After five years of abnormality, many features of the post-crisis investment landscape seem to be in the early stages of reversion toward pre-crisis conditions. Growth has generally improved in the developed world, fiscal deficits have narrowed and, most critically for bond investors, the Federal Reserve has begun tapering its asset purchase program. Some investors, perhaps not surprisingly, have allocated away from core bonds toward the extreme ends of the risk spectrum: cash, which is insulated from the negative impacts of yield increases; and equities, which would likely benefit from accelerating economic growth.
Yet we believe cutting core bonds is an explicit de-optimization of a broader portfolio. It can only be justified as a market-timing strategy: Not only must the investor be right about rising yields, but yields also must rise within a narrowly predicted window of time. Those who flee to cash yielding near 0% may experience a “death by a thousand cuts” if yields don’t rise as soon as expected. Those who increase equity concentrations may be overexposed to negative economic surprises – capital gains from bonds have tended to kick in just when equities sold off, helping to stabilize overall portfolio performance. So while it may seem intuitive to reduce core bond holdings in anticipation of yield normalization, many investors, if not most, may end up worse off by trying to time the market.
A more thoughtful response is to consider retaining core bonds and diversifying the specific risk factor of concern, in this case duration. In the past, global bonds have captured most of the upside but avoided a significant amount of the downside relative to domestic-only bonds; indeed, we believe the threat from yield normalization is a somewhat nonsensical concept for investors with a global perspective (Figure 1). This is portfolio efficiency at its best.
In our view, investors can do more.
The market offers the opportunity for compensation against rising rates through roll down in an upward sloping yield curve market, with the amount of compensation fluctuating dramatically over time and across countries. Active investors can potentially enhance the average level of this compensation in their portfolios; and if market expectations are not fully realized, then those investors may generate capital gains (in addition to coupon returns) even in a rising yield environment, as we explain below.
Constructing a framework and managing the forwards
Generating capital gains from bonds in a rising yield environment requires defining concretely what yield normalization means – where yields are going and when they will get there – and setting these expectations against forward market pricing, country by country.
So what does yield normalization mean?
Nominal GDP growth anchors nominal yields (see Figure 2). This makes sense: Lenders’ expectations for a respectable return on capital, as well as borrowers’ expectations for a reasonable cost of capital, are both conditioned by economic growth.
It is most common for nominal yields to be bounded from the top by nominal GDP growth. The disinflationary regime of 1980 to 1999 (as well as the last few post-crisis years) is the exception. This was a unique phase when central banks set policy rates above equilibrium to lower nominal GDP growth (mostly via inflation) and, with a lag, nominal yields. The reverse occurred during the post-war reflationary regime when policy yields were set below equilibrium, which in turn propelled nominal GDP growth higher while yields adjusted to this new reality with a lag.
We are more likely to see reflationary than disinflationary efforts by central banks in the coming years. We do not think nominal yields will be as far below nominal GDP growth as they were during past reflationary regimes – central banks have far more credibility and operate in a more liberalized financial system. Still, nominal GDP growth should serve as an upper bound for nominal yields.
The key question then becomes the outlook for nominal GDP. The International Monetary Fund (IMF) projects nominal GDP growth over the coming three to five years will average 5.2% in the U.S., 4.25% in the UK and 4% in the eurozone. However, beyond the next few years, trend growth is projected to decline due to demographic shifts and a slower pace of capital accumulation. Macroeconomic Advisors, for example, projects U.S. nominal GDP growth will decline to 4.4% by 2020, consistent with the U.S. Congressional Budget Office’s own forecasts. While imperfect, these figures serve as a good upper-bound estimate for yields over the next few years.
While real yields have tended to be a relatively small and stable component of nominal yields, their assessment should be more consequential in the years ahead. Real yields in the U.S. averaged between 1.5% and 2.5% during the post-war era, depending on the maturity of the instrument; but there are persuasive reasons to believe that real yields could be far lower going forward.
The supply of loanable funds, for instance, should be higher and the demand for loanable funds should be lower in an environment without balance sheet leveraging, both of which would put downward pressure on the market-clearing (equilibrium) level of real yields. It is not just domestic households in the U.S., UK, and Europe that are striving to save more, adding to the supply of loanable funds. Lenders in the emerging world will likely continue to supply loanable funds to globalized capital markets.
Taking the argument a step further, there has been quite a bit of academic research, summarized recently by Larry Summers’ forceful “Secular Stagnation” speech at the IMF in November, showing that equilibrium real yields could be close to 0%, or perhaps even negative, in an economy without balance sheet leveraging. If true, this would explain disinflation in countries where policy rates have been set far below inflation for more than five consecutive years. It would also suggest that nominal yields may peak at a much lower level than they did in 2007, when they briefly touched 5.25% in the U.S. According to secular stagnation theory, these levels were reached only because of an historic borrowing binge and a housing bubble. Assuming borrowing-binge conditions are not repeated, nominal yields could peak at a much lower level the next time.
Putting it all together, we think that the 10-year government yield (in the U.S. and UK at least) over the next three to five years should not likely go higher than 4.0%-4.5%, with real yields slightly below average and inflation premiums slightly above average.
Setting expectations against forwards
Forward yields re-priced far more than spot yields during the 2013 bond market selloff, so bond investors not only enjoy higher yields today but also greater cushion against further yield increases. It depends on the maturity of the yield and the geography of the market, but longer-dated yields already help protect investors against a full normalization over the next three to five years, at least according to our framework. If yields peak at a lower level, or take longer to get there, bond investors would theoretically realize capital gains on top of their coupon returns even as yields increase.
Specifically, forward markets anticipate that 10-year U.S. yields will rise to 3.6% over the next three years, and to 3.8% over the next five years (see Figure 3).
So, theoretically, if the U.S. 10-year yield normalizes at 3.75% over the next five years, core bond investors would realize a capital gain to augment their coupon return. The normalization priced in to French forward markets at the moment is greater; in Canadian markets, worse. And by shifting exposure to those forward markets in different regions when they offer a compelling cushion against rising rates, we believe quite a bit of excess return potential can be harvested.
Markets have historically offered a structural cushion against rising short-dated yields – a relatively small but stable uncertainty premium that active investors can potentially harvest as excess returns if spot yields do not increase to ”validate their forwards” (see Figure 4). This is quite common during phases when central banks are on hold.
This uncertainty premium varies over time and across countries. For example, last May, when the Fed first mentioned tapering its asset purchase program, this premium increased more in Europe than in the U.S. This didn’t make sense: Why would forward yields increase to fortify investors against ECB policy rate hikes in Europe when it was the Fed that had changed its stance on U.S. monetary policy? This opportunity began to disappear nearly as soon as it had surfaced, and was fully snuffed out when the ECB invalidated forward pricing completely by cutting its policy rate in November 2013. Many instruments sensitive to these dynamics – short-dated bonds but especially money market futures contracts – increased in price.
Back to the future
Many investors subscribe to the notion of symmetry – the idea that there must be an equal and offsetting bear run in bonds as the natural counterpart to the bull run experienced during the past 30 years. But, in fact, this has already played out: The bull run from 1980 to the present was the counterpart to the bear run from 1950 to 1979, during what was an abnormal and temporary aberration from a centuries-old trading range (see Figure 5).
Rather than selling bonds in preparation for the return of such an abnormal environment, then, we believe investors should focus instead on what is likely in the current macroeconomic environment: a rise in yields, perhaps, but one that is gradual and ultimately limited in scope. And because market prices adjust to anticipate future yield trajectories, one could even claim that yield normalization already occurred last May, when forward markets re-priced to reflect the expectation of rising yields.