MINISTRY OF EDUCATION AND TRAINING STATE BANK OF VIET NAM BANKING UNIVERSITY HO CHI MINH CITY NGUYEN HOANG CHUNG ANALYZING THE RESPONSE OF MACRO VARIABLES BEFORE MONETARY POLICY SHOCKS THROUGH A NEW KEYNES MODEL TO IMPROVE THE QUALITY OF VIET NAM’S MACRO ECONOMIC FORECAST THESIS SUMMARY Major: Finance - Banking Code: 34 02 01 HO CHI MINH CITY - 2019 MINISTRY OF EDUCATION AND TRAINING STATE BANK OF VIET NAM BANKING UNIVERSITY HO CHI MINH CITY NGUYEN HOANG CHUNG ANALYZING THE RESPONSE OF MACRO VARIABLES BEFORE MONETARY POLICY SHOCKS THROUGH A NEW KEYNES MODEL TO IMPROVE THE QUALITY OF VIET NAM’S MACRO ECONOMIC FORECAST THESIS SUMMARY Major: Finance - Banking Code: 34 02 01 ACADEMIC ADVISOR Assoc Prof., Ph.D NGUYEN DUC TRUNG Ph.D LE DINH HAC HO CHI MINH CITY - 2019 LIST OF AUTHOR’S PUBLICATION* No Name of publications Analysis of Monetary Policy shocks in the New Keynes model for Viet Nam’s Economy: Publication’s Publication’s Publication’s number year address ISSN Springer 2019 1860-949X Rational Expectations Approach Forecasting model for Vietnam's small and EconVN Paper 2019 ISSN opened economy Methodology approach: 2017 1859 - 1124 BVAR-DSGE The impact of monetary policy and macro ISSN prudential policy on financial stability in Viet 2018 1859 - 3682 Nam – regarding credit growth Viet Nam’s macroeconomy – Analysis and forecasting: 2018 – Actively respond to shocks to maintain macroeconomic stability 978-604-971- 2018 2019 before the decisive turn Determinants of debt to market value ratio: the case of Viet Nam of 2017 monetary ISSN 1859 – 4050 companies in Vietnam effect ISSN 1859 - 3682 Factors affecting the capital structure of listed The policy Banking Technology Review Banking University HCMC Conference 485-6 forecasting: Viet Nam’s macroeconomy 2019 Development ISBN Viet Nam’s macroeconomy – Analysis and Journal of Economic and 2017 0866 - 7802 0866 – 7802 Note: The publications (1), (2), (3), (8) relate directly to this thesis (*) Each study presented includes: cover page, table of contents, articles Banking Technology Review External Economics Review Economics Technology Journal of ISSN A New Keynes model without the LM curve HCMC Workshop Journal of ISSN macroprudential policy to financial stability in Viet Nam 2017 Banking University 2018 Economics Technology TABLE OF CONTENTS CHAPTER 1: INTRODUTION 1.1 The necessity of the study 1.2 Setting research issues 1.3 Objectives of the study 1.3.1 General objectives 1.3.2 Detail objectives .2 1.4 The study’s questions 1.5 Objectives and scope of the study 1.5.1 The study’s objectives 1.5.1.1 The factors of monetary policy and macroprudential policy affect credit growth 1.5.1.2 Macroeconomic variables of the New Keynes model .3 1.5.2 The scope of this study 1.5.2.1 Factors affect credit growth .3 1.5.2.2 Factors affect a small opened economy 1.6 Methodology Research 1.6.1 The method of estimating variables affects credit growth 1.6.2 The method of estimating macro variables of the SVAR new Keynes model 1.6.3 The method of estimating macro variables of the DSGE new Keynes model .4 1.7 The scientific and practical significance of the thesis 1.8 Research gap 1.9 The structure of study 1.10 Summary Chapter CHAPTER 2: THE THEORY BACKGROUND OF THE NEW KEYNES MODEL 2.1 Theoretial framework and literature review 2.1.1 The theory of macroprudential policy 2.1.1.1 Financial Stability 2.1.1.2 Macroprudential policy 2.1.2 The theory of monetary policy 2.1.2.1 Conceptual framework 2.1.2.2 Monetary policy transmission 2.1.2.3 The basic principles of monetary policy 2.1.2.4 The purpose of the monetary policy 2.1.3 The theory of general equilibrium 2.1.4 The theory of aggregate supply – aggregate demand 2.1.5 The theory of money supply and inflation 2.1.6 The framework of the inflation targeting policy 2.1.6.1 The conception 2.1.6.2 Characteristics and key factors of inflation targeting policy 2.1.6.3 The