SFAT's red flag on Moody's chills negative research
8 April 2016
On 31-Mar-2016, the Securities and Futures Appeals Tribunal (SFAT) handed down its Determination in the appeal of Moody's Investors Service Hong Kong Ltd against the SFC's decision to reprimand it and fine it HK$23m. The Tribunal partially upheld the SFC's findings and reduced the fine to HK$11m, as announced by the SFC on 5-Apr-2016. The findings relate to a Moody's media release titled "Accounting and governance warning signs for emerging market companies" and an accompanying report published on 11-Jul-2011 titled "Red Flags for Emerging-Market Companies: A Focus on China".
Webb-site agrees with part of the judgment, that the SFC had jurisdiction as licensor, and that the licensing system imposes a constitutionally necessary restriction on free speech for a legitimate purpose. However, we find other parts, and the SFC and SFAT's general approach to exercising this restriction, deeply disturbing. It sets far too high a floor on the acceptable standards of critical research and will have a chilling effect on negative criticism of companies in Hong Kong, in a market in which there is little enough of that already. This approach is part of a pattern that indicates a regulatory bias against negative criticism and short-selling. That pattern includes a reporting threshold of 0.02% for any short position, on a weekly basis, versus a 5% disclosure threshold for long positions.
If Moody's (or anyone) had published a "Green flags report" singling out those companies (positive outliers) which it thinks have exceptionally good accounting and governance, causing their share and bond prices to go up, would the SFC have clobbered them for minor errors in the flags?
As we explain below, the SFAT also displays a lack of understanding of what drives credit ratings and what investors expect from licensed researchers. Perhaps this was to be expected: the SFAT was Chaired by a senior Judge, Michael John Hartmann, with 2 lay members, Dr Billy Mak Sui Choi (Dr Mak), an academic, and Ms Ding Chen (Ms Ding). Dr Mak also sits on the SFC's Academic and Accreditation Advisory Committee and several government bodies.
Of the three, only Ms Ding might have market experience, as CEO of mainland-controlled CSOP Asset Management Ltd, which sells ETFs in and out of the mainland market. Here she is on CNBC on 16-Jun-2015 (Tuesday morning, HK time) at the top of the mainland bubble, promoting CSOP's newly-launched HK-listed (3147) ETF on the Shenzhen ChiNext market as the "future of China". She remarked that the PEG (P/E to earnings growth) ratio was "not that expensive - only 4.6". Perhaps the SFC should look into whether that was a false and misleading statement. Could any licensed professional, with due diligence, reasonably call a PEG ratio of 4.6 in a market trading at over 100 times historic earnings "not that expensive"? But then, the SFC has never acted against anyone who has been overly bullish in their opinions.
The ETF is down 45.6% since then. Ms Ding also sits on the high-level SFC Advisory Committee (appointed by the Financial Secretary), the SFC Process Review Panel (appointed by the Chief Executive of HK) and the SFC Products Advisory Committee (appointed by the SFC), amongst other bodies.
So what about the Moody's report?
About 1.2% wrong
Investors don't expect perfection in research, or anything close to it. They understand that any analyst whose buy or sell recommendations are correct more than about 51% of the time can out-perform the market and thereby add value to a portfolio. They understand that analysts can make mistakes. The Moody's report was an application of binary "red flag" framework to 61 companies across 20 separate binary tests - each of which produced a flag or no-flag result.
Moody's did not claim to have done any multiple regression analysis, statistical significance tests or back-testing of flags as predictors of subsequent credit ratings or stock performance. It was a point-in-time study. Moody's did not attempt to weight the flags, and in our view that was sensible, because with only 61 companies in several groups, the sample sizes would be too small to produce statistically significant results. It was a fairly simple exercise and was presented as such, highlighting "risks that may warrant further investigation".
That's 1220 tests and results, although the report only published the detailed flags for the 49 "high yield" (rated below Baa3) companies, of which 26 were property developers. And how many mistakes in this matrix did Moody's make? Just 12, as shown on page 67 of the report, including 4 mathematical errors, 6 input errors, and 2 cases where a newly created management role was mistaken as a change of senior management. Of course it would have been better to be perfect, but on the 49 companies with 980 tests, that is about a 1.2% error rate which any reasonable reader can live with.
