Foreign exchange swaps in the Brazilian economy dataset, 2008 - 2019
Description
This dataset contains all the information utilized in the paper 'foreign exchange swaps: a substitute for the use of international reserves?'. In the B3 related data, the main contribution is the information gathered by data-scraping the daily information available at the B3 exchange to build a time series of the institutional investor's net open positions in this exchange. Furthermore, the time series of the future BRL/USD obtained at the B3 exchange website is made available. In the swaps related data, there are different sets of data. The swaps raw dataset was collected by gathering information made available by the Brazilian Central Bank through its monthly open market notes. Furthermore, data scraped from the Braziian Central Bank norms' search engine provide the information on the type of the swaps contracts. With the datasets, it is possible to reproduce: 1. The descriptive statistics presented on the paper and to reproduce; 2. The logistic model that calculate the log-odds ratio of the spread between the coupon and the libor being larger than the EMBI+ risk-oremium measure be associated with expected or unexpected swaps. 3. The generalized autoregressive conditional heteroskedasticity models presented by the end of the paper. The null hypothesis in these model is that the inclusion of the external regressors has no statistical relationship with the conditional mean of the returns of the BRL/USD future' exchange rate in its first maturity and the alternative hypothesis is that the statistical relationship between the set of external regresors and the dependent variable is different from zero. All of the control variables are publicly available and were obtained at the FRED (St. Louis Federal Reserve database), at the Brazilian Central Bank database and at the IPEADATA database. They are reproduced here to facilitate the reproduction of the paper.
Files
Steps to reproduce
The steps to reproduce are explained in the R scripts and are divided by the subsections of the paper.