goals of inflation targeting policy 2.1.6.4 Conditions for applying inflation targeting policy 2.2 Basic theories of the new Keynesian model .8 2.2.1 The overview of the New Keynesian theory 2.2.1.1 The New Keynesian model structure with equations 2.2.1.2 The New Keynesian model structure with equations 2.2.1.3 The model structure applied in Vietnam (developed by IMF) 2.2.2 An SVAR new Keynes model 2.2.2.1 IS equation 10 2.2.2.2 AS equation 10 2.2.2.3 Uncovered Interest Parity (UIP) 10 2.2.2.4 A forward – looking monetary policy 10 2.2.2.5 Rational expectations econometrics 10 2.2.3 A DSGE new Keynes model 11 2.2.3.1 The theory of DSGE model 11 2.2.3.2 The structure of the DSGE model 12 2.2.3.3 The framework of policy analysis and theory of economic forecasting 12 2.2.3.4 The variety of DSGE model 13 2.2.3.5 Frictions cost of DSGE model .13 2.2.3.6 Advantage and disadvantage of DSGE model 13 2.3 Literature review .14 2.3.1 Studies of monetary and macroprudential policy .14 2.3.1.1 Foreign studies 14 2.3.1.2 Domestic studies 14 2.3.2 Studies of the SVAR new Keynes model .14 2.3.3 Studies of the DSGE new Keynes model 15 2.4 Summary chapter 16 CHAPTER 3: METHODOLOGY RESEARCH 17 3.1 For monetary and macroprudential policy .17 3.1.1 Model 17 3.1.2 Data research .17 3.1.3 Methodology research 17 3.1.4 Research procedures 17 3.2 For the SVAR new Keynes model 18 3.2.1 Econometric model 18 3.2.1.1 VAR model 18 3.2.1.2 SVAR model 18 3.2.2 Data research .18 3.2.3 Methodology research 19 3.2.4 Research procedures 19 3.2.4.1 Diagnostics of the reduced form VAR model .19 3.2.4.2 Contemporaneous structural parameter analysis 19 3.2.4.3 Impulse response function (IRF) 19 3.2.4.4 Variance decomposition 19 3.3 For the DSGE new Keynes model 19 3.3.1 Model 19 3.3.1.1 The IS equation .19 3.3.1.2 The New Keynes Phillips curve equation .20 3.3.1.3 Monetary policy equation 20 3.3.1.4 Other equations .20 3.3.2 Data research .20 Source: Summary of the author from many different studies 20 3.3.3 Methodology research 20 3.3.4 Research procedures 21 3.3.4.1 Procedures estimation for BVAR – DSGE model 21 3.3.4.2 Priors for the DSGE model 21 3.5 Summary chapter 22 CHAPTER 4: RESEARCH RESULTS AND DISCUSSION 23 4.1 Descriptive statistics 23 4.1.1 Micro variables data 23 4.1.2 Macro variables data 23 4.2 Empirical analysis .25 4.2.1 The first model .25 4.2.1.1 FGLS results and discussions 25 4.2.1.2 Robustness of FGLS model 25 4.2.2 The second model 25 4.2.2.1 Contemporaneous structural parameter estimation .25 4.2.2.2 Impulse response function .26 4.2.2.8 Variance decomposition 28 4.2.3 The third model 30 4.2.3.1 Selection of  and the lag length 30 4.2.3.2 DSGE model 30 4.3 Summary chapter 32 CHAPTER 5: CONCLUSIONS, POLICY IMPLICATIONS AND LIMITATIONS 33 5.1 The key results 33 5.1.1 Affirming the role of monetary policy in Vietnam‘s macroeconomy stability 33 5.1.2 Macro variables react dynamically to policy shocks 33 5.1.3 The new Keynesian forecasting model has a meaningful analysis of policy 33 5.2 Policy implications 34 5.2.1 The role of SBV in using monetary policy tools 34 5.2.2 Operating monetary policy 34 5.2.3 A New Keynesian model for macro forecasting 34 5.3 Limitations 34 5.3.1 Data and variables .34 5.3.2 Methodology and Viet Nam economic characteristics .34 5.3.3 Research results 35 5.4 Summary chapter 35 CHAPTER 1: INTRODUTION 1.1 The necessity of the study The scenario of the global financial crisis has changed the perception of Central Banks in the world that the aim of price stability is not enough to ensure financial stability Therefore, central banks need to implement measures and policies aimed at financial stability through monetary and macroprudential policies Accordingly, the evaluation, analysis and forecast of economic development, especially some key macro indicators, play a very important role in planning economic strategies and macroeconomic policies