Artificial time-series data can be created using formula such as:

Xt=aXt–1 + Normal (mean=b, SD=b/5)

These series include an initial transient and variability. Using the formula above generate ten series of data, each with 500 observations, in a spreadsheet. Set a = 0.9, b = 1 and the initial value of X (X0) = 0.1.

Determine the warm-up period, number of replications and the runlength for a single long run. Hint: use the ‘NORM.INV’ function in Excel to obtain samples from the normal distribution.

Note: an example of these data is available in a spreadsheet (Exercise9.

xls) on the companion web site.