Critique my model - Valuing a Chinese Stock accused of fraud

Hi Guys,

The company is accused of LARGELY overstating its sales, profitability ratios, cash position, acquisitions, since 2011(at least). The actual situation could be worst and its stock price plummeted significantly after an activist group published a report with the accusations.

It is expected to be suspended for a period of 12-18 months and I was hoping to get your input to validate the model I developed as I’m a newbie yes here!

Since it stopped trading a little while ago, my firm conducted research on the proper financial fraud accusation discount to be used. They ended up using the median excess return (against an internally built composite using various Hang Seng indexes) of comparables (Chinese companies accused of financial fraud) and they determined the financial fraud accusation discount, at this time, to be about 25%.

Now my job is to value the security going forward.

After conducting research, I decided not to use the cie’s beta and multiples linked to the accusations (sales, earnings, current stock price, etc…) Beta would be obtained from legitimate comparables in the same market.

Here’s the formula I developed. Explanations follow:

Fair Value_T= Fair Value( t0) *[1+(β*R_benchmark )]+ ε_T

Explanations of variables:

〖Fair Value〗_( t0)=Halted Market Value * (1-〖Discount〗_(Fraud Accusation) )

〖Fair Value〗_T =Current Fair Value

β= Leverage Adjusted Beta of company dertermined using comparables^’ unlevered adjusted beta

R_bench= Cumulative Return on Benchmark since 〖Fair Value〗_t0

ε= Cumulative Error term since trade halt

To re-lever the beta, I would use the following to get the D/E ratio for Cie in question:

Equity value would be determined using: Shares Outstanding *estimated share value after discount for fraud.

Debt value using total debt and accounting for impact of convertible debt, which RexLot and many small cap comparables issue. Equity portion of convertible could be moved to equity value.

All the work is really preliminary and goes as a project to value a broader set of securities, unfortunately.

Your input would be greatly appreciated!!!


The fundamental flaw in their model is that you just assume all fraud accusations are the same, whereas some are clearly going to me more serious (is it a total bankruptcy, $1B writedown, etc) and more likely than others.

I would imagine you’d be better spending your time trying to figure out the degree of fraud via fundamental analysis, channel checks, supplier checks, etc than trying to decide if one enron is more favorably priced than another worldcom.

Thanks a lot Swan,

Haven’t seen you around in the longest time I hope you’re doing well.

If we focus about the stock price post investigation (research done on degree of fraud and discount was applied). Anyone else has something to add?

A MD said:

If the allegation is that sales were overstated, then it would stand to reason the assets and equity are overstated causing doubt on the D/E. They would select an index whose historic beta is within a standard error of 1 and move the stock 1 for 1 against that index after taking a discount - index return *(1- fraud discount)

*** Newbie here:

I see two weaknesses in his suggestions:

The historic beta is no more safe than my D/E suggestion, if not worst because I adjust the E to reflect the fraud discount. He suggests using the historical beta when historical stock return was fueled by materially distorted financial data. No fraud accusations were made toward its debt reporting.

Also, taking a discount to the index return implies that the greater the discount, hence the fraud, the lower the volatility. It does not make sense to me.

Practically though, I have no idea which would result in the most accurate result. If anyone can help that would be greatly appreciated.

Have you tried back testing against existing data? might be tough to get said data, but you could try…

Not really,

Its tricky because the past 5 years of data is biased by the likely fraud and then comes the recession, etc…

I will need to figure out a methodology for testing though. Not there yet however.


For what it’s worth, my fairly limited experience with bankruptcy situations suggests that it is much more beneficial to closely review the precise legal avenues available to various agents to lay claims against the company’s assets, and the current market values of those assets, as well as the exact nature of the financing banks’ security packages, rather than attempt to value the stock using some type of modified CAPM approach. I am sure that once the going concern criteria washes away, at least some of the assumptions behind the Markowitz portfolio type approach do not hold any longer.

Interesting case, and a shame that bromion isn’t here to talk about something like this. This sort of stuff is his bread and butter.

One way to attack it is to value the company at the liquidation value of its tangible assets, then value the company in terms of its claimed figures. See what the disparity is there. From there you have two probabilities to consider - the probability that the fraud claims are correct, and the degree of impairment that they represent. This will give you the range of values for varied scenarios.

Then do channel checks and whatever research you can to come to some guesses as to what probabilities are correct, or at least plausible.

This seems like a good case for valuing the company based on tangible assets, then valuing the company based on earnings power (valuing current or modified current earnings as if they were a perpetuity with no actual growth), and valuing the present value of future growth opportunities. See where the current price fits into those valuations and go from there.

Thanks bchad,

For liquidation value, I was actually forced into it as the cie may go bankrupt in november:

On August 20th, REXLot’s Board of director reported that the delisting of the stock for over 30 days triggered a “Relevant Event” convenant on REXLOT’s entirety of its debt (HK$ 2.25B of convertible bonds). Effective November 2nd 2015, the bondholders will have the right to require the company to redeem all or some of the principal amount together with interest accrued.

I’ll work on other valuation methods once I’m done valuing another stock.

I think this only reinforces the point that in bankrupcty/fraud cases what matters is not largely theoretical valuation models but rather the nitty gritty of legal agreements. All that modelling goes out the window if someone can yank the rug out from under you.

Why not just buy a stock that probably isn’t a fraud? There are only a few real fraud short sellers and it would be foolish to bet against any of the good ones. It depends on who made the claim and which company this is but most likely you should just find another long.

I generally avoid accounting fraud shorts because:

  1. They are hard to prove

  2. They are capable of persisting for long periods of time (and are especially hard if they appear to be or are producing a lot of cash flow)

  3. I’m not a CPA

I do a lot of “this product is will never work” or “this company is backed by a criminal syndicate”. I find those to be much easier. I track the fraud short selling space closely, know most of the publishers personally, and would not take the other side of any of the good authors. But that’s just me.


I think this only reinforces the point that in bankrupcty/fraud cases what matters is not largely theoretical valuation models but rather the nitty gritty of legal agreements. All that modelling goes out the window if someone can yank the rug out from under you.


Legal helps in this case but so does accounting, finance, market analysis, and general management skills. i.e. When you read the accusations, you need serious accounting and business skills to know what the management is trully trying to do.

Sent you a PM

Thank you all or your feedback.

The analysis is now mostly done with good feedback from managers. It gave me much leverage to work on more interesting cases and report to the head of the group at my firm.

You guys are awesome.