I wondered what it might look like if one were to trade the top-25 cities based on momentum. Again this is based on OFHEO data, which pertains to single-family residential.
The Test:
1. Rank the 25 cities based on price appreciation over the prior three quarters.
2. Invest 10% of the portfolio into each of the top-10 ranked cities.
3. Whenever a city falls out of the top-10, sell the investment (with 6% transaction cost) and buy whatever city has moved up to the top-10. These trades are modeled using a two-quarter lag, so as not to incorporate any hindsight bias. Two quarters is probably the minimum amount of time necessary to receive the price appreciation data from the government pertaining to the prior quarter and then make the trades.
Results:
1. With transaction costs, the momentum trading strategy results pretty much match the results achieved by simply buying and holding all 25 cities. The benchmark of buying and holding all 25 cities is without any re-balancing and the associated transaction costs.
2. Without transaction costs, the momentum trading strategy is far superior, generating much higher returns with approximately the same volatility.
Conclusion:
As with many quantitative strategies, transaction costs become the limiting factor. However, the insight is useful if one is deploying new money and would have to incur the transaction costs in any event.
One day I'd like to see how closely the OFHEO data tracks the appreciation of apartment properties and other commercial property types. Since those larger transaction sizes typically have much lower transaction costs (say 1.5%), I'd like to see if this would be a viable investment strategy, or at least a helpful augmentation.
