The elusive one-Greek geek

In the land of volatility, is the one-Greek (“alpha”) geek still king?

How often has the phrase “it’s a stock-picker’s market” been used? This begs the question: are some conditions more favourable than others? And are these conditions more favourable with regard to competitors or versus the market as a whole? Guy Fletcher, Head of Institutional Solutions & Research at Sanlam Investments, unpacks it for us.

The “One-Greek” is an allusion to the Greek symbols used by market practitioners to define their chosen outcomes and variables, with the most important one being “alpha”, or returns in excess of the market. We celebrate alpha like no other Greek, if one observes the annual Raging Bull awards. But are there contributory factors that impact alpha? And is that alpha dependent upon levels of market volatility?

Intuitively, says Fletcher, one would expect the alpha seekers to appreciate higher return dispersions and volatility since this should promote the opportunity for the more discerning managers to demonstrate superior stock-picking skill.
Indeed, for funds that have outperformed, it is postulated that a positive correlation between high or rising CSV (dispersion in returns between stocks) and outperformance exists. Sanlam Investments has tested this theory through their own proprietary research to explore the outperformance of one of their best-performing equity funds.

Regarding this research, Fletcher notes the following points:

1. As a general rule of thumb, statistics indicate that higher volatility (as measured by absolute and cross-sectional variables) provides greater opportunity for active managers to outperform. However, it has a downside in that it is accompanied by a greater probability to underperform.
2. The better active managers do not necessarily require higher volatility to deliver alpha, but can execute in any environment.
3. Logically, there may be some merit in moving from an active to a passive strategy as volatility declines (in particular CSV); however, given that “active” is most often wedded to a specific manager or group of managers who will probably not have a consistent relationship with volatility, the benefits will largely be dissipated in transactional costs.
4. The ultimate goal is balance – to deliver consistent excess returns through various market conditions, requires investing with outperforming active managers that are largely uncorrelated, and incorporating them with other uncorrelated strategies such as smart beta and portable alpha.

We started by looking at the two major measures of volatility to describe market conditions:

1. Returns volatility – represented by the SAVI (South African Volatility Index) – the premise for this indicator is two-fold:
   1. Volatile movements in the market (both up and down) should allow market-timers to benefit.
   2. Similarly, aggressive movements in price, unrelated to earnings, should reveal significant mis-pricings, thereby allowing the active stock-picker to benefit.
2. Cross-sectional volatility (CSV) – this is a measure of the dispersion in returns across all the constituents within an index, weighted by the market cap of each constituent.
   1. A CSV of 0 would imply no difference in returns between the constituents and thus no ability to either pick an out-performing stock, or avoid an underperforming one.
   2. b) In contrast, a high CSV should be beneficial and allow for greater differentiation of “true skill” versus “luck”. Since it is a size-weighted measure, it should also support investability, i.e. the ability to execute on the plan.

Taking the above into account, we will now profile one of SIM’s favourite long-term outperformers (represented in both markets) to see whether we can glean any insights into whether its excess performance can be attributed to particular market conditions.

Let’s summarise the data:

Period: July 2007 to June 2017: 10 years

The table above confirms a number of observations:

1. 1. 1. The SWIX is a tough target to outperform
      2. Costs play a meaningful role in relative performance, i.e.
         1. The institutional performances are shown BEFORE management fees
         2. The retail performances are AFTER all fees
      3. The proliferation and turnover in retail funds has led to significant underperformance of the average retail fund against the average institutional fund (even after adjusting for higher fees)

A chart has a far more meaningful impact than the raw numbers, so below we have a comparison over time. We include both retail and institutional averages for interest sake; however, we will focus on the better performing institutional environment in the further analysis:

Are there certain environments that support outperformance?

First, let’s look at the institutional averages and add the charts of the normalised measures in the figure below:

Excess return is plotted on the left hand scale, whereas the normalised measures are on the right hand scale. Tabulating the relationships reveals the following:

What these tables indicate is that the bulk of the excess returns occur when the indicators are above average and the bulk of the underperformance occurs when the indicators are below average. These numbers are similarly borne out by the significant correlations of 52% and 58% of the excess returns with the volatility measures respectively. Intuitively, we appear to be on track.

Finally, let’s see how our long-term outperformer delivers excess returns. First, the chart:

Visually, there appears to be no real relationship between excess returns and the level of volatility – this is similarly borne out by the tabulated figures:

Excess return appear to be more randomly distributed, and this is also supported by the correlations that are 29% and 12% respectively. In particular, SAVI is now no longer significant.

Out of interest, if we run the same analysis on our favourite equity fund, but this time against the Average Institutional / Retail manager returns, we exacerbate the situation with the correlations now at 12% and -16% respectively!

Could SIM’s best-performing equity fund be an outlier?

In truth, not in respect of its relationship to volatility. An analysis of our major competitors (both outperformers and underperformers) in the institutional space, demonstrates correlations of excess returns to CSV that range between 69% and -14%; similarly, correlations to SAVI range between 52% and -49%. Demonstrably therefore, this suggests that volatility would not be the sole, or even major, driver of excess performance for institutional managers.

What about return dispersions?

Let’s look at the chart below – it shows the institutional 12-month rolling range between the best and worst managers, excluding outliers:

Visually, this would appear to hold the greatest interest so far. But the statistics don’t really support this:

With correlations to the measures of 41% and 10% for CSV and SAVI respectively, we seem to have retrogressed!

Conclusion

In general, we can confirm that CSV and SAVI are positively correlated with the Average Equity Managers ability to deliver excess return, with CSV having the more powerful association. However, we can also indicate that a high CSV is certainly not a pre-requirement for alpha delivery.

And our one-Greek geek?

It would certainly appear that the irregular land of volatility is not an impediment to the single-minded pursuit of the alpha, and so our one-Greek geek remains the king of the performance stakes. And we must simply unpack new relationships in new lands to isolate the additional drivers of alpha, other than volatility.