Impact of Macroeconomic Variables on Stock Price Index: Evidence from Vietnam Stock Market

Vietnam stock market after nearly 20 years of operation has undergone many ups and downs. In addition to the contribution to the development of the national economy, the Vietnam Stock market still has some certain shortcomings. Goods on the market have increased in size, quantity, and categories but not much of diversity. The potential risks due to many different causes, always hurt investors. Therefore, the study of Macroeconomic factors affecting the volatility of stock indexes has been paid special attention by society. Research results show that, in the long term, the VNI stock price index is affected by the oil price, money supply, interest rate, SJC gold price, and exchange rate. In the short term, the stock price index has a causal relationship with oil prices, interest rates, and the M2 money supply. When there are impacts of market shocks, short-term corrections affect extremely slow to the long-term balance.


Introduction
In the structure of the financial market, the stock market plays an essential role in providing an additional channel of long-term capital mobilization for the economy (Mishkin, 2004). People often compare securities price indexes such as mirrors to reflect the market health, through which we partly understand the business performance of corporations. The stock price index is a quite sensitive indicator in the volatile macroeconomic environment. In general, many different causes affect the volatility of the stock price index. In particular, the analysis of macroeconomic factors is an essential requirement of the primary analysis method when assessing Vietnam's stock market in the current complicated period.
Vietnam stock market was born after several decades of Asian countries and after hundreds of years of developed countries market. However, after nearly 20 years since its inception (commencing operations on July 20, 2000), the Vietnam Stock market has gone through many ups and downs. Although there are many different indicators such as VN index, HNX index, VN index -30, Upcom index, Baoviet index, etc., the VN index is always considered the most critical and typical stock price index when evaluating the overall volatility of Vietnam's stock market. At one point, VNI had a fantastic breakthrough, hitting 1,170.67 points (March 12, 2007), partly reflecting the overheating of the market. Besides contributions to the development of the national economy, the Vietnam stock market still has certain limitations. Commodities on the Vietnam stock market have increased in size, quantity, and types but not much of diversity. There are only five trading sessions a week, two times a day. Besides, the potential risks due to many different causes on the stock market always hurt investors. Therefore, the study of factors affecting the volatility of the stock index has always been of particular interest to investors and society. The stock market factors found in empirical studies have been published to include various factors stemming from macroeconomic factors and from the company itself. The study mainly focused on reviewing and analyzing macroeconomic factors affecting the stock market.
Macroeconomic Variables are indicators that signal the current trends in the economy, like, the gross domestic product, rate of employment, consumer price index, the balance of payments, broad money supply, lending interest rate, household consumption, etc. stock market comprehensively during the past ten years after the global economic crisis to date. At the same time, the article further considers the relationship between the domestic gold price (SJC), world oil price, consumer price index, and money supply to the Vietnam stock market.

