Class 11 Stat701 Fall 1997

Introduction to SPlus.

Todays class. Fit Nerlove data + Bootstrap demonstration.

Read in the Nerlove dataset.

Note that I have carried out the necessary transformations.

We are fitting Equation (3.20) the cost function associated with the Cobb Douglas Production function.

Knowing, prices, costs and output we can estimate the alpha's and the returns to scale r.

Note that the estimate of r comes from the relationship .

Checking if r = 1 is the interesting question, so we will find a bootstrap CI for r.

Also, to get an impression of the stability of R-squared we will find a bootstrap CI for R-squared - this we will do for the returns data set that includes the 87 crash.

The code needed to do this is here.

Bootstrap comments.

Produces valid estimates of standard errors and confidence intervals for just about any estimator

Relies on resampling data to produce many ``new'' datasets.

Extremely useful, especially for complicated statistics and complicated sampling designs

Is not a panacea. Does not save bad data. Does not work if resampling done incorrectly.

A link to Bob Stines bootstrapping notes.