At times when I can’t seem to figure out what is happening or why it’s happening, I find great comfort in theory and logic, which has been the mother’s milk of my professional life. This same proclivity, I must confess, has also at times been a great source of discomfort in my personal life: love ain’t necessarily logical, and no amount of theory can explain what the heart feels, rationally or irrationally.
As the great philosopher Kris Kristofferson wrote, in matters of love:
I know someday that I may wake up all alone
When the love I believed in has died
But at least I won’t wonder what I have won
From the lifetime that I watched going by
‘cause I’ll never know ‘til it’s over
If I’m right or I’m wrong loving you,
But I’d rather be sorry for something I’ve done
Than for something I didn’t do.
Not so in matters of investment management, in particular running money against a benchmark: it is better to be sorry for something you didn’t do, than for something you did. Not doing something is taking the same risk as the benchmark, generating index-like returns. This is not, of course, a path to long-term success for active managers, who are supposed to beat their indexes.
But in the short run, managers are rarely fired for periods of index-like performance. Thus, the burden of proof for making a bet relative to the index should be high; in the absence of such justifying proof, it is rational to mimic the index’s risk characteristics. To wit, closet indexing should be in the real-time tool kit of an active manager: deciding not to take a bet is, in fact, taking a bet, which can be a rational bet in real time.
An active manager need not apologize for periods of betting by not betting. Unless a bet can be taken with conviction, ‘tis better to be sorry for what you didn’t do than for what you did do.
No, this ramble is not a preamble to defending closet indexing of PIMCO clients’ portfolios. We are currently running quite large deviations from our benchmarks, as evidenced by the “tracking error” of our returns versus benchmark returns. 1 More on this at the end. Accordingly, I am not apologizing for risk aversion here at PIMCO, because we’re not in the index closet. Rather, I’m starting out with a kind conceptual word for tactical indexing, completely out of the closet, as part of strategic active management: not betting is an active bet!
At the moment, active managers could be forgiven for wanting to cleave close to indexes, because global yield curves are acting very, very funky relative to history: bull flattening rather than bear flattening in the face of Fed tightening. We’re all searching for “answers” as to why, and if an active manager doesn’t have conviction as to probability of various competing answers (eventually!) proving to be correct, then she should rationally not place a bet on the answer, but rather peg her duration and curve positions to the benchmark.
A Condo With Three Floors
The search for answers as to why global yield curves are doing what they are funkingly doing must begin with a review of the dominant theories of global interest rate “equilibrium.” After all, if the market is not doing what it is “supposed to be doing,” there must be some set of theories that the market is “violating.” Accordingly, logic demands that those theories be dissected, so as to determine if:
They still hold (in which case the market truly is whiffing something wacky!); and/or
- There has been either (a) some change in the weights the market is applying to orthodox models and/or that (b) some new model – implying structural change! – is waiting to be discovered.
Here are the three dominant theories of the determination of the level and slope of global yield curves:
- Uncovered interest rate parity hypothesis .
This theory is founded on the notion of “one global price” for money/credit (high church market efficiency), and therefore, posits that global interest rate forward curves are one and the same as global forward curves on currencies. To wit, expected changes in currency determine global interest rate differentials, as market operators demand compensation (or give up compensation) based on their expectations about currency values.
It's most elegant theory, with the unfortunate real world characteristic of not only not holding empirically, but frequently being rejected empirically. Global interest rate differentials are not an efficient predictor of future currency values.
- Expectation of future policy rates hypothesis.
This theory is founded on the notion that investors always have the option of holding cash, which always trades at par. Thus, the slope of the yield curve should be viewed as a forward curve on future policy rates, plus a risk premium for both uncertainty and volatility
This theory is quite robust empirically in the short run, particularly in the front end of the yield curve, particularly if the central bank is both highly credible and highly transparent. In the long-run, however, the term structure of interest rates is an empirically inefficient – i.e., lousy! – predictor of policy rates.
- Market segmentation hypothesis .
