Who needs theory when you have so much information? But this is categorically the wrong attitude to take toward forecasting, especially in a field like economics where the data is so noisy.

The Signal and the Noise: Why So Many Predictions Fail - But Some Don't