September 09, 2012
Woodford and QE3
Columbia University Professor Michael Woodford's paper at the Fed's Jackson Hole conference last week made the case that more large-scale asset purchases by the Fed would by themselves do nothing, and suggested that instead what really matters is the Fed's communication of its future intentions. There's a fair bit in Woodford's analysis that I agree with. But unlike Woodford, I think that asset purchases can be an important part of what the Fed could do in the here and now. Here I explain why.
Let me begin with Woodford's theoretical analysis of why large-scale asset purchases by the Fed wouldn't be expected to have any effects. The foundation of modern finance theory is the belief in what is called a pricing kernel, which one can think of as the value that the market places on getting $1 if some specified relevant future event were to occur. If investors were risk-neutral, that value would just be the probability of the event times a risk-free discount rate. With risk aversion, the value might be higher or lower than that based on how important it is to investors to have funds in that particular state of the world. Standard theory further teaches that this value ultimately is determined by the real resources that investors end up having available in that state of the world. Hence Woodford's conclusion:
Insofar as a mere re-shuffling of assets between the central bank and the private sector should not change the real quantity of resources available for consumption in each state of the world, the representative household's marginal utility of income in different states of the world should not change. Hence the pricing kernel should not change, and the market price of one unit of a given asset should not change, either, assuming that the risky returns to which the asset represents a claim have not changed....
All that one needs for the argument are the assumptions that (i) the assets in question are valued only for their pecuniary returns-- they may not be perfect substitutes from the standpoint of investors, owing to different risk characteristics, but not for any other reason-- and that (ii) all investors can purchase arbitrary quantities of the same assets at the same (market) prices, with no binding constraints on the positions that any investor can take, other than her overall budget constraint.
Under these assumptions, the irrelevance of central-bank open-market operations is essentially a Modigliani-Miller result, as noted by Wallace (1981). If the central bank buys more of asset x by selling shares of asset y, private investors should wish purchase more of asset y and divest themselves of asset x, by exactly the amounts that undo the effects of the central bank's trades.
Another implication of this view is that it should not matter what maturity of debt the Treasury chooses to issue. If one takes the future path of spending and taxes as given, whether the Treasury chooses to finance its current deficit with short- or long-term debt should be completely irrelevant. In a recent paper with University of Chicago Professor Cynthia Wu, I discussed this theory. We noted that 3-month and 10-year Treasury securities are treated by the private market as very different investments. Based on a very long historical record we can say with some confidence that, if the U.S. Treasury were to borrow $10 B in the form of 3-month T-bills and roll these over each quarter for a decade, it would end up on average paying a substantially lower total borrowing cost than if it were to issue $10 B in 10-year bonds. If it's really true that nothing in the world would change if the Treasury did more of its borrowing short-term, the natural question is why does the Treasury issue any 10-year bonds at all?
I think if you ask that question at a practical, institutional level, the answer is pretty obvious-- the Treasury believes that if all of its debt were in the form of 3-month T-bills, then in some states of the world it would end up being exposed to a risk that it would rather not face. And what is the nature of that risk? I think again the obvious answer is that, with exclusive reliance on short-term debt, there would be some circumstances in which the government would be forced to raise taxes or cut spending at a time when it would rather not, and at a time that it would not be forced to act if it instead owed long-term debt with a known coupon payment due.
The implication of that answer is that the assumption underlying Woodford's analysis-- that changing the maturity structure would not change the real quantity of resources available for private consumption in any state of the world-- is not correct.
Now, I presume that Woodford's response to this argument would be to say, well, yes there might be some effects, but the effects arise not from changing the maturity structure of the debt per se, but instead from changes in the timing of future spending and distortionary taxes that the changed maturity structure would lead to. But if the question we're asking is, would changing the maturity structure of Treasury debt cause anything else to change, I feel the correct answer is, yes, it surely would.
If the Fed buys 10-year Treasury bonds with newly created reserves, my view is that it is performing a similar operation to that just described, replacing long-term government debt with what is in effect the shortest-term government debt, namely Fed deposits that the holders could ask for in the form of cash immediately. I think the Fed's reluctance to do more has to do with the same kind of risk aversion exhibited by the Treasury, namely, large-scale asset purchases tie the Fed into a situation in which, under some possible future scenarios, the Fed would be forced to allow a larger amount of cash in circulation than it would otherwise have chosen.
