The effects of expansionary fiscal policy on macroeconomic development: A study on tax competition in an emerging market.
Description
This dataset was assembled to examine the macroeconomic effects of expansionary fiscal policy implemented through tax incentives under the state value-added tax on the circulation of goods and services (ICMS) in Santa Catarina, Brazil. The underlying research hypothesis is that tax benefits granted by subnational governments, as part of a strategy of fiscal competition, are capable of stimulating macroeconomic development, particularly economic activity and employment. The data consists of monthly macroeconomic time series spanning the period from January 1997 to December 2020. The variables capture key dimensions of the state economy, including the real exchange rate, real interest rate, electricity consumption (used as a proxy for economic activity), exports, imports, formal employment, ICMS revenue, and inflation. These series were collected from official Brazilian institutions and state agencies, ensuring consistency, reliability, and comparability over time. The dataset supports the estimation of a vector autoregression (VAR) model designed to trace the dynamic responses of macroeconomic variables to shocks in ICMS revenue, which reflect changes in both economic activity and fiscal policy, including the granting of tax incentives. The empirical results reported in the associated article indicate that ICMS shocks generate short-term effects on the exchange rate and imports but produce limited or statistically insignificant impacts on economic activity, employment, inflation, and state tax revenue over time. These findings challenge traditional Keynesian expectations of broad-based stimulus effects and are more consistent with neoclassical and public-choice interpretations of tax incentives as potentially inefficient and weakly connected to employment generation. All series were seasonally adjusted, deflated when appropriate, and transformed into logarithmic or percentage variations to ensure stationarity and suitability for time-series econometric analysis. The repository includes the original dataset used in the estimations, as well as a PDF log documenting the Stata commands and outputs employed to estimate the VAR model. Together, these materials allow users to understand how the data were gathered, processed, and analyzed, and to replicate or extend the empirical analysis in future research on fiscal policy, tax competition, and regional economic development.
Files
Steps to reproduce
The dataset was constructed by collecting monthly macroeconomic indicators for the state of Santa Catarina, Brazil, from official public sources, including national statistical agencies and state-level fiscal authorities. The variables were selected to capture key dimensions of macroeconomic performance and fiscal policy, with particular emphasis on the role of tax incentives granted under the state value-added tax on the circulation of goods and services (ICMS). After collection, the raw data was organized into a unified time-series database covering the period from January 1997 to December 2020. All series were inspected for consistency, missing values, and structural coherence. Seasonal adjustment procedures were applied where appropriate, and nominal variables were deflated using standard price indices to obtain real measures. Data preparation and econometric analysis were conducted using Stata software. The series were transformed to ensure stationarity, with most variables expressed as first differences of logarithms and interest rates and inflation expressed in percentage variations. A reduced-form vector autoregression (VAR) model was then estimated using ordinary least squares, with the optimal lag structure determined by standard information criteria. The analytical workflow included diagnostic tests for stability, autocorrelation, and residual normality, followed by Granger causality tests, impulse-response functions, and forecast error variance decompositions. The complete sequence of commands and outputs generated during the estimation process is documented in a PDF log included in the repository, allowing other researchers to understand, replicate, and extend the empirical analysis.
Institutions
- Universidade Federal de Santa Catarina