this rate begun to slow and decline. It would appear that the current trend in the velocity of money is one that reflects optimistic consumer behavior. (Appendix 5a shows the trends in the velocity of money over the past few decades.) Meanwhile the M2 money stock has been increasing at a fairly consistent rate for some time, with very little variation about its' trend. (A.5b) Although in the second quarter the M2 money stock increased by a somewhat larger margin than was originally expected. The above considerations were important when I attempted to create a forecast for inflation by applying techniques discussed in Economic Forecasting 470. In order to attain the most accurate forecast I tried several different methods; including a bivariate, a multivariate, a multivariate with dummy variables, an automatic forecast, and a combination of techniques model. The Bivariate model was based on regressing inflation against the cyclical and irregular behavior of gross domestic product in order to see how the business cycle affected the rate of inflation. This model produced a significant regression statistic near 32%. In other words, roughly one-third of the variation in inflation can be explained by the stage of the business cycle. Both of the multivariate models contained the following predictor variables; detrended seasonally adjusted GDP, changes in the M2 money stock, changes in the velocity of money, changes in the Ppi, and changes in real wages. The most highly correlated variable being percentage changes in the velocity of money (76)%, and the least correlated being changes in the Ppi (4%). The multivariate model was able to produce a regression statistic of approximately 46%. The multivariate with dummy variables actually produce a lower R^2 value, and thus a less dependable model. The automatic forecasting method with Smart software produced a model, which could explain 79% of the data. The software chose a single exponential smooth...
