liquidity traps, I never mean to suggest they will have desirable real effects. As you note, that is an entirely different issue.

    Kevin. Thanks. from a quick look it seems that Keynes is unclear. But when he wrote this (March 1933) the U.S. and France were still on gold. So from his reference to those two having less efective policies than Britain, I’d guess that either consciously or subconsciously it may have played a role in his analysis. That saves him from contradicting the IS-LM model, but at the cost of discrediting the only example of a liquidity trap in the GT, which was the 1932 OMOs, which failed due to a gold outflow triggered by a loss of confidence resulting from both expansionary monetary and fiscal policy. I may do a post on this sometime.

  14. Gravatar of StatsGuy StatsGuy
    5. May 2009 at 18:45

    Adam, the point about expectations from the paper is intuitive (and doesn’t require a recursive model to explain), but not conclusive.

    Specifically, how does it deal with various risk premium, behavioral dynamics, and liquidity constraints?

    Such as the default risk premium. Or, as Mark Twain put it, “I am more concerned about the return OF my money than the return ON my money.”

    And the household solvency risk premium. View this as a liquidity constraint. As we all know, the market can stay irrational longer than we can stay solvent. Even if I do believe that in 10 years we will be in hyperinflation, that’s 10 years from now. I’ve got to make it through 10 years.

    Consider the following problem: much of the spectacular collapse of stock prices was caused by widespready retail investor flight from mutual funds (and forced redemptions). Maybe I’m quite certain in 5 years it will come back, but in the meantime I need a reserve cushion (the labor market doesn’t come back that fast, you know). So now we have home finance gurus (Suze Orman) telling consumers that they now need an 8 month cash cushion (up from 6 months) and that they should be living on just one income, and we see consumers realizing how irrational their consumption patterns were leading up to the present (and committing to change).

    Recursive models that require hyper-rational back-tracking consumers run into real-world liquidity constraints, media management of perceptions, behavioral modifications, etc. They are knife edge solutions, easily broken.

    Here’s another interesting example: Maybe I believe with 100% confidence that prices will be much higher in 5 years, but I also believe that “things will get worse before they get better” – that is, before the hyperinflation we’ll see more deflation for another 2 years. I’ll therefore delay my investments until things have gotten worse, but if everyone does this then things keep getting worse. We can construct various arbitrage markets that would smooth out this trend, but that’s one frickin huge arbitrage market. Where does that liquidity come from (even if everyone has 100% confidence in a central bank commitment AND they believe the bank can actually do what it is trying to do).

    And finally, let’s not forget the velocity problem. From the market’s perspective, it’s not that the Central Bank can’t commit to 3% inflation. It’s that 3% inflation isn’t an option. It’s either plug along and allow deflation, or jack up base money so far that inflation is inevitable, at which point we see a velocity rebound and bam… inflation surges WAY past 3%.

    If this is the real choice central bankers confront, and investors/consumers know it, then the central bank has a much harder objective – convince the world that it really really prefers 15% inflation to 2% deflation. If that really was the choice, I personally have a hard time believing that even Ben Bernanke would make that choice. The Fed’s 6 month delay in using QE cost it a LOT of credibility.

  15. Gravatar of ssumner ssumner
    6. May 2009 at 05:24

    Statsguy, I strongly agree with your last line, but I don’t believe people expect deflation now and hyperinflation later. They say they do, but where is the evidence in the long bond market?

  16. Gravatar of Scott Lawton Scott Lawton
    6. May 2009 at 16:10

    @ssumner: The bond market may (in some absolute sense) be one of the best guides to policy, but that’s not universally agreed as true. The market is not the only source of info. that affects decision makers.

    It’s difficult to assess what “people” believe. Which people? There’s plenty of worry about future hyperinflation in the mainstream press and on blogs — including from people with some financial savvy.

    There are even (apparently) good reasons for fearing hyperinflation, e.g. if one looks at the money supply … without seeing how much seems to be trapped by excess reserves. Interest on reserves is a new policy, so many people don’t know about it or haven’t fit it into their understanding of the situation. I certainly didn’t until I discovered this blog. (It seems like a technicality rather than something fundamental.)

