Do you know your thematic stock basket returns ?

Well, kinda….

Yes quantitative investing can be amazing – it can help you get higher risk adjusted returns but not even remotely close to the kind that folks are promising these days. We regularly see strategies promising 5-10% alpha. We get these backtests, as the managers want to promote their quant models to our investors. We would happily do if they were any good.

Below is the standard set of questions I ask them and few if any care to respond. You can use them to evaluate your thematic manager too.

Ideally I want the manager to come back and say they are using 50 – 75 bps for bid-ask spread and an additional 50 – 75 bps for impact cost.

Most will come back and say bid-ask in the Indian market context is 10-25 bps. While that may be true for the mega caps, it is certainly not true for other large caps and certainly not true for mid caps and small caps that a lot of these strategies will end up investing in.

It gets even more interesting for impact cost. Most don’t assume any impact cost. A large order will move the market – which is what is captured by impact cost. The argument is that since each retail order will be so small (Avg Rs 10 – 25k) there will be no impact. This is a bad argument.

If you are following a thematic basket and 1k other are too, it is your combined order volume that matters for impact cost. How will one 10k buy order move the market argument doesn’t hold if their are thousands more 10k buy orders also in queue 🙂

For a high churn strategy, properly attributing just these two costs can dock 3-5% per annum from theoretical strategy returns. If you are using a full service broker or a fixed fee broker you need to add your actual broking charges too – it all adds up.

This one is a biggie. Stock churn in India will invite short term and long term capital gain taxes on returns. While all stock based thematic strategies churn a lot, not a single manager includes this in the backtest or theoretical performance numbers. They all hand-wave around it. As a rule deduct 10% or 15% from the promised returns for capital gains taxes based on how frequently the strategy churns. For a strategy promising 20% returns that is 2% or 3% paid per annum to the taxman that you won’t get.

A quant MF or ETF where stock churn is tax free has a distinct advantage here.

If they say they took the stock universe as of today and ran the backtest, run away as fast as you can. You will be surprised how many fall for whats known as “survivorship bias”. On most strategies it can easily dock 3-4% per year on realised strategy returns if not handled well.

Say your manager runs a momentum strategy and claims to use 5 day momentum in the backtests. Ask them to show results for 2 / 5 / 7 / 10 / 15 / 20 day momentum as well. This ensures that they didn’t just send the best performing backtest to you to invest in. You would think most managers claiming a quantitative strategy would have this handy or would be happy to share it – they don’t.

Q How does your strategy perform in other markets? How does different specifications perform in other markets? 

This is a test of robustness. Is your manager picking a “signal” or a “noise”? If you push statistics and data enough it will confess to anything. So it is important to run the strategy on markets where it is not formulated.  It is the easiest out of sample test a manager can do, but they don’t.

Read more.

The sad part is most investors don’t ask the above and don’t have the tools to compare their realised returns to promised strategy returns.

But, if a manager answers all the above satisfactorily then you need to answer one last question. Is tracking a strategy and rushing to rebalance when something changes the best use of your time? If it is, then go ahead you have likely found a thematic strategy / manager that will not disappoint you.