Abstract
In early 2020, the global economy was severely hit by the COVID-19 pandemic. Governments, including China, have adopted expansionary fiscal policies to stimulate demand and quantitative easing to increase the money supply to boost the economy by lowering lending rates. This paper starts from the influence of monetary policy on housing price mechanism, tries to explore the monetary policy on interest rate and money supply, finally establishes the VAR model, empirically analyzes the effectiveness of the impact of monetary policy on housing price changes, and makes a policy suggestion to demonstrate the necessity of consideration when stimulating the economy, to avoid housing price rise too fast, which affects the people's livelihood, and to have a negative impact on social sustainable development.
1. Introduction
Since the abolition of the housing welfare distribution system in 1998, the housing transaction market has shown a rapid upward trend [1]. Specifically, the total amount of real estate investment surges year by year, the supply and demand of the real estate market are booming, and the real estate prices have repeatedly reached new highs [2]. Since 2004, housing prices have basically increased by about 18% year on year, and even housing prices have increased by 25% in 2009. Since 2010, with a series of favorable policies and economic stimulus policies loosening the property market, the real estate industry has shown a rapid recovery and shown a relatively obvious polarization state. The first- and second-tier cities led by North, Shanghai, Guangzhou, and Shenzhen have begun to flourish, repeatedly reaching new highs. Even though China has begun to introduce a variety of housing price control policies since 2003, the housing price is still in the dilemma of being higher and higher. The high housing prices will lead to the redistribution of social resources, involving a series of issues such as financial stability, economic development, and social equity [3, 4]. At present, the high housing price in some economically developed areas has exceeded the consumption level of many residents, and even produced a large number of “mortgage slaves”. It can be seen that the rising speed of housing price rates is not only an economic issue, but also a livelihood issue closely related to every people [5].
2. Empirical Analysis
2.1. Variable Selection and Interpretation
This paper is based on analyzing the macroeconomic data for 44 periods between the fourth quarter of 1999 and the fourth quarter of 2010 [6]. Considering the availability of data, the variables selected the five-year nominal loan interest rate (X1), M2 (X2) of broad money demand, and real estate sales price index (Y) after eliminating the impact of seasonal changes, and were processed logarithmically [7, 8]. Table 1 is the list of variables affecting real estate prices. All the data came from the National Research Network, China Economic Network database, and the website of the National Bureau of Statistics.
2.2. Empirical Inspection and Data Analysis
2.2.1. ADF Stationarity Test
Stationarity means that the target time series is generated by a random process; that is, each value is derived from a random probability distribution. Neither the mean nor the variance varies with the time t, and the covariance between the values is only related to the time t. Before establishing the VAR model, the stationarity test [9] is required, and the time series can only be estimated using the OLS method; however, if the nonstationary time series is used, then the problem of pseudoregression is likely [10] because the Dickey–Fuller test assumed that time series is generated by first-order autoregression (AR) process with random error of white noise; however, in the actual test, time series may be generated by higher order autoregression process or the random error term is not white noise, and since the OLS method still shows the random error autocorrelation problem, the DF test is invalid. The author uses the ADF method for the stationarity test of three variables; the test equation is as follows [11].
Model 1(equations with the trend terms and the intercept terms):
Model 2 (the equation for which no trend term has an intercept term):
Model 3 (the equation for the intercept term without a trend term) [12]:where α is the drift term, and βt is the trend term. The null hypothesis H0: and the alternative hypothesis H1: are tested step by step from models 1 to 3. If the test rejected the null hypothesis, the sequence does not have a unit root and is a stationary sequence. If the null hypothesis cannot be rejected, the sequence is considered to have at least one unit root and is a nonstationary sequence, so differential processing is required to ensure its stationarity [13]. The intercept and trend terms are also determined in the equation according to the minimum information criterion (AIC) [14]. Table 2 shows the results of the ADF test.
The test results show that the ADF value of the variable LNY is less than 5% critical value, the value is less than 0.05, and the log of the house price index is a stable sequence [15]. At the significance level of 10%, the raw data of interest rate (LNX1) and money supply (LNX2) are not significant, indicating that the two variables correspond to the nonstable sequence, indicating that these data have a unit root. Regarding the loan interest rate and money supply log of the first-order difference processing, it is found that D (LNX1) ADF value is significantly less than 5% critical value, so the log of the interest rate index LNX1 is a first-order single sequence; however, for the generalized money supply M2, the test equation with intercept term form of value < 0.1 (further check the equation ) estimate result is negative and value = 0.0096 is significantly less than 0.05, so there is a enough reason to think that the difference of money supply index D (LNX2) is stable and that LNX2 is also first-order single sequence [16].
