June 8, 2004
T hank you, Terry 1. I am honored to be here at the 10th anniversary of this annual seminar. You and I were here together at the inaugural in 1995. It is truly amazing what you and your colleagues have accomplished over the last decade. Congratulations, Terry! And compliments to my many good friends in senior management at UBS, many of whom are here today, for your wise and continuing support of Terry in this endeavor. Thank you for inviting me to speak today, playing warm-up act for a return visit by my friend Alan Blinder, who was the keynote speaker nine long – or is that short? – years ago.
The timing of this year’s event is impeccable in that the world’s central banks are on the cusp of a major reversal in policy, shifting from blatant reflationary policies to a stance more consistent with winning the peace of effective price stability. Note, please, that I don’t say shifting to fighting inflation. That is a war that does not need to be fought, because it has already been won – fought opportunistically, business cycle battle by business cyclical battle, during the last twenty years of the last century.
Indeed, during the first three years of this new century, the predominant concern of global central bankers was to ensure that the war against inflation was not over won, taking the global economy into a debt-deflationary spiral, the nefarious malady that infected Japan during the last decade of
the last century.
Fortunately, we can now declare that the world’s central bankers, led by the Federal Reserve, have won the cyclical reflationary battle to avoid over winning the secular war against inflation. It troubles me when I hear pundits declaring that the world’s central bankers, and particularly the Federal Reserve, have somehow now “fallen behind the inflation curve.” By definition, avoiding a debt-deflationary spiral, or it’s less noxious sibling, “unwelcome disinflation,” requires that central bankers accommodate a cyclical increase in inflation once secular price stability has been achieved. This is success, not failure!
The doctrine of preemptive tightening, a key cyclical armament during the secular war against inflation, must be retired once that war has been won. And so it has been, at least at the Federal Reserve. Winning the peace of secular price stability means that central bankers can and should let inflation cyclically fluctuate. And fluctuate means move not just down, but also up!
This is a very different world than 1994 – one year before the birth of this seminar – which was a most nasty year for the bond market on the back of a most nasty Fed tightening process. I remember many things about that year, as I’m sure many of you do. But in my mind, one phrase uttered by Fed Chairman Greenspan stands above all others from that year:
“The hallmark of a successful monetary policy will be an inflation rate that does not rise.”
To me, Greenspan’s use of the word “hallmark” was perfectly consistent with the secular battle plan of “opportunistic disinflation”: the Fed would not purposely induce recessions to bring inflation down, but would opportunistically welcome them when they happened, because recessions pay disinflationary dividends; and then would lock-in those dividends by preemptively tightening in recoveries, aiming for an inflation that did not rise. Such a strategy pursued long enough and successfully enough would take us to the promised land of effective price stability.
It worked, as the recession at the beginning of this decade was the last recession “needed” to win the opportunistic war against inflation. Thus, by definition, at least to me, it can no longer be the case that the hallmark of successful monetary policy is “an inflation rate that does not rise.” For this to be true, we would need to believe that somehow the business cycle has been repealed. How so?
The whole concept of “opportunistic disinflation”, the Fed’s battle plan during the war against inflation, was founded on the proposition that recessions inexorably happen, even though policy makers don’t necessarily want them. As Greenspan has expounded many times, the business cycle will always be with us, because as economic agents, we are all human, given to cyclical swings between irrational exuberance and irrational doom. Thus, he’s intoned, unless policy makers can somehow repeal human nature – which they cannot! – then policy makers cannot repeal the business cycle.
I agree with Mr. Greenspan, as would most rational thinkers. Accordingly, it seems to me, rational thinkers should also agree that once secular price stability has been achieved, it would be irrational for policymakers to cling to the notion that the hallmark of a successful monetary policy is an inflation rate that never rises. By simple logic, we can conclude that if such a cyclical goal were to be achieved by policy makers, the secular consequence would be renewed risk – if not the eventual long-term certainty – of a debt-deflationary spiral. Why?
Because stuff happens, otherwise known as negative shocks to aggregate demand, which beget recessions. Thus, if policy makers were to refuse to allow inflation to rise during cyclical upturns, they would be assuring that it would fall too low during cyclical downturns.
