How to make US$100 million with almost useless knowledge

I've got the brains, you've got the looks
Let's make lots of money
You've got the brawn, I've got the brains
Let's make lots of money

You can tell I'm educated, I studied at the Sorbonne
Doctored in mathematics, I could have been a don
I can program a computer, choose the perfect time
If you've got the inclination, I have got the crime

---- Pet Shop Boys

Part of Joseph Wang's China quant pages.

As mentioned elsewhere, I'm trying to develop a skill (quantitative analysis of Chinese securities) which is almost useless knowledge. But almost useless knowledge might still be worth about $100 million if you can execute.

One of the key points that the keynote speaker made at TCFA 2005 is that in order to make money you need three things.

• Good concept
• Good execution
• Good pricing

My experience is that the second is always the hardest, and one of the corollaries of that is that one shouldn't be afraid of sharing good concepts with other people, because the hard part is not coming up with the idea, its executing it.

Here is a US$100 million idea which might make use of my almost useless knowledge.

China is in the process of unifying tradable and non-tradable A-shares. The institutional framework which has already been applied to 40 companies and will be eventually applying to all 1600 companies is to have each company come up with a plan to unify the tradable and the non-tradable shares which must be approved by each class of share holder. Some of those plans may involve stock splits, other plans might involve option like features. In any event, each plan will be a tailored for each company.

1600 valuations each with highly complex and invidividualized situations. WOW!!!!

You can make money in two ways.

• The first is by offering consulting services to the 1600 companies in order to mediate the transition. Assume US$500,000 for this service times 200 companies. That will get you $100 million.

• The second is to take advangtage of mispricing by arbitrage. Keep in mind that unlike LBO's the negotiation and the approval of these plans is going to be rather public. Also for situations where there are option like conditions, you can still take advantage of mispricing. Assume 5% mispricing, of which you can capture 20%, assuming each company has a capitalization of US$10 million. $100,000 times 1000 gives you about US$100 million.

Note: If you have any passing familarity with finance, you will realize that these numbers are almost totally meaningless. Saying that you can make US$100 million is meaningless unless you also specify a) the amount of working capital needed and b) the risk involved. Also you may well ask where I got the numbers from. The answer to that question is that I made the numbers up. The thing about the question of working capital and risk is that they required some pretty detailed calculations, and if I only came up with US$1 million with wild made-up numbers, there is no point in making those calculations.

Now the hard part ....

Execution......

What you need to do this is to be able to trade A shares in China. This eliminates most hedge funds who don't have QFII status and aren't interested in getting it. Better than QFII is being able to directly trade A shares. There are a lot of landmines here. First, there are all these subtle legal issues and corporate details. Second, you have to figure out how to take advantage of mispricing given that all of the tools for traders aren't available (in particular short selling.)

So the ideal people to try to do this would be a large investment bank with a Chinese joint venture or else an ambitious Chinese securities company. The team that you will want to do this would include a Chinese securities lawyer, a Western equities trader, a Chinese equities trader, and you'd also need a quant with the ability to read Chinese and some passing knowledge of Chinese institutional frameworks. He or she wouldn't need heavy trading experience and it wouldn't have to be the mathematically sharpest quant.

Now the real challenge is to assembly line the process. Looking at each 1600 company and trying to evaluate it one by one is going to be painful. So lets look at what the mechanized process looks like and what the challenges are at each stage..

(I'll go into more detail later.....)

1) Getting the data.

2) Setting up the ontological schema

This is probably where the real "secret sauce" is. You have 1600 companies each of which handle the unification of tradable and non-tradable shares differently. How do you classify and represent the variety of ways so that you can store it in the database and do a quant calculation?

3) Doing the quant calculation

4) Executing the trades.

So suppose that the model says that something should be worth CNY 30 when it is worth 40 or vice-versa. What you do with that information? If a share is undervalued you can try a buy and hold strategy. Buy what? How much? Hold for how long? Do you leverage? What are the limits at which you cut losses? What are your capital limits? When you you start suspecting that the model is wrong? Suppose something is overvalued. What do you do? You can't short sell A-shares. So can you come up with a clever way of doing arbitrage without short selling? Risk? Capital requirements? What if the regulations suddenly change on you? How many traders? What micro-strategies? What are the transaction costs?

5) Worrying about what can go wrong.

Also, you'd need to do this quickly. As with most arbitrage opportunities it disappears when people start taking advantage of it. Of course you'll need someone with some quantitative skills and China background. Even if you already have one, it wouldn't hurt to find another one since this strategy is going to be quite labor intensive. Remember 1600 listed companies. Lots of work.

Hmmm... Where can we find person with quant skills that also has at least passing familiarity with Chinese economic institutions. Hmmmmm.....

Goldman-Sachs? Morgan-Stanley? JPMorgan-Chase? Citigroup? Lehmann Bros? Guosen Securities? Anyone interested?

