“Impact of the US unconventional monetary policy in times of crises on Vietnam”

The study investigates the effects of the US unconventional monetary policy on the Vietnamese economy, focusing on the output and price index. Besides, the differences in the response of these indicators to the shock related to non-traditional tools implemented by the US in the Global Financial Crisis and during the pandemic crisis are also explored. By applying the Structural Vector Autoregressive model on the monthly dataset from January 2007 to April 2022, the results show that an expansionary monetary policy of the US by unconventional measures leads to a rise in the output and price index of Vietnam. In addition, risk-taking channels and portfolio rebalancing channels are important in the transmission of the US monetary policy to Vietnam, they cause booms in asset prices via the surge of capital flows. Moreover, another important finding is that the impact of US UMP on the Vietnamese economy during the Global Financial Crisis was generally two times higher than during the epidemic crisis.


INTRODUCTION
Under normal economic conditions, monetary policy is implemented by altering short-term interest rates by Central Banks to attain their economic goals. However, given recent financial and debt crises, and the pandemic's devastating effects on the global economy, conventional monetary policy instruments have not had the desired impact on financial stability. To get around the zero lower bounds on short-term interest rates, the US, and other advanced economies (AEs) have also implemented new measures, which are called unconventional monetary policy (UMP). The implementation of a huge number of nonstandard monetary policies during the global financial crisis (GFC) has resulted in significant changes in financial conditions and the real economy in both AEs and the rest of the world, especially when it comes to emerging economy markets (EMEs) and developing countries. For instance, a range of large-scale asset purchase programs in AEs have lowered the bond yield, raised real GDP and stock price, and increased the CPI. It also encouraged an increase in capital flows to EMEs, resulting in a rise in real output growth and positive response from financial markets such as the growth of real equity return, etc. The instability in global finance has not stopped since the 2008 financial crisis. The outbreak of the global pandemic crisis since COVID-19 started at the end of 2019 has brought a major shock to economies around the world. Asian economies are affected by the GFC and the non-traditional monetary policy of AEs, especially the United States in different ways (Punzi & Chantapacdepong, 2019). Developing coun-tries are also deeply affected by economic uncertainty, especially when it comes to sudden monetary policy adjustments from AEs. New US monetary policy operations are always an important signal for the developing countries' Central Banks to adjust their monetary policy implementation, especially in the GFC and during the pandemic crisis. Consequently, different conducts of such policy have raised a big concern about whether it has international spillover effects on real economies in developing countries and whether there are differences in the response of the real economy of these countries to the shock of the US UMP in the GFC and during an epidemic caused by COVID-19.

