Monetary Policy Transmission Under Conditions of Financial Integration

(𝐸𝑋+𝐼𝑀) 𝑖𝑡

After measuring each index of the Impossible Trinity in Vietnam, the study continues to examine the change in Vietnam's CSTT independence separately by measuring the coefficient 𝛽 according to the following equation:

∆𝑖𝑑 𝑡 = 𝛼 + 𝛽∆𝑖𝑓 𝑡 + 𝜃(𝑖𝑑 𝑡−1 − 𝛾𝑖𝑓 𝑡−1 ) + 𝜇 𝑡 (2.15)

The CSTT independence in the study is measured by the response of the 3-month interest rate of Vietnam (id) and the US (if). In addition, the study also considers the change in HNTC with a separate equation by regressing the size of foreign direct investment inflows (cap) on the domestic interest rate (id), the US interest rate (if) and the level of exchange rate depreciation (dep) according to the equation:

𝑐𝑎𝑝 𝑡 = 𝛼 + 𝛽 1 𝑖𝑑 𝑡 + 𝛽 2 𝑖𝑓 𝑡 + 𝛽 3 𝑑𝑒𝑝 𝑡 + 𝜀 (2.16)

Using the autoregressive distributed lag (ARDL) model, the research results show that in the period 2001 - 2006, Vietnam's monetary policy independence was quite low and there was a significant improvement in the period 2009 - 2015. Vietnam's exchange rate stability throughout the period was quite high. In general, Vietnam's level of monetary policy independence in the post-crisis period was higher than before the crisis. The regression coefficient of Vietnam's interest rate on the US interest rate is not statistically significant in the short and long term, showing that Vietnam still maintained monetary policy independence in the research period. Synthesizing by pair of indexes, exchange rate stability and monetary policy independence in the period 2001 - 2005 were very high, which also means that Vietnam chose exchange rate stability and monetary policy independence as priorities in this period. However, in the period 2009 - 2015, the pair of indexes of monetary policy independence and exchange rate stability had higher values.

Nguyen Kim Thai Ngoc & Le Phuong (2020) tested the Impossible Trinity theory for 7 developing countries in Southeast Asia including Cambodia, Myanmar, Indonesia, Malaysia, Philippines, Thailand and Vietnam in general without evaluating Vietnam specifically. The OLS method was used for the period 2000 - 2017 with annual frequency. The study used similar measurement indicators to Aizenman et al. (2008). The research results confirmed that, for the countries in the observation sample, the three policy variables are linearly correlated with each other, or in other words, there is a trade-off between these three goals as the constraints in the Impossible Trinity theory.

In general, through a review of previous studies on measuring the impact of HNTC on CSTT independence, it is found that:

First , this relationship is often approached indirectly by testing the existence of the Impossible Trinity, that is, whether the sum of the coefficients measuring the three variables corresponding to the three policies is equal to a constant or not, this constant is usually 1 or 2. This approach only shows whether there is a trade-off when increasing the proportion of one or two of these variables with a decrease in the remaining two or one variables or not, however, it does not indicate whether HNTC or other variables have a positive or negative impact, and to what extent, on monetary policy independence. Foreign exchange reserves in studies approaching this direction are not considered independently but are often associated with different combinations of the Impossible Trinity to consider the impact on the ultimate goal of monetary policy, so the role of monetary policy independence cannot be determined separately. There are few studies that investigate the direct impact of HNTC on monetary policy independence, detailing the level and direction of impact as well as assessing the role of foreign exchange reserves on the country's monetary policy independence.

Second , the variables used in the model also have many differences. In addition to the method of calculating the variables similar to Aizenman & ctg (2008), many studies when measuring the CSTT independence of countries have used quantitative models to determine the system.

The coefficient shows the degree of response of domestic interest rate to the central country interest rate, the higher the estimated coefficient shows the lower the monetary policy independence.

For the HNTC variable, many studies use measurement methods based on legal regulations such as the KAOPEN index or similar recalculations, however, recently the method based on measuring actual HNTC is more commonly used.

The level of exchange rate stability is also calculated based on actual exchange rate developments in the market or determined according to the reported exchange rate mechanism.