Accordingly, it befuddles me that those who are accusing the Fed of being “behind the inflation curve” (including many dear friends and colleagues!) do not embrace this obvious inference. To me, it should be viewed as a tautology! But then again, I may be too harsh, as I’m not sure Fed policy-makers fully grasp that a regime of winning the peace of price stability requires a fundamentally different reaction function than a regime of war against inflation.
To be sure and to their credit, Fed policy-makers have embraced the notion that the doctrine of preemptive tightening deserves a proper funeral. And indeed, holding the Fed funds rate at 1% over the last year has been just such a memorial service. Preemptive tightening is now resting in peace, even as inflation warriors rant about the tardiness of the Fed’s anti-inflation mojo .
The Fed has not, however, had a full transformation to a proper peacekeeper for price stability. Most important in this regard, and echoing Fed Governor Ben Bernanke, the Fed has not embraced the obvious need to quantify its definition of price stability. Note, please that I said definition, not target. Before you can logically have a target, it seems to me, you must have a definition!
Where’s the Firebreak?
Informally, of course, the Fed does have a definition: 1-2% for the core PCE deflator. But the Fed has not officially “owned” that definition. What is more, there has been no public discussion whatsoever about an acceptable cyclical pattern for inflation around that 1-2% zone, leaving the presumption that the Fed thinks it should and can cyclically hold inflation within that 100 basis point band.
I don’t buy that. And, indeed, I don’t believe that Fed policymakers do either. I recall, as I’m sure many of you do as well, a famous (or infamous) comment that Mr. Greenspan made almost exactly a year ago (June 3, 2003) on this subject:
“One important issue here is that we’ve worked with inflation over the years. We know how it functions, we know its dangers, we know how to address it. It’s a difficult job to contain inflation, but we’re not all that uncertain about it as we are with the issue of deflation. We’re far more unclear on the issue of deflation, and as a consequence, we need a wider firebreak, in logging and forestry terms, because we know so little about it. So we lean over backwards to make certain that we contain deflationary forces.”
Firebreak! That was a magical word when Greenspan uttered it at the time, turbo-charging the on-going bubble in the Treasury market, founded on the proposition that the Fed was headed towards the unconventional anti-deflation easing step of buying longer dated Treasuries, with the explicit goal of driving up their price so as to drive down their yields. It was not to be, of course, as Greenspan told the world a month later, popping the bubble that he rhetorically had inflated.
In retrospect, I find Mr. Greenspan’s decision to nurture the bubble, rather than preempting it by rhetorically correcting the market’s faulty presumptions to be indefensible. But that’s not my topic today. Rather, I want to compliment Mr. Greenspan for declaring the need for a firebreak between price stability and deflation risk.
He didn’t put it that way, of course; he simply declared the need for a firebreak, rather than saying the Fed should aim for a firebreak between (implicitly) targeted inflation and putative price stability. I don’t think, however, that my inference is wrong: the Fed should not seek price stability per se, but rather price stability plus a firebreak of inflation above price stability!
That was the case a year ago and remains the case today, even as nattering nabobs of negativity rant about the Fed’s killing of the doctrine of preemptive tightening and spew lather about the Fed’s somehow being behind the inflation curve. In my view, the Fed is to be congratulated for its self-described “very accommodative” policy over the last year. It’s been about building the firebreak of inflation that the economy needs, so as to be able to absorb a negative shock to aggregate demand – remember, stuff does happen; sometimes opportunistically, and sometimes not! – without the resulting recession-driven cyclical disinflation jumping into the fire of “unwelcome disinflation.”
Quite frankly, the firebreak is not yet wide enough, despite the Fed’s yeoman work with the 1% Fed funds axe over the last year. That is not to suggest that the Fed should not lift that rate. I’m on public record stating that the Fed should and will start a tightening process soon, most likely at the FOMC meeting at the end of this month. But as and when the tightening process begins, I believe strongly that the Fed should forthrightly discuss not what it is doing, but why it is doing it. The time has come for the FOMC to articulate more clearly its definition of price stability and objectives. The time has also come for the FOMC to communicate more descriptively its reaction function in the new world with an objective of maintaining price stability, rather than fighting inflation.
