One Market - Two Perspectives
Dear reader,
this article introduces two perspectives on markets. The first is a classic economical view on a market - the Capital Asset Pricing Model. The second is from a view point of a gambler - the Parimutuel betting. Both models are based on the idea of transactions - the price of an asset (or bet) is defined by the price of the last transaction of that bet.
Caprital Asset Pricing Model
The fundamental assumption of Capital Asset Pricing Model (CAPM: http://en.wikipedia.org/wiki/Capital_Asset_Pricing_Model) is that all Investors are (1) rational, (2) risk-averse and (3) cannot influence market prices. The result is, that market prices move in unpredictable manner and equals a gauss distribution. This chaotic price movement is only influenced by (4) new information hitting the market, which should be unpredictable.
I want to argue against CAPM. Therefore I have to disproof at least one of the assumptions above.
(1) Rational investors: Behavioural finance suggests that humans are not totally rational and some are less rational than others. In some situations apes are better than humanes in assessing probabilities. Once I broke my arm and I got to hospital. Who stood in the smokers area in front of the building and smoked cigarettes? Surgeons. These are highly intelligent people and know what happens if they smoke and they do it anyway. Thats how irrational humans can be even if they are highly intelligent.
(2) Risk-averse investors: I think this argument is valid for most people most of the time, but on occasion when envy is controlling the investors mind they become risk-prone. As Warren Buffett puts it "What really drives [people] is envy, not greed. You pay someone $2 million and they might be quite happy until they hear that someone else got $2.1 million." (http://buffettfaq.com/#b65)
(3) Only information influence market prices: This one is idiotic! The assumption is, that at every moment every security is tradable in every amount possible. First no human can decide to buy or sell a security in sub-second timespan. Even computers cannot make decisions in an infinity-small timespan (I think 100s of pico-seconds is the minimum timespan possible today). The amount of all stocks tradable at one moment is limited, because an investor cannot follow all stocks all the time and making an optimal portfolio decision every moment. If a new information hits the market all investors optimize their portfolio and send their demand and supply of securities to the exchange. Because supply and demand of securities mismatch the price is change to meet the supply and demand equilibrium. The problem is, in real world this process takes time and not every investor can buy the stock at any given price. This leads to liquidity issues of investors which has massive impact on price building. (http://en.wikipedia.org/wiki/Flash_Crash)
(4) Real world observations: Flash crashs happened many times in history. In bull markets are long strings of growing growing market capitalization. A single string is very unlikely but not impossible, but this happened in every bull market. In bear market the market price drops are big and therefore very unlikely. But they happen in nearly every bear market. All these observations cannot be explained through the CAPM. The CAPM indicates that all price moves are normally distributed. The real world indicates that they are not some times.
Another argumentation against CAPM is, that it cannot observe the models consequences in the real world all the time.
I leave the pro argumentation for CAPM to business schools, because it is basic knowledge for every investor.
Is the CAPM right or wrong? I would say it is right and wrong. It is right if the assumptions are true which is most of the time. It is wrong if the assumptions are broken. I think there is a need for society and science to have more than one model for working markets. If CAPM assumptions are broken another model has to explain what happens.
A side note: In physics are multiple models about forces (gravity, magnetic fields, momentum, ...) which explains the movement of particles. Why do we have in economics only one major model for price movements in markets? In my eyes this is ignorant.
Parimutuel System (Parimutuel Betting)
First I hear of parimutuel betting as a market model in a talk of Charlie Munger (http://www.scribd.com/doc/75389403/Charlie-Munger-Art-of-Stock-Picking) in which he explains his view on financial markets. Wikipedia article on parimutuel betting: http://en.wikipedia.org/wiki/Parimutuel_betting.
Parimutuel betting is a system used for horse races, poker and lottery. In such a system you bet on the outcomes (a,b or c) of an event with some money, say 5$ on (a). The total bets are: (a) 10$, (b) 20$, (c) 30$. After all money is collected (60$) the betting fees (eg. 10% = 6$) are subtracted from the pot (60$-6$=54$). If (a) wins, 5.4$ (=54$/10$) is paid out for every $ betted on (a). Therefore you get back 5 * 5.4$ = 27$ and make a profit of 27$ - 5$ = 22$.
