Essays in Applied Economics

Underlying this doctoral thesis is the growing importance of the study of the relationship between cause and effect. This relationship can relate to policies, events, actions and processes. The analysis of causal relationships has developed over many years and still remains a central issue today. Indeed, in economics, the measurement of a particular causal effect is often one of the objectives of empirical analyses. When researchers are unable to conduct a randomized controlled experiment, they must necessarily rely on the observation of non-experimental data, i.e. data from the real world, as is the case with sample surveys for example. A major problem in the use of this data arises from the fact that they are not derived from an experiment specifically designed for the purpose. Therefore, there may be variable factors whose effects are difficult to separate from the specific treatment effect. The group of statistical techniques which have been developed for the assessment of causal effects from non-experimental data is therefore a very important field of applied econometrics. In this dissertation, I will face the the relationship between cause and effect and several impact evaluation methods, presenting a collection of three applications of these techniques, adapted from three different works: “Price Matching and Platform Pricing”, presented at the First NetCIEx Workshop (EU Joint Research Centre - Ispra); “Roads to Innovation: Evidence from Italy”, presented at the First NetCIEx Workshop (EU Joint Research Centre - Ispra); “Public Funded R&D as a Device for Local Innovation? Evidence from Italian I.I.T. Foundation”. In the first paper, the effects of Price Matching Guarantees (a PMG refers to the price strategy where different retailers commit themselves to match any lower price offered by competitors on the same item or product category) on U.S. online consumer electronics prices are empirically investigated, by means of a unique dataset developed through sophisticated and computerized scraping procedures. In such paper, joint use of daily price data, product characteristics derived from User Generated Contents (UGC) and the construction of a control group with a novel approach allow to implement a Difference-in-Difference analysis, aimed to assert the existence of a causal effect of PMG on prices. The paper finds evidence in favor of price reductions occurring after the PMG policy is repealed. The analysis further investigates if such effect is heterogeneous according to products characteristics, by exploiting UGC (products popularity and quality) and online search visibility measures (Google Search Rank). Estimates suggest that for high quality (visibility) products PMG policies harms competition by keeping prices high, while for low quality (visibility) products, prices decrease during the policy validity period. The second article deals on the literature on the economic impact of transport infrastructure, and in particular on the role that road infrastructure can have on innovative regional capacity. The seminal contribution by Agrawal et al. (2017) is followed to estimate a model of "roads and innovation" where the innovative activity in 1988 is linked to the length of motorways system in 1983, in order to investigate the impact of motorways endowment on the innovative capacity in each Italian NUTS-3 region. The main challenging issue about the estimation of the model arises from the possible endogeneity of highways stock. To deal with this problem, the "historical instrumental variable" approach is followed, by using the length of the ancient Roman roads system dating back to 117 AD as an instrument for the length of current motorways. Overall, the Instrumental Variable estimates indicates that 1983 highways network has a positive and significant impact on 1988 innovative capacity. Moreover, the analysis find a declining role for highways over time. Furthermore, results suggest a spatial reorganization of economic activity rather than a pure net economic effect. The third paper concludes this dissertation. In such final analysis, the effects on the regional economy of a prominent Italian place-based policy, the institution of Istituto Italiano di Tecnologia (Istituto Italiano di Tecnologia, IIT, is a scientific research centre established in 2003 that conduct scientific research in the public interest), are investigated by means of a novel identification strategy, the Synthetic Control Method (SCM). Such identification approach, unlike other counterfactual impact evaluation techniques, is based on the creation of an artificial control unit that not only follows the same pre-treatment trend as the treated unit, but even overlaps the same one. In particular, through the SCM approach, the innovative and economic development (measured by patents per capita, number of local inventors and per capita GDP) of the treated region, namely Genoa, is compared with a set of Italian NUTS-3 control regions with the aim to estimate the causal effect of the location of IIT in 2006. Estimates show significant effects of IIT presence on local patent activity, highly specialised human capital endowment in research and GDP per capita, suggesting the existence of local spillovers from public research.