Bond Yields Are Too Low Somewhere by Zach Pandl, Columbia Management
- Long-maturity bond yields are determined at a global level.
- Abnormally low forward rates are not just a U.S. phenomenon: there’s been a similar shift in the relationship between rates and growth across developed markets.
- If global rates remain persistently low, financial conditions will eventually need to tighten in other ways to offset this unexpected stimulus.
The big surprise in bond markets this year has been the low level of long-maturity rates despite continued economic growth. To see these changes most clearly it is often best to focus on distant forward rates, such as a five yield, five years forward (“5y5y”). After a big increase in 2013, these forward rates have declined across a diverse set of countries year-to-date (Exhibit 1). This comovement should come as no surprise: changes in forward rates are always highly correlated across countries. And we can use these cross-country relationships to shed some light on today’s low bond yields.
Exhibit 1: Cumulative change in 5y5y yields since December 2012
There’s a well-known relationship between nominal growth and forward interest rates, which holds in both theory and practice. Countries with higher nominal growth expectations tend to have higher forward rates and vice versa (Exhibit 2). It’s frankly hard to say much about global bond yields without starting with this reliable pattern—it’s the foundation for macro analysis of interest rates.
Exhibit 2: Nominal growth expectations and 5y5y yields in G10 economies (June 2014)
While we almost always see this upward-sloping line in history, it does not always have exactly the same shape. Mechanically speaking the pattern can shift in two ways: through changes in the slope or changes in the intercept. The flatness or steepness of the slope represents the degree of differentiation in markets between countries with higher and lower nominal growth expectations. We therefore call the slope of this relationship the growth premium. The intercept represents the level of forward rates for a given level of nominal growth expectations. Changes in the intercept term reflect parallel shifts in interest rates across all markets (Exhibit 3). We call the intercept term the cross-country discount rate.
Exhibit 3: Hypothetical shift in cross-country discount rate
The first chart below plots a monthly estimate of the growth premium (the slope term) over time. For G10 economies this figure averages a little over 1.0 and is generally pretty stable. However, we see a marked decline—a flattening of the slope—starting in the middle of 2011 (around the beginning of aggressive forward guidance from the Fed) and a big reversal during 2013 (during the taper tantrum). Thus, one way to describe the low level of yields during that two year period could be as a low growth premium: the market was pricing historically low differentiation among economies with high and low nominal growth. Today the growth premium is close to its long-run average, implying a normal level of differentiation among countries (e.g. U.S. and Australian forward rates are currently a normal distance above German and Japanese forward rates, given different prospects for nominal growth).
Exhibit 4: Estimated growth premium (slope term)
The next chart plots our estimate of the cross-country discount rate. To aid interpretation, we have expressed this variable as an interest rate (i.e. in percent per annum, rather than the raw regression coefficient; see chart footnote for details). The important observation is that for the last few months the estimate has been at historic lows. In other words, controlling for the market growth premium (the slope term), 5y5y yields have never been lower in major economies during modern history. Bond yields are historically low (or yield curves are historically flat) across the board.
Exhibit 5: Estimated cross-country discount rate (in percent per annum)
There are a variety of popular explanations for the low level of distant forward rates—or in our terminology, the cross-country discount rate. These include high debt burdens and deleveraging, greater financial intermediation costs, and price-insensitive buying by pension funds and foreign central banks. Rather than debate the causes here, we will simply point out what we see as the main implications.
First, to the extent that yields are inappropriately low for the big economic regions—i.e. that we are not in a stable equilibrium—then the adjustment will simply be higher forward rates. In our view this is at least partly correct, and we expect higher G10 forward rates in general over time. However, this conclusion is admittedly not obvious in light of the challenges facing the eurozone and Japan. In both cases persistently low rates might be required to achieve a full economic recovery.
Second, even if low rates are warranted in some areas, this will certainly not be true for all economies—bond yields are too low somewhere. All else equal, a decline in the cross-country discount rate lowers bond yields in all markets, stimulating output and credit growth. This may be welcome in the eurozone but will not be in developed and emerging markets with positive output gaps, excessive credit growth and/or overheating property markets. In these economies we are therefore likely to see a combination of higher short-term rates (and thus flatter yield curves), exchange rate appreciation and macroprudential tightening—i.e. tighter financial conditions through other means.
What about the U.S. in particular? Given solid growth in real activity, limited slack in labor markets and generally easy financial conditions, we believe long-term Treasury yields should be considered too low. If, for whatever reason, long-term rates remain low globally, policymakers will need to offset this unexpected stimulus in some way—likely via larger increases in short-term rates once they begin.