Accounting delay of WRDS Factors
It’s common practice to include several, if not dozens of, pricing factors in your empirical model. The computation of these factors is by no means an enjoyable experience. To solve the issue, WRDS launched a product named “Factors by WRDS” or “Backtester”" which provides 134 pre-computed pricing factors on a monthly basis.
WRDS offers a comprehensive manual for this product, but there’s one thing it doesn’t talked about: How do they deal with accounting delay?
To illustrate the problem, let’s consider the factor “Book-to-Market ratio (BtM)”. To get the book value, we need to refer to 10Q/K. Assume that the fiscal year end for a company is Dec 31, 2000, and the 10K is available on Feb 28, 2001 (the next year). Since there’s a delay between the fiscal year end and the release of 10Q/K, at day Dec 31, 2000, the market won’t be able to know the book value of Q4. To compute Book-to-Market ratio, we have to use the value of Q3.
In finance/accounting papers, researchers typically applies for a uniform delay to the accounting data. For example, in Sloan’s famous accrual paper (Sloan, 1996), he used a four month delay.
The problem is, WRDS doesn’t tell the users how much delay (if any) they applied to the computation of factors. So I asked them.
The replay from WRDS is prompt. The ansewr is two month. That is, at 2000Q4, they’ll only use 10K/Q of 2000Q3. The two-month lag is reasonable to me, because SEC requires public firms should report their 10K/Q within 45 days.
- Sloan, R. G. (1996). Do stock prices fully reflect information in accruals and cash flows about future earnings? Accounting Review, 289–315.