The Value factor has been exploited for decades by equity investors and academic research shows it can also be applied to fixed income, in particular to government bonds. But this requires more than a basic selection process based on individual value measures.
From a practical point of view, investing in value government bonds means precisely defining the value concept for this kind of security. At Robeco, we believe it can be seen as the sum of the yield and the capital gains over time, minus the currency hedging costs.
A value strategy focusing on government bonds will look for the best issuer countries and the most attractive maturities based on this measure of value. However, our backtests show that allocating naively according to value may lead to highly volatile performance.
Using market data from JP Morgan, covering a period from January 1985 to May 2016 for a sample of seven issuer countries and six maturity buckets, we simulated the returns of a ‘generic’ value strategy. Every month, this strategy would have invested half of its holdings in bonds of the best country and the other half in the best maturity of each country.
We compared its returns with those of a reference index investing in all 42 country-maturity combinations according to the bond market capitalization. Our calculations showed that the ‘generic value’ portfolio would have achieved higher but much more volatile excess returns than the index. The main reason is that the best maturity is often the longest maturity.
To address this issue, we designed a ‘smart’ value selection process. The idea is that investing only in the best country-maturity combination according to its value measure would maximize exposure to this factor, but would also mean ignoring risk and trading costs.
Meanwhile, being forced to buy at least one maturity bucket from each country would guarantee some diversification but may also lead to suboptimal exposure to the value factor. That is the case, for example, when a particular country only has bond maturities with poor or negative value measures.
In the long-run, investors can benefit from the bond risk premium and double it by exploiting the value factor
Our bond selection process seeks to find a balance between value, risk and costs. These three aspects are managed at portfolio level, leaving enough flexibility to select high value bonds. The starting point is to maximize the value exposure minus a penalty for transaction costs. Only when the increase in the value exposure exceeds these costs will the trade be executed.
To manage the risk, we require the portfolio duration to equal that of the market cap weighted portfolio of our universe. We also make sure that the portfolio is invested in at least two countries and four maturities, to ensure a minimum level of diversification.
Backtests using the dataset mentioned above show this ‘smart value’ approach efficiently harvests the value premium for government bonds. We calculate that, on average, this premium is 0.9% per annum. In the long-run, an investor can therefore not only benefit from the bond risk premium, which is also estimated at 0.9% per annum, but he can also double it by exploiting the value factor.
For investors who consider government bonds as a diversifier for equities there is also good news. In the seven worst equity drawdowns since 1990, Robeco smart value returned 4.5% compared to 4.1% for the reference index. Hence besides the diversification offered by safe bonds, the value premium also contributed 0.4% per drawdown.
This article was initially published in our Quant Quarterly magazine.
Dit document is niet beschikbaar voor gebruikers uit landen waar het aanbieden van buitenlandse financiële diensten niet is toegestaan, zoals ‘US persons’.
Uw gegevens worden niet aan derden verstrekt. Deze informatie is uitsluitend bedoeld voor professionele beleggers. Alle aanvragen worden gecontroleerd.