We analyse four leading indicators in the US economy using autoregressive models and find strong evidence in favour of GARCH effects. All series remain fat-tailed after controlling for GARCH effects, suggesting that non-Gaussianity of the innovations should be accounted for.