Ireland's Economic Growth: Impact of Rates, GDP, and CPI - Empirical Analysis, Ejercicios de Macroeconomía

¡Descarga Ireland's Economic Growth: Impact of Rates, GDP, and CPI - Empirical Analysis y más Ejercicios en PDF de Macroeconomía solo en Docsity! Mario Iñiguez Mangado 100407581@alumnos.uc3m.es My Empirical Project Does the Interest Rate, Real GDP and the CPI Growth affect the Gross Domestic Product, and so the economic growth? The case of the Celtic Tiger, Ireland. Introduction and literature review Economic growth is an increase in the production of economic goods and services, compared from one period to another. In economics, this production is known as the Gross Domestic Product (GDP). It can be measured in nominal or in real (inflation adjustment) terms. Interest rate is the amount a lender charges for the use of assets, its volatility is very related with inflation rates, and so the economic growth. In the business field, the interest rate can be known as the profitability of an inversion, where there are two ways of use, simple interest, or compound interest rate. Consumer Price Index (CPI) is a measure of the average variation of prices of a basket of consumer goods on a time and place determined. The Growth of CPI is the percent annual change of CPI, meaning that is the best approach to inflation and deflation. The Celtic Tiger is a nickname for Ireland relating to the rapid economic growth it experienced during the 1988 until 2007, with an average annual Real GDP growth of 7% (reaching at most 14% during the period 1995-2005). In addition, the unemployment rate fell from 16 per cent in 1994, to 4 per cent in 2000, and the non-agricultural employment rise from 33 per cent of the population in 1993 to 46 per cent in 20071. The purpose of this study is to analyze the effect of basic macroeconomic factors such as interest rate, real GDP, and growth of CPI on real economic growth of Ireland in the period of the Celtic Tiger and later by answering the following questions: 1- Is the effect of interest rate on real economic growth statistically significant? 2- Is the effect of consumer price index growth on real economic growth statistically significant? 3- Is the effect of real gross domestic product on real economic growth statistically significant? The papers in which I will support my analysis and conclusions are the following: - “What went wrong in Ireland”, Patrick Honohan (2009). This paper is more a theorical paper than an econometric paper. However, it will be helpful due to the author’s review of the Celtic Tiger and the fiscal crisis period for Ireland. - “Analysis of the effect of inflation, interest rates and exchange rates on gross domestic product (GDP) in Indonesia”, Hatane Semuel and Stephanie Nurina (2015). This is the first of my two papers of econometrical analysis reference. This 1 “What went wrong in Ireland”, Patrick Honohan (2009) Mario Iñiguez Mangado 100407581@alumnos.uc3m.es paper will be helpful due to the analysis between the relationship of inflation (in my case CPI), interest rate and gross domestic product. - “The effect of interest rate, inflation rate, GDP, on real economic growth rate in Jordan”, Abdul Aziz Farid Saymeh and Marwan Mohammad Abu Orabi (2013). This is my main econometrical reference paper. As well as the paper mentioned before, the analysis of this paper will support my analysis. However, this paper is more related with my empirical project than the previous paper, due to this article focuses on the relationship between interest rate, inflation rate and GDP on real economic growth. Hypotheses tested. In order to answer the questions mentioned before, the hypotheses tested on my analysis are: H1): There is no significant effect of interest rate on real economic growth. H2): There is no significant effect of consumer price index growth on real economic growth. H3): There is no significant effect of real gross domestic product on real economic growth. Methodology Data The data was obtained from Fred and Central Bank of Ireland for the period 1997Q1- 2015Q4 in quarterly frequency. The growth variables are calculated by the following formula: 𝑋𝐺𝑟𝑜𝑤𝑡ℎ = 𝑋𝑡−𝑋𝑡−1 𝑋𝑡−1 *100% Augmented Dickey-Fuller Test and Lag Length Criteria (Based on Schwartz Criterion) To test if the variables are stationary or not (null hypothesis: have a unit root), the study use the Augmented Dickey fuller Test (ADF) which eliminate error term correlations. There are three ways of testing for unit root: A. No intercept and no trend: Δyt = φyt−1 + ∑ ∅𝑘 𝑗=1 𝑗 Δyt−j + 𝑎𝑡 B. Intercept and no trend: Δyt = α + φyt−1 + ∑ ∅𝑘 𝑗=1 𝑗 Δyt−j + 𝑎𝑡 C. Intercept and trend: Δyt = α + βt + φyt−1 + ∑ ∅𝑘 𝑗=1 𝑗 Δyt−j + 𝑎𝑡 For the Lag Length Criteria, the study is based on Schwartz Criterion. Mario Iñiguez Mangado 100407581@alumnos.uc3m.es Table-3. Johansen Unrestricted Co-Integration Rank Test (Maximum Eigenvalue) Hypothesized Number of Max-Eigen Cointegrating Equations Eigen Value Statistic 5% Critical Value P-Value None * 0.467667 45.39503 27.58434 0.0001 At most 1 0.250008 20.71389 21.13162 0.0571 At most 2 * 0.205992 16.60766 14.26460 0.0209 At most 3 0.003778 0.272534 3.841465 0.6016 *Denotes rejection of the hypothesis