In the prevailing low yield environment, some German institutional investors trying to meet the needs of their liability demands deem it necessary to increase the risks of their portfolios. For most investors however, increasing portfolio risk is not an easy thing to do. Firstly, risk budgets - against the background of a tight regulatory environment - are generally quite low. Secondly, risk premiums in the markets tend to be lower than they used to be. Thirdly, increasing risk - according to modern portfolio theory - leads to higher expected returns. However, there is no guarantee that higher returns will actually be achieved,
This is why active management will have
to play a more important role. In our view, some markets are not suitable for active portfolio management, but others can produce some extra return, and picking an active manager should make good sense for investors.
In traditional modern portfolio theory, however, there is no room for active management, the only source of risk/return is beta. But many years of experience show us, that active management not only creates alpha but in some cases leads to consistent outperformance.
Therefore, it should be possible, to analyse both sources of risk and return, alpha and beta and come up with conclusions on how to make efficient use of them. As we will show below, achieving higher overall performance without increasing risk is not impossible.
ALM approach
The breaking down of risk and return into beta and alpha implies, that there should be two separate risk budgets for both sources, beta and alpha. Significantly, both risks don't just add up. If correlation is negative or low, beta risk can be replaced by alpha risk without increasing the overall risk level of the portfolio.
This concept is shown in chart above right: At a given information ratio of 0.2, market risk can be reduced from 8.3% to 6.4%, leaving room for increasing the tracking error up to 5.3%. Despite the substitution of beta by alpha, overall portfolio risk remains stable at 8.3% and at the same time the return increases from 5.5% to 6.1%.
Surprisingly, despite the fact that many investors in Germany do employ active portfolio managers, many decision makers so far have not quantified the impact of active portfolio management on their portfolios. This is because most investors define their strategic asset allocation (SAA) on the basis of ALM-studies which exclusively use general market (i.e. beta) data.
Thus many investors do not
have a clue what active portfolio
management is doing to their portfolio and in which way risk budgets should be adjusted. We think, ALM-studies should use alpha-adjusted data, at least for those (apparently less efficient) markets where proven track records and high and robust information ratios can be found (equities, high yields, emerging markets etc). As a result, we would expect that investors get a better sense which part of their overall risk budget should be employed for active vs. passive portfolio management. Secondly, we would expect ALM modelling to start replacing part of beta risk by alpha risk, i.e.shifting a higher portion of the portfolio into markets with high information ratios at the expense of markets with low information ratios. All this we expect to happen without increasing the overall portfolio risk.
Manager allocation and high-alpha
Asset manager selection has a tremendous impact on the effective use of real risk budgets. However, not many investors in Germany make efficient use of manager selection. And this is despite the fact, that employing active portfolio managers is quite fashionable among institutional investors. But does it really make sense, to employ up to four or five active portfolio managers - as we often see in Germany - for similar or identical mandates with in general quite low tracking error goals?
While on the one hand there is some smoothing of tracking error among similar managers, in some cases - if active managers follow the same style - active management can even produce cluster risks, pushing overall portfolio risk into unwelcome territory. This calls for constant monitoring, as some investors we know have not detected the accrued cluster risk in the past.
Therefore, investors should use manager allocation: ie start analysing the correlation of risk profiles for each manager employed and how the overall risk budget is affected.
Investors might for example add managers with a different investment style and low correlation to existing managers and thereby gradually move to a more efficient more diversified portfolio. Finally, the combination of several managers with different investment philosophies and a low correlation of alphas should be able to lower the overall tracking error of the fund. Because the tracking error for each manager can be increased at the same time, the return of the overall portfolio will increase.
Investors should be aware that picking the best portfolio manager per segment is not enough as it does not avoid cluster risks per se. Therefore, prior to manager selection, investors are advised to analyse the correlation of potential mandates, investment styles etc and quantify their contribution to the overall portfolio, before they derive specific goals (tracking error etc) for specific mandates and manager selection. This is where manager selection at least has to be accomplished by the analysis of the risk adjusted return of each existing manager on an ongoing basis.
The concept of high alpha investing "aims at replacing beta by adding alpha in order to achieve high performances without increasing the overall risk.
The idea of high alpha investing combines strategic asset allocation, manager allocation, manager selection and controlling to a new holistic approach. Performance attribution analysis, risk analysis and style analysis have to be in place for monitoring and revising strategic goals of institutional investors on an ongoing basis.
As in many cases, the scarcity of one factor (in this case the problem of generally low risk budgets) leads to revising existing concepts in asset management. High alpha investing seems to be an appropriate answer to the problems of a low yield environment.
Rainer Buth is partner at FAROS Consulting in Frankfurt
No comments yet