As for the foreign exchange reserve variable, it is measured relatively consistently, calculated by the size of foreign exchange reserves (cumulative value) minus gold as a percentage of GDP, however, many studies do not include the foreign exchange reserve variable in the model.

Third , the commonly used estimation method is OLS regression in indirect approach studies. However, when measuring direct impacts, studies use more diverse methods and pay more attention to the nature of time series data, most notably Taguchi & ctg (2011) using GMM method and taking differences to ensure the stationarity of the data series, besides Klein & Shambaugh (2013) using panel data with fixed effects or Law & ctg (2019) considering the threshold level of HNTC affecting the measurement results.

Fourth , the results of indirect studies generally find that there is a trade-off when combining the Policy Triad, increasing the proportion of one or two variables will reduce the proportion of the other two or one variables. With the direct approach, it clearly shows the impact of HNTC on monetary policy independence but the results are not consistent. Taguchi & ctg (2011) found that HNTC has a negative impact on monetary policy independence in the countries in the research sample, but Law & ctg (2019) only found a statistically significant opposite impact when HNTC is above a certain threshold, when HNTC is below this threshold, the impact is not statistically significant. Foreign exchange reserves are also found to play a supporting role for monetary policy independence in the study of Taguchi & ctg (2011) but only found a statistically significant impact when HNTC is above the threshold.

with the study of Law & ctg (2019) and the negative impact in the study of Glick & Huchison (2009).

Fifth , in Vietnam, there has been no research measuring the direct impact of HNTC on monetary policy independence for the Vietnamese economy. This relationship is mainly considered through the trade-off between the Impossible Trinity policy variables. Monetary policy independence and exchange rate stability are often measured similarly to Aizenman & ctg (2008), while Pham Thi Tuyet Trinh (2010) uses a regression model to assess the response of Vietnamese interest rates to US interest rates to measure domestic monetary policy independence. HNTC mainly uses measures based on legal regulations, Ho & Ho (2018) measures based on actual results but uses the calculation method according to Ito & Kawai (2012), this calculation method does not show the total size of assets and liabilities of financial flows compared to the size of the economy but considers the aggregate with trade openness. At the same time, studies have not considered the role of foreign exchange reserves in this relationship.

Studies often use the OLS method to measure the linear combination of the Impossible Trinity, however, the data series characteristics corresponding to the three policy variables can be stationary at different levels, so Ho & Ho (2018) performed a stationarity test of the series and determined the appropriate regression model to be the ARDL model.

Most research results show that Vietnam's HNTC is increasing along with the trend of financial globalization, however Pham (2016) believes that Vietnam is prioritizing the goal of capital control. Exchange rates are assessed in different directions, Le Phan Dieu Thao (2010), Pham Thi Tuyet Trinh (2010) and Pham (2016) found that the VND and USD exchange rates are tightly controlled or stable at a high level, while Dinh Thi Thu Hong (2012), To Trung Thanh (2012) found that the exchange rate has become more flexible. Vietnam's monetary policy independence is a concern for Le Phan Thi Dieu Thao (2010) or found that the results of monetary policy independence are not completely consistent with the research of Pham Thi Tuyet Trinh (2010), but Dinh Thi Thu Hong (2012), To Trung Thanh

(2012) found that the independence of monetary policy has reached a moderate level. Ho & Ho (2018) found that the independence of monetary policy in Vietnam has improved significantly.

Research on the Vietnamese economy has been conducted in many different periods, some studies were conducted before the global financial crisis and many studies extended after this period, but the most recent studies are Ho & Ho (2018) or Nguyen Thai Kim Ngoc & Le Phuong (2020) which only access data up to 2015 or 2017.

2.5.2 Monetary policy transmission under conditions of financial integration

Foreign research

Research on the transmission of monetary policy under the influence of HNTC in recent times, especially since the global financial crisis in 2008-2009, has found that a number of external shocks dominate the transmission process from domestic monetary policy to financial conditions, and thereby affect the ultimate goal of monetary policy. External shocks have been detected through many studies, including interest rates of central countries, especially US interest rates, quantitative easing policies, announcements of changes in monetary policy of central countries and global risks.