This theory is founded on the notion that investors are not homogenous, with many if not most, constrained to a “natural habitat”, borne of regulation and/or heterogeneous risk appetites. Most simply, this theory is a refutation of any grand efficient markets hypothesis, resting on the simple observation that global market participants do not view debt in various currencies and at various tenors as perfect substitutes, via forces of arbitrage. And, ironically, the proof statement for this hypothesis is the absence of robust empirical proof for the other two theories! 2
All active managers, in some fashion or another, rely on all three theories: we just know that somehow, the first two theories help inform market price action, even if the data don’t confirm that to be the case. Put differently, just because a theory fails empirically ex post doesn’t mean that investors don’t apply it ex ante. After all, one half of all marriages fail ex post, but people keep getting married and re-married with ex ante optimistic expectations.
The Here and Now: Too Low, Too Flat?
Consensus market opinion presently holds – and has held for well over a year! – that the yield curve is “too low and too flat.” Applying a bit of reverse engineering, we can conclude that this viewpoint is grounded on some combinationof the uncovered interest rate parity hypothesis and the expectations of future policy rates hypothesis. This combination gives voice to two strains of talking-head commentary:
- The yield curve is both too low and too flat because Fed policy is too accommodative, generating excessive domestic demand growth and an ever-rising U.S. current account deficit, implying a falling dollar, which will beget rising U.S. inflation and with a lag, nastier than expected Fed tightening.
- The yield curve is too low and too flat because the “failure” of long rates to follow short rates up is generating easier financial conditions, fostering above-potential domestic demand growth, which will tighten U.S. labor markets and put upward pressure on U.S. unit labor costs, generating accelerating inflation, and with a lag, nastier than expected Fed tightening.
These two “justifications” for the yield curve being too low and too flat are, of course, complementary not competing: they both lead to the same conclusion, just arriving at it through different channels, the former international and the latter domestic. And most elementally, these rest on the notion that the yield curve is under-estimating either future inflation or future real short rates; or both!
I must confess that I have some sympathy with this view to the extent that it is based on inflation; it does seem to me that the yield curve is perhaps carrying a too-skinny risk premium for inflation. After all, reflation is as reflation does, in the words of the famous economist Forrest Gump, and there is no question that the Fed is biased towards reflation as opposed to disinflation. And I viscerally believe that central bankers with printing presses who get their ink for free, as is the case with the Fed, ultimately get what they want.
I do not, however, have sympathy with the consensus notion that the market is underestimating future real short rates, notably the “equilibrium” real Fed funds rate. As I’ve been preaching for over two years now, I believe that the “equilibrium” real policy rate is somewhere in the 0.5%-1.0% range, very close to where we are now with the prevailing 2.5% nominal Fed funds rate and core PCE inflation running a touch below 2%.
Thus, unless there is a breakout to the upside on inflation, certainly a risk but also certainly not a present reality, I’m not nearly so sure as the consensus that the yield curve is too low and too flat. Maybe it is just a matter of the consensus having a too high and too fast trajectory in the back of their minds for the real Fed funds rate!
“Not so!”, I hear many of you retorting, notably those who vehemently believe in a “neutral” real short rate in the 1-3% range (the midpoint of which is the constant term in ubiquitous renditions of the famous Taylor Rule). Which, of course, leaves such naysayers with the burden of explaining why the yield curve is too low and too flat: to wit, why is the market wrong and you are right? And the common rejoinder is – you guessed it! – the market segmentation hypothesis: there must be buyers of duration somewhere who are irrationally buying – from the standpoint of either the uncovered interest rate parity hypothesis or the expectations of future policy rate hypothesis.
Bretton Woods II
And who are those buyers? For years, foreign central bank buyers have borne the burden of being labeled the fish in the game, the “stupid” buyers, I often hear. Once these putative non-economic buyers come to their economic senses, I also hear, letting their currencies appreciate against the dollar, reducing any need to buy dollars and dollar-denominated assets, the yield curve will finally levitate upward to levels that “make some sense.”
To my way of thinking, it is a cop out to “blame” the low, flat yield curve on foreign central bank buyers, even if they are in fact irrational. In the game of active investment management, understanding the motives of all buyers, rational and irrational, is called doing your homework. This is particularly the case when the presumed irrational market participant has more money than you do, which is tautologically the case with foreign central banks, who can print the stuff up for free.