This model captures much better the risks about the future of assets (= outcomes of bets). Second, this system shows how price of an asset influence outcome of an asset. If all 60$ is betted on (a) and only 54$ is paid out, than we have only losers! The difference of 6$ can be interpreted as transaction costs. The relatedness to markets are: (1) it is possible to overpay an asset. (2) the outcome of an transaction is determined by the price paid. (3) the price is determined by the last transaction. (4) the price is expressed by the outcome $ per 1 $ betted. (5) Demand and supply is expressed as the price relationship between bets, so that demand and supply is expressed by the expected probabilities of outcomes. (6) the expected probabilities are the assumptions of the participants. (7) the intermediate revenue is independent from the outcome.
Opportunities appear if the outcome probability is hard to know or unknown and therefore asset price in outcome $ per 1 $ invested is very high! The price is over proportionally influenced by a change in expected probabilities. The math behind this relation is lim p->0 = 1/p = price. A small miscalculation in p causes huge price differences. This non-linearity creates opportunities.
The consequence of such a system is to bet (1) rarely on big opportunities (reduce transaction costs) and (2) do not bet too much (do not get killed). The payout ratio can change till "rien ne va plus". (3) The Kelly criterion helps to determine the bet size - http://en.wikipedia.org/wiki/Kelly_criterion but be careful(!!!)*.
Conclusion
In summary, a parimutuel system can explain some behaviour seen in extreme market conditions, where CAPM cannot provide a satisfying answer. The price being paid defines the investment outcome for an individual and contains probability assumptions about the outcome. This non-linear influence of probability on price and return on investment creates opportunities. Therefore a valuation strategy is to ask what must happen to a company to justify its market price and how probable is that.
Implications on my valuation process
I expect a company is under valued and is steadily making money, than I reassemble the companies business units earnings (on income statement) and cash and near cash items (on balance sheet) so that it will justify the market price. Everything which is not assembled, I get for free. If nothing is left the company is fair or over valued. This tactic works best if the company is heavily undervalued because of problems in few business units - the market overestimates the impact of the problematic BUs on the overall company.
I construct multiple company models to justify the market price and estimate how likely they are. I play around with these models to extract the value drivers for the company (eg. deferred revenue (software companies)). I even try to destroy the company to see what must happen so that I have a total loss of capital. This stress testing gives some fun insight how the economic machine works for that company. For me its like playing bridge builder with companies (http://www.bridgebuilder-game.com/ highly addictive!!!). I will write an example valuation with that method in some time.
* Mohnish Pabrai burned his fingers with a Kelly criterion portfolio strategy and now applies a 75%-10%-10%-5% portfolio diversification strategy. Recommend to see his talk:
http://www.youtube.com/watch?v=xa1CH2nK1cM
http://www.youtube.com/watch?v=z2YhBr_HRCw
this article introduces two perspectives on markets. The first is a classic economical view on a market - the Capital Asset Pricing Model. The second is from a view point of a gambler - the Parimutuel betting. Both models are based on the idea of transactions - the price of an asset (or bet) is defined by the price of the last transaction of that bet.
Caprital Asset Pricing Model
The fundamental assumption of Capital Asset Pricing Model (CAPM: http://en.wikipedia.org/wiki/Capital_Asset_Pricing_Model) is that all Investors are (1) rational, (2) risk-averse and (3) cannot influence market prices. The result is, that market prices move in unpredictable manner and equals a gauss distribution. This chaotic price movement is only influenced by (4) new information hitting the market, which should be unpredictable.
I want to argue against CAPM. Therefore I have to disproof at least one of the assumptions above.
(1) Rational investors: Behavioural finance suggests that humans are not totally rational and some are less rational than others. In some situations apes are better than humanes in assessing probabilities. Once I broke my arm and I got to hospital. Who stood in the smokers area in front of the building and smoked cigarettes? Surgeons. These are highly intelligent people and know what happens if they smoke and they do it anyway. Thats how irrational humans can be even if they are highly intelligent.
(2) Risk-averse investors: I think this argument is valid for most people most of the time, but on occasion when envy is controlling the investors mind they become risk-prone. As Warren Buffett puts it "What really drives [people] is envy, not greed. You pay someone $2 million and they might be quite happy until they hear that someone else got $2.1 million." (http://buffettfaq.com/#b65)
(3) Only information influence market prices: This one is idiotic! The assumption is, that at every moment every security is tradable in every amount possible. First no human can decide to buy or sell a security in sub-second timespan. Even computers cannot make decisions in an infinity-small timespan (I think 100s of pico-seconds is the minimum timespan possible today). The amount of all stocks tradable at one moment is limited, because an investor cannot follow all stocks all the time and making an optimal portfolio decision every moment. If a new information hits the market all investors optimize their portfolio and send their demand and supply of securities to the exchange. Because supply and demand of securities mismatch the price is change to meet the supply and demand equilibrium. The problem is, in real world this process takes time and not every investor can buy the stock at any given price. This leads to liquidity issues of investors which has massive impact on price building. (http://en.wikipedia.org/wiki/Flash_Crash)
(4) Real world observations: Flash crashs happened many times in history. In bull markets are long strings of growing growing market capitalization. A single string is very unlikely but not impossible, but this happened in every bull market. In bear market the market price drops are big and therefore very unlikely. But they happen in nearly every bear market. All these observations cannot be explained through the CAPM. The CAPM indicates that all price moves are normally distributed. The real world indicates that they are not some times.
