Alexandre Gaillard will defend his thesis on Tuesday 7 June at 10:30 AM by zoom meeting.
«Essays in Macroeconomics with Heterogeneous Agents».
To get the zoom link, please ask the secretariat Elvire Jalran
Supervisor: Professor Patrick Fève
- M. Patrick FEVE, Université Toulouse 1 Capitole, Supervisor
- M. Sang Yoon (Tim) LEE, Queen Mary University of London, Moderator
- Mme Axelle FERRIèRE, Paris School of Economics, Moderator
- M. Christian HELLWIG, Université Toulouse 1 Capitole, Assesor
This thesis is divided into four essays on the broad implication of heterogeneity in macroeconomics. It puts a special emphasis on (i) the relation between the sources of wealth inequality and the effects of wealth redistribution, (ii) the effects of unemployment insurance on the selection into self-employed, (iii) the role of the market for businesses in transmitting the intangible value of firms across generations of entrepreneurs, and (iv) the role of international trade linkages with many interconnected countries in generating cross-country GDP correlations.
In my first chapter, together with Philipp Wangner, we show that the aggregate and welfare implications of redistributing wealth depends on the underlying forces behind capital investment and wealth inequality. In the data, rich households invest a higher fraction of their wealth into risky assets. This is the result of two distinct channels. Wealthy households may invest differently due to heterogeneity in specific skill or risk tolerance (type-dependence) or because wealth itself induces wealthy households to undertake riskier investments (scale-dependence). We first clarify the role of type and scale dependence and we argue that a number of existing frameworks studying the macro consequences of micro investment heterogeneity rest on a particular combination of type and scale dependence. Second, we show that their distinction is crucial for assessing the effects of wealth taxation. In an incomplete markets quantitative model calibrated to the US using micro datasets, both channels are found to lead to opposite predictions regarding wealth taxation. Under type-dependence, rich individuals with low capital returns dissave faster than those with high capital returns. By taxing the stock of wealth of the richest households, only the fittest survive at the top which reinforces the selection of agents with high investment skills among the rich. When returns to wealth reflect capital productivity, taxing wealth at a high rate is optimal because it raises productivity. Under scale-dependence, a wealth tax reduces productive investments, such that subsidizing wealth becomes optimal. In a benchmark model calibrated to take into account both channels, it is optimal to tax wealth at 0.8% above an exemption threshold of 550K with little effects on overall productivity. This tax rate remains robust when returns reflect rents instead of productivity. However, in such a case, the optimal wealth tax increases under scale dependence but decreases under type dependence such that both effects on the size of extracted rents almost cancel each other in the benchmark economy.
In the second chapter, Sumudu Kankanamge and I study how entrepreneurship contributes to the micro and macro-level patterns of gross labor market flows, and explains why the selection into that occupation is highly responsive to unemployment insurance (UI) changes. Our framework merges search models in the spirit of Mortensen et al. (1994) with an occupational choice models with entrepreneurship along the lines of Quadrini (2000) and Cagetti and De Nardi (2006). We show that our model is able to replicate key facts regarding occupational flows at the aggregate level and along the ability and wealth distributions. Higher UI is associated with strong disincentive to start businesses out of unemployment and employment, consistent with CPS data. Intuitively, higher UI generosity changes the riskiness of self-employment relative to paid-employment. In turn, this has important aggregate implications on occupational masses. Surprisingly, we find that an increase in UI generosity leads to an increases in the unemployment and employment rates, but substantially decreases the self-employment rate.
In the third chapter, Sumudu Kankanamge and I study the role of the market for small and medium-sized enterprises (SME) for the transmission of the intangible value of businesses within and across generations of entrepreneurs. In the data, a large fraction of entrepreneurs enter this occupation by acquiring an existing business instead of founding a new one. In contrast, the share of inherited businesses accounts only for a small fraction of business transfers. Using a new dataset, we also document that the market for SME is subject to important selling frictions. Motivated by these facts, a large-scale life-cycle model with entrepreneurs is developed to understand and quantify the role of the market for SME. A key attribute of the model is that newly established businesses are, on average, less productive and face higher failure risks than older ones, consistent with the selection of the best firms over time. The market for SME lets those businesses be transferred across individuals. Younger potential entrepreneurs would like to buy an existing business, but selling frictions and borrowing constraints prevent them from doing so. Shutting the market down leads to a substantial drop in aggregate productivity and output, and alters the pool of firms, incentives to enter and exit, and the wealth distribution.
In the fourth chapter, Françcois de Soyres and I study the role of international trade linkages in increasing GDP synchronization. Empirically, trade linkages are associated with higher cross-country GDP correlations. However, international real business cycle models fail to generate the magnitude of this relationship. This puzzle is known as the Trade Comovement Puzzle (TCP) since Kose and Yi (2006). We argue that the way GDP is measured by statistical agencies, using double deflation, may largely account for this puzzle. When base period prices are used in real GDP construction, any distortion between the price of imported input and their marginal revenue product generates a link between input usage and movements in real GDP, which in turn increases the strength of propagation of shocks across countries. Focusing on two common sources of such a distortion, markups and love of variety, we construct a many country international business cycle model with imperfect competition and an extensive margin of imported goods. The model is shown to quantitatively replicate the strong relationship between trade linkages and cross country GDP correlation when GDP is measured using the same method as in statistical agencies. Each component, markups and love of variety, accounts for a significant fraction of the relationship. The results highlight that, when comparing a macroeconomic model to the data, it is key to define aggregate variables in a way that is consistent with statistical agencies’ procedures.