Data and methodology 3.1 Data description
The study focused on analyzing six factors affecting VNI, namely: SJC gold price, world oil price, M2 money supply, interest rate, exchange rate, and consumer price index. Research data were collected over ten years from January 2009 to December 2018. -VNI stock price index: VNI was the first stock index on HOSE in July 2000 and is considered one of the most representative indicators for the whole market. With the capitalization scale, operating time, several listed companies, as well as a large number of investors, participating, so VNI is more representative than some other indicators like HNX index, VNI 30, SSI index, BaoViet index, etc. The stock price index Variable is taken in units of points and is the closing price of the last trading day of each month. Data collected from the website of the State Securities Commission. Most companies with stocks listed on HOSE are companies with large capital sources, healthy financial situation, effective and transparent business operations, and disclosure of information according to regulations. Periodically, HOSE will conduct delisting procedures for certain companies on the floor. These are the indications that VNI has a good ability to represent the Vietnam stock market. Data of VNI Variables collected from Ho Chi Minh City Stock Exchange. https://www.hsx.vn/ -Gold price Variable: Like all precious metals, gold is calculated by troy weight and in grams. When it combined with other metals, the term kara or karat is used to indicate the amount of gold present, with 24 karas being pure gold and lower ratios represented by lower numbers. A decimal number can also express the purity of a gold coin or gold coin from 0 to 1, called the golden age in parts per thousand, as 0.995 is very pure. Gold prices determined by trading in gold and derivative markets, but a process called Gold valuation in London, which began in September 1919 provides a benchmark for the industry. Later on, the same valuation was introduced in 1968 to provide gold prices when US markets opened. The topic of consideration on the Vietnam stock market and gold price collected according to the selling price of 1tael SJC gold, the unit of millions of dong calculated on the last trading day of each month. Gold price data collected from the Saigon Liberation Online website. http://www.sggp.org.vn/ -Oil prices: The oil industry divides "crude oil" according to the area in which it originates (e.g. "West Texas Intermediate" (WTI) or "Brent"), usually by weight and its relative viscosity (light, average or heavy). Crude oil also classified as "sweet" or "sour" depending on the level of sulfur. Sweet crude oil contains 0.5% sulfur or less, and sour crude contains about 1% sulfur and takes more steps to produce it according to current parameters. In general, the heavier the crude oil, the greater it's sulfur content. Excess sulfur is removed from crude oil during refining because sulfur oxide discharged into the air during oil combustion is a major contaminant. The barrel is the unit used to measure the volume of crude oil in trade. One can hold 42 US gallons~158.9873 liters, 07 barrels (barrels) of oil~1.113 tons, 1 gallon in the US~3.785cm 3~3 .785 liters. Collected data on world oil prices collected by the average monthly value of WTI crude oil traded on the NYSE exchange. (https://www.investing.com/currencies/wti-usd-historical-data).

-Money supply M2:
The money supply is the amount of money provided to satisfy the payment and reserve needs of entities in the economy. The money supply must correspond to the demand for money. The shortage or surplus of money supply against demand can all lead to negative impacts on the economy. Since the level of financial market development varies from country to country and changes over time, the composition of the measurements mentioned above maybe covariance across countries and change over time. For example, The US Federal Reserve (Fed) also added traveler's checks and other forms of deposits that can be issued to other checks on M1 and M2 measurements they split term deposits into term deposits. In small quantities, leaving M2 with long term deposits then moving to M3 and adding M2 money market deposit accounts, shares of money market mutual funds (not organized). Vietnam added M2 commercial bank bills. The selection of a precise measurement of the actual amount of money will base on the measurement of the amount of money that helps to predict the best economic Variables that the currency has as much impact as the inflation rate, business cycle. Countries often use M1 or M2 and Vietnam measurements using M2 measurement. Money supply data collected from the Vietnam Securities website. (https://finance.vietstock.vn/).

-Interest rate Variable:
The rate by which the borrower pays for the money they borrow from a lender. Interest rate targets are an important tool of monetary policy and are taken into account when dealing with Variables such as investment, consumer price index, and unemployment. Central banks or reserve banks of countries, in general, tend to reduce interest rates when they want to increase investment and consumption in the country's economy. However, a low-interest rate as a macroeconomic policy can be risky and can lead to the creation of an economic bubble in which a large amount of investment poured into the real estate and stock market. The data on the interest rate used by the project calculated according to the percentage of refinancing interest rates prescribed by the State Bank and published on the State Bank Website (https://www.sbv.gov.vn/).

-Exchange rate Variable:
The exchange rate is the rate at which a currency will exchange for another currency. It is also considered the currency price of a country expressed by another currency. Comparing to other currencies, the USD most commonly used and used by most businesses in international payment contracts. Exchange rate data collected from the website of the State Bank, taken from the exchange rate between USD and VND, this Variable represents the foreign exchange market. (https://www.sbv.gov.vn/).