Another argumentation against CAPM is, that it cannot observe the models consequences in the real world all the time.
I leave the pro argumentation for CAPM to business schools, because it is basic knowledge for every investor.
Is the CAPM right or wrong? I would say it is right and wrong. It is right if the assumptions are true which is most of the time. It is wrong if the assumptions are broken. I think there is a need for society and science to have more than one model for working markets. If CAPM assumptions are broken another model has to explain what happens.
A side note: In physics are multiple models about forces (gravity, magnetic fields, momentum, ...) which explains the movement of particles. Why do we have in economics only one major model for price movements in markets? In my eyes this is ignorant.
Parimutuel System (Parimutuel Betting)
First I hear of parimutuel betting as a market model in a talk of Charlie Munger (http://www.scribd.com/doc/75389403/Charlie-Munger-Art-of-Stock-Picking) in which he explains his view on financial markets. Wikipedia article on parimutuel betting: http://en.wikipedia.org/wiki/Parimutuel_betting.
Parimutuel betting is a system used for horse races, poker and lottery. In such a system you bet on the outcomes (a,b or c) of an event with some money, say 5$ on (a). The total bets are: (a) 10$, (b) 20$, (c) 30$. After all money is collected (60$) the betting fees (eg. 10% = 6$) are subtracted from the pot (60$-6$=54$). If (a) wins, 5.4$ (=54$/10$) is paid out for every $ betted on (a). Therefore you get back 5 * 5.4$ = 27$ and make a profit of 27$ - 5$ = 22$.
This model captures much better the risks about the future of assets (= outcomes of bets). Second, this system shows how price of an asset influence outcome of an asset. If all 60$ is betted on (a) and only 54$ is paid out, than we have only losers! The difference of 6$ can be interpreted as transaction costs. The relatedness to markets are: (1) it is possible to overpay an asset. (2) the outcome of an transaction is determined by the price paid. (3) the price is determined by the last transaction. (4) the price is expressed by the outcome $ per 1 $ betted. (5) Demand and supply is expressed as the price relationship between bets, so that demand and supply is expressed by the expected probabilities of outcomes. (6) the expected probabilities are the assumptions of the participants. (7) the intermediate revenue is independent from the outcome.
Opportunities appear if the outcome probability is hard to know or unknown and therefore asset price in outcome $ per 1 $ invested is very high! The price is over proportionally influenced by a change in expected probabilities. The math behind this relation is lim p->0 = 1/p = price. A small miscalculation in p causes huge price differences. This non-linearity creates opportunities.
The consequence of such a system is to bet (1) rarely on big opportunities (reduce transaction costs) and (2) do not bet too much (do not get killed). The payout ratio can change till "rien ne va plus". (3) The Kelly criterion helps to determine the bet size - http://en.wikipedia.org/wiki/Kelly_criterion but be careful(!!!)*.
Conclusion
In summary, a parimutuel system can explain some behaviour seen in extreme market conditions, where CAPM cannot provide a satisfying answer. The price being paid defines the investment outcome for an individual and contains probability assumptions about the outcome. This non-linear influence of probability on price and return on investment creates opportunities. Therefore a valuation strategy is to ask what must happen to a company to justify its market price and how probable is that.
Implications on my valuation process
I expect a company is under valued and is steadily making money, than I reassemble the companies business units earnings (on income statement) and cash and near cash items (on balance sheet) so that it will justify the market price. Everything which is not assembled, I get for free. If nothing is left the company is fair or over valued. This tactic works best if the company is heavily undervalued because of problems in few business units - the market overestimates the impact of the problematic BUs on the overall company.
I construct multiple company models to justify the market price and estimate how likely they are. I play around with these models to extract the value drivers for the company (eg. deferred revenue (software companies)). I even try to destroy the company to see what must happen so that I have a total loss of capital. This stress testing gives some fun insight how the economic machine works for that company. For me its like playing bridge builder with companies (http://www.bridgebuilder-game.com/ highly addictive!!!). I will write an example valuation with that method in some time.
