Using economic indicators to create an empirical model of inflation

Inflation affects all aspects of the economy, and it is important for traders, economists, and monetary authorities to understand the behavior of inflation to predict economic growth in the future. In our research, we investigated and calculated the correlation between various economic indicators and inflation over history. We then used the correlation information to develop a rolling linear regression model to predict inflation on an out-of-sample basis. For the measure of inflation, we used the Month over Month Consumer Price Index Seasonally Adjusted (CPI), released by the U.S. Bureau of Labor Statistics (BLS) every month. CPI measures inflation by calculating the change in the price of a basket of goods that consumers pay for. We chose 50 of the most important economic indicators followed by the market, and we hypothesized that by using out-of-sample data, the CPI of the next month could be reasonably predicted by using a regression model of a subset of economic indicators. We concluded that the average gasoline price, U.S. import price index, and 5-year market expected inflation have the most significant correlation with CPI. By using these indicators, we predicted CPI using a linear regression with a mean absolute error of about a tenth of a unit of CPI, where CPI is measured as a percentage. Furthermore, we discuss the possible public policy implications of our study and how inflation may be reduced by focusing on the three economic indicators highly correlated with it.