Bringing theory down to reality at the moment (in order to forecast!), we must take as a given that China has a fixed exchange rate with the U.S. Accordingly, China does not have an independent monetary policy, but rather “imports” U.S. monetary policy, which is “complemented” on the domestic front with regulatory tools. This will not always be the case, of course, but for right now, it is reality.
We must also take as a given that the rest of Asia, and much of the EM world, does not want to see their currencies appreciate very much (or at all) against China, and thus shadows China’s fixed exchange rate regime with the United States . These are facts, not opinions. To wit, we must take as a fact that, for the moment, what is frequently called the “Bretton Woods II” arrangement, does exist.
It is not enough for active investment managers to simply say that the arrangement creates a universe of stupid buyers of duration and yield curve risk, who will ultimately come to their senses. In the fullness of time, that might well be the case. But it is also true that in the fullness of time, we are all dead. Active management is about putting alpha up on the board between now and then.
We must also take as a given that the present “Bretton Woods” arrangement is fundamentally different than the original version, in that the dollar is not convertible into gold, even at the official level, as it was in the original version. This means that the U.S. has moreflexibility now than during the Bretton Woods era to pursue a domestically-oriented monetary policy: there will be no Charles de Gaulle demanding that the USA send all the gold in Fort Knox to Paris .
Yes, America does, as custodian of the global reserve currency, have “exorbitant privileges,” as de Gaulle bitterly bellyached 40 years ago. Indeed, they are even more exorbitant now than then! America presently has the necessary degrees of freedom to pursue an accommodative, domestic demand-oriented monetary policy, aimed at domestic job creation, secure in the knowledge that:
- Inflationary pressures will be defused by a portion of domestic demand “leaking” out to the rest of the world via the current account deficit, and
- Foreign central banks will act as a “put” under U.S. debt prices or, if you prefer, act as a ceiling on the level and term structure of U.S. interest rates.
This is reality: the market segmentation hypothesis of the level and term structure of U.S. interest rates dominates the other two hypotheses. Not that the other two theses are irrelevant. They remain very relevant, but will systematically overestimate the level and slope of the U.S. yield curve, as long as the Bretton Woods II arrangement remains in force, which could be for years .
This reality also tells us that the Euroland yield curve will be lower and flatter than what “conventional” models would suggest. Euroland has chosen not to follow America’s reflationary monetary policy lead, as is the case with China and her fellow travelers. Euroland rejects the famous Connelly Doctrine 3 that while the dollar is America ’s currency, a weak dollar is the rest of the world’s problem.
Euroland not liking what the Dollar Bloc is doing, however, is no reason to expect the Dollar Bloc to stop doing what it is doing, starting with America’s domestically-oriented policy objective of more robust job growth in America . Monetary imperialism is as monetary imperialism does, as Forrest Gump might also say. And America unquestionably has imperialist monetary objectives.
An externality of this state of affairs is a perpetually sick Euroland economy, in particular Germany , which is also demographically challenged in the context of under-funded pension schemes. This exigency is begetting regulatory efforts to “insure” the integrity of pensions by demanding de facto immunization of their long-duration liabilities.
This brings the market segmentation theory of rates squarely to the forefront in Euroland: if the ECB is going to insist on de facto deflationary monetary policy for Euroland, then regulators will insist that pension fiduciaries (and life insurers) “insure” against the implied deflationary impact on equities by shifting towards credit risk-free, long-duration bonds.
Dutch regulators are leading the charge on this “defeasance” front, with others looking in the same direction. Thus, there is scope for the Euroland yield curve to move lower and flatter than orthodox business-cycle models would suggest.
The United States, Too?
Many are asking the same question about the United States: Might the U.S. yield curve be driven lower and flatter by the same regulatory force, as the market has been romancing ever since January 10, when Labor Secretary Chao called for regulatory changes that would push American benefit schemes in the defeasance direction?
