top of page
Search

What is the optimal way for a market neutral quant factor fund to reach a volatility target?

Updated: Feb 14, 2023

Before answering this question, we need to consider why it is important. Headline performance if a function of return volatility. Without volatility, it’s not possible to generate attractive returns. Put simply: no risk, no reward.


The OQ Asia Absolute Alpha Fund targets a return volatility of 6% to 8%pa. Given the predictive power of the systematic investment process and the alpha generation from the non-systematic overlay, this should result in attractive fund returns.


What are the options?

There are several portfolio parameters than can be adjusted to achieve the desired return volatility:

  • Leverage: adjusting gross exposure

  • Stock specific risk: taking more or less concentrated stock bets

  • Risk factor tilts: adjusting net exposure to risk factors such as country, sector and size

  • Alpha factor tilts: adjusting net exposure to alpha factor families such as value and momentum

As a general rule, quant factor funds should have a relatively large number of stock positions and tight risk constraints. Alpha should be generated based on relative stock positioning. This is due to the fundamental law of active management: IR = IC & sqrt(Breadth).


If Breadth is low (eg. market timing or country selection models), the model’s predictive power needs to be extremely high. It is also hard to avoid data mining when there are limited degrees of freedom.


Having high breadth doesn’t simply equate to having a large number of stock positions. It is also important that the stock poisoning, rather than any risk tilts, drives portfolio performance. To illustrate: if most of the long positions are in one country and most of the short positions are in another country, it doesn’t matter how strong the stocks selection process is as relative country performance will likely drive fund performance.


The stock positions should also provide the desired alpha tilts. For our investment process, this means having a positive bias towards value, growth, momentum, sentiment, and certainty factors. Some of these alpha factor families are negatively correlated (eg value tends to be negative correlated with momentum). Hence, when value factors perform poorly, momentum factors tend to perform relatively strongly. This dampens return volatility and performance. Taking an outsized bet on one of these factor families using a factor timing model is potentially another way to increase return volatility. However, the typical breadth caveats apply: it’s hard to avoid data mining and you need to have relatively high conviction.


Is high leverage the answer?

All of this points towards having high leverage. Based on quant theory, stock specific risk, risk factor tilts and alpha factor tilts should be avoided. Hence, the only remaining lever to pull is leverage. While this sound sounds good in theory, in practice it is dangerous.


The main problem with high leverage is you can be forced to unwind positions at the most inopportune time. Potentially the decision making it taken out of your hands and you’re at the mercy of your prime brokers. And given recent events with Archegos Capital, these brokers are unlikely to show much mercy. It was the brokers which acted quickly and decisively that escaped unscathed.


It can be argued that Archegos Capital, which had a concentrated high conviction portfolio, is totally different to quant factor funds. However, this doesn’t mean that quant factor funds are immune from forced de-leveraging. The Great Quant Unwind in August 2007 clearly illustrates the dangers of high leverage for quant funds. The losses suffered by numerous quant factor funds which were forced to unwind positions were extreme. Quant funds which didn’t have high leverage and hence were able to take the opposite side of the unwind trades were able to benefit (at Macquarie, we generated a positive return in August 2007).


Extreme spikes in cross-sectional return volatility, such as what occurred in late 2008 and March 2020, can also result in forced deleveraging.


In addition, high leverage increases financing charges which can erode after-cost performance.


What’s the optimal mix between leverage, risk tilts and alpha tilts?

The extra conviction afforded by the “micro” non-systematic overlay allows us to take slightly more concentrated stock bets than could be done using a purely systematic methodology. As a guide, the number of positions in the portfolio would be approximately 30% higher using a purely systematic methodology (850 vs 650).


Similarly, our analysis of the market environment as it relates to our investment process allows us to take slightly larger risk and alpha tilts. This is done as part of our “macro” non-systematic overlay.


Currently, net risk tilts comprise:

  • Country: Net long Chinese equities and net short United States equities. Many Chinese stocks look attractive based on our stock selection process and we are currently getting some short exposure via US index futures contracts.

  • Sector: Net long Telecommunication Services, Industrial Conglomerates, and Construction & Engineering; and net short Beverages, Food & Staples Retailing, Office REITs, Restaurants and Leisure.

Current alpha tilts comprise a larger than usual net long Value exposure and smaller than usual net long Momentum and Growth exposures. This is based on our value spread analysis.


Gross exposure is currently around 240%. At this level, we will never be forced to unwind positions due to being overleveraged. It also leaves us with enough “dry powder” to take advantage of any liquidity driven price distortions during market dislocations.


We believe these are currently the optimal leverage, risk and alpha tilts to achieve our volatility and return targets. We adhere to the maxim that quant factor funds should have high breadth and tight risk constraints, but not to the point where they need to use extreme leverage. We also believe that our non-systematic overlay allows us to take risk and alpha tilts in a smart way which gives us an edge over systematic quant funds.


26 views0 comments

Recent Posts

See All

Investor sentiment & AI as a risk factor

“In the short run, the market is a voting machine but in the long run, it is a weighing machine.” This is a quote from Warren Buffett in 1987 who paraphrased a passage from Graham and Dodd’s seminal b

Comments


bottom of page