“Impacts of monetary policies on the real estate bubble in Hanoi, Vietnam”

The development of the real estate market always goes hand in hand with the fluctuation of the economy. In recent years, this market has experienced many recessions and «freezes» associated with the appearance of a real estate bubble. To approach this issue, this paper studies and gives an overview of the real estate bubble and the impact of monetary policies on the real estate bubble in Vietnam. This paper’s purpose is to identify and measure the influence of monetary policies, including interest rates, credit and money supply, on the real estate bubble in Ha Noi. The vector autoregression model (VAR) is used to test the interaction of the variables in the model. Dickey-Fuller test (DF) is applied to determine the stationarity of the variables, while the Akaike information criterion (AIC), Likelihood Ratio (LR), Final prediction error (FPE), Hannan-Quinn information criterion (HQ) and Schwarz criterion (SC) are used to find optimal lag of the model; then Granger causality test is utilized to determine the two-way correlation between variables. The results showed that the real estate bubble reacted quickly to shocks from macroeconomic factors representing the monetary policy, consisting of interbank interest rates, credit growth, and money supply growth. Thus, it is concluded that monetary policy is not only the cause of formation, but also one of the effective solutions to deflate the real estate bubble.


INTRODUCTION
A real estate bubble (housing bubble) is an economic phenomenon when real estate prices surge to an unreasonable level in a short period, accompanied by the optimism of investors about the future.History has shown that this phenomenon may last for a long time with the risk of a bubble bursting.A common explanation for this phenomenon is the occurrence of the supply-demand gap.The influence of monetary policy on the property bubble is explained by the fact that interest rates and credit, affecting the money supply, cause the housing supply-demand curve to change, from which the contractionary or expansionary monetary policy can either increase or decrease housing prices, thereby stimulating or bursting the bubble.
In fact, until now, Vietnam has not had any research exploiting deeply and comprehensively the impact of extremely important factors (monetary policies) on the real estate bubble, as well as, focusing only on analyzing the housing bubble in Ho Chi Minh City but ignoring this phenomenon in Ha Noi.
As a result, being aware of the fact that completing content for other scholars, learning and evaluating the existence of the real estate bubble in Ha Noi and explaining the impact of monetary policies on the real estate bubble is undoubtedly necessary and urgent, theoretically and practically, this paper is done to research on the impacts of monetary policies on the real estate bubble in Ha Noi.
The research is aimed at (1) constructing a specific framework and theory of the real estate bubble and monetary policy; (2) pointing out and analyzing the existence of the real estate bubble in Ha Noi and Vietnam; (3) evaluating the influence of monetary policy tools on the real estate bubble through quantitative models; and (4) discussing the facts of the Hanoi real estate bubble and governmental solutions to ease the situation.
This study contributes to providing managers and policy makers with a more comprehensive view of the real estate bubble in Vietnam.Vietnam's real estate market in general and Ha Noi's real estate market in particular are currently in a period of instability: prices in many areas have skyrocketed, investment demand has been greatly influenced by the COVID-19 epidemic.Monetary policy is not only the cause of formation, but also one of the effective solutions for deflating the real estate bubble.For managers in the real estate business, it can also be utilized as a source of reference to be able to assess the advantages and disadvantages of the real estate market initiated by alterations of monetary policy.

