Wong Chin Yoong, who helped us throughout the research and writing of this thesis. The purpose of this thesis is to examine how the fiscal adjustments affect the national debt, exclusively in Malaysia.
RESEARCH OVERVIEW
- Research Background
- What motivates us?
- What’s going on in Malaysia?
- Current Issues
- The Solutions
- Problem Statement
- Research Objectives
- General Objective
- Specific Objectives
- Research Questions
- Significance of the Study
To investigate the dynamic effects of changes in government development expenditure on Malaysia's national debt. To investigate the dynamic effects of changes in government wage expenditure on Malaysia's national debt.
LITERATURE REVIEW
- The Emergence of Fiscal Consolidation?
- Does spending tax or tax hikes matter?
- The effects through different time horizon
- Using VAR model to identify fiscal shocks
Findings from Alesina and Perotti (1997) showed that the composition of fiscal adjustments matters regardless of the probability of success and macroeconomic consequences. Also, in the presence of fiscal consolidation and financial crises, it increases the Keynesian effects of total tax revenue. In Akanbi (2013), the author using explicit and robust macroeconometric models, the author analyzed the macroeconomic effects of fiscal policy in South Africa, found that the implementation of expenditure-based fiscal consolidation will be.
In addition, Alesina and Ardagna (2009) examined the composition of fiscal arrangements in OECD countries, tax cuts are likely to stimulate growth compared to those in the expenditure perspective. In other words, it is very important to identify which cuts in which components of government spending, because cuts in different components will result in different consequences, not just cuts in overall spending. 2013) quantified the macroeconomic effects of fiscal consolidation in the US. Erceg and Linda (2013) informed that credibility plays a crucial role for the success of fiscal consolidation.
Meanwhile, Fats and Mihov (2006) investigated the dynamic effects of fiscal policy on consumption and employment. Also, there is a negative relationship between the positive effects of fiscal consolidation and a country's stage of development.
METHODOLOGY
- Data Collection
- The Vector Autoregressive (VAR) Model
- Cholesky Ordering
- Toda and Yamamoto (1995) Procedure
- Unit Root Test
- Stability Conditions
- Breusch-Godfrey serial correlation LM test
The public debt, also known as the national debt, is the total amount of money owed by any branch of government at any given time. If we plot our data in graphs, we can see the trend of the data as shown below:-. Then, regardless of the order of integration of the series, we add VAR-in-levels before the non-stationary data (if integrated to first order or higher).
It should be noted that the additional lag coefficient "m" is not included when the Granger causality test is performed. It relies on a parametric transformation of the model that removes the serial correlation in the error term and leaves asymptotic distributions of the various τ statistics. The PP test does not need to determine the form of the serial correlation.
First, we are not estimating the coefficients, but instead estimating the dynamic effects of the variables, so we tend to allow the unit root problem to exist in the data so that we can better capture the dynamic effects. Next, we run only the unit root test to determine the maximum order of integration, which is one of the key steps in applying the Toda and Yamamoto (1995) procedure.
DATA ANALYSIS
Unit Root Test
OG_GDP = Government operating expenditure normalized by GDP DG_GDP = Government development expenditure normalized by GDP NW_GDP = Government non-wage expenditure normalized by GDP W_GDP = Government payroll expenditure normalized by GDP Direct Taxes_GDP = Government Direct Taxes normalized by GDP.
Lag Length Selection
Therefore, we reduce the lag length to 2 to ensure that the model is dynamically stable. In estimating the effects of DG_GDP and W_GDP on debt, the suggested lag length is 4. Therefore, when we apply the Toda-Yamamoto (T-Y) procedure, we add m=1 additional lag length for each of the independent variables in each of the equations.
Just to clarify, when we estimate our VAR model using EViews, the last added lag length will not be immediately included in the endogenous variable, but instead treated as "exogenous" variable. This is because according to Dave Giles, Professor of Economics at the University of Victoria, if this extra lag were included in the endogenous variables and all the coefficients were counted to perform Granger Causality afterwards, it would be wrong, i.e. the test statistic will not have its usual asymptotic chi-square null distribution.
Diagnostic Checking
Since the proposed lag length is 5 in this baseline model, the p-value of LM tests is 0.1482. Based on figure 4.2, since the inverse characteristic roots are all below 1, the second model (government operating expenditure, net taxes and government debt) is dynamically stable. From table 4.5 we can see that the second model has no autocorrelation problem as the lag length suggested is 2 in this model.