More specifically, the time has come for the Fed to communicate both its definition of price stability and its estimate of the size of the firebreak needed above that definition. For example, if we assume that a 1-2% range for the core PCE deflator is the working definition of price stability (which many FOMC members presumably would), how high should the Fed be willing to let inflation go cyclically, so as to make the economy “safe” for the next recession? 3%? 4%? 5%?
I don’t know the answer. What I do know is that the answer is categorically not what it was in 1994, when Mr. Greenspan declared that “the hallmark of a successful monetary policy will be an inflation rate that does not rise.” If inflation is not higher than it is today before the next recession hits, the Fed will have failed to win the secular peace of price stability .
Let’s Get Real
The Fed and we, as market practitioners, also need to re-think the matter of the “neutral” real short rate – the constant term in Taylor’s famous rule. Professor Taylor assumed 2% when he specified his Rule over a decade ago, mid-way through the two-decade war against inflation. And in reflection of what can only be called intellectual laziness, most analysts and most Fed policy-makers continue to use 2% (or higher!) as their assumption for the “neutral” real Fed funds rate.
Most famously, retiring San Francisco President Bob Parry has opined that the Fed should use a 2.7% assumption for the “neutral” real rate, as that was the realized real Fed funds rate for the 1966-2003 period.
I applaud Mr. Parry for taking a stand! I really do. It drives me nuts that most Fed officials are willing to wax wonkishly about the concept of “neutral” but then religiously refuse to put any numbers on the table. I wish Mr. Parry all of life’s blessings in his retirement. At the same time, I think he is wrong.
Not that I know precisely what the “neutral” real Fed funds is. Nobody does. What I do know, however, as Professor Lucas taught us long ago in his famous Critique, is that it is wrong to extrapolate policy parameters across policy regimes: model constants should not be assumed constant in the face of structural change. And, I submit, shifting from fighting a war against inflation to striving for the peace of price stability is a huge structural change!
And intuitively, this structural change tells us that the “neutral” real rate should be lower than it was over the last two decades (and, thus, the three and one-half decades that span Mr. Parry’s data series). How much lower we don’t know precisely, but logic tells us that the real rate average of the 1980s and the 1990s was not neutral, because the 1980s and 1990s were a long sojourn of secular disinflation.
One of the hallmarks of that period was that no-risk money served not just an effective store of real wealth, but a generator of real wealth. Indeed, a positive real rate of return on no-risk money was, I believe, a critical force behind secular disinflation: dear money makes for cheap goods, as goods bought with levered money are negative-carry investments, a disincentive for speculative hoarding of goods (just the opposite, of course, of the 1970s!).
Thus, as a starting point for thinking about the “neutral” real Fed funds rate, I believe we need to return to first principles: what rate is consistent with neutrality between holding money and holding a basket of goods? There are shades of the gold standard in that question, of course. And for the record, I have never been – and never will be, I trust! – a gold bug.
It is useful to remember, however, that with a gold standard, money does not buy more ounces of gold over time, but the same number of ounces. Under Bretton Woods, $35 bought one ounce of gold, which didn’t magically grow, but rather remained one ounce of gold. As such, money held its real value in terms of ounces of gold, which was a proxy for a basket of goods and services. The dollar was “as good as gold” but not better than gold, which pays no real rate
The modern day equivalent of the gold standard is, I submit, a nominal rate of interest on money that makes the holder whole for the two taxes that our government imposes on money: the explicit tax on nominal interest and the implicit tax of inflation . Put differently, money is “neutral” if the after-tax real rate of interest is zero! That implies, of course, a positive before-tax real rate of interest, so as to generate sufficient nominal income for the holder to pay the explicit tax on nominal interest. And the higher the inflation rate, the higher the “necessary” real rate. More specifically, I believe that the 2% constant in the Taylor Rule should be changed to:
Economy wide marginal tax rate ÷ (1 - Economy wide marginal tax rate) ´ x Inflation
Thus, if inflation is at 2% and the economy-wide marginal tax rate is 20%, then the “neutral” real Fed funds rate would be ½% and the “neutral” nominal Fed funds rate would be 2½%. In which case, the holders of money, after paying 20% of 2½% of nominal interest income in explicit taxes, would have just enough nominal interest income to make them whole for 2% inflation. If the inflation were 5%, the real rate would have to be higher, of course: 1¼%, so as to allow the holder to pay his 20% income tax on a 6¼% nominal rate, leaving 5% left to cover for the implicit 5% inflation tax.