2.2.2. Integration Test
The classical regression model is based on the assumption that all variables are stable data. For nonstationary data, the classical regression model cannot be used, but if there are long-term stable relations between variables, the method of classical regression model can be used. The logical significance of the cointegration test is as follows [17]: there is a long-term change law of multiple nonstable sequences; if there exists (d, d) order cointegration, their sequence is under a linear combination of long-term and equilibrium stable proportional relationship, namely, that variables will not produce false regression results; and even if there is a variable after a period of time interference from the equilibrium point, the equilibrium mechanism will also cause the next adjustment regression near the equilibrium point [18]. Two variables selected in this paper are first-order single integration sequence, and there may be a cointegration relationship, which needs to be tested. Here, the more commonly used Johansen’s method for the cointegration test are selected [17–19].
Selection of the lag order: The lag-order posterior term of the VAR model is designed to exploit the limited sample size to eliminate the time-series correlation of the residual term. The larger the number of model lag periods, the smaller the degree of freedom, and the greater the sample loss. However, the lag order unable to eliminate sequence correlation may lead to error in the established model [20]. In general, the optimal lag order is determined according to the principle of minimum AIC and SC values. If AIC and SC are not minimal at the same lag order, then trade-offs are performed using the LR test.
As can be seen from Table 3, the AIC and the SC information criterion appear at the third order, which indicates that the minimum value appears at the lag of 3 periods, namely, −15.35634 and−14.20464, respectively, so the optimal lag period is 3.
Johansen’s coconsolidation test. The cointegration testing for multivariates and occurrence represents the rejection of the null hypothesis at a 5% significance level, and that for not occurrence represents the nonrejection of the null hypothesis. As can be seen from Table 4, there is a cointegration relationship between the trace test and the maximum value test at the significant level of 5%; that is to say, there is a long-term and stable proportional relationship between China’s property price index, benchmark five-year loan interest rate, and broad money supply.
2.2.3. Vector Autoregression (VAR) Model
Vector autoregressive (VAR) models are often applied to the time-series data associated with each other and to analyze the dynamic impact of random noise on the whole variable system. Therefore, the following content is based on the VAR model architecture proposed by C.A.Sims (1980) to analyze the impulse response and variance decomposition between variables, so as to study the dynamic interaction and effect between monetary policy variables such as real estate price and interest rate and generalized money supply in China [21].
Model theory form:
Organizing the above formula, the following can be obtained:
In the formula, is the vector of m-dimensional endogenous variables, is the vector of d-dimensional exogenous variables, and (i = 1,2,3. p.; s = 1,2,3. r) are the parameter matrix to be estimated, endogenous and exogenous variables have p and r order lag periods, respectively, and is the random error term. In this paper, the Y variable refers to the processed property price, and is the logarithm of the interest rate and money supply of the exogenous variable.
Because in the traditional VAR model, only in the variable or variable first-order difference is divided into stationary sequence, or the stability of the model is stable, to pulse response analysis and variance decomposition, the variables, in the case of the real estate sales index is the first order single whole nonstationary sequence, so they first order difference into the model, and test the stability of the sequence and the model [22].
Model stationarity test:
As shown in Table 5 and Figure 1, the characteristic roots of the VAR model all fall in the unit circle, and the corresponding modules of each characteristic root are less than 1, indicating that the characteristic roots of the model are stable; that is, the model is stable. Thus, the pulse response function analysis with a standard deviation for this VAR can be proceeded [23].
Pulse response analysis: The pulse response function describes how the residual of the shock responds to the endogenous variable. The pulse response applied in this paper is an effective method to analyze the correlation between domestic loan interest rate, money supply, and real estate prices; by depicting the pulse function, the dynamic influence on the response element can be seen [24].
First, with two monetary policy factors as the impact element, the property price is treated as the response element. The solid blue line in Figure 2 shows the pulse response of property prices to the positive impact of one unit of standard deviation of interest rate and money supply as the number of forecast periods increases. The red dashed line indicates the confidence region of 2 standard deviations from the pulse response image.
From Figure 2, we can know that in the short term, an impact of interest rate has a positive effect on the housing price. After about 4 issues, the highest value is about 0.025, the fifth period becomes negative, and the negative effect of the seventh period reaches the maximum of about -0.050, and then, it begins to gradually stabilize. This shows that the 5-year loan interest rate still has a negative impact on the housing price, but the effect of interest rate policy adjustment has a lag. In the short term, the rising interest rates will cause the real estate developers to transfer the increased costs of capital and materials to the sales prices, and the housing prices will tend to rise [25, 26].
