The 4 key questions to ask equity quants


(Wednesday, Jan, 09, 2019)
|   5 mins

Quantitative equity strategies or “quants”, have become a staple of most investor’s portfolios over the last decade. But how effective are manager research processes at assessing them?

Most research processes were designed to assess fundamental, “stock picking” managers, but given quants have vastly different investment methodologies and processes, one might question their efficacy when applied to the quant universe.

We would argue that traditional manager research processes are not appropriate and to assess quants, researchers need to have a very different mindset.

So, at Redington, we focus on 4 key questions.

1. How do you derive and test your investment hypotheses?

The danger for quants, which is less prevalent in the fundamental world, is that with the right data mining skills almost anything can be turned into a factor (shown expertly by the famous butter in Bangladesh example).

Overfitting the S & P 500

Source: David Leinweber, “Stupid Data Miner Tricks: Overfitting the S&P 500” (1995) 

It is very easy to mine seemingly explanatory factors and apply a narrative to it ex-post – especially with desirable factors the market is after (e.g. ESG). Therefore, new factors must have an intuitive rationale behind them underpinned by behavioural or economic reasoning.

There is also a temptation for quants to fight last year’s war. Removing what has “stopped working” while simultaneously adopting what has worked well recently. Avoiding managers chasing their tails with excessive factor turnover is usually a good strategy.

Hypotheses are followed by testing. In the highly evidence-based quant world, this is where ideas live or die.

Despite its importance, many gloss over the testing process and take backtests at face value. However, even sound hypotheses can be easily manipulated.


Source: Dilbert

Therefore, understanding the process behind tests is crucial. Those running backtests need to be properly incentivised and challenged, backtests should have statistically significant time frames and efficacy out of sample as well as in. The assumptions used should also be conservative and realistic.

We’re only interested in backtests that are created with intellectual honesty and reflective of the real world.

2. Do you have the resources to keep up with your competition?

When quants create and combine factors to integrate them into their model they need to source, interrogate and test the underlying data. However, this is not cheap and not easy.

Managers should have the resources to access sufficient data and allow experimentation with newer or differentiated data streams. They also need the manpower to clean and investigate the data thoroughly.

We focus on this due to the need for evolution of models. When looking at fundamental managers “process drift” is a dirty phrase but in the world of quants it is a race to stand still.

As a result, we are seeing increasingly creative methods of factor creation helped by an exponential growth in data and methods of analysis. Things like natural language processing, satellite imagery and artificial intelligence are now common sights on quants’ research agendas.

Finding managers with sufficient resources and willingness to evolve their process should lead you to more innovative managers. If quants don’t want to join this arms race, then they should be pricing themselves appropriately.

3. How are factors incorporated to create a final portfolio?

Which factors make it into the model and how they are combined will ultimately drive performance. We want to know why certain factors are selected and how they are going to be weighted.

For factor selection, solid justification and understanding of what factors are contributing to the portfolio are crucial. Adding seemingly differentiated factors (to justify higher fees) which are highly correlated with existing factors and contribute nothing is not desirable.

How the selected factors are then weighted will also make a huge difference. The below chart shows the performance and volatility of the exact same value factor, just constructed in different ways.

Performance of portfolios based on “High Cash Flow to Price” in US Large Caps over 35 years

High Cash-Flow to Price Chart

Source: Style Analytics

Given its importance, we want to understand how the weighting decision has been made. Ideally looking for more nuanced thinking than just selecting what’s worked well in the past.

Strategies may also look to dynamically weight factors. This “factor timing” is notoriously difficult, so one must dig into arguments and evidence regarding how a manager can add value by doing this.

We also spend time on the risk model and its influences on the portfolio. An inappropriate risk model may confuse alpha for risk and dampen the efficacy of a quants’ research. Whether managers are using off-the-shelf models or have built their own, we need to be sure it’s appropriate.

4. What is the approach to implementation and trading?

All this differentiated research, testing of factors and portfolio construction is useless if it isn’t effectively implemented. Reducing transaction costs and market impact will have a material influence on returns over the long-term.

Quants are often higher turnover than fundamental strategies and can be trading relatively illiquid names. Therefore, managers need to have effectively incorporated transaction costs in their model and have strong trading resources and experience.

Assets and capacity also have some unique aspects. With quants there is a key trade-off between scale and use of shorter-term factors which are often the most differentiated. Many quant strategies profess to be very scalable, but this can be at the expense of some of the most powerful factors.

So where do traditional manager research processes go wrong?

Parts of the manager research process are effective across the whole universe – firm culture, alignment, capacity, risk management and fees to name just a few.

However, there are elements relevant to fundamental managers that aren’t important with quants and vice versa. If researchers do not shift their mindset and fail to focus on these key questions they may miss the aspects that will define a quant’s success.

Obviously, this list of questions is not exhaustive and there’s plenty more to be asked. However, we’ve found it a valuable starting point to help our clients find some of the best quants in the world.


Pin It on Pinterest