In this thesis, I study the local employment multiplier effect; the effect of employment in the tradable sector on employment in the non-tradable sector of the same region. Using a reduced form regression with a shift-share instrument I find a significant local multiplier effect in Metropolitan Statistical Areas in the USA. I show that this result is robust to many different regional definitions, controls and ways of classifying tradable industries. I find larger multipliers for high-wage or high-skilled workers in the tradable sector and I find that most of the jobs created in the non-tradable sector are fulfilled by high-skilled workers who already reside in the region. A replication of the most influential paper in this literature, by Moretti (AER; 2010), demonstrates the sensitivity of his results to six idiosyncrasies of his analysis. To better understand how these local multipliers work, I develop an efficiency wage model with rural-urban migration for the non-tradable sector. In this model, I consider the impact of a shock to employment in the tradable sector in the city and find a positive local employment multiplier effect. The model predicts that attracting tradable jobs to a city has a bigger positive impact on employment in the non-tradable sector in the same city when the unemployment rate is higher. The model also predicts that this increase is driven by a larger multiplier for current inhabitants and that there is no, or even a negative, effect of the unemployment rate on the multiplier for movers. Both these predictions are reflected in the results of my non-parametric analysis of the data. I find similar results for European TL3 regions. Policies that try to increase growth in less favoured regions by stimulating tradable firms to locate in areas with high unemployment, will both reduce disparities between regions and efficiently reduce unemployment across the board.