Use of Economic Indicators in Correlation and Prediction Analysis of S&P 500

Abstract

This paper covers the correlations of the macroeconomic indicators of GDP, unemployment rates, CPI (Consumer Price Index), interest rates, expected inflation, and home. What was seen was correlations that were both strong and weak, and that correlation never means that they can predict the index, as while using these to predict for the index what was found was that the p-values were exceedingly high, and could thus not justify the indicators in the forecasting of the S&P 500 price, and the model used to predict the index through the indicators was an ARIMA Timeseries forecasting model. The S&P 500 also had positive correlations and negative correlations to indicators that had certain trends occurring, such as Pearson coefficient was positive, but the response and Timeseries model, along with the logistic coefficient was negative, thus showing that correlation does not work as an actual way of showing responses or forecasting for the S&P 500. With Timeseries as well, the Mean Squared Error (MSE) was far too high to show the S&P 500 could be predicted by the indicators. Thus, we can conclude the null hypothesis of the paper, which was that the S&P 500 could not be forecasted or linked to the economic indicators used was proven correct via the high p-values.