Regarding the backtesting parameters in SignalDoc.csv (e.g., Stock Weight, LS Quantile, Start Month, Portfolio Period, Filter):
Are these values extracted directly from the original factor papers, or are they standardized implementation choices made by Chen & Zimmermann for their replication framework?
For example, I noticed that many monthly-rebalanced factors (Portfolio Period = 1) have Start Month = 6, which doesn't seem to come from the original papers. The Notes column also contains entries like:
BM: "OP actually forms portfolios at the end of March… But for consistency with FF1992, we use June as the port form month."
Mom6m: "OP doesn't skip the current month… We follow McLean and Pontiff in skipping the current month instead."
BidAskSpread: "We use Corwin Schulz spread following MP, but OP uses Fitch's Stock Quotations."
This suggests SignalDoc records C&Z's standardized replication settings rather than the original papers' methodology. Could you clarify which columns reflect original paper choices vs. C&Z's implementation decisions?
Regarding the backtesting parameters in SignalDoc.csv (e.g., Stock Weight, LS Quantile, Start Month, Portfolio Period, Filter):
Are these values extracted directly from the original factor papers, or are they standardized implementation choices made by Chen & Zimmermann for their replication framework?
For example, I noticed that many monthly-rebalanced factors (Portfolio Period = 1) have Start Month = 6, which doesn't seem to come from the original papers. The Notes column also contains entries like:
BM: "OP actually forms portfolios at the end of March… But for consistency with FF1992, we use June as the port form month."
Mom6m: "OP doesn't skip the current month… We follow McLean and Pontiff in skipping the current month instead."
BidAskSpread: "We use Corwin Schulz spread following MP, but OP uses Fitch's Stock Quotations."
This suggests SignalDoc records C&Z's standardized replication settings rather than the original papers' methodology. Could you clarify which columns reflect original paper choices vs. C&Z's implementation decisions?