| Resumo: | In this research, I show that aggregate information from financial statement analysis helps in predicting real economic development. Further, I show that using the top 100 U.S. public companies, ranked by market capitalisation, represents a convenient method to proxy for the entire portfolio of traded companies. I then show that aggregate accounting information of the same 100 biggest companies has predictive information for next quarter real Gross Domestic Product (GDP) growth, after controlling for the traditional stock market returns, and explains a portion of professional macro forecasters’ revisions and errors. Konchitchki and Patatoukas (2014a) provide an intuitive framework for these findings. Yet, I contribute by finding that aggregate accounting drivers from the Alternative Breakdown provide greater predictive power when compared to DuPont and that introducing financial and nonfinancial data split reduces heterogeneity. Another contribution of mine is introducing out-of-sample analysis. Although, I find that current methods used by professional macro forecasters exhibit slightly lower root-mean-square error (RMSE), I only use annual stock market returns and aggregate accounting profitability drivers in my model
|
|---|