Apart from the theory, Woodford separately discusses the empirical evidence of whether the Fed's large-scale purchases so far have had any effect. He raises a number of doubts about the event-study methodology that has been employed by many researchers in this area, doubts by the way with which I am sympathetic. But there is a second way to look for evidence about these questions using historical data, and that is to examine the relation between the maturity structure of publicly-held Treasury debt and the interest rates that investors earn on different Treasury securities. What one finds in the data is that periods in which there was a larger proportion of long-term debt are also periods in which long-term bonds paid a higher yield relative to short-term. This finding was reported by Gagnon, et. al. (2011). Woodford's paper discussed the event-study evidence from the Gagnon paper at length, but I did not see any mention of this important alternative form of evidence. Gagnon, et. al.'s finding has also been confirmed by a number of other researchers using very different data sets and methods, including Kuttner (2006), Greenwood and Vayanos (2010), Doh (2010), and Hamilton and Wu (2012), none of which are cited in Woodford's paper.
Having said all this, I should clarify that my position is not really that far from Woodford's. Cynthia and my estimates of the size of the effects are pretty modest. As I have said on a number of occasions, I believe that the primary significance of large-scale asset purchases is not their direct effects, but rather their usefulness as a supplementary signaling device, adding an exclamation point, if you will, to the Fed's communication of its future intentions. Woodford would like the Fed to say something such as it will restore nominal GDP to trend, but that begs the question of exactly how. The public could easily respond to the Fed, ok, so you don't like the current situation, but what are you going to do about it? And if the Fed's answer is, well, nothing right now, but trust us, we're going to do something a few years down the road, then the theorists may be surprised to find that the Fed's announcement doesn't change all that much, either.
The financial press writes with a perspective that is deeply rooted in American culture-- actions speak louder than words. I think it's necessary for the Fed to accompany its forward communication strategy with a specific plan of action in the here and now. And even if, as Woodford and I both believe, the effects of those immediate actions are of necessity limited, I think including those immediate actions in the Fed's overall plan can be helpful.
Posted by James Hamilton at September 9, 2012 07:54 AMdigg this | reddit
When the short end of the yield curve is locked in ZIRP, where is the investor appetite coming from (and heading towards)?
The daily news flow may shape the "instantaneous" perception of risk in which the shallow end of the treasury pool beckons the wary (Scaredy cats) so that the demand for capital preservation -- intensified in a deflationary milieu -- motivates the purchasers who see turmoil ahead.
This flight to safety benefited buyers of the US long bond by having a risk free liquid asset with better than money market returns, plus a dollop of price appreciation for the principal.
Treasuries returned quite an ROI as the long end demand boosted prices for the diminishing coupon.
Ironically, debt purchases by the Fed (T's or MBS) seem to be having the perverse effect of flattening the yield curve in a manner that speaks to a long-ish term suppression of interest rates.
Is it artificially stimulative in the sense that investors must now engage riskier assets for income? Or is it the opposite where a Fed-flattened yield curve dissuades observers from believing future growth is forthcoming?
Given laymens' propensity to anchor expectations in "rules of thumb" (yield curve slope) Fed balance sheet expansion via Twist purchases telegraph a long term economic prognosis of low growth, stagnant income and moribund profits.
That picture of restrained economic activity anchors expectations in the disinflationary doldrums! This is more than an instantaneous perception -- it is an expectation.
It may be perceived as safer to add treasury debt at the short end of the curve to minimize carrying costs to the Treasury, but what happens when the Fed is compelled to de-lever?
The notion that the Fed can sell its long dated paper and create a steeper yield curve suggests a future scenario where prescient investors will rotate out of gov't paper and back into the real economy.
So it seems that Twist is like a foundation for an exit strategy in the making, as much as an accommodation in the present...
Posted by: person1597 at September 9, 2012 10:50 AM
What about open-ended LSAPs tied to a NGDP target? I believe that if the Fed announced it would purchase securities each week until a specified NGDP target was hit, that would be a loud signal with sizable, immediate results.
I wrote about this awhile in response to a similar question you raised about Fed efficacy: Monetary Policy Change Jim Hamilton Can Believe In
Posted by: David Beckworth at September 9, 2012 11:32 AM
I am pretty sure if they bought a dozen or two $1T platinum coins from Treasury rather than the private sector, it would be very significant.
Posted by: Lord at September 9, 2012 11:50 AM
No it does not beg the question
Posted by: Anon pedant at September 9, 2012 12:03 PM
"... the shortest-term government debt, namely Fed deposits that the holders could ask for in the form of cash immediately. ... the Fed would be forced to allow a larger amount of cash in circulation than it would otherwise have chosen."