Vietnam's economy is growing rapidly and deeply integrating into the world economy so opportunities and challenges for the process of economic development are increasing, requiring rapid improvement research capacity, macroeconomic forecast in Vietnam In Vietnam, when macroprudential tools are incomplete, the macroprudential policies have not been able to fulfill the role of financial stability, so tools’s monetary policy still play an important role in stabilizing and regulating macroeconomy (Nguyen Duc Trung & Nguyen Hoang Chung, 2018) From that role, this study is based on the new Keynes model for small and opened economies as Vietnam through rational expectations of agents in the economy This approach aims to evaluate the response of important macro variables in monetary policy such as output gap, inflation, exchange rate, and policy interest rate before the shocks of themselves to identify key variables, thereby contributing to improving efficiency in policy analysis In recent years, the construction and application of economic models to forecast the economy in Vietnam has improved significantly in recent years and has made an increasingly larger contribution to the policy - making procedures However, there are not many in - depth studies on the development of Dynamic Stochastic General Equilibrium (DSGE) based on the new Keynesian framework in Viet Nam’s macroeconomic forecast Therefore the DSGE model has a stronger theoretical foundation than traditional models For these reasons, the application of DSGE and SVAR models under the new Keynes framework is becoming more popular at central banks, gradually adding and replacing classical econometric models, especially in central banks which pursued the mechanism of inflation targeting policy 1.2 Setting research issues The thesis finds evidence that the monetary policy tools have a stronger influence than the macroprudential tools in maintaining financial stability in Vietnam Besides, the study confirms again that Vietnam's economic growth depends heavily on credit growth Therefore the effectiveness of monetary policy is shown through controlling credit growth is still mainly so that it is showing the important role of the State Bank in contributing to macroeconomic stability in Vietnam This is an important point for Keynes's theory (Keynes, 1936) to be applied and developed in this research First, the SVAR new Keynes model includes the dynamic IS equation based on the marginal utility optimization representation of the subjects in opened and small economies, the aggregate supply equation (AS) or the new Keynes Phillips curve (NKPC) is based on Calvo's (1983) research on a staggered price model, the Uncovered Interest rate Parity (UIP) and a forward-looking monetary policy rule So, the study simulates the reaction of macroeconomic variables including output gap, inflation, exchange rate and a policy interest rate for four structural shocks - aggregate demand shock, aggregate supply shock, exchange rate shock and monetary policy shock Second, the study of the new Keynes DSGE estimates forecasts for a small and opened economy like Vietnam The model is built and adjusted to be consistent with the forecasting target for macroeconomic variables such as output gap, inflation, policy interest rate, exchange rate fluctuations, and terms of trade 1.3 Objectives of the study 1.3.1 General objectives Testing the suitable of the new Keynes model to confirm the role of monetary policy in macroeconomic stability in Vietnam 1.3.2 Detail objectives First, this study assesses the importance of monetary policy in macroeconomic stability in Vietnam through controlling the credit growth; Second, this study evaluates the relevance of the new Keynes model in explaining macroeconomic fluctuations; Third, this study uses the SVAR new Keynes model to evaluate the response of macro variables to the shocks of