Equal-Weighting-O-Clasm

I landed in a town in Western Sumatra called Padang a few days after an earthquake hit. 7 or 8 on the Moment Magnitude scale.

Just as in finance there are various ways of measuring quakes. The Richter scale measures ground motion whereas the more modern Moment Magnitude scale measures energy released.

In any case my favourite measure of earthquake size is the number of deaths.

For example a 7 magnitude quake may mean 1 death in Canterbury, New Zealand (heart attack, not necessarily directly related) but >1,000 in West Sumatra and 1,200 seriously injured.

When I landed, I met up with my local pal from a nearby town and we set about finding a hotel. The rather morbid joke we had going at the time was that if a hotel didn't answer it the phone had succumbed to the quake and we rang up the next one.

I admit - the blackest of humour.

The largest hotel in town which I originally planned to stay in was a pile of rubble along with many other large buildings with questionable planning permission.

Part of the problem we have in finance is that financial engineering doesn't kill anybody. We have the questionable planning permission without the rubble.

Perhaps then we would take far more care in building financial edifices.

Imagine if momentum strategy funds could kill investors?

Actually cause a fatality or three! Perhaps there would be fewer momentum funds leaching their client's fees (oh! and fewer momentum evangelists beating me up!).

In any case, as part of my series of strategy iconoclasm I decided to take a look at equal weighting.

For example the Guggenheim RSP fund resets each constituent of the S&P 500 to a 1/500th weight every quarter or so.

Previously I have likened equally weighted portfolios to the ultimate random allocation benchmark but of course their bigger claim to fame is consistently beating the market.

For example RSP returns 9% per annum versus 7% SPY returns since 2003. The Sharpe is also an OK improvement, 0.44 versus 0.35.

Just OK - i.e. there are many other simpler more cost effective ideas out there.

Nevertheless, here is a rarely mentioned downside to Equal Weighting.

The skew of the S&P 500's daily returns is -0.1. Equal weight those 500 bad boys and skew drops to a whopping -0.26!

I.e. expect large drops and small gains on average, your upside is more constrained than your downside. Far more so than even the S&P 500.

We see this if we look at the 'downside beta' of the RSP with respect to the S&P 500.

The beta is about 1.07 when the S&P 500 drops, which means we will lose 7% more than the S&P on average.

When the market does well RSP does 5% better on average - or a beta of 1.05.

Your eye's aren't deceiving you. We have a 2% (5-7) deficit, and that coincidentally matches the average difference in returns.

In effect that 2% bump in returns which the RSP gives you is compensation for holding more of the crappiest S&P 500 stocks than you probably should.

This is yet another return / volatility 'anomaly' which dissolves away when you introduce an extra dimension.

Here are a bunch of other anomalies which I have looked into: