Each year the Fund Mangers perform a detailed study of the state-of-the-art in finance theory. Each strategy is critically analyzed for theoretical soundness, implementability, and performance history. This study results in new candidate strategies, that are evaluated against previously implemented strategies. The new strategies are then carefully executed in parallel with surviving existing strategies. (Click the Model Name to download the academic research papers on which our strategies are based).
- Variant of Joseph Piotrosky's Value Investing: The Use of Historical Financial Statement Information to Separate Winner from Losers working paper
- Uses published accounting numbers to separate future winners from losers; significantly improves mean returns and variance
- Selects stocks based on nine factors measuring profitability, changes in capital structure and operating efficiency
Earnings Announcement Return (EAR)
- Based on the academic research paper Earnings Announcements are Full of Surprises by Santa Clara, Brandt, Kishore and Venkatachalam
- Variation of an earnings drift strategy which seeks to exploit the earnings drift anomaly; attempts to capture all information, quantitative and qualitative, surrounding a company's earnings release
- Takes long positions in companies whose stocks have experienced the highest quintile of excess returns around earnings announcements
- Seeks to exploit market price anomalies related to changes in economic variables found to be significant in explaining asset price returns, as described in Advanced Theory and Methodology of Tactical Asset Allocation by Wai Lee.
- Tactically rebalances asset class weightings away from a pre-determined Strategic Asset Allocation (SAA), which is mean variance optimized using long-dated historical time series of key macroeconomic indicators
- Tactical asset allocation decisions are made via factor signals shown to be predictive of excess returns for a particular asset class; relative weightings are scaled relative to aggressiveness factors related to the strength of the factor signal
- The BCD model estimates a stock's fair market value through a closed form extension of the Gordon growth dividend discount model.
- The model explicitly relates the stock’s fair value to its observable fundamental variables. The fundamental variables vital to the BCD model are the means, standard deviations and rates of mean reversion of the risk-free interest rate, EPS, and EPS growth rate.
- The model assumes that expected EPS growth follows a stochastic process and is hence dynamically changing over time, whereas growth is constant and perpetual in the Gordon model.
- Empirical results suggest that modeling the EPS growth process properly is of highest importance, which is what the BCD model attempts to do.