Not logged in.
Quick Search - Contribution
Contribution Details
| Type | Working Paper |
| Scope | Discipline-based scholarship |
| Title | Deep Learning for Search and Matching Models |
| Organization Unit | |
| Authors |
|
| Language |
|
| Institution | University of Zurich |
| Series Name | SSRN |
| Number | 4768566 |
| ISSN | 1556-5068 |
| Number of Pages | 44 |
| Date | 2024 |
| Abstract Text | We develop a new method for characterizing global solutions to search and matching models with aggregate shocks and heterogeneous agents. We formulate general equilibrium as a high dimensional partial differential equation (PDE) with the distribution as a state variable. Solving this problem has previously been intractable because the distribution impacts agent decisions through the matching mechanism rather than through aggregate prices. We overcome these challenges by developing a new deep learning algorithm with efficient sampling in a high dimensional state space. This allows us to study search markets that are not “block recursive”. In applications to labor search models, we show that while block recursivity may approximately hold under symmetric shocks, it fails to capture the dynamics when shocks have an asymmetric impact. Business cycles have a “cleansing” effect by amplifying positive assortative matching in recessions, and the magnitude of the countercyclicality depends on the bargaining process between workers and firms. |
| Free access at | Official URL |
| Official URL | https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4768566 |
| PDF File |
Download from ZORA
|
| Export |
BibTeX
EP3 XML (ZORA)
|
| Keywords | Deep learning, Search and Matching, Block Recursive, Sorting, Business Cycles |