    Then there’s the huge projected deficits, which seems to me a genuine problem. Pro-growth policies might be able to overcome that, but I doubt many growth experts are fans of this administration or Congress.

  17. Gravatar of ssumner ssumner
    7. May 2009 at 05:10

    Scott, Good question. It motivated me to do a new post, so that everyone could see my answer.

  18. Gravatar of Clayton Clayton
    7. May 2009 at 17:08


    “But then, as I never tire of saying, not all interest rates are at zero. Just because the Fed can’t carry out its traditional practice of targetting the federal funds rate by trading T-bills (and even repurchase agreements with T-bills) doesn’t mean that it can’t lower nominal interest rates by other means.”

    I don’t think this logic works (as you intended anyway) if I’m a rational, risk averse investor…

    Let’s assume the real return on cash is around zero (slightly negative rates don’t change the overall argument). Your answer is to drive risky rates towards zero as well.

    So you’re basically forcing the market to pay me (an investor) a *lower* rate of return for the same risk. On the margin, private investors will disinvest in these risky assets and substitute cash. At the margin, you’re only encouraging increased liquidity preferences.

    Now… on the other side, driving down the risk premium has real effects in the market. Additional investment is necessarily made at the margin to drive down these rates. Increased investment does drive up R- and N-GDP.

    [Note that this may be the “gap” that the Austrians say you’re filling with money. Private investors disinvest in risky assets (preferring riskless cash due to the lower premium) while capital investment increases in response to a lower risk free rate (theoretically distorting capital allocation as a result of incorrect time preference signals)]

    You’re right that this fixes the economy, but not in the same sense as QE — quite the opposite as it actually increases money demand.

    What does (artificially) is drive down the risk premium. I’ve argued since my first post that I believe irrational risks tolerances (or aversions) are a key problem and cause of “real” cycles. I agree that this policy will work, but this policy is not really in Sumner’s camp.

    What Scott is advocating (indirectly perhaps) is an increased inflation to drive the real return on cash negative. This permits the real rate of return on risky assets to fall to the same nominal rates (as Bill proposes) without compressing the risk premium (thus avoiding the increased monetary demand). This is why it’s so easy for me to jump on the inflation targetting bandwagon because it mitigates one quite real “zero bound” that, combined with risk premiums, almost certainly is having real consequences.

  19. Gravatar of Bill Woolsey Bill Woolsey
    7. May 2009 at 18:09


    Quantitative easing involves having the Fed make enough open market operations to increase base money the amount needed reach the desired nominal target. Scott favors keeping nominal income at a trend growth rate of 5%.

    If the Fed buys up all the T-bills, it has to buy other things.

    That will reduce their nominal yields.

    When a 5% nominal income growth rate, the lower bound on the real interest rate is -2%. If the target is missed, and the price level gets below its long run growth path, then inflation will be higher, expected to be higher, and the lower real bound on interest rates will be more negative. But trying to make real interest rates sufficiently negative to get nominal income back to target is not the goal.

    You are confusing an argument about why the liquidity trap cannot prevent an expansion in base money from raising nominal income with how quantitative easing actually does raise nominal income.

    Well, I am glad you convinced yourself that creating money to purchase risky assets doesn’t create a liquidity trap at high interest rates becauase the risk premia compensates people for not holding money. You know, not only does it create an incentive to finance investment, there might be a subsittuion to other risky assets, or even current consumption.

    I think that assuming that risk premia are solely a matter of a decision between money or securities is a mistake. Why not buy equities? Or real estate? Or go out to eat more?

    So, risk premia causes people to work more or less? Or is it the accumulation of capital that impacts prodction? Personally, I don’t think “real” cycles are anything to worry about. If God isn’t telling us the true risk premia, the market is all we have, and if people work to invest or the want to consume rather than take risk, so what?

  20. Gravatar of ssumner ssumner
    8. May 2009 at 04:50

    Clayton, You said the following:

    “Let’s assume the real return on cash is around zero (slightly negative rates don’t change the overall argument). Your answer is to drive risky rates towards zero as well.