I just do not see what the practical difference between reserve deposits and FRB notes outstanding is, other than that the FRBs now pays nominal interest on the deposits. Recall that the member banks can count vault cash towards their reserve requirements.
Posted by: Walter Sobchak at September 9, 2012 02:54 PM
Apparently, someone found a video of
young Federal Reserve chairman in action.
Posted by: B Turnbull at September 9, 2012 07:33 PM
I think the Fed's reluctance to do more has to do with the same kind of risk aversion exhibited by the Treasury, namely, large-scale asset purchases tie the Fed into a situation in which, under some possible future scenarios, the Fed would be forced to allow a larger amount of cash in circulation than it would otherwise have chosen.
This is funny, because it translates into the Fed being worried that by engaging in stimulus whose effect ranges from weak to uncertain, they may be forced to engage in stimulus that will absolutely work. :)
Posted by: John B at September 10, 2012 07:40 AM
I think US unemployment is coming from structural failure of education system rather than worsening economic condition. If we look at unemployment rate, we will see major driver of high unemployment coming from teenagers with nearly 25%. If we look at education we will see that the below Bachelor's degree will be the major of high unemployment. The unemployment of Bachelor's degree is at 4.1%. We can see the education will help reduce unemployment with more skills. The skill mismatch of young workers and aging uneducated workers would be the major reasons of high unemployment. The QE cannot solve the problem becasue high skill labors can find jobs and the improvement of productivity and technology will need more skill labors not unskilled labors. The more QE with more expectation of QE will not solve high unemployment but QE will destroy the root of problem of structural shifts in US/global economy on the need of high skill workers but more QE, the higher inflation and the more support of unproductive economy to go bigger and one day the unprodcutive economy will destroy the major economy from speculative bubble in stocks, real estate or subprime lending in real estate, auto etc. QE is very crazy policy. The global economy cannot survive with unproductive economy but QE will support unproductive economy to survive with the support of higher unskill labors to get jobs. Only way to solve unemployment in US is bring uneducated and unskill workers to get education and work productively in productve economy. Surely the structural unemployment will be lower from education not from QE but QE that will be annouced on 13 September will hide the root of high structural unemployment in US. Fiscal policy is important to drive growth but stupid policy that we have seen like Bush tax cut or sustaining welfare support for unemployed will destroy economy. Bring budge to improve worker skills, support cheap education for Bachelor's degree and technology not hike tuition fee like this.
Posted by: Young Economist at September 10, 2012 08:03 AM
Should Janus be an economist and to speak with both months,this could give the following
George A. Akerlof "If he wants to sell theses assets, why should I purchase them?"
"If the central bank buys more of asset x by selling shares of asset y, private investors should wish purchase more of asset y and divest themselves of asset x, by exactly the amounts that undo the effects of the central bank's trades."
"Based on a very long historical record we can say with some confidence that, if the U.S. Treasury were to borrow $10 B in the form of 3-month T-bills and roll these over each quarter for a decade, it would end up on average paying a substantially lower total borrowing cost than if it were to issue $10 B in 10-year bonds."
The other half of Janus would add:
Cauchy or Fourrier have demonstrated that mathematical sequences may converge towards a limit that is a certainty, but the limit is not known.
Rarity,all participants are pari passu that means, neither the issuers nor the Central Banks and or the markets know where is the limit,but there are mathemetical proof of this limit.
Sleep well, good people Janus is watching.
Posted by: ppcm at September 10, 2012 11:55 AM
Without having read the paper, I would assume that Woodford’s argument about the inefficacy of Quantitative Easing rests on the assumption of a representative agent. If everyone were identical, then it would not matter if the asset purchases that comprise QE were advantageous to the central bank because everyone would be an equal shareholder in the central bank’s windfall from QE. In this framework, it makes sense to claim that QE is ineffectual. In reality, however, QE should be an impetus for investors to recycle bank reserves into investments with positive rates of return, which should help spur nominal GDP growth.
Posted by: MDueker at September 11, 2012 08:07 PM
Hmmm... More twist and more MBS. For the (foreseeable) future.
Will that flat-line the yield curve?
Not for the retail credit consumer!
Posted by: person1597 at September 13, 2012 12:07 PM
US Credit Downgraded because of Bernanke new QE.
Egan-Jones is not as influential as other rating agencies but in a way that insulates them from government intimidation. But such a downgrade opens the door for other rating agencies for not being the first.
Posted by: Ricardo at September 14, 2012 02:37 PM