Sharpe versus Sortino - It does not seem to make a difference

In a very simple paper "Sharpe & Sortino - Does it Matter", Linus Nilsson present his findings on comparing the Sharpe and Sortino ratios. The correlation between the two risk measurement approaches is very high regardless of the strategy set used. You may think there is added value with using more complex risk measurements, but you will still get the same answer. This does not mean that you should throw-out other measures, rather the conclusion is that you will not get any more information from these alternatives. This similar risk comparison is even the case when there is positive skew with a strategy. The sample must be large to make a distinction.

Nevertheless, this works does not change our earlier view:

How should you rank hedge funds

Algo anxiety and aversion - Investors do fully trust AI

 

The FT recently posted an article on algo anxiety based on a paper titled, "Man vs Machine: The influence of AI forecast on investor beliefs". This is a topic that has been associated with technological change in finance for years. We have discussed this issue in the past as the problem of algorithm anxiety. Humans wants to embrace models, but they often don't trust models. 

The authors look at incentive experiments with over 3500 participants to determine whether investors are more likely to change their beliefs in response to a forecast generated by AI. The researchers find that investors are less responsive when an analyst incorporates AI to help generate forecasts. For all the talk about the AI revolution, there is still the issue that many do not trust the AI forecast, or at least there is a degree of caution or anxiety associated with these forecasts.

It was found that those who are more AI-literate will be more responsive, and women are more open to AI generated forecasts. New technology is feared less when it is better understood. New ways of generating forecasts may be effective but there is learning curve before it is accepted.

We have discussed this issue in the past. Here are a few references:

We forgive humans but not machines - Bias against algos

"Algorithm aversion" and managed futures

Algorithm aversion or just a desire for low cost optionality?

More on algorithm aversion - it can be eliminated if given some choice

Should you invest in the SG CTA index?

The SG CTA index has reached its 25th birthday and as a celebration, the people at Transtend offered their ode to the index called "Just Buy the Index". However, their paper does not offer the celebration that the index creators would like. There is no question that the SG CTA and Trend indices have become somewhat of a benchmark for the industry. It is popular, yet there are significant issues once you start to dig into the construction of the index. 

The SG CTA index is supposed to be investable through being an equal-weighted index of the largest 20 CTA funds that are open for investments. Hence, you can, in theory, replicate the index. Nevertheless, the components of the index will change every year with fund joining and being dropped from the index. The composition of the index today is not the same as the composition five or ten years ago. Since the index is for CTAs, it is not a trend-following index although many of the funds in the index are trend-followers. The index behavior will change with the composition of the managers and the changes in the strategies of the managers. It is hard to compare the behavior of the index today with performance ten years ago other than to say that both represent the behavior of the largest investable CTAs at a given point in time.

There is an argument that the index is not that different than a portfolio of stocks. These companies change their behavior through time like managers. The stock composition of the index will change like the CTA index. Hence, there is nothing special with investing in a manager benchmark.

Transtrend shows that if you invest with just the largest CTA listed in the index you will have a higher return and lower risk. You can beat the index by holding a subset of the index. Unfortunately, the work does not focus on what would happen if you randomly invested in a few of the index components. Is there a benefit with holding the index? That question was which was he title of the paper was not addressed. 

Volatility laundering and private equity - A big potential problem

An ongoing issue with private equity is that these investments may have more volatility than suggested in the data. It is not the volatility by itself that is the problem but rather the likelihood that there will be more downside risk with these investments than suggested with the standard deviation estimates. With no mark-to-markets on a daily basis, and prices determined usually at year-end by a process not based on sales and purchases, it is not easy to say what these investments are worth. This illiquidity and lack of pricing transparency suggests that the return for private equity should be higher than those of similar public investments. Yet, sellers have not marketed the infrequent pricing as an advantage given the smoother prices versus public investments. This is not really a benefit.

Cliff Asness describes this private equity volatility issue as "laundered volatility", generating lower volatility than the true investment will lead to excessive allocation or a strong demand by investors who want to have a lower overall portfolio volatility. Investors love return-smoothing and really don't want to know how managers got it. 

If volatility applied to private equity is too low, then for a given expected returns versus other assets, an optimizer will generate significantly higher allocations than what would be expected from an unlaundered volatility. The volatility should be roughed-up or at least made more realistic.

The argument is that since these are long-term investments, there does not need to be daily pricing, yet there are other investments that are long-term and are public with daily pricing. Should we smooth the prices of any long-term investment? These pricing issues have major ramifications for portfolio construction and asset allocation. We should get this right. 

For more on this issue see: "Demystifying Illiquid Assets:

Expected Returns for Private Equity"