1. Levy, Solomon and Levy's Microscopic Simulation of Financial Markets points us towards the future of financial economics." Harry M. Markowitz, Nobel Laureate in Economics

2. microscopic element = individual investor interaction =buying / selling of stocks / bonds Discrete time investment options: Riskless: bond; fixed price return rate r : investing W dollars at time t yields W r at time t+1 Risky: stock / index (SP) / market portfolio price p(t) determined by investors (as described later )

3. Returns on stock Capital gain / loss If an investor i holds N i stock shares a change P t+1 - P t => change N(P t - P t-1 ) in his wealth Dividends D t per share at time t Overall rate of return on stock in period t: H t = (P t - P t-1 + D t )/ P t-1

4. Investors divide their money between the two investment options in the optimal way which maximizes their expected utility E{U[W]} =. To compute the expected future W, they assume that each of the last k returns H j ; j= t, t-1, …., t-k+1 Has an equal probability of 1/k to reoccur in the next time period t.

5. INCOME GAIN N t (i) D t in dividends and (W t (i)- N t (i) P t ) r in interest W t (i)- N t (i) P t is the money held in bonds as W t (i) is the total wealth and N t (i) P t is the wealth held in stocks Thus before the trade at time t the wealth of investor i is W t (i) + N t (i) D t + (W t (i)- N t (i) P t ) r

6. Demand Function for stocks We derive the aggregate demand function for various hypothetical prices P h and based on it we find P h = P t the equilibrium price at time t Suppose that at the trade at time t the price of the stock is set at a hypothetical price P h How many shares will investor i want to hold at this price? First let us observe that immediately after the trade the wealth of investor i will change by the amount N t (i) ( P h - P t ) due to capital gain or loss

7. Note that there is capital gain or loss only on the N t (i) shares held before the trade and not on shares bought or sold at the time t trade Thus if the hypothetical price is P h the hypothetical wealth of investor i after the t trade P h will be W h (i) = W t (i) + N t (i) D t + ( W t (i) - N t (i) P t ) r + N t (i) ( P h - P t )

8. The investor has to decide at time t how to invest this wealth He/she will attempt to maximize his/her expected utility at the next period time t As explained before the expost distribution of returns is employed as an estimate for the exante distribution If investor i invests at time t a proportion X(i) of his/her wealth in the stock his/her expected utility at time t will be given by t-k+1 E{U[X(i)]} = 1/k  ln[W ] j=t W= (1-X(i)) W h (i) (1+r) +X(i) W h (i)(1+ H j ) bonds contribution stocks contribution

9. The investor chooses the investment proportion X h (i) that maximizes his/her expected utility  E{U[X (i)]} /  X(i) | X (i)= X h (i) =0 The amount of wealth that investor i will hold in stocks at the hypothetical price P h is given by X h (i) W h (i) Therefore the number of shares that investor i will want to hold at the hypothetical price P h will be N h (i, P h )= X h (i) W h (i) / P h

10. This constitutes the personal demand curve of investor i Summing the personal demand functions of all investors we obtain the following collective demand function N h (P h )=  i N h (i, P h )

11. Market Clearance As the number of shares in the market denoted by N is assumed to be fixed the collective demand function N h (P h ) = N determines the equilibrium price P h Thus the equilibrium price of the stock at time t denoted by P t will be P h

12. History Update The new stock price P t+1 and dividend D t+1 give us a new return on the stock H t = (P t+1 - P t + D t+1 )/ P t We update the stocks history by including this most recent return and eliminating the oldest return H t-k+1 from the history This completes one time cycle By repeating this cycle we simulate the evolution of the stock market through time. Include bounded rationality: X h * (i)= X h (i) +   (i)