Using the Panel Study of Income Dynamics, we find that true wage changes have many fewer nominal cuts and more nominal freezes than reported nominal wage changes. The data overwhelmingly rejects a model of flexible wage changes and provides some evidence against a model of perfect downward...