-Consumer price index Variable:
This indicator is used to reflect the level of inflation, i.e., the increase over time of the general price level of the economy. In an economy, inflation is a loss of market value or a decrease in the purchasing power of the currency. When compared to other economies, inflation is the currency devaluation of a currency compared to other currencies. Usually, in the first case, it is understood that the inflation of the monetary unit within the economy of a country, in the second case, it is understood as inflation of a currency in the global market school. The scope of these two components is still a controversial topic among macroeconomists. In contrast to inflation is deflation. The level of inflation is zero, or a small positive index is called "price stability." Data, The consumer price index, is taken from the percentage change in CPI of each month compared to the previous month and taken from the General Statistics Office's data source (https://www.gso.gov.vn/).

Methodology
This study uses a quantitative analysis method with a monthly time series of 120 observations from January 2009 to December 2018 as well as uses a due model (Hanif and Bhatti, 2019) to propose. On time series data, research rescue using additional DF test method is ADF (Augmented Dickey-Fuller test) for corpse stationery, Cointegrated Test by the technique of Johansen and Juselius to consider whether there exist. Long-term relationships between the variables studied to assist in selecting the use of the least squares multivariate regression analysis method (OLS) to determine the Variation Macroeconomic factors in the long term and using the VAR method to determine the optimal lag. Granger Causality Test test to assess the degree of influence of short-term variables; while tissue Error Correction Model (ECM) will help track the adjustment process of Vietnam stock market from the state Short term towards long-term balance. The research model looks like this: VNIi = β1 + β2SJC + β3OIL + β4M2 + β5IR + β6 EX+ β6CPI + Ui ( Y = β1 + β2X2 +β3X3 + β4X4+β5X5+ β6X6+ β7X7 + Ui ) Where: Y: (Dependent Variable) is a stock price index of VNI X: (Independent Variable) are Variables of SJC gold price, world oil price, M2 money supply, Interest rate, exchange rate, and Consumer price index (representing inflation). Ui: As residuals, is assumed to be random  (Hanif and Bhatti, 2019), (Do, Mcaleer and Sriboonchitta, 2009), (Smith, 2001), (Gokmenoglu, 2015) Oil (Price of oil) USD/barrel β3 + (Chittedi, 2012), (Kumar and Narayan, 2010), (Tursoy and Faisal, 2017 At the time of the highest price index, it reached 1174,460 points, many times higher than the stock index at the lowest time was 245.7400 points. Similarly, for gold prices, there is also a big gap in the last ten years. The consumer price index over the past time has a quite different growth rate with the increase and decrease shown clearly at 0.47 to 4.32%. World oil prices have a big gap, sometimes reaching the highest monthly average value of $113.7300/barrel, and sometimes falling by only $33.75000/barrel. The refinancing rate averaged 7.9% of this value quite far from the highest and lowest interest rates. This shows that in the recent period, the interest rate adjustment is quite a lot, especially after the global economic crisis in 2007 and 2008, the credit market became too hot. Skewness statistics of VNI have a Positive value of 1.186920 shows that VNI has a positive deviation distribution, which is common for emerging stock markets. In the world, most of the distribution of profitability ratios of financial assets has no normal distribution.

Unit root tests
One of the assumptions of the classical regression model is that the independent Variables must be non-random. If we estimate a time series model where the independent Variables are not stationary, then the assumption of OLS is violated (expectation of mathematics, Variance, and covariance is constant over time) leading to Inefficient t and p tests (also called fake regressions). A unit test is a criterion to test Variable is stationary. Dickey-Fuller gave the following inspection criteria: Ho: p = 1 (Variable is not stationary) H1: p ≠ 1 (Variable is stationary) We estimate the model: t = p/se (p) with distribution according to DF rules If │t│> │tα│ reject the hypothesis Ho. In this case, the Variable is stationary. With test results of table 3, we realize that the CPI price index Variable has unit root and stationary at level. The remaining Variables of the original data series (at the level) are non-stationary (or unit roots) when the firstdegree difference has the calculated absolute value more significant than the table lookup value at the significance level 5% and 10%. This means that data from VNI Variables, Gold price, oil price, money supply, exchange rate, and interest rate are non-stationary series, the integration level of Variables is 1 or I(1). With purpose avoid false correlation regression, next test is necessary to verify whether there is a co-integration between the Variables in the regression model.

