The pro and con of fundamental indexing.
Last Tuesday the London Quant Group sponsored a boxing match between forces for and against fundamental indexing. Adam Olive was in the pro corner. Ed Fishwick was in the con corner.
Round 1: FI comes out swinging
Fundamental indexing is an alternative to market capitalization indexing.
The optimal thing to do is to buy the assets proportional to their fair value. But — certainly in the case of equities — we don’t know fair value.
A mathematical toy that has been produced is the Noisy Prices Model. This says that prices wander around randomly but mean-revert (on some time scale) towards fair value. Assuming this model is true, cap-weighted portfolios over-weight assets that are over-priced, and under-weight assets that are under-priced. Hence their returns are lower than they would have been if the weights were proportional to fair value.
A fundamental index attempts to find weights that are closer to fair value than cap weighting. The ingredients of the Robert Arnott et al version includes:
- book value
- cash flow
- dividends
- sales
Round 2: AFI strikes back hard
The counter-attack was more powerful than I had anticipated. This was not a frontal asault, as there was admission that the idea has some merit. This was going for the kidneys.
The real issue is how it was brought to the world. There are patents on the idea, there were possible conflicts of interest, there seems to have been over-selling.
The word “incensed” was used by more than one person.
There are also dissenting voices regarding the actual technique. This view talks of fundamentally flawed indexing.
Round 3: Does “fair value” exist?
This round played out entirely in my head.
The question is, of course, silly — of course any stock will have a fair value. We don’t know what it is, but it’s there somewhere.
But wait a minute. That assumes that fair value is exogenous to the market. The socially useful aspect of a stock market is to provide capital to those companies that can best use it. If that is to work, then fair value must be endogenous to the market.
It is easy to imagine that some small companies without capital get a boost in their fair value when the market takes a liking to them.
The opposite is probably the case as well. I could believe that Apple is getting fat and lazy because they are the darling of the market — that their fair value might actually increase if their stock fell out of favor.
Round 4: FI punches back
There came the claim that the mathematics in the fundamentally flawed paper were themselves fundamentally flawed.
Round 5: Is FI active?
In “What does ‘passive investing’ really mean?” I suggested that there are two dimensions: turnover and effort. Ed added another dimension: depth. If lots of people were to take up your passive strategy, would you need to change? If yes, then you actually have an active strategy.
Aside: I suggested that some subset of low volatility investing can be thought of as passive. How well does this pass the depth test? Given the popularity of lotteries, I’m doubtful that there will ever be so much money invested in low volatility funds to make their returns unreasonably low.
There was the claim that fundamental indexing is really an active strategy because it will lose its outperformance if everyone does it. I think this was an opening that FI failed to spot.
The way fundamental indexing was first sold was as a way to outperform — as an active strategy. But if you think of fundamental indexing as merely a way to get a reasonable return, then there is no problem with everyone else doing it as well. Cap weighting and fundamental indexing will converge.
Round 6: FI scores on volatility
The wide adoption of fundamental indexing would be a damper on volatility. While cap weighting encourages momentum, fundamental indexing encourages mean reversion.
The next event
The next LQG event will be on 2012 April 24 on the topic of news analytics.
Epilogue
Gentlemen, you can’t fight in here. This is the war room.
A reader who wishes to remain anonymous writes:
I don’t think the noisy prices model is a “toy model” – if it is then so are APT linear factor models in which stock prices follow random walks! There are several academic papers on Noisy price models – the first ones I know about are by Poterba and Summers in 1988.
There are papers on how noisy price models can generate the value, small cap and momentum effects. The first one was by Markowitz and some of the Research Affiliates, the first version came out in 2007 I believe and it has been revised several times since.
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=936272
Actually I do rather think of APT as a toy model. Toys can be useful. But point taken.