    So you’re basically forcing the market to pay me (an investor) a *lower* rate of return for the same risk. On the margin, private investors will disinvest in these risky assets and substitute cash. At the margin, you’re only encouraging increased liquidity preferences.”

    To me, this seems to confuse shifts in demand with movements along a demand curve. Analogy: suppose easy money causes house prices to rise—does that make people buy fewer houses? No, because the only way easy money can cause house prices to rise is if it shifts the demand curve for houses out to the right. But in this case both the price and quantity of housing rises. Similarly, if monetary policy raises the price of risky assets like stocks and junk bonds, it will reduce the yield on those assets. But that would not cause people to shift to cash, because it was an outward shift in the demand for such risky assets that caused their yield to fall in the first place.

    I’m not sure if this relates to the point you’re making, let me know if I misunderstood your argument.

  21. Gravatar of Clayton Clayton
    9. May 2009 at 19:50

    I want to come back and address the whole issue in depth… but I think this is the basic argument:

    The Fed enters the market and buys up houses. If you don’t successfully change expectations about future home prices, the demand curve doesn’t shift. Instead, you put a floor under housing prices. The general public sells the Fed some of the existing inventory (falling from Q* to Q-floor on the demand curve) and some additional houses are built (going from Q* to Q-floor on the supply curve). You “fill the gap” with printed money.

    Now… you definitely create some GDP by employing construction workers (as I pointed out) and that can help. I’ll subsequently explain why it may not help enough.

    So… it’s then a fair question to ask what home sellers, etc. do with the money. Again, if you don’t change expectations and risk premium are rational, the perceived value of financial assets (like the houses) are all fixed. In order to buy a risky asset, home sellers have to offer an above-market price for that asset so they don’t buy these assets.

    The only unlimited asset in the economy — the only asset whose return doesn’t fall as you increase demand — is cash money. The reason why cash is “unlimited” is that the Fed pushes it into the economy to buy the houses… they create a precisely offsetting amount of cash that can be held at the current (zero) rate of return.

    So the only “outlet” where intervention can’t be 100% neutral comes from the increased construction activity. Normally this demand would drive inflation and break the “if you don’t successfully change expectations” that keeps everything else neutral.

  22. Gravatar of Clayton Clayton
    9. May 2009 at 20:13

    So how could increased activity not be inflationary…

    **Under current conditions** there’s effectively a zero bound on (a floor under) the real risk free rate of return because cash is risk free and inflation is (for the sake of argument) zero. If the risk premium is 100% rational, this also places a floor under risky rates.

    By definition, our various rate floors cause (1) an excessive supply of savings and (2) an inadequate demand by borrowers for investment capital. Naturally, below equilibrium demand for savings means inadequate investment… which means an inadequate contribution by this sector to aggregate demand (AD).

    The reduction in AD puts downward pressure on prices… which helps reinforce low inflation (or even deflation) and thus the “zero bound” on the risk free rate. This vicious cycle is a “liquidity trap” of sorts. It’s not necessarily the preference for liquidity directly, but liquidity arising from the gap between the supply and demand of investment… thanks to the real rate floor provided by cash.

    So buying houses created excess AD via construction. Unfortunately our zero bound on real interest rates has created a shortfall in investment and ultimately AD. If the construction activity from the last post is less than the shortfall in AD in this post, prices (once they become un-sticky) will actually fall. The resulting deflation exacerbate the “zero” bound by moving the floor up to a (say) 1% return on “risk free” cash. [We haven’t even discussed what happens when the houses are sold back to the market to unwind the Fed’s position]

    Liquidity trap… expectations trap… zero bound… whatever the accepted term. Your intervention in prices — your floors — will *eventually* drive up GDP either through construction (by buying houses), investment (by lowering the risk premium), consumption and investment (by distorting longer term interest rates) or a comparable activity.

    However, the magnitude of intervention may need to be *quite* large depending on how big the hole.

