A comparison of approaches for identifying minimum wage workers: Direct question versus administrative earnings data

The ability to identify minimum wage employees is essential for minimum wage research. When the minimum wage is set at an hourly rate, such as in Ireland, Germany, the United States and the United Kingdom, researchers often combine monthly or weekly earnings data with self-reported hours worked to calculate an hourly wage to determine if a worker is on minimum wage. This may lead to misclassification, whereby higher paid employees are misclassified as minimum wage employees or vice versa. An alternative, but less common, approach for identifying minimum wage employees is to use a direct survey question. Ireland is the only country in Europe whose labour force survey includes both linked administrative earnings data and a direct minimum wage question, thereby facilitating a comparison of both approaches. When studying the incidence, characteristics and outcomes associated with minimum wages, we show that the choice of approach (administrative data vs. survey question) can produce different results. We discuss the limitations of each approach when it comes to implementing difference-in-differences analysis to study the impacts of a minimum wage increase. We then combine useful features from both approaches to come up with a difference-in-differences estimate that overcomes some of the limitations that exist when using either approach on its own. We argue that the addition of a standardised minimum wage question in labour force surveys across Europe would allow researchers to provide more robust estimates on the impacts of minimum wage changes, while also facilitating cross-country minimum wage research.