at the 0.05 level. These tests are done to analyze if variable have long term equilibrium relationship. The results show that with both trace and eigenvalue statistic have at most two cointegrating equations, meaning all variables have long term relationship. Table 4: Engle-Granger co-integration test, where the null hypothesis is series are not cointegrated. Table-4. Engle-Granger Co-Integration Test. Value P-Value Tau-statistic -11.64341 0.0000 Z-statistic -94.52716 0.0000 *MacKinnon p-values. As we have seen before, the variables are co-integrated, because we reject the null- hypothesis that series are not-cointegrated (p-value is 0) Table 5 shows the pairwise Granger causality tests, to analyze the granger causality relations between Real GDP (RGDP), Real GDP Growth (G), Interest Rate (IR) and CPI Growth (CPI). Table-5. Pairwise Granger Causality Tests Null Hypothesis: Obs F-Statistic P-value RGDP does not Granger Cause G 70 1.82546 0.1217 G does not Granger Cause RGDP 1.51183 0.2000 IR does not Granger Cause G 70 0.59132 0.7066 G does not Granger Cause IR 0.79677 0.5564 CPI does not Granger Cause G 70 2.20797 0.0654 G does not Granger Cause CPI 2.75193 0.0266 Mario Iñiguez Mangado 100407581@alumnos.uc3m.es (0.07798) (0.00007) (0.11556) (0.11451) (0.08406) (0.14449) (0.08276) IR does not Granger Cause RGDP 71 0.93870 0.4627 RGDP does not Granger Cause IR 1.20014 0.3201 CPI does not Granger Cause RGDP 71 1.27311 0.2874 RGDP does not Granger Cause CPI 2.98460 0.0180 CPI does not Granger Cause IR 71 0.98490 0.4346 IR does not Granger Cause CPI 0.38489 0.8572 Looking at the p-values, we notice that Real GDP Growth does granger cause CPI Growth, and Real GDP does granger cause CPI. In addition, CPI does not granger cause Real GDP Growth and CPI does not granger cause Real GDP. Otherwise, the remainder relations between the variables do not granger cause each other. After running different least squares estimation, the best least square model is shown on table 6. Then, it is tested for ARCH effects, to check if there is autocorrelation or heteroskedasticity. After that, we look for the best ARCH/GARCH/TARCH/EGARCH model, to consider the heteroskedasticity problem (in case of need), and after all, we will have a regression that adjust the coefficients to a more real situation. Table 6 is the estimation output of the least square model and table 7 for the EGARCH model. Table-6. Least square model 𝐶 𝑅𝐺𝐷𝑃 𝐼𝑅 𝐼𝑅(−𝟏) 𝐶𝑃𝐼 𝐶𝑃𝐼(−𝟏) 𝐶𝑃𝐼(−𝟐) Coeff. 0.01875 0.00249 0.04668 -0.03973 0.08289 -0.11197 0.06513 P-Value 0.8107 0.0000 0.6876 0.7297 0.3276 0.4413 0.4341 R-Square 0.93078 Schwarz info. Criterion 1.57066 *Standard error in parenthesis. Values in bold indicates they are significatively at 5% Table-6.1. Heteroskedasticity Test (ARCH) for LS model. Null Hypothesis: there is no arching effects. F-Statistic 8.529755 P-Value 0.0047 Mario Iñiguez Mangado 100407581@alumnos.uc3m.es (0.01468) (0.00000) ) (0.02267) (0.01828) (0.02074) (0.03020) (0.00985) Table-7. EGARCH (1,1) model 𝐶 𝑅𝐺𝐷𝑃 𝐼𝑅 𝐼𝑅(−𝟏) 𝐶𝑃𝐼 𝐶𝑃𝐼(−𝟏) 𝐶𝑃𝐼(−𝟐) Coeff. -0.04755 0.00226 0.04342 -0.00220 0.06663 -0.10279 0.07973 P-Value 0.0012 0.0000 0.0555 0.9041 0.0013 0.0007 0.0000 R-Square 0.95684 Schwarz info. Criterion 0.41336 *Standard error in parenthesis. Values in bold indicates they are significatively at 5% Table-7.1. Heteroskedasticity Test (ARCH) for EGARCH (1,1) model. Null Hypothesis: there is no arching effects. F-Statistic 0.03033 P-Value 0.8622 As we can see on Table 6, the only statistically significant variable for the Real GDP Growth is the Real GDP which makes sense, with a coefficient of 0.00249. Moreover, we observe a R-Square of 0.93078, meaning the independent variables explains the 93,07% of the dependent variable and a Schwarz info criterion of 1.57066. In addition, we observe on Table 6.1. that LS model has arching effects (p-value of 0.0047), so there must be done a conditional heteroskedasticity model. After some testing, the best ARCH model obtained is the model shown on Table 7. That model is an EGARCH, which its principal characteristic is that this model consider the asymmetry and the volatility. Notice that as much as the constant as the Real GDP and the CPI Growth are statistically significant for the Real GDP Growth (all these variable’s p-values are lower than 0.05). However, Interest Rate is not statistically significant for the Real GDP Growth, (highlight that the p-value of IR is 0.0555, which is very close to 0.05, while the p-value of first lag of IR is 0.9041). Also, the EGARCH model has a R- Square of 0.95684, slightly bigger than the LS R-Square. In addition, the Schwarz info criterion is 0.41336, which is better than the 1.57066 of the LS model. To sum up, on table 7.1, we notice that with this model the heteroskedasticity problem is corrected, due to the p-value is 0.8622, so we don’t reject the null hypothesis, and we have a better adjusted coefficients to a more real situation. Mario Iñiguez Mangado 100407581@alumnos.uc3m.es Figure 3. Real GDP (RGDP) and First Difference Real GDP (DRGDP) Graph Figure 4. CPI Growth Graph Mario Iñiguez Mangado 100407581@alumnos.uc3m.es Figure 5. Least Square Model Residuals Figure 6. EGARCH Model Residuals