* Mohnish Pabrai burned his fingers with a Kelly criterion portfolio strategy and now applies a 75%-10%-10%-5% portfolio diversification strategy. Recommend to see his talk:
http://www.youtube.com/watch?v=xa1CH2nK1cM
http://www.youtube.com/watch?v=z2YhBr_HRCw
I took a look over a view of your posts and found it quite interesting. I've added your blog to my blogroll (http://commentsonpositions.blogspot.co.at/, on the right side ;-)
AntwortenLöschenOn CAPM: my feeling was always that this model is fundamentaly flawed. it is ok to say people are risk averse. but the question is: what is risk?
if you say risk is the fluctuation in share price, then it is ok to use beta. but I think risk is not fluctuation in share price (to the contrary: it is chance), but uncertainty in the underlying business that stands behind the stock.
when using CAPM, and consequently beta, to value shares, then the fluctuation in share price INFLUENCES your estimation of fair value.
so, this model is mathematically very nice (why academics love it), but fundamentally flawed (not usable in practice).
your approach is more promising, as I see it (although it is much more work ;-).
I hope I was able to make my point clear.
Best wishes for your blog, entrepreneural and, of course, investing activities.
TomB
Thank you very much!
LöschenI did not know the blogroll feature. I copied that idea and added you to my blogroll too :-)
What I love about this article is that, we have multiple models to view markets. The CAPM is not always true but many times it is. Lets face it, most companies are fairly valued, but the rest is not. And the rest are the interesting ones ;-)
What I hate about my professors at the business school, is that they teach CAPM in finance, Keynesian in macro economics and market equilibrium theory in micro economics. But they do not teach other important perspectives on the markets like austrian school, pari-mutal systems, evolutionary systems (even physics study evo. systems!). I asked them why? And their response was that it is not in the syllabus. What kind of ignorance is that!? Its like medicine in mediaeval times - they practice it but never opened a body and looked inside! And these people educate my fellow students and myself who will run our economy in future. Not too hard to predict what will happen!
You are totally right about the approach. At first, it was ok, because you get a very detailed view how business models work on the quality side - I follow the solgan "I am a better investor because I am a businessman, and a better businessman because I am no investor." On the down side, this approach is exhausting. But its not the modelling and the "probabilities", but the information gathering of the parts is.
At the moment I play around with earnings power value by Bruce Greenwald (http://www.oldschoolvalue.com/blog/valuation-methods/bruce-greenwalds-earnings-power-value-epv-lecture-slides/), Königsanalyse by Max Otte and Devensive and Entrepreneurial Valuation by Hewitt Heisermann (http://www.amazon.de/Its-Earnings-That-Count-Long-Term/dp/0071463992). The first and last method can be used for capturing quality companies, where as Köningsanalyse is easy and fast. Valuation is an art!
What I love about this article is that, we have multiple models to view markets. The CAPM is not always true but many times it is. Lets face it, most companies are fairly valued, but the rest is not. And the rest are the interesting ones ;-)
Löschenthat's true. On Greenwald: he wrote a few interesting books, which are quite readable (I have already done):
http://www.amazon.com/Value-Investing-Graham-Buffett-Beyond/dp/0471463396/ref=sr_1_2/176-0525300-3897706?s=books&ie=UTF8&qid=1403333728&sr=1-2
http://www.amazon.com/Competition-Demystified-Radically-Simplified-Approach/dp/1591841801/ref=pd_sim_b_1?ie=UTF8&refRID=030JVKJW0AFHBTFAAKBP
a short summary/paper of the 2nd one: I do not know if I am allowed to link it, but you will find it with google: 'all strategy is local greenwald kahn'
I do not know Otte an Heisermann, but I can imagine you will like the ones I've mentioned above
TomB
and by the way: i've studied economics too. I've learned this CAPM up and down, and it was always hard for me to write down this stuff to pass the examn, when I always knew that it is hard to apply in the real world...
Löschenbut like you said before: it is often ok to use it. but it's the ones where you can't apply it, that are interesting.
I have read those books of Greenwald too. They are great!
LöschenThanks for the Greenwald paper! Didnt know it. Its added on my reading list :-)
I didnt like the real world aspects of CAPM, too. As I learned it, it felt wrong from the begining. Than I discovered the math aspects, which are great - similar to a philosopher which discusses a totally hypothetical problem, its fun but no value added there.