Reading through the 82 page ruling, we note that the SFC and the Tribunal repeatedly emphasised that Moody's had said that there was "limited correlation in China between lower credit ratings and larger numbers of red flags" and sought to conclude from this that the report should not have been issued, or that those companies with the most red flags should not have been singled out. The word "correlation" appears 18 times in the SFAT report. Moody's noted that:
"In China, property companies with lower ratings do not tend to have a greater number of red flags. However, a degree of correlation exists for non-property firms..."
But here's what the SFAT seems to have missed. Moody's was looking at "accounting and governance warning signs". A badly-run but well-capitalised company (or government) will naturally have a better credit rating than a badly-run under-capitalised company (or government), because what credit ratings are supposed to measure is the probability of default on the debt and interest obligations, not the governance or accounting quality. Put simply, it is the probability of running out of cash. So it is unsurprising that there was only a small correlation between current credit ratings and red flags. Of the 5 areas into which the flags were grouped, only 2 of them could be said to relate to credit worthiness, involving big capital investments or a mismatch between cashflows and accounting profits.
Secondly, it is unsurprising that property companies had no visible correlation - because they have real assets (land and buildings) which tend to provide more comfort for creditors than say, a mineral exploration company or a retailer.
For these reasons alone, it is obvious to an educated reader that the red flags were a very different exercise to then-current credit ratings and that the correlation between flags and credit ratings should be low, but it does not follow that the flags were meaningless and that the report was somehow misleading.
Credit ratings are not static over time. A badly-run, well-capitalised company may eventually burn through its capital and find it harder to raise more equity to refill the coffers. Remember that debt ranks ahead of equity on a liquidation, so the red flags are of even greater concern to equity investors than creditors.
How did they do?
If the Moody's report was so shoddy, you would think that the red-flagged companies would have rebounded from the criticism by now, and that the market commentators who quickly disagreed with Moody's (as they were entitled to do) would have been proven right. Nearly 5 years later, let's take a look at whether the pudding was worth eating.
Moody's identified 6 companies which were "negative outliers" with the greatest number of red flags, of which 5 were in the non-property group which, Moody's had noted, had some correlation between flags and credit ratings. Of the 6 companies, 5 were listed in HK (not 4 as the SFAT says): West China Cement Ltd (2233), Winsway Enterprises Holdings Ltd (1733, then "Winsway Coking Coal Holdings Ltd"), China Lumena New Materials Corp (0067), Hidili Industry International Development Ltd (1393) and Longfor Properties Co Ltd (0960). So how did they do - was there any predictive value? Click here to see the graphs, or look at this table of returns from 11-Jul-2011 to 7-Apr-2016:
|2233||West China Cement||-38.24%|
|2800||The Tracker Fund of HK||+5.14%|
The performance is actually worse than it looks. Did you spot the asterisks? That means the stock is suspended from trading.
Winsway and Hidili were both suspended on 1-Apr-2016 for failing to publish their 2015 results on time. Winsway defaulted on its offshore bonds on 8-May-2015 and is now in the middle of a debt restructuring exercise. Hidili began defaulting on its onshore debts in Jun-2015, but didn't say so until 30-Oct-2015, and is now in negotiations for debt restructuring.
China Lumena has been suspended since 10:56 on 25-Mar-2014 as it was unable to produce its 2013 results on time and to answer criticisms in two research reports. Webb-site warned readers about this company back on 25-May-2010, over a year before the Moody's report. On 4-Mar-2015, provisional liquidators were appointed, and on 2-Jul-2015, the INEDs all quit. The 2013 results are still not published, and the second stage of delisting has just expired, so it is one step away from delisting.
So that's a pretty impressive hit-rate for Moody's with 4 of the 6 "negative outliers" defaulting on their debts within 4 years, and the other 2 having massively under-performed the stock market, even though one of them, West China Cement, is now under a takeover offer. Tell us again, what was wrong with the red flags report?
