A scrutiny into fiscal policy in the South African economy: A Bayesian approach with hierarchical priors

Figures & data

Figure 1. Graphic analysis of the mean of economic growth and fiscal variables, 1990Q1-2020Q2.Source: Author’s illustration based on South African Reserve Bank (2020).

Figure 2. Graphic analysis of the trend of economic growth and fiscal variables, 1972Q1-2020Q2.Source: Author’s illustration based on South African Reserve Bank (2020).

Table 1. Posterior marginal likelihood

Optimisation concluded. Posterior marginal likelihood: −1716.465. Hyperparameters: lambda = 1.49452; soc = 0.246702; sur = 0.35986. ||==================================================| 100%. Finished MCMC after 20.3 secs.

Source: Author’s calculation based on South African Reserve Bank (2020).

Table 2. Summary of the BVAR model

Bayesian VAR consisting of 238 observations, 4 variables and 4 lags. Time spent calculating: 16.89 secs. Hyperparameters: lambda, soc, sur. Hyperparameter values after optimisation: 1.513, 0.126, 0.476. Iterations (burnt/thinning): 25,000 (8000/1). Accepted draws (rate): 3933 (0.393).

Source: Author’s calculation based on South African Reserve Bank (2020).

Figure 3. Trace and density plots of all hierarchically treated hyperparameters and the ML.Source: Author’s calculation based on South African Reserve Bank (2020).

Figure 4. Generated impulse responses of the Bayesian VAR.Source: Author’s calculation based on South African Reserve Bank (2020).

Figure 5. Unconditional forecasts for all variables employed in this study.Source: Author’s calculation based on South African Reserve Bank (2020).

South African Reserve Bank. (2020). Retrieved October 1, 2020, from https://www.resbank.co.za

 Google Scholar