    More importantly from an efficiency standpoint, as a result of these interventions, the Austrians ring in and complain about misallocation of resources:

    – If you buy up houses, you cause overproduction of homes.
    – If you lower the risk premium, you encourage over-investment in risky ventures
    – If you lower longer term interest rates, you encourage too much investment in long-term projects (discount rates on the profits from long term investments are too low)

    And the core of their complaint… is that none of this would be possible without the artifical “zero bound” that can only occur inside a fiat money scheme.

  23. Gravatar of ssumner ssumner
    10. May 2009 at 05:21

    Clayton, Taking your last point first, liquidity traps are just as likely under gold standards, as in the 1930s.

    I am confused about the purpose of your example of buying houses. Is this something you think should be done? Are you showing a way out of liquidity traps? Clearly there are much better ways out of the liquidity trap, just adopt an expansionary monetary policy that raises inflation expectations.

    Almost nothing in macro is “by definition,” including your assertion that zero rate floors cause excess saving. Savings equals investment regardless of the level of nominal rates. That is truly “by definition.” I know the Keynesians make a big distinction between ex ante and ex post saving, but I don’t find those distinctions to be useful.

    One other point, GDP is much bigger than any of us can imagine. You do “create some GDP” by employing construction workers. But the numbers would be trivial (a few 100,000 at best) and that doesn’t even account for indirect job losses in other sectors. Now if you do this by printing money then yes, it is possible (not certain) that it could sharply boost AD, and lead to much more employment.

  24. Gravatar of Clayton Clayton
    11. May 2009 at 05:07


    I’m not advocating the policy. I used houses (though increasing the price of any real or financial asset can wokr the same way) to respond to your analogy:

    “Analogy: suppose easy money causes house prices to rise””does that make people buy fewer houses?”

    To which the answer is clearly “yes” certain types of intervention can cause housing prices to rise while private “consumption” of houses fall. It’s a distinction from Fed action that does shift the demand curve right and Fed action that operates like a price floor.

  25. Gravatar of Clayton Clayton
    11. May 2009 at 05:23

    “just adopt an expansionary monetary policy that raises inflation expectations”

    If the Fed has lost credibility, the only way to restore inflation expectations is to generate inflation. Technically, they could commit to trend inflation targetting (i.e. if we can’t inflate now we’ll inflate when we can), but, as you’ve pointed out, their public statements contradict this possibility. So I’m running on the assumption that they must generate inflation to restore credibility.

    However, my “real rate of return floor” argument illustrates how the zero bound can actually lower AD below equilibrium even if there is adequate money in the system.

    The post before that illustrates how expansionary monetary policy (delivered through forms of OMO) *can* be almost completely converted to liquidity… with only a small boost to AD. [I believe this is most salient to your core QE arguments, but *only* when complemented with the functional zero bound.]

    The net-net is that even modest expansionary policy can result in falling AD, falling prices, and a vicious cycle that raises the “real rate floor”. At some point, expansionary monetary policy will exceed the AD shortfall and restore inflation expectations, but it may require FAR more expansion than the simple “QE causes inflation” would imply.

  26. Gravatar of ssumner ssumner
    11. May 2009 at 17:46

    Clayton, Now I see your house example. But I was thinking of a market-clearing situation. If expansionary monetary policy causes more houses to be built, there is no actual government subsidy. (Although I suppose you could say foolish private bankers provided a sort of subsidy in this case.)

    I’ve addressed your second point in many posts. let’s just say there are all sorts of ways to restore policy credibility, not just one way as you suggest. I just did a few posts on an NGDP futures market that would do the job quite effectively. I certainly don’t think they need a lot of inflation to get out of the liquidity trap.

    I also don’t understand your “public statements contradict this possibility.” Right now they are doing exactly what they publicly state, i.e. near zero inflation, and the market believes them. If they changed their public statement, I presume the market would also believe their new statement–especially if they did what they said they were going to do. Has there ever been a central bank that promised to reflate, and wasn’t believed? Not the BOJ, they never promised to inflate. I know of no such example.

Leave a Reply