Morningstar article on asset pricing anomalies was published today.
Cheers,
Rich
The paper simply illustrates the well-worn phenomenon that factor returns have been lower in the last 20 years. And the paper gives a wrong explanation: information technology. However, Edgar was born in 1993, but the dividing line in the paper is 2004/2005.
the phenomenon actually emerged because the dot-com bubble mania had largely subsided by this point, and so there were no longer as many dumb money to trade with. After the frenzy of 2020 we now have many more of these.
Then why haven't factor returns bounced back?
It has bounced back
From the paper: "…..excludes anomalies based on market variables (e.g., momentum, size, trading volume) and valuation multiples (e.g., market-to-book, earnings-to-price).
No analyst data either, I think.
That's a bit limiting as far as relevance to many P123 ranking systems. Although the authors may still have a valid point about how information is disseminated more quickly now.
Value ratios (which are not discussed in the paper) can continue to provide some information long after an earnings announcement if the price (also enterprise value or market-cap) has changed recently. Limiting, to some extent, any conclusions about time-decay of value factors, I think.
Claude seems to agree after uploading the paper and puts it this way which may be clearer, if redundant: "The exclusion of valuation ratios like market-to-book and earnings-to-price is particularly noteworthy because these metrics are dynamically updated as prices change, even when the accounting components remain static. For example, a stock's earnings-to-price ratio provides constantly updating information as the price moves, even if earnings haven't changed since the last report….The conclusion that anomaly returns are concentrated in the period immediately following information releases may be specific to pure accounting anomalies rather than hybrid metrics that combine accounting and market data"
Thus, “anomaly returns are at least partly due to delayed information processing by investors.”
I agree. FactSet is relatively fast in processing US stocks but also surprisingly fast for large European stocks. For example, most recent ACP:POL earnings were processed within 24 hours.
On the other hand, it is quite normal to wait +10 days to process data for European small-caps + errors, missing/lost data, etc.