Normal Distribution and Stock markets

 You toss a coin , 100 times, what is the number of heads you expect? 500 employees of a company, appear for a test(multiple choice Qs with 1 right ans out of 4 choices) having 100 questions. Employees attend without preparation and have no clues on the subject (say quantum mechanics). How many will get 100/100? How many will get zero?

Can we connect the results with stock markets? We pick up 100 stocks randomly and construct a portfolio. What would be the returns after 20 years?

 First 2 events are random events. Probability theory suggests, the results of the first two will follow normal distribution if we do experiment for a large number of times. What normal distribution curve says? You can expect a value clustered around the mean(=no of experiments* probability of success). Most of the times you would end getting 45-55 heads if you toss a coin for 100 times. Getting 100 heads/tails are extremely rare, Even 80 heads will be quite rare event.

 Similarly, many people would get around 25 marks. Most of the people would score between 16-34 marks. Number of people getting 50 and above would be quite less.

 Eugene Fama, Robert Shiller along with Peter Hansen shared Nobel prize for economics in 2013. Fama’s works revolves around efficient market, which says that capital market prices in, all the information quite quickly. So all stock prices are always rightly priced. For example, if a company like Infosys comes with stellar results during  market hours, within seconds, it’s stock price will go up. This news and possible improved prospects would be priced in fast. You cannot make easy money. The future price will be determined by the future news flows which could be positive or negative, which are random in nature.

 To model stock prices in such situation, random walk model can be used. The returns of a stock goes up by say, 1 percent, say if the coin turns head(good news). It goes down by say ½ percent if it’s a tail. The stock returns in such cases will follow normal distribution over long period. This could be the reason some risk/return /pricing model assume Normal distribution for the underlying asset.( In real world, perfect normal distribution cannot explain wild swings resulted by black swan events. So daily returns has fatter tails than ND. Ref Fama)

Robert Shiller, a behavioural economics proponent, believes markets are not always efficient. There could be bubbles. Over a long period it could be predictable. He has successfully predicted dotcom, housing bubbles and off late, gold bubble.

 

Comments

Popular posts from this blog

5:Continuous compounding formulas and Euler's number

3-Jackson Hole 2005

Yield curve and sensitivity