themselves; Fourth, this study proposes the DSGE new Keynes model that serves to analyze, forecast and communicate macroeconomic’s policies in Vietnam 1.4 The study’s questions First, whether the effectiveness of monetary policy in Vietnam is reflected through controlling credit growth primarily? Second, why is the new Keynes model suitable for explaining macroeconomic fluctuations in the short term? Thirdly, how does the shock of macro variables affect these variables through the new Keynes model? Fourth, how is the model predict simulate by the DSGE new Keynes model for macro variables of Vietnam? 1.5 Objectives and scope of the study 1.5.1 The study’s objectives 1.5.1.1 The factors of monetary policy and macroprudential policy affect credit growth Factors affecting the credit growth of joint-stock commercial banks (JSBs) in Vietnam during the period 2000 - 2017 1.5.1.1.1 Dependent variables This study uses the TTTD (credit growth - CRD) indicator as the power index for financial stability 1.5.1.1.2 Independent variables The index of credit growth (crd) may reflect the level of risk of banks and these indicators will reflect the ability of sustainable and stable development of the financial system in the future In addition, independent variables representing monetary policy are reserve requirement ratio (rrr) and discount interest rate (dr); The 21 According to Hodge et al (2008), the DSGE model is used to provide information about the parameters for the VAR model The way to this is to simulate data from the DSGE model and combine it with real data when estimating the VAR model A relative weight on DSGE model simulated with real data is called λ (lamda), to control the density of the prior information Then build the prior information for VAR parameters VAR p(Φ,Σu |θ ) 3.3.3.2 Priors for parameters in DSGE model Prior distribution plays an important role in DSGE estimation (An & Schorfheide, 2006) With p(θ) is a priori probability (Prior Beliefs) for parameters in the DSGE model Joint Prior Density of sets of parameters is: p(Φ,Σu,θ) = p(Φ,Σu|θ)*p(θ) 3.3.3.3 Posterior estimation in VAR model According to Hodge et al (2008), the posterior distribution of parameters Φ in the VAR model and VAR and Σu , p(Φ,Σu|Y,θ) The study carried out the posterior distribution simulation for the parameters in the VAR model drawn from the vector set θ the parameters of the experiment in the DSGE model and then sampled from these distributions 3.3.3.4 Selection of  and lag length This study uses the marginal data distribution function to select the lag length of the VAR model, p (del Negro & Schorfheide, 2004) Besides, the important purpose of this study is to select λ and the lag p based on prediction technique out – of - sample (Hodge et al, 2008; Nguyen Duc Trung & Nguyen Hoang Chung, 2017) 3.3.4 Research procedures 3.3.4.1 Procedures estimation for BVAR – DSGE model Step 1: Declare endogenous variables and observed variables in dynare; Step 2: Declare exogenous variables; Step 3: Initialize the parameters of the DSGE new Keynes equations; Step 4: Set up prior information for model parameters; Step 5: The study established the DSGE new Keynes equations which showing the relationship between observed variables and endogenous variables Step 6: Setting initial values for macro variables and observed variables in the model Step 7: Declaring prior information for estimation parameters as well as shocks in the model such as the distribution function, average value and standard deviation Step 8: Declaring lag length as well as a relative weight (λ) to estimate models; Step 9: Making estimates with the parameters which declared and simulated 3.3.4.2 Priors for the DSGE model The prior information for the parameters in the DSGE model is shown in Tables 16.1 and 16.2 of this thesis 22 3.5 Summary chapter Chapter presents the research methods to identify micro and macro factors