Macroeconomics of Labour Markets: The Role of Modern Technologies and Search Frictions

This research work covers three essays that deal with the impact of technology on labour markets worldwide (chapter 1 and 3) and the macroeconomic role of search frictions (chapter 2 and 3). Chapter 1 provides an analysis that contributes to the current discussion concerning automation and insists on the following highly debated questions: Do robots reduce employment? Does the impact of automation on labour market differ between developed and developing countries? Is the global value chain affected by the automation wave? I show that the worldwide impact of robots on employment is negative, largely driven by the effect in developing countries, for which I quantify a loss of employment of more then ten percent. Chapter 3 contributes to the literature by providing a theoretical equation for the labour share depending on the price of ICT capital. Moreover, given the rising interest in search frictions and the matching process, it provides the answer to the following question: in a context where technological capital becomes (at constant lower cost) substitute to labour, do labour market frictions play any role? For eight European countries and the US, we estimate, first, the aggregate elasticity of substitution between capital and labour, then, the elasticity of substitution between ICT and labour. Both with and without the search costs. We find that the aggregate elasticity is unable to explain the decline in the labour share, as the majority of the estimates in the previous literature point to. Instead, the elasticity of substitution between ICT and labour is statistically larger than one and robust to a number of alternative specifications. When we account for the presence of search costs in the labour market, we find that the ICT-labour substitution is still key for the decline in the labour share, but that search costs acts also in this sense by decreasing labour demand. Search costs represent a key ingredient of modern micro- and macroeconomics of the labour markets. Economic models embedding the search process have been successful in explain several job market facts, such as the presence of frictional unemployment, the evolution of job creation and job destruction, the presence of employer’s and employee’s rent. However, the business cycle properties of search models proved to be poor. The analysis of chapter 2 aims at providing an empirical assessment to the hypothesis of Pissarides (2009). First, I document the existence of a volatility puzzle also in Germany (see also Gartner et al., 2009). Second, I use two waves of the German Job Vacancy Survey (2014 and 2015) containing novel information on the duration of the vacancy and the cost the employer bears to fill the vacancy (among others, headhunter, advertising, travel reimbursement for the interview). Together with precious data on plant and vacancy characteristics, the analysis assesses the dependence of search costs on search duration via an OLS and IV approach. In both cases, the search costs appear to be weakly, if not linked at all, with the duration of the vacancy. This result gives a first empirical validation to the hypothesis of Pissarides (2009)