LITERATURE REVIEW
The real estate industry is an important part of the national economy, so the development of this market is always closely tied to the changes in the economy.In recent years, beside growth, the real estate market also has experienced many recessions and "freezes" associated with the appearance of a real estate bubble.To make the issue more approachable, this study provides an overview of the real estate bubble and the impact of monetary policy on the real estate bubble in Vietnam and around the world.
Many studies have been done to confirm the existence of real estate bubbles in different countries.Kim et al. (1993), through analyzing the influence of speculation on the House Price/Rent Index, show the existence of real estate bubbles in Korea and Japan.Different studies in China demonstrate unsimilar findings about the time and the venue of bubble existence and explosion (Chen et  Many other studies have been conducted to find the impact of monetary policies on the housing market in different countries.Xiaoqing Eleanor Xu and Tao Chen (2012) used a quantitative vector autoregression (VAR) model to measure the impact of monetary policy on the Chinese housing market from 1998 to 2010.The study showed that variables of the long-term benchmark bank loan rate, money supply growth, and mortgage credit policy indicator moved in parallel to the formation of a real estate bubble.Besides, expansionary monetary policy tended to accelerate house price growth, while contractionary monetary policy tended to slow down house price growth.Monetary policy tools that have a strong impact on the real estate market were given, including: (1) Required reserve ratio; (2) Interest rates, and (3) Open Market Operations (OMO).
Na Yan (2019), through the results of using the VAR regression model, showed that a country's monetary policy played a very important role in the real estate market, whether the impact was observable or not.It is shown that both money supply and interest rates had an impact on real estate prices.Therefore, the central bank could use interest rates as a tool to regulate the behavior of real estate market participants, thereby controlling supply -demand and real estate prices.Besides, the government should concentrate on regulating the money supply to the equilibrium point of the real estate market because ill-implemented policies will widen the gap between supply and demand of real estate.
Besides, Subramaniam S. Pillay also studied the Chinese real estate market (2010) and built a perfect model to verify the existence of the real estate bubble in Singapore, while Stefan Gerlach (2003) built a model to evaluate the influence of bank credit flows on the growth of real estate prices in Hong Kong.
Mark Thornton (2009), John F. McDonald and Houston H. Stokes (2011) all agreed that the FED's low interest rates (specifically, below 2% from the end of 2001-2002 to less than 1% in 2004) conditioned the formation of the 2008 US housing bubble, as well as the ineffective implementation of monetary policy.Tsatsaronis and Zhu (2014) found that lending rates explain 10.8% of the change in house prices.When interest rates are floating, short-term rates have a stronger impact on house prices than long-term rates.
In Vietnam, Le Thanh Ngoc (2014) used modeling and inductive methods to find out the causes of the real estate bubble and the reverse effect of the real estate bubble on the economy in Ho Chi Minh City.According to the research, the total bank loans for real estate increased the most drastically in 2007, which was also the year the housing fever occurred in Ho Chi Minh City.In the following years, this rate decreased along with the slump of the real estate market.The real estate bubble in Ho Chi Minh City is also affected by money supply and capital into the real estate market.M2 money supply growth peaked in 2007 and then recession, along with the depreciation of the real estate market.During the period from 2010 to 2012, the money supply was low and the real estate market fell into a frozen state.
In addition, Doan Thanh Ha (2013) also used a self-regressive vector quantitative model to intensively analyze the contemporary situation of housing prices in Ho Chi Minh City in particular and in Vietnam in general based on data gathered from different surveyors and reliable sources: Real Estate Market Division -Ministry of Construction, National Financial Supervision Commission, General Statistics Office.By the method of data analysis, two periods have been demonstrated clearly signaling the boom -slump of the real estate bubble in Vietnam.Do Duy Tan and Phuong Lan Le (2022) found evidence of contagion between stock bubbles and real estate bubbles in Vietnam by Granger causality test, however, contagion effect was found to move from the real estate market to the stock market, not the backward way.
Most studies have shown that the real estate market is associated with monetary policy and is directly affected by interest rates and money supply of the central bank.Studies also show that the cause of the housing crisis is largely attributed to the unreasonable monetary policy of that state.And to solve the problem of the real estate bubble, the most effective tool is still monetary policy.
Given that literature review, some hypotheses for this paper are constructed as follows:

RESEARCH RESULTS
After doing the stationarity test, the results are shown as follows.It can be seen that the variables SPPI, R1, R6, TM, TM2 in the original series are non-stationary.However, when taking the first-order difference, all variables were stationary at the 5% confidence level.
It is determined that the lag at p = 4 was selected according to three criteria: FPE, AIC and HQIC.
Granger causality test results are shown below.The results showed that there was a Granger cause-oriented relationship between the variable SPPI and R1, as well as between SPPI and R6.Besides, SPPI had an impact on the credit growth variable (TM) and the money supply growth variable M2 (TM2), while in the opposite direction, these variables did not affect the SPPI variable in the period 2009-2019.

DTM2
The model is represented as a matrix: the residuals, it can be concluded that the research model satisfied all the above conditions of econometrics at the 5% confidence level.Next, after estimating the VAR model, the impact of various factors on SPPI can be clearly seen as follows: the overnight interbank interest rate variable (R1) and 6-month term (R6) along with the real estate price index SPPI had a positive impact on the bubble variable with the lag of one to three period.The remaining variables, including credit growth (TM) and money supply growth M2 (TM2), were not statistically significant.

