Working Paper: CEPR ID: DP14448
Authors: Henrique Alpalho; Marta Lopes; João Pereira dos Santos; Jos Tavares
Abstract: We investigate the causal relationship between local government expenditure and private firm performance, using the quantity and naming of civil parishes within each municipality as an instrumental variable. Religious denominations are taken as a proxy for strong local identity, which likely increases competition for resources between neighboring parishes. We explore a dataset on the universe of private firms, local government expenditure categories and socio-economic indicators for all mainland Portuguese municipalities, in a period encompassingboth normal and crisis times. The number of parishes per municipality, as exogenously set by the central government, and the number of parishes that display religious denominations are both used as instruments that explain local government spending, indirectly impacting firm performance. We find that both display considerable power in determining total primary and current spending, which then positively impacts private firms’ sales and value added. Using religious denominations is found to yield a particularly potent instrument, confirmingand expanding the baseline results. In a field that mostly relies on natural experiments for instrumental variable frameworks, our proposed instruments are both easily obtainable and powerful.
Keywords: instrumental variables; local governments; local fiscal multiplier; firm performance; fragmentation; local identity; religion
JEL Codes: D72; E62; H72
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| local government expenditure (H76) | private firm performance (L25) |
| number of civil parishes (R59) | local government expenditure (H76) |
| number of parishes with religious denominations (Z12) | local government expenditure (H76) |
| number of parishes (Z12) | local government expenditure (H76) |
| local government expenditure (H76) | firm sales (L10) |
| local government expenditure (H76) | value added (D46) |