Interest rates of central countries, especially US interest rates, are a factor that is recognized to have a great influence on domestic financial conditions of many other countries. One factor that has also been observed and recognized to play an important role, especially in developing and emerging countries since the global financial crisis of 2008-2009, is global risk. This risk is often measured by the volatility index of options prices on the Chicago Board Options Exchange (Volatility index - VIX). In many studies since the crisis, VIX has been considered together with US interest rates as external variables affecting domestic variables while previous studies have rarely mentioned it.

Pradhan & ctg (2011) measured the factors affecting domestic bond interest rates of emerging market economies, including Brazil, Indonesia, Korea, Malaysia, Mexico, Finland, Thailand and Turkey during the period 2000-2010.

It can be seen that the risk premium measured by the credit default swap (CDS) spread or by the global risk (VIX) has a positive impact coefficient, with a 1 percentage point increase increasing the domestic bond yield by 4 percentage points.

The study uses 10-year government bond interest rate data to represent the long-term interest rate variable, while for Türkiye, 5-year interest rate data is used instead, and policy interest rates are used for the short-term interest rate variable. In addition, the research model also includes GDP growth and inflation variables to control for other factors that also affect long-term interest rates. In particular, industrial production growth data is used to replace GDP growth because the study is conducted with monthly frequency. Panel data regression models are used for this study. In addition to the main results mentioned above, the study also found that periods of higher capital inflows significantly reduced bond interest rates, but domestic policy interest rates still had an effective impact on long-term interest rates. In addition, inflation expectations and GDP growth also had a statistically significant positive impact on long-term bond interest rates in these countries.

Rey (2013) in a broader study on the issue of global financial cycles and CSTT independence with a large sample of countries from the US, Canada, Latin America, Europe, Asia, South Africa from 1990 to 2012, with quarterly frequency, found that changes in the FED real interest rate led to an increase in VIX after 5 quarters and lasting up to 11 quarters. In addition, a decrease in VIX increased cross-border credit flows lasting up to 6 quarters, an increase in VIX reduced the FED real interest rate, and an increase in the FED interest rate reduced total credit flows after 12 quarters. Stock prices were also found to be negatively correlated with VIX. At the same time, through examining the correlation between the increase/decrease in financial flows and VIX, the study found that flows other than FDI were highly and negatively correlated with VIX. Indirect credit and debt flows increase as the VIX declines, but debt flows to Asian countries are positively correlated with the VIX. FDI flows are positively correlated with the VIX in all regions in the sample, meaning that as global risk increases, capital flows

FDI inflows also increased. The correlation was only found for gross flows and not for net flows. Panel regressions with fixed effects and vector autoregressive (VAR) models were used to empirically test these relationships.

Jinjarak (2014) measures the impact of financial globalization on international capital turnover over the period 1989-2006 with monthly frequency data by estimating the response of capital prices to oil prices and US interest rates, and also identifies the factors driving this response. The study uses the OLS method to estimate an international asset pricing model to measure the response of excess international capital turnover to oil prices and interest rates. The vector autoregressive (VAR) model is then used to identify the factors driving the response of excess international capital turnover, the two factors of interest are trade integration and HNTC. The results show that trade integration increases the response of international capital turnover to oil prices, while HNTC increases the response to the Fed interest rate. A 1% increase in HNTC increased the response of capital revenues to changes in the Fed funds rate at 10.90%.

Jain-Chandra & Unsal (2014) evaluate the effectiveness of monetary policy transmission under the impact of international capital flows in Asian emerging economies during the period 2000-2010 with monthly data frequency. Using SVAR model to estimate the response of domestic long-term and short-term interest rates to US interest rates, in which the vector of endogenous variables in addition to the two key variables of domestic short-term and long-term interest rates also includes changes in exchange rates, GDP growth and inflation. Exogenous variables include US interest rates, VIX and world demand. The study uses 3-month interest rates to represent domestic short-term policy rates and 10-year government bond interest rates to represent domestic long-term interest rates. In addition, the study also examines the response of domestic interest rates to the average world interest rate, calculated by averaging interest rates of the US, the European Union and Japan.