Sources of the current bond volatility
The return of bond volatility in Europe was expected. However, its severity was a big surprise indeed. During the Fed's three quantitative easing programmes, an increase in US long-term interest rates was observed during the twelve months after the initial purchases. This movement is explained by a rebound in inflation expectations, which confirms the effectiveness of this monetary policy, followed by an increase in the term premium, which reflects the uncertainty of medium-term inflation prospects.
In Europe, QE was deployed against a backdrop of worries about liquidity, resulting from the fact that the ECB's programme was meant to absorb 40% to 50% of risk-free bond issues. As a result, liquidity problems have reached a new historic high, if judged by the performance of the bid-ask spread for Bund futures (see graph 2).
Furthermore, the latest act in the Greek tragedy we are witnessing at the time of this writing has only reinforced these initial imbalances. One-year forward rates in Europe first went up from -0.7% to -0.5% between April and early June before easing up after Mario Draghi1 clearly indicated that the ECB would be very vigilant about "unwarranted tightening" of the eurozone's monetary conditions.
Paradoxically, an extension of the ECB's QE programme is not possible under the current structure without exacerbating liquidity problems in European bond markets. Against this backdrop, it is not surprising that the ECB announced it would expand the range of assets eligible for purchase to include investmentgrade corporate bonds.
The Fed's normalisation strategy will determine the level of medium-term volatility
Across the Atlantic, the situation is totally different, as Treasury bond volatility depends on Fed communications. As shown in Graph #4, expected interest rate trajectories have once again become a decisive factor in the volatility of long-term interest rates. This correlation is also strengthened by the fact that market expectations regarding the long part of the curve, as measured by five-year forward rates, are becoming increasingly sensitive to changes in expected monetary policy.
Furthermore, the US yield curve went through an anomaly last year that is now diminishing, with a very elevated correlation (0.8) between US five-year forward rates - a proxy for market expectations regarding long-term rates - and European one-year forward rates - a proxy for market expectations regarding ECB monetary policy. Since then, this correlation has been cut in half, and the correlation between the expectations related to the short and long segments has recovered from a low of -0.7 one year earlier, and now stands at +0.2.
This trend is a good short-term sign because it means that the sensitivity of US fixed income to European bond volatility is on the decline. Therefore, if we were to try to draw a theoretical curve of expected bond volatility, it would be reversed, with a volatility expected to be higher in the short term (over the next 1 to 3 months) than in the medium term (3 months - 12 months) or long term (> 1 year). However, that curve would have a positive slope on its long end, consistent with the prospects of Fed rate normalisation.
Overall, in the short term, disagreements between the Greek government and the Eurozone will add to liquidity problems in for European fixed income as long as the two parties fail to reach a viable deal. However, in the longer term, bond volatility is highly dependent on the uncertainty inherent in the Federal Reserve's normalisation process.
Credit markets face the risk of a steepening US yield curve
The strengthening of the Fed's communication tools limits economic uncertainty. We have witnessed a substantial decrease in indicators of economic uncertainty2, like the one developed by Scott Baker, Nick Bloom and Stephen Davis in use at the St. Louis Fed. However, this is not necessarily reassuring. Over a long period, economic uncertainty behaves cyclically over cycles lasting about ten years. The Baker-Bloom-Davis indicator is currently at a 30-year low. It is therefore reasonable to expect a rise in uncertainty over the next decade, though there is no need to fear that it will reach the heights seen during the decade between 2005 and 2015.
Economic uncertainty results in the risk of more volatile expectations. This has consequences on risk-free yields. By definition, a term premium reflects economic uncertainty. It is defined by the yield-to-maturity of long-term bonds that cannot be explained by expected changes in short-term bonds. In practice, the long-term rates are broken down into a component that depends on the future trajectory of short-term rates and a term premium tied to the uncertainty inherent in the act of projecting short-term rates.
The higher economic uncertainty climbs, the more the stability of investor expectations falls, leading to a rise in the term premium, and therefore a rise in the term spread, more commonly known as the yield curve slope.
A steepening of the US yield curve due to an increase in long-term rates ("bearsteepening") is not favourable to credit assets, as it strengthens the appeal of holding risk-free bonds.
