Monday, February 11th, 2013

The Problem with Value

We don’t think for a minute that John Authers (FT, The Long View, February 9th 2013) consciously set out to endorse our approach in his recent review of annual investment returns, but his final paragraph neatly frames many of the ideas we use. “So buying into funds that keep costs low by following disciplined, quantitative strategies to invest in value, high-dividend, or small-cap stocks, or to harness the momentum effect, looks like a great idea.”

Our probability-based models explicitly use past returns as an indicator of future returns. We focus on the momentum of total returns rather than capital gains, which means that we capture the high yield effect. We recommend that clients use low-cost passive instruments like ETF’s or futures to implement our model. We don’t invest in individual stocks, so we cannot harvest the small-cap effect, but in terms of asset allocation we do the next best thing, which is to identify small (emerging) markets as a separate asset class.

There is no place for value in our process, not because it is a bad idea per se, but it’s much harder to define in practice in real time, than it is in theory with hindsight. At the very least, the raw data need to be adjusted for the cycle (thank you Professor Shiller). We would also argue that different definitions of value (price to book, price to cashflow, price to earnings) have different impacts on performance at different stages of the cycle.

There are two other issues which need to be discussed. First, it can be difficult to adjust a value style for risk. The best equity value tends to emerge when the company (or sector or country) is most distressed. This is often accompanied by higher than average volatility, variable liquidity, and a breakdown in previously-established correlations (which makes nonsense of VAR-based control techniques).

Second, most back-tests assume there is unrestricted liquidity at all times, but this may not be true for distressed equities. The assumption can give rise to a counter factual paradox. A stock is cheap because no-one wants to buy it. If an investor had offered to buy it, restricted liquidity would have forced him to pay a higher price, maybe not for the first bargain, but certainly for subsequent ones. Each model has its own set of problems, but restricted liquidity matters more to a value style than it does to momentum.

Leave a Reply

Full Blog Archive:


Other Recent Posts