In both cases, money would retain its real value, but would not appreciate in real value. And this would be fair and just, I submit, as money is a zero risk instrument, since it is guaranteed to always trade at par.
If savers wanted to generate a real rate of return on their stock of wealth, they would have to take principal risk – duration risk, credit risk, equity risk, and a plethora of other risk types. In such a low-inflation world, the real yield curve would be steep, as investors received compensation for assuming the real risk of losing principal in real time in the event of rising inflation (or a positive shock to long-term real interest rates stemming from a positive shock to the real rate of return on productively-employed real capital).
And if inflation were to accelerate sharply, the real yield curve would bearishly flatten, as the “neutral” real short rate would rise more-than-arithmetically with accelerating inflation . Indeed, the more I think about it, I’m actually suggesting to not just replace the real-rate constant in Taylor’s Rule, but also the first “active” term: the “penalty” term for over-target inflation, in which Taylor proposed an upward adjustment in the real short rate by one-half of the amount that inflation is over target. My alternative would produce the same result as Taylor: real rates rising faster than inflation, pulling inflation – by slowing the economy – back towards price stability!
But What About the Carry Trade?
But, some of you might ask: would not a perpetually-steep real yield curve beget excessive (and therefore inflationary) leverage in the economy, as investors would be incentivized to play the “carry trade?” The short answer is no, if the Fed is willing to allow inflation to cyclically-oscillate, rather than preempting the cyclical upside, as was the case during the war against inflation. In such a world, nominal interest rates (as well as TIPS breakevens) would oscillate with inflation, thereby imposing the real risk of cyclical losses for the carry trade . But if that risk was not sufficient to tame irrational exuberance on the behalf of carry-traders, the Fed would still have a very viable tool to check such excesses. It’s a tool that the Fed doesn’t presently use, but is tailor-made for the job: variable overcollateralization requirements for Repos!
Stay with me here, ‘cause I’m going to break some new ground! Right now, the Fed requires 102% over-collateralization for lending liquidity via its open market desk. The price for borrowing from the Fed (indirectly via the open market desk, in contrast to directly via the discount window) is the Repo rate, which is the Siamese twin of the Fed funds rate. Thus, a levered investor can hold Treasuries with only 2% equity capital: the amount of collateralization that is required. Put differently, you can lever Treasuries 50 to 1!
The Fed’s 102% over-collateralization requirement is, of course, the same one that we all use: it’s written into all the PSA Repo agreements that we sign. Indeed, at PIMCO, we lend out some $10 billion a day in the Repo market, collateralized by 102% of that amount in Treasury securities, which the borrower must post to our clients’ custodian accounts. I don’t need to teach this group anything about the mechanics, as all central bankers live and breathe in the Repo market, the conduit for monetary policy actions to set the short-term policy rate. We all view the 102% over-collateralization rule to be a prudent safeguard against the risk that the borrower fails to meet a margin call, in which case we have the legal right to sell the collateral, without having to go through the bankruptcy legal process.
Accordingly, the 102% rule has never been viewed as an instrument of monetary policy, but rather simply the Fed-endorsed standard for protection against default risk on the part of the borrower. In contrast, the central bank’s reserve requirements for banks’ deposits are viewed as an instrument of monetary policy. At least that’s what we all learned in our first money and banking class! The right answer on the exam about the Fed’s tool kit included three instruments: open market operations, the discount window, and reserve requirements.
As a practical matter, however, changes in reserve requirements as a tool of monetary policy have become an extinct tool in the United States, as the Fed has secularly reduced reserve requirements to the bare minimum, because reserves held at the Fed don’t pay interest, acting as a tax on banks, distorting the playing field for banks relative to their non-bank competitors. Thus, the old text notion that the Fed can stimulate or restrain bank deposit growth (the money supply!) via cutting or hiking banks’ reserve requirements is an anachronism. No tears from me, because policy instruments must evolve with the changing financial system architecture.
