Small vector autoregressions are commonly used in macroeconomics for forecasting and evaluating shock transmission. This requires VAR parameters to be stable over the evaluation and forecast sample or modeled as time‐varying. Prior work has considered whether there were sizable parameter changes in the early 1980s and in the subsequent period until the beginning of the new century. This paper conducts a similar analysis focused on the period since the recent crisis. Using a range of techniques, we provide substantial evidence against parameter stability. The evolution of the unemployment rate seems particularly different relative to its past behavior. We also evaluate alternative methods to handle parameter instability in a forecasting context. Copyright © 2016 John Wiley & Sons, Ltd.