Several lines of work have been undertaken since the beginning of the project. For ease of exposition I shall break up the work-done to-date in categories below:
1. Firm-Level and Aggregate Business Cycle Correlation: Empirical Work
Following from the first part of my proposal, I have constructed a dataset at the firm-level using French administrative data, which has annual firm-level growth rates, along with export, import, and multinational exposure to countries around the world. Using these data, I have shown that firms with foreign linkages experience a larger correlation of their annual value-added growth (as well as sales growth) with a trading/multinational partner country relative to a firm that does not have such linkages. These results are found using large panel level regressions, which control for country, sector, and even firm-level fixed effects, thus controlling for potential endogeneity issues that have plagued the literature thus far. I then use the regression results to compute the impact of international trade/multinational linkages on aggregate business comovement between France and its partner countries (its ten largest trading partners). I show that shutting down all trade and multinational linkages will decrease the aggregate correlation of GDP growth across France and its partner countries by one third on average. I have also run a similar exercise using only the one-hundred largest firms in France and have shown that they and their trade/multinational linkages can account for a non-trivial portion of France’s aggregate business cycle comovement with other countries.
2. Firm-Level and Aggregate Business Cycle Correlation: Theoretical Work
I have begun work in constructing a multi-country, multi-sector, heterogeneous firm model of trade, which will allow me to examine the impact of the transmission of a variety of foreign shocks on a country’s GDP. Notably, the model incorporates firm-level heterogeneity, which can be quantified using the rich firm-level dataset assembled for part (1) discussed in the previous paragraph. A crucial aspect of this is that I have data on firms input shares and the origin of intermediate imports and embed this information into the modelled production function. I also take advantage of knowing the destination of firms’ exports, as well as the relative size of firms. These different levels of heterogeneity are used to explore how shocks are transmitted across borders in the model, and the role of firm heterogeneity along different dimensions in contribution to the overall impact of shocks on the aggregate domestic economy. The main code has been written to solve the model quantitatively, and some counterfactual exercises have been conducted. Overall, I find that heterogeneity matters in that by making all firms within the French economy (in a given sector) look the same in terms of size, imported input usage, and export exposure, leads to a smaller impact of foreign shocks (particularly productivity) on France real GDP movements.
3. Firm-to-Firm Linkages and the Transmission of Shocks: Data Acquisition
An important part of my grant proposal was to purchase and exploit the AMADEUS-ORBIS historical database, which is a product of Moody’s (which recently purchased from Bureau VanDijk, which was previously running the database). Using firm-level tax codes, the goal is to then merge these data with French exporters to examine the impact of shock transmission across borders. One example is to study how credit shocks to Spanish firms affects their demand for goods from their French suppliers. I have made progress in the acquisition of the necessary data:
a) I have purchased the AMADEUS-ORBIS data, and French customs is currently working to merge these data with the exporter database.
b) I’ve also obtained credit shock data for Spanish firms from my Spanish co-author, and these data have been merged with the firm-level data and are also been processed by French customs.