Co-integration testing according to VAR method
When estimating a regression model with Variables that time series is not stationary if the model is co-integrated, then there will be no counterfeiting regression, and standards-based tests of t and F are significant because the trend in the chains has eliminated each other. Currently, there are several co-integration testing methods, such as Engle-Granger, Johansen testing, etc. In which, Johansen testing based on a VAR model, including two tests, called trace test and maximum eigenvalue test. + Trace test -H0: the number of co-linked vectors in the system is r, smaller or equal to ro with ro < p (p is the number of Variables in the system); H1: The impact matrix is synchronous. If the trace test < critical value then accept H0 (not co-integration) and vice versa. + Eigenvalue test considers hypothesis H0 is ro co-integration vector for the hypothesis H1 is that there is a ro + 1 co-integration vector. EViews 10 performs co-integration testing according to Johansen's VAR method. This method makes sense when we play time series, not stationary. Hypothesis H0: none means there is no co-integration. Assumption H1: At most, 1,2,3,4,5 means having a co-relation between the price of gold, interest, consumer price index, exchange rate, oil price to VNI. To decide to reject or accept H1, we proceed to compare the value of "Trace Statistic" with critical value at the significance level defined in MHM (here we choose 5%).
-If the Trace Statistic < Critical Value, we accept the H0 hypothesis -If Trace Statistic > Critical Value, we reject the H0 setting   (1990) provide that the matrix traceability test (trace) and the maximum value control of the matrix (Maxeigenvalue) all claim to have at least one co-integration at 5% significance level. This proves that there is a healthy long-term relationship (cointegration) between the study Variables. Accurately, we see Trace Statistic> Critical Value, (165.8732> 125.6154) => we reject the assumption H0 => meaning the model has co-integration, gold price Variables, oil prices, money supply, price index consumption, exchange rates, and interest rates have a co-relationship with the VNI. This proves that we can perform OLS regression without worrying about fake regression.

Select the optimal lag
The selection of the optimal lag for the model will make by applying the VAR model to the original data series of Variables with a maximum lag of 5. The VAR model will automatically select the optimal lag, based on the standards: Akaike Information Criterion -AIC, Schwarz information criterion -SC and Hannan Quinn information criterion -HQ to select the optimal lag for the model. The optimal lag is determined based on the results consistent with the most standards. Schwarz information standards and information standards Hannan-Quinn proposed to choose the maximum model lag of 1, meaning the values of the current Variables will be affected by the value of the late Variables as a month earlier.

Co-integration regression model
The use of a regression model aims to consider the long-term relationship of independent variables affecting the VNI stock price index. The study uses the p-value to verify whether Variables affect VNI. The results show that SJC gold price, oil price, M2 money supply, exchange rate, and interest rate are statistically significant at 1% level. Mainly, the consumer price index is not statistically significant. We will conduct further verification to eliminate the consumer price index Variable. , we can see that R2 = 0.878080 indicates that the six macroeconomic Variables affecting the stock price index from the original data explain 87.80% of changes in the VNI stock price index. This suggests that in addition to being influenced by the above factors, VNI is also affected by other factors such as total industrial production output, trade balance, or herd phenomenon, etc. and VNI stock price index of the previous periods. With an emerging stock market like in Vietnam, the stock index is affected by psychological and asymmetric information, and securities investment is mainly in the form of speculation. This makes the stock price index quite complicated in recent years.
However, with the processed result, for the Consumer Price Index with the elasticity coefficient 11.82328, however, the value of P-value = 0, .2978> the significance level of 0.05. So the consumer price index is a Variable that is unlikely to affect.

Testing of redundant variables
To verify the inheritance of CPI, we build the following hypothesis: H0: β4 = 0 (redundant CPI Variable) H1: β4 ≠ 0 (CPI Variable is not redundant) The result of processing from table 7 gives us the value Prob = 0.2978 > α = 0.05 so we accept the hypothesis H0 means CPI consumer price index, model. The results of model processing after eliminating the consumer price index Variable with R2 = .876899 indicate that the model has a significant level of explanation in the rate of 87.68%.