Based on Figure 4.3, the inverse characteristic roots are all below 1 and therefore the third model (government development expenditure, net taxes and government debt) is dynamically stable. Since the suggested lag length is 4 in this model, the p-value (0.6135) is greater than all three levels of significance and therefore we can conclude that there is no autocorrelation problem in this model. Therefore, based on table 4.7, it showed that the probability of lag length 4 is greater than 10%, 5% and 1% significance level and it is not autocorrelation in the fourth model.
From Table 4.8 we can observe that the fifth model is no autocorrelation as the probability of lag length 5 is greater than the significance level of 10%, 5% and 1%. Based on Figure 4.6, the sixth model (government expenditure, direct tax and government debt) is dynamically stable as the inverse characteristic roots are all below 1.
Granger Causality
In this section, we are focusing specifically on the results in which we set debt as the dependent variable. We have placed the remaining results in the appendix since the empirical results of debt seem to be our main concern in this research paper. Based on the table below, we can see that government operating expenditure and non-wage expenditure Granger cause debt at all three levels of significance due to the low P value (0.0000<.
In contrast, Granger's net taxes do not create debt at all three levels of significance, due to the higher P-value (0.8175> when measured together with government household expenditures). Meanwhile, measured by non-wage expenses, what net taxes Granger causes debts comes out differently at 10% and 5% significance levels. Somehow we found that government spending only caused Granger debt with a significance level of 5% and 10% when we applied it with direct taxes rather than net taxes due to lower P value (0.0171<. do not cause 1% debt due to the higher P-value (0.0171>0.01).
Looking at the table, government development expenditures do not create debt at all three significance levels due to the higher P-value (0.6055>. OG = government expenditure normalized by GDP DG = government development expenditure normalized by GDP NV = non-government expenditure - Wage expenditure normalized by GDP W = Government wage expenditure normalized by GDP Direct taxes = Government direct taxes normalized by GDP.
Impulse Response Function (IRF)
- The Role of Different Spending Component
- The Role of Direct Tax
After that, the response of government debt to government spending was again positive in the long run. Figure 4.8 showed that the shock of government operating expenditure has a positive effect on the short-term impact on the share of debt. In contrast, the tax shock appears to be insignificant in terms of its effect on the debt ratio.
Figure 4.10 shows that the wage expenditure shock has a positive effect on the debt ratio in the short and medium term. On the other hand, the tax shock has no significant impact on the debt ratio. The debt ratio will increase in the short term, driven by non-wage expenditure.
After that, the response of the share of debt in public financial expenditure was again positive in the long term. Figure 4.12 shows the response of government expenditures and direct tax revenues to the government debt.
Variance Decomposition
Finally, when we include direct taxes (table 4.16) in our estimation, they do not play an important role in explaining the volatility of debt, while government spending plays an important role. Almost 70% of the contribution of government spending can be used to explain the variation of debt over the long term. NW_GDP = Government non-wage expenditure normalized by GDP T_GDP = Net taxes normalized by GDP.
Direct Taxes_GDP = Government Direct Taxes normalized by GDP D_GDP = Government Debt normalized by GDP.
CONCLUSION
Summary of Result
Therefore, we are unable to determine whether net taxes play an important role in debt reduction. Regarding government spending, we found that government spending is insignificant in debt reduction in the short term, but appears to be significant in debt reduction in the medium term. In other words, a decrease in government spending will reduce debt in the short term; somehow it will result in an increase in debt in the medium and long term.
In the meantime, we have concluded that a reduction in government spending on business will lead to a reduction in debt in the short term. After that, we also observed that the reduction of expenditure on wages and other expenses will result in a reduction of debt in the short term. Ultimately, we will conclude that changes in the spending adjustment would be more effective in reducing debt compared to the tax adjustment.
However, we found that reducing government spending will only be effective in reducing debt in the short term. Based on our calibrated results, we cannot determine the effect of fiscal policy on debt reduction in the medium and long term.
Policy Implication
Limitation and Future Recommendation
An empirical characterization of the dynamic effects of changes in government spending and taxes on production. Tax reforms and labor market performance in the euro area: a simulation-based analysis using the New Area-Wide Model.