Impulse response function then is conducted
The analysis results showed that the SPPI variable responded to all the shocks from the selected variables in the model right from the first period in a positive direction and at different levels, as well as lasted for different time periods.However, compared to its response to itself, TM, and TM2, the SPPI variable had a stronger response to two variables R1 and R6.

Variance decomposition
The results of variance decomposition by SPPI showed that the endogenous shocks from the real estate market played a relatively large role in explaining the change in SPPI (67 -79%).Then, the variable R1 explained about 15-17% of the variation, while the remaining variables R6, TM, and TM2 only explained a very small part with the values respectively: 0.3-5%, 0.4-4% and 2-7%.

DISCUSSION
Based on the results of the model, further details can be discussed.
Firstly, the results obtained from the VAR model and the tests of overnight interbank interest rate (R1) and 6-month term (R6) variables are consistent with the theory studied by the group.In the first period of 2008, when inflation showed signs of increasing, the State Bank of Vietnam promptly implemented a flexible monetary policy and lowered interbank and credit interest rates.This http://dx.doi.org/10.21511/imfi.20 (1).2023.20 made it easier for businesses to access credit capital, leading to an excess of credit in the economy and a rise in real estate prices.From 2009 to the first half of 2011, with signs of a bubble flashing, the State Bank of Vietnam tightened the monetary policy by increasing interest rate followed by the need for loans between banks to offset the required reserve ratio, as well as a decline in lending, which made it more difficult for businesses and investors to access loans as banks became more cautious in their lending policies.Besides, in explaining for results of the variance decomposition test showing that R1 explains more of the volatility of the bubble than the variable R6, the group argues that long-term interest rates are often behind the short-term interest rates on managing the bubble (Doan Thanh Ha & Le Thanh Ngoc, 2013).
Secondly, clarifying the lack of statistical significance of the credit growth variable (TM) in the model and test results obtained, it is believed that the use of the general credit growth instead of real estate credit growth as a variable can result in inaccuracy in interpreting the impact of credit on the real estate bubble because, from 2009 to 2011, the decline in real estate credit growth was slower compared to the decline in total credit in the previous year.However, theoretically, the variable TM has been found to have an impact on the real estate bubble because an increase in credit will, in turn, trigger an increase in the capital flowing into both the economy and the real estate market, raising the amount of money in circulation and escalating the housing prices.In this situation, the State Bank of Vietnam implemented a tightening monetary policy to control the supply of credit capital, as a result, from the second half of 2011, credit growth declined, which cooled down the housing market.With prices anchoring high and the fear of loss infused among investors, a spike in demand for real estate speculation was recorded.This explains the results of the variance decomposition function, which showed that the impact of SPPI on itself accounted for about 67%-79%.Moreover, the monetary policy variable modestly contributes to an increase in SPPI.This implies that monetary policy is a catalyst for the real estate market, while the main cause of the bubble in the psychology and behavior of real estate investors.