which affecting the credit credit growth (CRD) through econometric models that handle the bias of the model by FGLS This method uses Stata 12 software to estimate Next, chapter continues to mention the SVAR method to build a new Keynes model to analyze the shocks and assess the impact of macro variables This method uses Eview software to estimate Finally, the DSGE estimation method through Matlab 2016a software, with Dynare 4.1.3 software preinterprets the prior parameters, combines the estimated data as the maximum likelihood function (MLE), Metropolis-Hasting algorithm has estimated the posterior parameters contributing to the construction of a predictive simulation model 23 CHAPTER 4: RESEARCH RESULTS AND DISCUSSION 4.1 Descriptive statistics 4.1.1 Micro variables data Table 4.1: Descriptive statistics variables crd Variables rrr dr car liq ldr gdp cpi Mean 0,3206 0,036 0,067 0,1201 0,236 0,645 0,060 0,081 Max 4,2375 0,08 0,125 0,4589 0,6457 1,084 0,068 0,198 Min -0,3149 0,03 0,0425 0,08 0,048 0,2142 0,0503 0,0063 0,4454 0,015 0,027 0,0697 0,1130 0,165 0,005 0,061 Std error 210 Number of obs 210 210 210 210 210 210 210 Source: The author calculates data from GSO and SBV (2018) 4.1.2 Macro variables data GDP (gdp): 8.00% 7.54% 7.55% 7.50% 7.13% 7.00% 6.90% 6.79% 6.81% 6.98% 6.68% 6.42% 6.50% 6.32% 6.00% 6.24% 6.21% 6.19% 5.98% 5.66% 5.50% 5.42% 5.40% 5.25% 5.00% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Fig 4.1 Vietnam's economic growth for 17 years (2000 – 2017) Source: WB (2018) Inflation – (cpi): 20% 19.87% 18.13% 15% 11.75% 10% 8.28% 7.76% 5% 3.22% -1.71% 0% 6.81% 8.30% 7.39% 3.83% 6.52% 6.04% 4.74% 2.97% -0.43% 0.63% 3.53% -5% Fig 4.2 Vietnam inflation rate in the period 2000 - 2017 Source: Summary and calculation of the author from IMF (2018) 24 Exchange rate: 23,000 22,000 21,000 20,000 19,000 18,000 17,000 16,000 15,000 14,000 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Fig 4.4 The exchange rate movements in the period 2000 – 2017 Source: IMF - IFS (2018) Policy rate: 16.0 14.0 12.0 10.0 8.0 6.0 4.0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Fig 4.5 Policy interest rate movements in the period 2000 - 2017 Source: IMF - IFS (2018) US Inflation and Fed interest rate: 6% 5% 4% 3% 2% 1% 0% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 -1% Lãi suất Lạm phát Fig 4.6 Evolution of the US interest rate and inflation (2000 - 2017) Source: Fed and IMF – IFS (2018) 25 4.2 Empirical analysis 4.2.1 The first model 4.2.1.1 FGLS results and discussions With the dependent variable crd, after using the FGLS method to overcome the defects of the estimation model, the model is significant at the 1% significance level (because of Prob = 0,0000) so the model result Table 4.2 Testing by the least squares method is generally feasible crd Hệ số hồi quy Độ lệch chuẩn z P>|z| Khoảng tin cậy drrr -8,0401 1,1070 -7,26 0,000 (-10,2099; -5,8702) ddr -1,9982 0,6067 -3,29 0,001 (-3,1873; -0,8091) car -0,4755 0,1799 -2,64 0,008 (-0,8283; -0,1227) liq 0,6568 0,1467 4,47 0,000 (0,3691; 0,9445) dldr 0,7271 0,0869 8,36 0,000 (0,5566; 0,8975) dgdp 9,2236 1,9282 4,78 0,000 (5,4443; 13,0028) dcpi 1,1549 0,3129 3,69 0,000 (0,5416; 1,7682) Hệ số chặn 0,1186 0,0410 2,89 0,004 (0,0381; 0,1991) Note: ***, **, * respectively significant at 1%, 5%, 10% Source: Author’s collection data from SGO & SBV and estimation of data by Eview Firstly, the impact of economic growth (GDP), inflation target (CPI) has a positive impact on the credit market (crd) with the significance level of 5% and 10% respectively, this result is consistent with works of Le Thi Man & ctg (2012), Nguyen Minh Sang (2014), Le Tan Phuoc (2016) Secondly, the ratio of compulsory reserve and rediscount rate has a negative impact on the credit market with a statistical significance of 1% Third, the capital adequacy ratio is statistically significant but the impact is negligible (-0.4735