In addition, this study places special emphasis on the relationship between bond spreads and fiscal indicators with additional consideration of the effect of control and dummy variables. This naturally raises the question of how and to what extent bond spreads and fiscal indicators can influence each other (Nickel et al., 2011). Furthermore, theoretical and empirical studies in the existing literature do not seem to provide a comprehensive understanding of the dynamic relationship between bond spreads and fiscal indicators.
In this study, I therefore examine the dynamic relationship between government bond spreads and fiscal policy indicators, e.g. In the aid literature, however, the connection between government bond spreads and fiscal policy indicators has been taken into account from both theoretical and empirical points of view. Before delving into the dynamic relationship between bond spreads and fiscal indicators, the reason why these two fiscal indicators were determined and the interaction between them will be discussed.
In light of the above, this paper aims to expand the scope of those discussed above and consider other factors that will influence the dynamic relationship between government bond spreads and fiscal indicators.
Data and Econometric Methodology
Variables Definition
It is not necessary to start from scratch, but the variables of interest consist of the ratios of primary budget balance and gross external debt over GDP as fiscal indicators, and EMBI+ as government bond spreads. Thus, a higher ratio of primary budget balance over GDP is expected to reduce bond spreads through higher credibility. In terms of gross external debt, it is the sum of short-term and long-term outstanding debt stock of the public sector, the private sector and the Central Bank of the Republic of Turkey (CBRT).
When it comes to EMBI+, it is the interest payment of Turkey's 10-year US dollar-denominated bond yield relative to the default-free instrument of the 10-year US Treasury yield. As a measure of borrowing costs for EMs, the Emerging Market Bond Index (EMBI) is often used in international financial markets (Özatay et al., 2009). Although EMBI and EMBI+ are highly correlated (0.98), EMBI+ is the extended form of the EMBI index, covering only Brady Bonds, and differs slightly from EMBI by also including Eurobonds, sovereign external loans and local instruments (Morgan , 1995).
On the other hand, some control variables have been included in the model to ensure that I am not ignoring the pull and push factors that would have a significant impact on the dynamic structure of the analysis. An increasing ratio of the current account balance to GDP would increase the government's resilience to risky external developments and competitiveness in servicing its debt, implying that bond spreads are narrowing. However, in terms of push factors, the three-month US Treasury yield, which is the short-term segment of risk-free US Treasury bonds, is used to measure the degree of global liquidity and short-term capital flow in the international market.
I then replaced the 3-month US Treasury bill rate with the 10-year US Treasury bond rate to check the robustness of the model. Summary statistics of the variables and where the data are collected from and how they are proxyed are presented in Table 1 and Table 2 below, respectively. Louis FED (FRED) https://fred.stlouisfed.org LVIXI Volatility Index (VIX) / Log of the Index St.
Methodology
National Bureau of Economic Research (NBER) https://www.nber.org/research/data/us-business-cycle-expansions-and-contractions. Since the order in the VAR model is important for a better understanding of the causality between the variables, I ordered the variables taking into account their relative exogeneity. In other words, the variables are arranged to align from more exogenous to less endogenous using economic theory.
According to this logic, I classify the variables as follows: Growth rate of real GDP (GDP) - ratio of current account balance to GDP (DCAB) - ratio of gross external debt to GDP (DGED)/. BILLRT is then replaced by the US 10-year bond rate (TENYR) and the IRF is rerun to check the robustness of the model and introduce a dummy variable to account for the effect of the 2007–2008 financial crisis. As an introduction, I first use the unit root test to test the stationarity of the variables with the Augmented Dickey-Fuller (ADF) method, because the accuracy of the prediction of the reaction of the variables to shocks strongly depends on the state of stationarity. variables.
Accordingly, the zero is not rejected when 𝜑 = 1 and the variables are said to be non-stationary, while in the case of 𝜑 < 1, the zero is rejected and the variable is said to be stationary by referring to Eq. However, this allowed us to make a short-term analysis, since taking the first differences between the variables causes some loss of information in the long run (Wooldridge, 2013, p.632). To estimate the VAR model and see the results of both IRFs and Granger causality tests, the correct delay length for the model must first be chosen.