The attractiveness of a buy and hold strategy for EM debt
What about emerging market debt assets in this environment? With the prospect of higher volatility for risk-free bonds, it is helpful to compare dollardenominated emerging debt to a dollar-denominated aggregate bond index, because this category of indices contains both loan instruments and nominal or inflation-indexed Treasury securities. As a result, this type of index is a good proxy for a non-emerging market allocation.
Based on historic data going back to 1988, the EMBI Global Diversified dollar-denominated emerging market sovereign debt index has an average excess return, or ex-post risk premium, of 3.7%, compared to the Barclays US aggregate index.
Additionally, duration-adjusted yield is, on average, 1.7% higher for emerging market debt compared to the aggregate yield over that period3. For a similar duration, buy and hold is unsurprisingly more attractive.
Because our concern is bond volatility, it must be emphasised that the durationadjusted yield ratio is currently at a level of volatility compatible with a 15% higher volatility ratio. Dollar-denominated emerging market debt is therefore attractive in the short term. Concretely, the volatility spread is 140 bp, which could increase to 160 bp without yields needing to go up.
The diversifying ability of emerging debt should increase
The volatility of the excess returns on dollar-denominated sovereign debt is, on average, 30% higher than the aggregate index. It has fluctuated between 4% and 6% since 2010. The result is that an allocation based on maximising the Sharpe ratio (tangent portfolio) leads to exposure to emerging debt of between 7% and 20% over the period from 2010 to the present (which leaves 80% to 93% allocated to non-emerging markets).
What would be the consequence of a 100 bp increase in volatility in the aggregate part of the allocation? The volatility of the excess returns on emerging-market debt would increase by 103 bp. That said, calculating the contributions to portfolio variance shows (see box) that the increase in volatility for excess returns is chiefly proportionate to that of the volatility of nonemerging market assets. Even more important, the proportionality constant is equal to the inverse correlation between the ex-post premium and the nonemerging component's yield.
It is interesting to note that the sign of this correlation is negative, which means that the excess return of emerging debt is generally changing in the opposite direction to that of the non-emerging component. This is because investors tend to trade Treasury bonds, or even corporate investment grade bonds, in exchange for emerging market debt.
The dynamic of this correlation largely reflects the market expectations regarding the term premium or spread on the US yield curve. It is also ahead of the curve's movements. This is because a term spread that increases due to bear-steepening corresponds to the renewed appeal of long-term risk-free bonds. In this case, the excess return on emerging market assets tends to decline, or even go negative.
That said, it is important to note that the correlation between the excess return on emerging market debt and the performance of non-emerging bonds and debt securities needs to drop. It is still too high, while the risk of future pressures on the term premium is only increasing. With the prospect that the Fed's rates will normalise, even gradually, the term premium should go up due to the rise in economic uncertainty inherent in such a process.
If that were to happen, the correlation between the volatility of excess returns and the performance of the non-emerging component would decrease. In other words, the excess return on the emerging component compared to the non-emerging part would become more stable. This is why emerging market debt should therefore gain diversifying power in the medium term, although in the short term, the rise in bond volatility can only have adverse effects on this asset class. The result is that in a strategic allocation, exposure to emergingmarket debt should increase despite a rise in its own level of volatility.
1 “We are closely monitoring conditions to detect signs of an unwarranted tightening of our stance, to which we would need to react.” Mario Draghi, President of the ECB, 15 June 2015, hearing at the European Parliament’s Economic and Monetary Affairs Committee, Brussels. (Hearing at the European Parliament’s Economic and Monetary Affairs Committee, Introductory statement by Mario Draghi, President of the ECB, Brussels, 15 June 2015).
2 The Baker-Bloom-Davis indicator is based on three components: the frequency of certain items referring to the notion of economic uncertainty appearing in newspapers, the dispersion of forecasts on infl ation and on growth, and provisions changing tax law at the federal level.
3 Currently, the yield of the Barclays US Aggregate index is 2.4% for a duration of 6, or a duration-adjusted yield of 0.4%, while the yield of the EMBI Global Diversified index is 5.8% for a duration of 6.9, which gives a duration-adjusted yield of 0.8%.
The return of bond volatility
A steepening of the US yield
The correlation between
An exploration of the marginal contribution of emerging