LITERATURE REVIEW
There are many empirical studies investigated the response of the economy to the shock of UMP came from AEs, especially focusing on EMEs  (Fic, 2013;Kiendrebeogo, 2016). They mainly concentrated on the real economy and financial market in these countries, and these findings emphasize that the expansion of AEs' UMP had a significant impact on EMEs, led to an increase in portfolio flows accompanied by the increase in real output and a range of movement in financial variables, such as the increase in the stock market or the appreciation of local exchange rate. Some studies explored the change in return of real estate (Gupta & Marfatia, 2018); other studies proposed the significant effects of one of the typical instruments of UMP in AEs such as QE on capital flows (Alper et al., 2020;Kiendrebeogo, 2016;Tillmann, 2016), the results emphasized that these policies in AEs boost capital flows to EMEs and developing countries. The implementation of UMP led to a significant increase in portfolio flows and cross-border credit flows in these countries. The other consideration in analyzing the effect of UMP in a more specific scope is focused on the Asia and Pacific region, however, a number of studies are quietly limited, they paid less attention to this area, although Asia is one of the markets that play a key role for the international investors and the industrialized countries. Most of them are in a working paper (Ganelli &  number of studies in this area, empirical results also demonstrated that countries in Asia and the Pacific region have strong responses to the shock of UMP in AEs, especially the US, when it comes to economic activities and financial variables. The expansion on a large scale of UMP in AEs led to pushing capital inflows to this area, resulting in an increase in the stock market and the appreciation of local currencies; secondly, it also had a positive effect on the real output of these countries. However, some studies seem to be different, finding evidence of this effect to a rather small extent. More specifically, some studies examine the impact of unconventional monetary policy on Asian countries. For example, Rafiq (2015) explored the effect of the US's QE on a group of countries in Asia. He found that this spillover effect was likely to be small, the cause of this different result may come from the group of countries belonging to the frontier economies in Asia. Lee and Choi (2010) also focused on the analysis of the Asia region. They evaluated the impact of QE originating from the US on return and volatility in selected Asian stock markets. They discovered that the QE policies of the United States have had a major impact on the correlations between the United States and several Asian nations, with a considerable progressive reduction in correlations during the most recent QE. Besides, they also found that greater stock market liquidity significantly enhances financial spillovers. Tran and Pham (2020) track the monthly reactions of Asian emerging market equities prices, long-term interest rates, and currency rates to US UMP. The first finding is that UMP shocks from the United States relate to an increase in equities prices, a decrease in interest rates, and an appreciation of the currency in Asian emerging economies. In contrast, traditional monetary policy shocks from the United States appear to have a negative impact on these recipient nations. These empirical findings showed that policymakers in Asian developing nations should be wary about the spillover effects of US UMP if it is implemented.
Moreover, the crisis-related factor also plays an important role in the study of non-conventional monetary policy. Because the starting point of the implementation of these policies is the crisis. However, most of the studies presented seemed to raise concerns about the Global Financial Crisis (GFC) but forgot that a pandemic crisis caused by COVID-19 is currently existing. Indeed, they paid little attention to examining the different effects of US UMP on other countries during two different crises: the GFC period and the pandemic crisis, even though COVID-19 is one of the crucial factors changing the monetary policy in AEs, and it is the biggest threat to the global economy. Few studies mentioned a little about two crises. The study by Cortes et al. (2022) only proposed that the outcomes of the COVID-19 interventions suggest a positive spillover of Fed policy for AEs and EMEs groups. The policy announcements made by the Fed during COVID-19 had pronounced effects on the short-term maturities of EMEs. Besides, when focusing on the equity markets of individual nations, the study also discovered particular spillovers for catastrophe risk that are primarily concentrated in Asia (e.g., South Korea, Taiwan, and Malaysia). Another study examined the effects of UMP in the midst of the COVID-19 outbreak-related global financial turmoil (Rebucci et al., 2022). They found the important results that when compared to previous interventions made during the Great Recession, the effect of QE announcements during the pandemic crisis on the EMEs was significantly greater although they only focused the analysis on government bond yields and exchange rates. In summary, some studies showed that EMEs are more sensitive to large-scale interventions (the COVID-19 QE announcements) whether they are domestic or international in nature.
The analysis of monetary policy for researchers and policymakers has been supported by the vector autoregression approach platform introduced by Sims (1980). VAR models are the most widely used approach for studying the consequences of traditional monetary policy. As a result, the international impact of UMP on other countries is usually analyzed by applying VAR models (Anaya et al., 2017; Bowman et al., 2015; Sugimoto & Matsuki, 2019). However, the greatest flaw in VAR is that it does not distinguish between independent and dependent structural shocks when assessing how variables respond to independent structural shocks. Therefore, Bernanke et al. (2004) suggested using a structural VAR model to address the issue because the impulse response function and variance decomposition for the studied variables might not be correct in the VAR model. This method allows the analysis of the effect of official announcements or speeches by the government related to non-standard monetary policy on the variables concerned. However, although it is significantly useful in applying the event study method, evidence showed that it also has several limitations. Variables expressing non-standard monetary policy like QE or forward guidance are likely to be endogenous and they cannot easily be modeled as dummy variables (Meinusch & Tillmann, 2016). Besides, the effect of UMP shock on other variables cannot be observed through the impulse response functions. Finally, because event research often uses high-frequency data (daily, intra-daily), it is suitable for financial market indicators such as interest rates, the yield curve, or asset prices. Therefore, the SVAR model is still the most suitable model when investigating the impact of UMP on the macroeconomics of developing countries.
The purpose of the study is to investigate the international spillover effects of the US unconventional monetary policy (US UMP) on the real economy in Vietnam, which is a small, open, and frontier market in Asia. More specifically, the study examines the response of the output and price index of Vietnam to US UMP shocks. Besides, this study also investigates whether there exist likely differences in the effects of US UMP on Vietnam during the global financial crisis and during the pandem-ic crisis. A structural VAR model will be used to study the reactions of the Vietnamese real economy to US UMP shocks based on the literature.

