The HKICS blackout study
8 June 2009
If you are not complete bored by the insider dealing blackout saga, then you may have read lovable local curmudgeon Jake van der Kamp ranting in a local newspaper a month ago that the Listing Committee (LC) had got it all wrong, that 94.6% of the respondents to an opinion poll on this site were also wrong, and that the information possessed by insiders ahead of results announcements (but after the year has ended) is of no advantage at all in their dealing decisions. He based these comments on a flawed study commissioned by the Hong Kong Institute of Chartered Secretaries (HKICS). What he didn't mention is that HKICS commissioned this study after having put their name in that famous tycoon-backed newspaper advertisement opposing the blackout extension in the first place.
The HKICS President is currently Natalia Seng Sze Ka Mee, who is executive director of Tricor Group, a subsidiary of The Bank of East Asia Co Ltd, the Chief Executive of which, David Li Kwok-Po, was a vocal critic of the blackout extension - and he should know a thing or two about insider dealing after buying his way out of SEC charges in the Dow Jones case. A Vice-President of HKICS is April Chan Yiu Wai Yee, Company Secretary of CLP Holdings Ltd, which also opposed the extended blackout. Another director of Tricor, Diana Chung Miu Yin, sits on the HKICS Council, as does Edith Shih, Company Secretary of Hutchison Whampoa Ltd, Polly Wong Oi Yee, Company Secretary of Dynamic Holdings Ltd, Seaman Kwok Siu Man, Company Secretary of S E A Holdings Ltd, and Richard Leung Wai Keung, INED of Asia Standard Hotel Group Ltd and Asia Standard International Group Ltd. so we count at least 7 Council members (out of 13) whose employers or companies openly opposed the blackout extension.
The study did not adjust for the size of transactions (you would think bets would be larger when the information advantage is higher), nor did it eliminate noise such as rights issue entitlements, open offer entitlements and bonus issues, all of which result in an involuntary increase in a director's nominal share interests without an active purchase. It is not even clear whether the study adjusted for dividends and other non-cash distributions, which add to returns.
The returns were also calculated relative to the Hang Seng Index, which is not necessarily the most reliable benchmark, particularly for the vast majority of companies (by number) which are not in it, handy though it is. It only covers about 69% of the market value. The choice of benchmark is important in market-adjusted performance studies, because if it is not in line with the average performance of all stocks in the survey then it can give skewed results.
The study examined post-results returns from day 1 to day 14 after the results announcement (it did not say whether this was trading days or calendar days) - but for companies which announce during the lunchtime halt, they should have been including day zero, when the information hits the price in the afternoon session. It may also be the case that positive earnings surprises are more likely to be released at lunchtime, to get the best media coverage, whereas negative earnings shocks are more likely to be released at night, avoiding publicity. So a safer study would have measured the returns from dealing date (i.e. at least 1 month before results) until 14 days after the results.
Furthermore, the study adopted an aggressive statistical significance level of 0.1%. That means that a finding was only considered significant if it could happen by pure chance only 1 in 1000 times. By analogy, the study would only consider a tossed coin to be biased if it came up with 21 heads or tails in a row (a 1 in 1024 probability) whereas most studies would look at the 5% (1 in 20, or more than 5 heads/tails in a row) or 1% (1 in 100, or more than 7 heads/tails in a row) levels. We assuming they are talking about two-tailed significance levels, but the study doesn't say.
Anyway, despite the deficiencies, the study showed that directors who sell shares before results announcements out-perform the market, and more so (by an average 7.8%) if they do it 3-4 months before the results announcement, before the rest of the market can get wind of the bad news. They also out-perform by selling after results announcements (before the next period-end), but not by as much, in our view because their inside information advantage is smaller. The study found no significant advantage for purchases, but that is probably because of the deficiencies in the methodology outlined above.
From any common sense perspective, information can never have negative value, and the greater the information advantage you have, the more likely you are to make money if you deal on it. It doesn't take a study for investors to know that.
The study's author
The study was conducted by Professor Chan Ka Lok of HKUST Business School. He also endorsed the methodology in a study of 5-year MPF returns from Apr-2001 to Mar-2006, which conveniently focused on dollar-weighted returns, which weren't as bad as the actual fund performance on a price basis, because the tech bubble was still deflating in the first 2 years of the study when the MPF had just started, and then the market rallied sharply from the SARS/ gulf war lows of 2003 when there was more money accumulated in the MPF system (it only started in Dec-2000).
That piece of MPFA propaganda, largely designed to deflect attention from criticisms of total expense ratios, also contained howlers such as "the annualized dollar-weighted return...was calculated by multiplying monthly IRR by 12". Will someone please explain compounding to this man, who incidentally, is the brother of Secretary for Financial Services and the Treasury Bureau, Ceajer Chan Ka Keung (aka K C Chan). The FSTB was of course heavily involved in lobbying the Stock Exchange and its Listing Committee to undo its extended blackout rule. Ceajer would enjoy reading his brother's report.
© Webb-site.com, 2009