CONCLUSION
In summary, the study has completed all the goals set out initially, however, there are still some limitations that cannot be resolved within the framework.Firstly, the sample size and research time scope are limited because there are many difficulties in accessing real estate price data and macro variable data in Vietnam.Besides, due to the limitation in finding statistical data of Vietnam, only the impact of overnight and 6-month interbank interest rates, credit growth (TM) and growth money supply M2 (TM2) to the housing price index in Ha Noi can be assessed.These variables represent monetary policies in Vietnam but are not the most optimal.Moreover, the research team has proposed a number of solutions to deal with the real estate bubble, but those solutions are only relevant to general market participants such as commercial banks, business enterprises and real estate investors.There are no clear solutions for people who are in real need of accommodation.
This study has met the aims of constructing a theory of the real estate bubble and monetary policy correlation, and analyzing the existence of the real estate bubble in Ha Noi and Vietnam so that the influence of monetary policy tools on the real estate bubble can be analyzed using a quantitative model and further issues of governmental control can be discussed.The results from VAR model have demonstrated that (1) there was a Granger cause-oriented relationship between the variable SPPI and R1, as well as between SPPI and R6.(2) Besides, the overnight interbank interest rate variable (R1) and 6-month term (R6) along with the real estate price index SPPI had a positive impact on the bubble variable with the lag of one to three period.(3) Moreover, this study documented that the SPPI variable responded to all the shocks from the selected variables in the model right from the first period in a positive direction and at different levels, as well as lasted for different time periods.However, compared to its response to itself, TM and TM2, the SPPI variable had a stronger response to two variables R1 and R6.(4) The endogenous shocks from the real estate market played a relatively large role in explaining the change in SPPI (67-79%), while the variable R1 explained about 15-17% of the variation.
Upon learning those findings, the government should take into consideration the need to have judicious monetary policies to prevent the emergence of a real estate bubble during the country's developing period.Among the resolutions made by the government, controlling overnight interbank interest rate as well as 6-month interbank interest rate is the most important monetary policy that should be well considered.
al., 2013; Coleman et al., 2008; Tie-Ying Liu et al., 2016).Many regions in the USA were proved to have bubble formations, such as the Northeast, Midwest, South and West regions from 2005-2006 (Zhou et al., 2006), and the whole USA (Nneji et al., 2011).Itamar Caspi (2016) conducted research on the basis of concerns about the bursting of the real estate bubble in Israel based on specific data from 2008 to 2013 and found the emergence of a housing bubble in this period.Coskun et al. (2017) conducted the first study to find a housing bub-ble in Turkey from Jan 2010 to Dec 2014 and from June 2007 to Dec 2014, but only concluded that the Turkish housing market has experienced some cases of overvalued, but not bubble formation. (1)p://dx.doi.org/10.21511/imfi.20(1).2023.20 The secondary data collected was in the form of time -series data from the first quarter of 2009 to the first quarter of 2019 from different sources, such as Savills Corporation Vietnam, the website of the State Bank of Vietnam and the website vi-etstock.vn.After collecting the data, the research team used Microsoft Excel to calculate the average quarterly data, and Stata 16.0 software to make descriptive statistics of the variables.SPPI in Ha Noi (SPPI -Savills Property Price Index): represents the quarterly fluctuations of real estate prices across different segments of the market as calculated by Savills Vietnam.Accordingly, for the Ha Noi housing market, the index is built based on a fixed sample basket of more than 161 projects in the primary and secondary markets.The basket, nevertheless, is updated regularly with new projects to ensure timely response to market changes and applies a "liquidity ratio" to adjust the asking and trading prices to get the most accurate results.The increase or decrease of SPPI shows the corresponding fluctuations of real estate prices traded on the market and this is also one of the manifestations of a real estate bubble.(Unit: point) The Savills Real Estate Price Index in Ha Noi (SPPI) in the previous quarters has an influence on itself in the present.http://dx.doi.org/10.21511/imfi.20(1).2023.20 VAR model (Vector Auto Regression) is chosen in this study based on the theoretical model studied by Na Yan (2019) and Le Thanh Ngoc (2014) to test the interaction among the variables in the model, especially the influence of variables representing monetary policies on the SPPI price index, which represents the real estate bubble.Accordingly, besides estimating the VAR model, within the scope of this research paper, other various tests are also used with a view to: • Determining the stationarity of the variables by Dickey-Fuller test (DF).Growth of money supply M2 (TM2): represents the growth rate of M2 money supply over the months, an important factor in forecasting ecohttp://dx.doi.org/10.21511/imfi.20(1).2023.20 where i = 1, p is a lag order; SPPI t , SPPI t-i : Savills Real Estate Price Index in Ha Noi in period t and period t-i; R1 t , R1 t-i : Interbank overnight interest rate in period t and period t-i; R6 t , R6 t-i : Interbank interest rate for 6-month term in period t and period t-i; TM t , TM t-i : Credit growth in period t and period t-i; TM2 t , TM2 t-i : Growth of money supply M2 in period t and period t-i; c jt , α jt , β jt , γ jt , δ jt , λ jt -coefficients, with j = 1, 2, 3, 4, 5 respectively; ε jtwhite noise errors that can be contemporaneously correlated, with j = 1, 2, 3, 4, 5 respectively.

Table 2 .
Granger causality test results Source: Compiled by authors from Stata 16.0.

Table 3 .
VAR model estimation resultsSource: Compiled by authors from Stata 16.0.