Since the VAR model will be applied twice to each fiscal indicator, I separately estimated the appropriate lag length for two different models. Granger causality tests allow us to analyze a possible causal relationship between the variables in question. Similarly, one can say that 𝑦𝑡 does not granger cause 𝑥𝑡 if ∅𝑖 = 0, and does granger cause if ∅𝑖 0.
Empirical Results
This specifically implies that increases or decreases in the ratio of gross foreign debt over GDP can be simply explained by the past values of the government bond spreads. The collective expression of the risk factors therefore appears to explain a large part of the changes in the ratio of gross external debt to GDP in the short term. On the other hand, Table 5 below reports that there is unidirectional causality from bond spreads (EMBI+) to the primary budget balance as a percentage of GDP (SMPBS), implying that deviations in primary budget balance-to-GDP ratio are quite probably driven by bond spreads in addition to the effects of certain country-specific factors.
At first glance, both gross external debt and the primary budget balance in Turkey appear to be driven by bond spreads. However, strangely, bond spreads are a common expression of various risk factors, so it is quite difficult to know exactly which factor has the most impact on fiscal indicators in Turkey. Over time, they would inevitably affect the country's riskiness through credit risk and the probability of default.
To better understand the dynamic relationship between government bond spreads and fiscal indicators, one must look at the effect of a positive shock on other endogenous variables in the model. To better assess the individual relevance of the fiscal indicators for bond spreads, I separately examined IRFs for each fiscal variable. The short-run evidence presented above indicates that both the ratio of gross debt and the primary budget balance over GDP is explained by changes in the country's risk premium.
This is because there is a unidirectional causality from bond spreads to fiscal indicators and the shock caused by bond spreads is more consistent and statistically speaking. Since the dynamic relationship between bond spreads and fiscal indicators is analyzed, I have reported only the IRF results belonging to these variables of interest. That is, the robustness of the model was confirmed by replacing one exogenous variable with another in a recursive order (Appendix 2).
Conclusion
These results draw our attention to the importance of examining the role of fiscal indicators in the medium and long term against external shocks. Although extensive research has been conducted on the determinants and significance of bond spreads, there is no single study that adequately addresses the use of fiscal indicators in a dynamic context in mitigating the disruptive effects of short-term external shocks. and immediate increases in bond spreads. The results of this paper show that the ratio of gross external debt to GDP responds positively to a positive shock in government bond spreads for four periods, statistically significant only in the third period.
On the other hand, the response of bond spreads to the gross external debt shock is insignificant but positive for about three periods. Regarding the relationship between the primary budget balance and government bond spreads, a positive shock to government bond spreads appears to worsen the equilibrium by about four periods. Overall, the results of the Granger causality test and impulse response functions show that the ratio of gross external debt to GDP appears to be more sensitive and important to changes in bond spreads compared to the ratio of the primary budget balance to GDP.
This eventually shifts our attention to strengthening the structure of fiscal fundamentals for our case, because countries that exhibit better fiscal policy performance appear to have lower bond spreads and better financial structures. Overall, this analysis has some policy implications for policymakers seeking to reduce the negative effects of bond spreads in the short term. I suggest that in the short term, fiscal indicators defend against the worsening effect of higher debt service.
This article is produced from the thesis "The dynamic relationship between government bond spreads and fiscal indicators: Evidence from Turkey". Review of the determinants of sovereign bond spreads in EMU (Working Papers Department of Economics, ISEG No. WP 08/2016). Determinants of sovereign bond spreads in the Eurozone: in good times and in bad.
Determinants of emerging market sovereign bond spreads: Fundamentals against financial stress (IMF Working Paper No. 2010/281). Global monetary conditions versus country-specific factors in determining emerging market debt spreads.
Overall Impulse-Response Results
The Robustness Check Results