Error correction model
The short-term relationship of the model is taking into account the transient nature of the time being studied and considering the volatility of the stock price index over the month (Variation by month) influenced by Variability. of macroeconomic factors (independent Variables) and the stock price index itself. After determining the results of co-integration between the Variables under study, the ECM was applied to consider the short-term relationship between the Variables. Since the first difference of the Variables is a stop sequence and a lag of one month impact, the study can use OLS technique, adding residuals with t-1 lag are included in the model to ensure long-term relationship system is satisfied. Looking at the data table 8, we see: in the short term, the VNI stock price index is affected by oil price factor (5% significance level) and interest rate factor (10% significance level).
When other factors remain unchanged, the Variation of oil price increases by 1%, the stock price index will increase by 1.84%. Besides, when other factors do not change the interest rate Variation by 1%, the stock price index will decrease by 11.67%. Furthermore, the coefficient of correction error (ECM t-1) is -0.441489 but not statistically significant, indicating that the short-term correction rate for long-term equilibrium is slow after the impact shock. Diagnostic tests of autocorrelation and Variance have also shown that the necessary conditions of econometrics on the reliability of regression results are guaranteed.  Table 9 describes the causal relationship between the Vietnam stock price index and macroeconomic factors with a lag of one month. This result shows that: Vietnam stock index has a one-way causal relationship with the oil price. The M2 money supply is significant at 10% and with interest rate at 5% significance level. This proves the fluctuation of oil prices, money supply, and interest rates affecting the variation of the VNI stock price index. Thus, the research results here also almost coincide with the model's test after adjusting the error according to the ECM model.

Conclusion
From the analytical results, it shown that the data series used in the model except for the consumer price index is stationary at level, the remaining Variables are stationary at first difference, meaning I(1). The co-integration between the Variables found. This is the basis for us to conduct co-integration regression, determine the relationship in the long term. Examine Variables to eliminate unnecessary Variables and little effect on Dependent Variables. At the same time, the model of error correction (ECM) and causality test (Granger) reflects the relationship between Variables in the short term.
In the long term, the results indicate that the VNI stock price index is affected by the following factors: SJC gold price and exchange rate affect negatively on the VNI stock price index. This coincides with the research results of (Hanif & Bhatti, 2019), (Akbar et al., 2019). Also, oil prices, money supply, and interest rates have a positive impact on the stock price index, which is also consistent with the stock market at Jordanian and Turkey (Tursoy & Faisal, 2017), (Mohamed & Ahmed, 2014). The consumer price index Variable is not statistically significant, meaning there is no impact on the stock price index. The removal of the effect of the consumer price index on the stock price index is consistent with the tests.
In the short term, the VNI stock price index positively related to the volatility of oil prices, and the relationship is contrary to the variability of interest rates. Granger causality test also shows that oil prices, money supply, and interest rates are the cause of volatility of the VNI stock price index. However, the verification of the error correction model indicates that the short-term adjustment speed of long-term balance is slow after the impact shocks.
Some of the following main ideas can summarize the findings of the study: Firstly, adding the world oil price into the model of macroeconomic factors affecting the stock price index. Supplementing the model of studying macroeconomic factors affecting the previous Vietnamese stock market (Hussainey and Khanh Ngoc, 2009), (Vinh, 2014). The fact that the Vietnam stock market is the emerging stock market, as well as the market size, is quite small compared to the market in developed economies, so the stock indexes are affected by Global macro factors are inevitable.
Secondly, the study has analyzed the SJC gold price factor in Vietnam as an independent Variable, which is a critical Variable affecting investors' psychology when investing in the Vietnam stock market. However, so far, there have been no in-depth studies on gold prices affecting the VNI stock price index.
Thirdly, To expand the scope of this study, we can use other factors such as world gold price, world-famous stock price index such as Dow Jones, Nikkei 225, Shanghai Composite Index..., In addition, the impact of macroeconomic factors derived from the company itself, such as earnings per share, seasonal effects, or herd behavior on the Vietnamese stock market can study.