No one?

How things are shaping up

It seems that most companies are keeping it simple and just offering to give more shares to the trading share holders. It seems that the institutional setup is very clever. The trading shareholders will not agree to any plan that causes the value of their shares to go done, while the non-tradable shareholders want to maximize the value of their shares. So the net result is that the non-tradable share holders will give just enough shares to the tradable holders to make their shares fair valued, but no more.

I think things get a lot more complex when you put factor in a company which has multiple non-tradable share holders. I'm still trying to figure out the plan for Minsheng Bank.

Also, a lot of these plans have option like pledges by the nontradable shareholders not to let the shares drop below $X amount. This leads to some interesting valuation issues.

So lets do some math

Let be the value of the company which you can calculate from discounted cash flow. Let and be the number of tradable and non-tradable shares respectively. Let be the current price of the shares.

Assume that share exchange ratio is . We end up with the equation

For this to work you need which gives

Blah this is wrong....

I'm assuming that the share price stays constant through the reform. That won't work.

Lets try again.....

After the reform, the share price should be evenly distributed between the tradable and the non-tradable shares

The criteria for a tradable share holder is that he or she does not lose money. This implies

which implies

Need to work on this on paper. See you later.....

The next step once we have a formula for p is to

• see if this makes sense with the data
• think of ways of gaming the system. One good way is that we've established above that share price stability is not possible with the new reform, but information about the share price reform is going to come out in drips.
• Also you have some companies pledging to set a floor on the share price. This is likely to have very interesting effects.
• Also the fair value argument assumes that state owned shares will be valued the same as non-state owned shares. What happens if this doesn't happen.
• What are the price earning ratios pre and post share reform?

Anyone out there interested in paying me to do this????

More about how to make money

The basic data you need is here

http://finance.sina.com.cn/stock/chinaggzw/index.shtml

It has a link to each of the 40 approved policies regarding the stock reform. Each of the documents is this long winded legal document with all sorts of tiny annoying details. As I mentioned before the key to make this work is to be able to distill each of the documents into something that you can put into a quantitative model.

You may be right, I may be crazy
But it just may be a lunatic you're looking for
Turn out the light, don't try to save me
You may be wrong for all I know, but you may be right
--- Billy Joel

Show attachmentsHide attachments

Topic attachments

I	Attachment	Action	Size	Date	Who	Comment
png	latex2733f0cd0f5ef280c75d59f0ad7ad443.png	manage	0.4 K	25 Feb 2006 - 23:09	UnknownUser
png	latexd19ccd76777ea49946d1474a356c5b87.png	manage	0.9 K	25 Feb 2006 - 23:09	UnknownUser
png	latexc7e456538991675f90b165b741810075.png	manage	1.2 K	25 Feb 2006 - 23:09	UnknownUser
png	latexa195dc751a1e1c2ffd688ec5d5ddf267.png	manage	0.8 K	25 Feb 2006 - 23:09	UnknownUser
png	latexdc5d3b8385dfc0970772286a5ff031bd.png	manage	0.9 K	25 Feb 2006 - 23:09	UnknownUser
png	latex6f7a0ad09abbf295e142df1316a75439.png	manage	0.3 K	25 Feb 2006 - 23:09	UnknownUser
png	latex369ed892a0b59468d6c45c4f72655722.png	manage	1.1 K	25 Feb 2006 - 23:09	UnknownUser
png	latex558d9b162de6d21058258f9af4625f78.png	manage	1.1 K	25 Feb 2006 - 23:09	UnknownUser
png	latex585884ca5f9bfe931014435efe696902.png	manage	0.4 K	25 Feb 2006 - 23:09	UnknownUser
png	latex16f901d91a62c7092ebcd494c150fd9e.png	manage	1.0 K	25 Feb 2006 - 23:09	UnknownUser
png	latexf7de41f4b55fe6229ddc38f8e57bd78b.png	manage	0.8 K	25 Feb 2006 - 23:09	UnknownUser
png	latex4c176cca97c5149d5a24f08d9fa29477.png	manage	1.0 K	25 Feb 2006 - 23:09	UnknownUser
png	latex702e61fdb79ad7b36d8e1103d775d8fc.png	manage	0.7 K	25 Feb 2006 - 23:09	UnknownUser
png	latex04f9cd2d59397eaab1182b18175891af.png	manage	0.3 K	25 Feb 2006 - 23:09	UnknownUser
png	latexb34e9498bad461671e99f26161798c77.png	manage	1.0 K	25 Feb 2006 - 23:09	UnknownUser
png	latex91683115dfe4b759a6a3caf034ecf8d6.png	manage	0.3 K	25 Feb 2006 - 23:09	UnknownUser
png	latex7d4bd6cc37f571a885c354ce9304456f.png	manage	0.3 K	25 Feb 2006 - 23:09	UnknownUser