METHODOLOGY
This study employs the SVAR model to analyze the international spillover effects of US UMP shocks on the real economy of Vietnam based on the approach of (Carrera & Ramírez-Rondán, 2020; Yildirim & Ivrendi, 2021) with relied on the pioneering research of Cushman and Zha (1997), which introduced the small-open economy assumption to the model. The benchmark SVAR model has the following representation: where Y t is a (n x 1) vector of n endogenous variables at time t; A(L) is a (n x n) matrix of lag polynomials that captures the dynamics of the endogenous variables; B(L) is a (n x j) matrix of lag polynomials that captures the contemporaneous effects of the exogenous variables on the endogenous variables; Z t is a (j x 1) vector of exogenous variables at time t; t is a (n x 1) vector of structural error terms.
The selection and order of variables for the model are bases on the international transmission mechanism of US UMP and research of Yildirim and Ivrendi (2021) To examine the effects of US UMP, this study uses an eight-variable SVAR model, with an indicator of US UMP and variables relating to domestic factors. It can be distinguished into two groups as follows: • Group 1: The indicator of US UMP and international transmission channels to Vietnam.
• Group 2: Domestic variable. In this group, five variables from the Vietnamese economy will be displayed, including an index of equity prices, a measure of market interest rate, an exchange rate, and a measure of real economy such as the output and inflation.
The following describes the vector of endogenous variables: In which UMP (a proxy for the unconventional monetary policy of the US), VIX (VIX index), CF (capital inflow to Vietnam), SP (stock price), IRR (interest rate), FOREX (exchange rate), GDP (gross domestic product), CPI (price index). Based on the Cholesky decomposition, this paper proposes a short-term recursive restriction on contemporaneous structural parameters, allowing suitable economic structures to be generated. Specifically, a recursive restriction assumes that one shock affects the system at time t, followed by another shock at time t + 1, and so on. It notes that the recursive matrix needs the endogenous variables to be ordered in the estimate Y t . Accordingly, the assumption in our model is that the US UMP has significant effects on macro-financial conditions in Vietnam while the impact of domestic economic conditions on global conditions is negligible.

RESULTS
The paper will present the main results of the study from the SVAR model. A robustness test will then be performed to further strengthen the study results.  The paper selects lag-length models based on the LR, FPE, and AIC criteria. The standard information test suggests adopting an 8-lag length for models. The estimated results of the stability proposed that the lags are completely suitable for the model.

Empirical analysis from the SVAR model
Using a 12-month restriction horizon, the responses of Vietnamese output and price index to US UMP shocks are presented in Figure 1. As shown in this figure, a shock to the total asset of the Fed balance sheet increases Vietnam's output and price index significantly although only the response of GDP is significant for the ninth month. More specifically, an increase in CBB leads to a positive and statistically significant rise in GDP in the ninth month after the shock, reaching its peak after around ten months at an increase of 4.2 pp. The reaction of CPI to a shock of US UMP is also close to GDP's response when a positive shock to US UMP leads to a stable upward effect on CPI of Vietnam in the short run, peaks at a 0.3 pp increase, and remains for around one year, however, this response is not statistically significant.
The VIX index also has a significant impact on the GDP and CPI of Vietnam ( Figure 2). The results show a negative response of GDP and CPI to a shock of VIX. An increase in the VIX index reflects rising risk aversion, that is, decrease risk appetite, and increases uncertainty in the financial markets. This resulted in a rapid go down of GDP after shock, reaching its peak after five months at a decrease of 3.1 pp. The response of CPI is similarly negative, peaking at a 0.26 pp decrease. The effect on CPI variables becomes significant from the beginning of the second month until the third month.   Although the responses of GDP and CPI to CF shock are not statistically significant, foreign portfolio flow is also having a relatively similar impact on the GDP and CPI of Vietnam to CBB shock ( Figure 3). However, the surge in inflows leads to a significant increase in equity returns in Vietnam. The immediate response of stock prices in Vietnam is significantly positive for a period of four months, reaching its peak after four months at a rise of 2.6 pp.
To compare the response of GDP and CPI, as well as the level of impact on these variables to the US UMP shock in different crisis periods, the study divides the sample into two periods: the GFC period (January 2008 to December 2014) and the pandemic crisis period (December 2019 to April 2022). Which, the exogenous variable is the number of monthly new confirmed cases of COVID-19 used in the phase of the pandemic crisis in order to clarify the impacts of the epidemic. Figure 4 presents the impact of US UMP shock on GDP in two different periods. The result shows that the response of GDP to the US UMP shock was stronger in the GFC than during the COVID-19 period. The magnitude of GDP increase for US expansionary monetary policy is larger over time, reaching 18.23 pp one year after the shock. Meanwhile, during the COVID-19 pandemic, the response of GDP to the Fed's balance sheet expansion shock is nearly half lower than during the GFC period, at 7.04 pp one year after the shock occurred.
Similarly, the effect of US UMP on CPI in Figure  5 also shows a difference between the two periods, but the response starts to differ markedly from the fourth month onwards. Specifically, during the GFC period, CPI gradually increased to the US UMP shock, from 0.18 pp in April to 1.44 pp in December. Meanwhile, during the epidemic crisis, CPI only increased slightly to the third month after the shock, at 0.17 pp, then the response gradually decreased and stayed close to zero in December.