To a limited extent, the answer must be yes, as the U.S. Pension Insurer, the PBGC, is moving in the direction of defeasing the long-term liabilities of plans that go broke and that are taken over with long-duration bonds. The PBGC’s logic is imminently sound and, indeed, suggests that the PBGC should have been following this path all along: busted plans are by definition no longer going concerns but rather liquidating plans and everything we know about financial theory suggests that they should be defeased.
Such is not the case, however, with the plans that are going concerns, viewed as a whole , which regulators should do. To be sure, some concerns that are presently going concerns will, in the fullness of time, become non-going concerns, with their pension liabilities “put” to the PBGC.
But if we assume that American Capitalism itself is a going concern – a fair assumption I think for America, even if a problematic one at the margin for Euroland – it makes no sense whatsoever for regulators to demand that the private sector, as a whole, defease its pension liabilities. To be sure, there is the smell of such a demand in the Labor Department’s regulatory trial balloon, but I seriously doubt when it will float in legislative space.
Indeed, I don’t think it should float for the simple reason that if American capitalism itself is a going concern, equity returns will, over the long run, exceed debt returns . This is a tautology! Accordingly, there is no sound reason to demand that going concerns manage their financial affairs as if they must pass a real time liquidation test.
Yes, companies on the cusp should be made to pass such a test, as the PBGC is suggesting with risk-based insurance premiums, an excellent way to formalize what the PBGC is already doing. But demanding system-wide defeasance – or immunization, if you prefer – would not make macroeconomic sense. Such a regulatory course would run head on into the fallacy of composition, similar analytically to the paradox of thrift.
Thus, while I think there are a lot of good reasons to conclude that the U.S. yield curve is not egregiously too low or too flat (most notably, my views on the “neutral” real Fed funds and the staying power of the Bretton Woods II arrangements), I believe chatter over the last month about a grand move to long-duration portfolios by American pension schemes is not a good reason.
To be sure, the notion has alluring appeal, especially to a certain segment of market participants: hedge funds, who are masters of the Keynesian doctrine that real time market prices are not determined by investors who are buying them “for keeps,” but rather speculators specializing in figuring out what consensus opinion will be about what consensus opinion will be about what consensus opinion will be (and even higher degrees, Keynes mused).
God bless them: hedge funds bring real time liquidity and efficiency to markets. But real time efficiency is not the same thing as rational discounting of enduring fundamental forces. Put differently, the invisible hand of the markets is actually very visible in real time, only revealing its invisible wisdom in the fullness of time.
Accordingly, while I buy the notion that the Bretton Woods II arrangement is a fundamentally important, and thus, enduring force that is generating a lower and flatter U.S. yield curve relative to history, I don’t buy the notion of a coming massive move to long duration portfolios by U.S. pension funds.
What we are witnessing right now is speculative buying romancing that scenario. And, ironically, the greater the romance – the lower and flatter the yield curve becomes – the less likely the romance is to be consummated. “Not so,” I hear some of you retorting, ready with the argument that the more hedge funds force the yield curve lower and flatter, the greater will become pension funds’ net liabilities, “forcing” regulators to force them to buy long-duration assets. Yes, I understand that argument. I just don’t buy it.
It reminds me of the summer of 2002, when hedge funds were indiscriminately shorting BBB credit spreads, on the presumption that if they could “force” the rating agencies to downgrade them to below-BBB (junk!), their selling would be a self-fulfilling profit machine. And, indeed, it was for a time.
But in the fullness of time, the fallacy of composition took over: policymakers simply could not stand idly by witnessing the BBB universe turned into junk: individual BBB companies, yes, but not the BBB universe, as such an outcome would represent a macroeconomic debt-deflation spiral. Accordingly, the Fed responded forcefully with easing, explicitly citing (irrational) widening of credit spreads as the reason.
And to back up that action, the Fed declared that it would, if necessary, use its printing press to ensure that the BBB universe was a going concern, most famously in Ben Bernanke’s November 21, 2002 speech, Deflation: Making Sure “It” Doesn’t Happen Here .
It worked. Essentially, the Fed said that the moral equivalent of a bank run was occurring, not in the banking system but rather in the corporate bond market, and declared that it would do whatever it took to stop it. And the right way to stop a bank run was, as always, with the overwhelming force of money: not one, but multiple Brinks trucks of cash...‘twas a most excellent time to buy BBB spread risk!
