Calibrate S&P 500 Returns to a Mixed Normal in Excel
Use both the random variable approach and the density method. For the random variable approach, also create a theoretical histogram.
- Download S&P 500 historical prices from 1950 to the present (using Bloomberg or Thomson Reuters if possible).
- Calibrate the two parameters σ1and q via the method of moments.
- As an initial guess, let σ2= 0.8%. The range of the other two variables will be σ1∈ [1%,10%], q∈[1%,10%]. Using another choice of σ2can change these ranges. Experiment with different values.
- Let σ2= 0.8%, w1= 1, w2= 10000, if using the Excel Solver.
- Choose the discretized x-space carefully (start with -25% to 25% in increments of 0.05%).
- Graph both empirical and theoretical distributions as in Figure 3-44. Also, graph a log version of this graph to see how well the distribution fits the tail data.