Moody's chose to keep its report to a relatively brief 25 pages (including disclaimers). Except for the 6 outliers, it did not provide a detailed explanation or analysis for each of the 1220 tests performed or the 980 tests on the High Yield companies. The SFAT wrote:
"the Tribunal has little difficulty in concluding that, in its failure to provide commentary on all of the red flags, Moody's not only made it impossible for readers to accurately assess the significance of the flags in context but, by that fact, created and unfair, unclear and misleading picture of creditworthiness of a material number of the companies allocated red flags. As such, its failure constitute a breach of General Principle 1 of the Code of Conduct..."
In our view, that strays into micro-management of how a research firm chooses to present its work. The methodology of the 20 tests was adequately described in Appendix 1 of the Moody's report. It is unreasonable to say to researchers who compile such surveys of the corporate landscape that they must either publish all of their workings on each test, or no report at all. Having seen the flags, given that they were all based on public disclosures, it was possible for any reader who wished to dig further to do so on her own, and to verify the findings herself. Moody's had also made it clear that it was not ascribing weightings to the flags, and that they were only areas warranting further investigation.
Jurisdiction and free speech
We agree with the Tribunal's conclusion that Moody's was carrying out a licensed function by publishing the report under its name and in connection with its credit ratings. If a different unlicensed company (perhaps a fellow subsidiary) or a publication like Webb-site Reports, had published the report instead, making it clearer that the flags were not themselves credit ratings, then the SFC would not have had jurisdiction as licensor, but that was not the case. It was unfortunate that Moody's chose to attack the SFC's jurisdiction rather than just confronting the allegation head-on.
We also agree that, if you are writing as a licensed entity, then you are held to higher standards than the general public in your comments. The freedom of speech is not absolute, and there is a legitimate purpose fulfilled by imposing minimum standards so that customers can rely on the licensing system as providing some basic level of informed and skilled research - otherwise there would be little point in a licensing system. So the licensing system, in itself, is not a constitutionally unnecessary restriction on Basic Law/Bill of Rights free speech. But that does not mean that research should be flawless or that standards should be set so high as to deter research for fear of minor errors causing regulatory sanctions.
Could the report have been better-written, and clearer in the limitations of its findings? Yes it could. Could the flag-tests have been better than 98.8% correct? Yes, they could. Was the report of a lower standard than all the other pieces of (mostly positive) research that the SFC has allowed to circulate without interference? Certainly not. That's what makes a market - and research firms rise and fall based on the quality of their output. We liked the Moody's report, and we want to see more of that kind of critical research - but what licensed firm will now dare to publish such a report if the regulator is going to pick it apart afterwards and then slam them with a fine and potential loss of licenses for the individuals involved?
If listed companies disagree with research reports, they are of course entitled to respond with rebuttals, clarifications of their past disclosures or explanations, to ask for corrections, or even to sue for libel or defamation. As far as we know, none of the companies involved has sued - the criticism wasn't that far wrong.
Not only has the SFC pursued a licensee's report, they have also gone after an unlicensed person in the Market Misconduct Tribunal for expressing his negative opinions about a company while putting his money where his mouth was and being short: Andrew Left, of Citron Research, writing about Evergrande Real Estate Group Ltd (3333). The verdict in that case (also Chaired by Justice Hartmann) is awaited. In our view, unless the SFC can show that Mr Left didn't believe what he was saying, then the statement of his opinion cannot be false - however wrong his opinion turned out to be.
The SFC, and now the SFAT, has done Hong Kong a disservice by chilling negative criticism of companies, thereby skewing the market even further towards positive research. "Sell-side" investment banks world-wide tend to withdraw coverage of a stock or use euphemisms rather than issue a sell note on a potential client. They will say "reduce", "hold", or "buy on weakness" (when it goes down) rather than say "sell". Hong Kong sits on the doorstep of a country which stamps out all forms of criticism. We need to strengthen and encourage, not weaken, freedom of debate and criticism of companies.
© Webb-site.com, 2016
Organisations in this story
- China Evergrande Group
- China Lumena New Materials Corp.
- E-Commodities Holdings Limited
- Hidili Industry International Development Limited
- Longfor Group Holdings Limited
- Moody's Investors Service Hong Kong Limited
- SECURITIES AND FUTURES COMMISSION
- WEST CHINA CEMENT LIMITED