Uncle Sam Will Be Rational, Not the Fool
In current circumstances, everybody “knows” that it would be impossible for the American defined benefit pension system to be “instantaneously” transformed into a fully marked-to-market world. Yes, individual companies could switch on the wire to such a regime; but the system could not, just as the banking system could not honor its deposits at par. The fallacy of composition rules!
But to the extent that speculative market participants, again notably hedge funds, anticipate that some companies will walk down the defeasance path, it is rational for speculative money flows to want to “front run” real money flow – a “run” on long-duration asset prices, if you will. In such circumstances, it would be irrational for regulators to speed up any structural changes in pension accounting, which would be the moral equivalent of the FDIC encouraging bank runs.
The rational thing for pension regulators to do right now, and I fully expect them to be rational, is openly declare that there will be a long, long transition period of any de facto defeasance requirement. And if speculative money flows don’t respond by curtailing their “front running” purchases of long-duration bonds, then it would make sense for the Treasury to increase the supply of long Treasuries – the moral equivalent of sending Brinks trucks to stop a bank run.
Yes, I know that the Treasury ceased issuing long bonds in 2001, and has no plans to start doing so again. But since it is the Treasury, otherwise known as the taxpayer, that is on the lender-of-last-resort hook for the PBGC, I believe the Treasury should be, and indeed would be, willing to change its mind about not issuing long bonds, if its regulatory siblings take actions that structurally increase the demand for long bonds.
Doing so would essentially shift the risk of falling interest rates from pension schemes (who are structurally short of duration in a mark-to-market world) to the Treasury, who has no mark-to-market requirements and has a claim, one step removed, on the Fed’s printing press.
I actually don’t think events will require such a road-to-Damascus conversion experience by the Treasury about issuing long bonds. Rather, I expect the speculative “run” on long-duration bonds to exhaust itself, as speculative money flows grow weary to real money not “showing up” to take them out of their bigger-fool purchases. Simply put, I expect pension regulators to play smart, rather than playing the fool.
The on-going bull flattening of the U.S. yield curve, a most unusual occurrence is, I think, resulting from a combination of:
The market segmentation thesis as it relates to the exigencies of the prevailing de facto Bretton Woods II arrangement – what Bill Gross calls the “Asia Put;”
- The policy expectations thesis as its relates to both (a) the “neutral” real short-term policy rate in America and (b) expected inflation, both in America and the world; and
· The market segmentation thesis as it relates to regulatory changes that would structurally increase pension plans’ demand for long-duration assets.
I buy the first two of the explanations; they are sound reasons for the curve to be “lower and flatter” than many people think it should be. I don’t buy the third, at least in the United States , even as I understand the dynamic of speculative flows trying to front run potential regulatory-driven real money flows. Uncle Sam may be naive at times, but he ain’t stupid. Therefore, to the extent that the most recent bull flattening of the curve has been at the hands of such speculative flows, I share the widespread belief that the yield curve is too low and too flat.
The bottom line for PIMCO clients’ portfolios? We are measurably below benchmark durations, while significantly underweight U.S. duration relative to European duration, notably in the long end. We are not married to these positions, but like them, conditioned upon events unfolding as we see them. If they don’t, then we will change both our minds and our positioning, as Keynes intoned all smart people should do when they get new information.
Meanwhile, on this Valentine’s day, let us all remember that love is the most special asset that exists, the only asset class in which 1+ 1 adds up to more than 2. Do something special for the special one in your life, remembering that in matters of love, it is better to be sorry for having risked and lost, rather than to have not risked at all.
February 14, 2005
1 The term “tracking error” has always befuddled me, because it is defined as performance variance against the benchmark, which ain’t an “error” when it is positive. What’s say we as an industry start using the expression “tracking variance”!
2 For a wonderful discussion of these empirical matters, see Chapter 3 of former Fed Vice Chairman Alan Blinder’s nifty new (little) book, The Quiet Revolution: Central Banking Goes Modern.
3 U.S. Treasury Secretary at the time of the break-up of Bretton Wood.