Categories
Sunday Papers

The Sunday Paper – Language Commonality and Sell-Side Information Production

Britain and America have been described as countries divided by a common language.

How then should we think of China? A country with 17-official language groups that break down into 105-language types; and that’s just official categorizations. Reality is messier.

Ruishen Zhang, of the Frankfurt School of Finance and Management, wondered if sell-side analysts covering companies where the CEO and the analyst shared a common dialect obtained a measurable informational advantage that subsequently came out in recommendations? They do, and significantly so.

To understand how important the use of dialects remains in China the researcher cites a 2006 study that revealed 86% of respondents using local dialects in preference to Standard Mandarin and a 2010 follow-up study that reported only 25%, 17% and 9% of respondents in Hebei, Jiangsu and Guanxi respectively using Standard Mandarin at home with families.

Public databases on both CEO’s and analyst’s ‘Native Place’ are available and by lining this information up with recommendations and then looking at the one year stock return the effect can be studied.

It seems where the CEOs and analysts share a common dialect it leads to stronger opinions from the analyst. Where this commonality exists there’s a 73% higher chance of the analyst rating a stock a STRONG BUY. Does implicit kinship though retard the production of SELL recommendations? Not at all, in fact it seems also to lead to their more frequent appearance.

What’s going on? It’s not entirely clear, but back-channeling of information seems unlikely. More likely, analysts with a language commonality with CEO’s are reading between lines and picking up dog-whistles better than those less well tuned to the dialect’s unique frequency.

Trading on both STRONG BUY and SELL recommendations from analysts with a shared dialect of the CEO of a covered company produced a return 6%~8% above market over the period observed. However, as the study window, 2005~2016, was a time when disclosure requirements were significantly improved it’s not a given this effect persists. I’d be prepared to bet though among smaller and less-liquid stocks, in A-share land in particular, it almost certainly does.

You can access the paper in full via this link Language Commonality.

Happy Sunday

print