Changes in reserve requirements were a logical tool of policy in a bank-lending-centric financial system, but we now live in a capital markets-driven world. The Fed can no longer constrain economy-wide credit growth by quantitatively changing the statutory demand for banks’ reserves relative to that which it supplies.
Well, it could do that, but the transmission mechanism would not be the textbook T-account process (remember that wonderful academic exercise?!), but rather a consequence of an increase in the price for reserves – the Fed funds rate and the Repo rate! – which would temper the economy-wide demand for credit. Simply put, in a disintermediated financial system, the Fed rations credit through the impact of changes in the price on the demand for credit, not through non-price restrictions on the availability of bank reserves. Or, as the great Henry Kaufman said over twenty years ago, in a deregulated world, the Fed rations credit not by restricting supply, but rather bankrupting the borrower by jacking the price.
This is most unfortunate, in my view. The Fed needs some means to restrict growth in economy-wide leverage besides hiking interest rates! And more narrowly, the Fed needs a tool beside interest rate hikes to restrict growth in levered financial market speculation, which imparts boom-bust proclivities to the real economy.
During the bubble years for the stock market at the end of the 1990s, I pleaded publicly, including testimony before Congress, that the Fed should hike margin requirements for borrowing against stocks, to communicate firmly its view that the bubble was indeed a bubble, while also reducing the need to excessively hike interest rates to get stock speculators’ attention, which was imposing collateral damage on the real economy.
The Fed didn’t take my advice, of course. Indeed, to this day, Chairman Greenspan periodically includes a footnote in his speeches on the bubble years, lambasting those who argued that a hike in margin requirements would have been a good idea. The fact that he still doth protest so vigorously suggests to me that he also doth protest too loudly.
But enough of that! My point today is that the Fed does have a tool that could become a quantitative policy tool to restrict growth in credit creation: the over-collateralization rule for Repos! And unlike the margin tool for stocks, there is no way that the financial markets could arbitrage around changes in the Fed’s over-collateralization requirements for Repo. If the Fed were to say, for example, that the requirement is now 104% not 102%, I assure you that PIMCO would follow the Fed’s lead, and demand exactly the same when lending via the Repo market; indeed, our clients would rightly demand that we follow the Fed’s lead. And I feel quite sure that all of you in this room would do the same thing!
I’m not suggesting that the Fed do that right now, I want to stress. What I am suggesting is that in a world of a low “neutral” real Fed funds rate and a steep real yield curve, which I think is consistent with maintaining the peace of price stability, the Fed needs a non-price tool to temper levered speculation in the “carry trade.” And to me, a variable over-collateralization requirement for Repos would be a most excellent tool, both mechanically as well as a signaling device.
Changes in the Fed’s over-collateralization requirement would become the modern day equivalent of changes in reserve requirements in the textbooks of our youth. Way cool, my 15-year old Jonnie would say.
Let me now draw to a close. Here are my four takeaways (I love that expression!):
- The Fed is now trying to win the peace of price stability, not fighting a war against inflation. Accordingly, the doctrine of preemptive tightening is dead, as well as the notion that inflation should not rise cyclically during recoveries.
- The present inflation rate is too low, because there is no buffer or firebreak of inflation to comfortably absorb the disinflation that would be wrought by a negative shock to aggregate demand and a resulting recession. Thus, the Fed should prudently allow the inflation rate to rise. And the Fed should communicate that intention, while more clearly identifying its definition of price stability and associated tolerance bands around that definition.
- The “neutral” real short rate should be appreciably lower in the years ahead than the average of the last two decades, a period of secular disinflation. Conceptually, the neutral after-tax real short rate should be zero, which would be consistent with no-risk money holding its real value. Accordingly, the 2% real rate constant in Taylor’s Rule is wrong.
- In a world of price stability and properly calibrated Fed policy, designed to protect the real value of money, the real yield curve would be very steep, inviting excessive leverage. To guard against that risk, the Fed should transform its over-collateralization rules for Repos into a policy instrument, changing over-collateralization requirements in a manner similar to the textbook role of changes in reserve requirements.
Thank you, Terry for inviting me here today. It is an honor that I will long remember and cherish.
1 Terrence Keeley, Managing Director and Founder of UBS’s Seminar for Sovereign Institutions