Robustness test
To check the robustness of the models, Wu and Xia's shadow short rate for the US is collected to replace the total assets of FED's balance sheet in proxy for US UMP. Wu and Xia's Shadow Short Response of CPI to CBB shock COVID-19 period GFC period Rate (SSR) is an estimate of the short-term interest rate that is not observable in the market. It is derived from the model that combines information from the interest rate term structure and macroeconomic variables. The SSR introduced by Wu and Xia (2016) is based on a so-called shadow-rate model that describes the relationship between the observable yields on longer-term bonds and the unobservable short-term rate that the Central Bank would like to set if the zero lower bounds were not a restriction (i.e., the lowest possible interest rate). When the policy rate is at or close to the zero lower limits, the SSR can be used to examine how monetary policy affects the overall state of the economy. The SSR, according to Wu and Xia (2016), provides a more accurate reflection of the stance of monetary policy compared to the real policy rate, which might not adequately account for the impacts of unconventional monetary policy measures like quantitative easing. Thus, SSR is one of the most effective measures besides the total assets of the Fed's balance sheet to investigate the impacts of US UMP on other countries. It can be understood that an expansionary monetary policy of the Fed via unconventional instruments will be shown by a decrease in SSR, and vice ver-sa, so the explanation for the results using the SSR variable will be explained in reverse compared to that of the using the variable CBB.
Although the response of GDP to the shock of SSR and VIX is not statistically significant ( Figure  6), it still reflects the negative impact of SSR and VIX on GDP. Specifically, when there is an increase in shock caused by SSR and VIX, GDP will decrease. Capital inflows have positive effects on the output of Vietnam. GDP is raised after the CF shock, peaks at an 8.7 pp increase after around six months, then it gradually decreases and tends to turn to baseline after one year. Figure 7 shows the impulse response of the price index to different shocks. The study finds a negative response of the CPI variable to a shock to shadow short rate and risk aversion, and a positive response of CPI to capital inflows shock. These results support more evidence of the research results when using the variable CBB proxy for the unconventional monetary policy of the US.
To summarize, both results in which US UMP is proxied by total assets of the Fed's balance sheet

DISCUSSION
The SVAR impulse response results for the impact of US UMP shock on Vietnamese output and price index reveal a positive relationship between these variables, although it is not statistically significant in CPI response. The results are consistent with those of Anaya et al. (2017) and Punzi and Chantapacdepong (2019). This implies that in the short run, the implementation of UMP in the United States can lead to an increase in investment flows into emerging markets such as Vietnam, as investors search for higher returns. This increase in investment, especially portfolio investment on the stock market can also contribute to an increase in GDP in Vietnam. The results are also supported by the response of stock prices to a shock of capital flows. This suggests that the expansion of unconventional monetary policy in the US led to a surge in capital flows to Vietnam, resulting in a temporary upward effect on the SP variable. These results suggest the importance of capital flows as a channel for the transmission of US monetary shocks (Rey, 2016). In addition, the findings also demonstrate that GDP and CPI react negatively to a VIX shock. These results are consistent with the literature on risk-taking channels Yildirim and Ivrendi (2021) and the changes in risk aversion to output and price index (Tillmann, 2017). This emphasizes the importance of risk-taking channels in the international transmission of the US UMP to Vietnam.
Finally, another important finding is that the shock brought on by the FED's balance sheet expansion during the global financial crisis typically had a more negative effect on the Vietnamese economy than it did during the pandemic crisis. It can be explained as follows. Firstly, the GFC was a more severe and prolonged crisis than the COVID-19 crisis. Moreover, the GFC led to a significant contraction in global economic activity,

CONCLUSION
This paper applies the SVAR model to examine the international spillover effects of US UMP on the Vietnamese output and price index between January 2007 and April 2022. The study also compares the differences in the response of these variables to US UMP in the GFC period and during the pandemic crisis caused by COVID-19. The results indicate that the unconventional measures of the US lead to an increase in Vietnam's output and price index. In addition, Vietnam's GDP and CPI are also significantly affected by risk aversion and capital flows. This shows the significance of the risk-taking channels and portfolio rebalancing channels in the international transmission of US UMP in Vietnam. Finally, the impact of the shock caused by the expansion of the FED's balance sheet on the Vietnamese economy was generally two times higher than during the epidemic crisis. The robustness check makes these above results stronger. By using Wu and Xia's shadow short rate for the US, which is proxied for US UMP, the results of the response of GDP and CPI to the UMP shock are consistent with the findings, which are estimated by using the total assets of the Fed's balance sheet. These findings imply the importance of closely monitoring and responding to global economic developments when making domestic policy decisions. As a result, the study recommends that the advantages of an expansive monetary policy must be carefully weighed against the possible hazards of inflation and exchange rate instability in Vietnam, while also working to promote long-term economic growth and stability.