ASAM Check List

ASAM Strategy Checklist

If a part of the signal is itself calculated from stock returns, should you use high-frequency data to calculate your signal? Did you?
Has the core strategy been clearly specified so that another party can replicate your stock picks.
- Did you explain all data selection criteria with timing?
- Did you clearly explain all variables. Did you confirm timing?
- Did you show an example stock calculation?
- Is there a print-out of all stocks that would have been selected last January, with investment weights.
Did you hand-check the rate of returns of your strategy for at least one year, calculating by hand from as-early-scratch-as-can-be?
Are you sure that your variable is known by investors in the market before January?
For ease of debugging, are you using yearly rates of return as your dependent variable?
Does the Fama-Macbeth (FM) procedure suggest a positive cross-sectional gamma coefficient on your strategy. Your strategy here would or could be your strategy investment weights. (For us, it would be zero for most stocks in the CRSP sample.) Does the FM regression include all stocks in the CRSP universe (perhaps fitting the marketcap requirement), and not just the stocks that your strategy selected.
Does the Fama-French procedure suggest a positive time-series alpha?
- Are both the independent and the dependent variables zero-investment portfolios?
- Does the FF regression use XMKT and/or XMKT/SMB/HML as controls?
Was the strategy reasonably stable, so that the last 10 years still gave a reasonable performance, or has it all disappeared?
Did you check the performance of your strategy if you start to invest on Jan 1 vs. if you start to invest later in Jan (say, Jan 10 or later). (Holding should always end on Dec 31.)
Has another group confirmed your key recommendation for one year of your sample and/or replicate the annual performances?
Do your backtests contain at least one bear market / recession (e.g., 2008-10)?
If you combine multiple signals, do the signals have comparable standard deviations? (Otherwise, your ranking weights change!)
At the end---have you checked your specific stock selections by eye for heavy weighting towards industries or other aspects that you would consider less than desirable?