Monday 9 November 2009

Airbrushing analyst history

Hat-tip to Aled for alerting me to this rather interesting paper. Here's the abstract:
We document widespread ex post changes to the historical contents of the I/B/E/S analyst stock recommendations database. Across a sequence of seven downloads of the entire I/B/E/S recommendations database, obtained between 2000 and 2007, we find that between 6,594 (1.6%) and 97,579 (21.7%) of matched observations are different from one download to the next. The changes, which include alterations of recommendation levels, additions and deletions of records, and removal of analyst names, are non-random in nature: They cluster by analyst reputation, brokerage firm size and status, and recommendation boldness. The changes have a large and significant impact on the classification of trading signals and back-tests of three stylized facts: The profitability of trading signals, the profitability of changes in consensus recommendations, and persistence in individual analyst stock-picking ability.

2 comments:

Nick Drew said...

a parallel study would be valuable into the true performance of 'professional' economic & market forecasters

there are a couple I follow in detail (no names ...) and their records are disastrous: if anyone ever acted on their conclusions - which I tend to doubt - it would be to invite catastrophe

you never ever hear them offering to submit to back-testing on their output, though they certainly boast long and loud on anything they 'get right' - pure stopped-clock syndrome

(I wouldn't normally cite Brian Appleyard but he was good enough to quote me here on the topic of forecasting)

Tom Powdrill said...

yeah definitely on the same page with you. forecasting is clearly a mug's game, as most economists seem to accept. they seem to do it purely because they get asked to/paid to.

and obviously what also tends to also happen is that people use the forecasts when they say something helpful, but then remember that forecasts are pointless when they come up with the wrong answer