In the medium and the long term, the supplier real estate developers are facing greater pressure of capital recovery, while the residents as consumers have more space for adjustment in the long term. They are more and more flexible to housing prices, thus leading to the decline of housing prices, and the effect of interest rate policy is gradually obvious. However, for the money supply, we can know that one unit of new interest rate can quickly act on the housing price, exerting a growing positive impact on the housing price. As can be seen in the figure, the first four housing prices experienced a period of rapid growth, by about the fifth period, the monetary policy has a maximum effect of about 0.01, and the 0 effect level line is outside the confidence domain, so the effect is significant. This effect lasts long until about issue 10. To Phase 12, a slight positive effect will be on housing prices after the impact on housing prices is gradually stable.
Therefore, it is believed that the positive effect of the change in the total amount of borrowing funds caused by the increase in the money supply on the real estate consumers and suppliers is greater than the negative effect of the increase in the real estate supply caused by the suppliers. In general, the government authorities can play a good regulation effect on housing prices through the two policy tools of money supply and loan benchmark interest rate.
Secondly, the real estate price is the impact element, and the two monetary policy factors are the response element. From Figure 3, we can see that the real estate price has a positive effect on the interest rate in the first four periods, but the effect gradually decreases over time. For the money supply, the real estate price fluctuates with alternating positive and negative effects in 10 periods, and then gradually stabilizes. It can be seen that interest rates and money markets are related to the real estate market.
Predicted variance decomposition: The variance decomposition [27–35] is conducted to attribute the variance of the endogenous variables, to analyze the contribution of each structural impact to its change, and to judge the importance of the impact of different variables. Therefore, the variance decomposition of the real estate price index can more clearly explore the importance of the variables affecting it. Table 6 presents the process of the 20-stage variance decomposition, and the data in the first column are the standard error after the prediction of each stage of the LNY. The last three columns represent the extent to which each variable contributes to the predicted standard deviation of property prices in each issue.
It can be seen that the impact of real estate prices is gradually declining and gradually stabilizing in Phase 11, around 67%. However, the explanatory power of interest rate on the standard deviation of housing price prediction was generally increased and gradually stabilizing at nearly 10% from the first phase (0) to Phase 10. Similarly, the explanatory power of money supply was basically gradually increased and gradually stabilized at 23% in the 15th period. Therefore, in general, the control ability of money supply policy is greater than the ability of interest rate to control housing prices.
3. Suggestions
3.1. Suggestions on the Real Estate Market
3.1.1. Monetary Policy Recommendations
Focus on the regulation of the money supply: China’s financial market is still gradually improved, the degree of market capitalization is relatively low, and a perfect interest rate liberalization mechanism has not yet been formed. Interest rates are less sensitive to monetary policy in the long-term and equilibrium stable relationship of the three. Compared with interest rates, the money supply and social and economic development and inflation levels have stronger correlation, so the monetary authorities should put the money supply as the best control target, pay close attention to housing prices, timely take countercyclical control strategy, do monetary control that is given priority to, appropriately raise interest rates, and curb house prices rise too fast. At present, due to the COVID-19 epidemic, China has released a large amount of liquidity in a short period of time, but due to the rapid effect of money supply, it will bear the rise of housing prices in a short period of time. Therefore, in order to ensure the healthy and sustainable development of the property market, some regulation policies can be introduced in time, such as the implementation of the maximum price for new buildings according to the time and price of developers, the restriction of the purchase of multiple commercial houses, and the differentiated down payment ratio of the first and multiple houses at the current stage to achieve the goal of promoting economic recovery, to meet the reasonable real estate demand, and to avoid real estate speculation and speculation again.
3.1.2. Interest Rate Policy Recommendations
Flexible use of various loan interest rate policies: For example, to further improve the differential interest rate policy of non-first-time homebuyers, there are various loan interest rate policies such as provident fund loan, commercial loan portfolio loan, and other related housing loan interest rate policies. We should further improve the effect of interest rate policy transmission channel, strive to promote domestic and regional interest rate liberalization, improve the sensitivity of real estate prices to interest rates, and determine the appropriate regulation time. The interest rate regulation policy has an obvious lag effect, in the medium- and long-term interest rate policy on the housing price, adjustment effect will appear. When the macroeconomic situation changes, China’s monetary authorities from the understanding of the problem (i.e., the understanding of the time lag) to the regulatory policy formulation and implementation (i.e., the decision time lag), and then to the effective interest rate policy (i.e., the external time delay), each link takes a certain time. This requires the central bank to make a timely and accurate forecast of the housing price and the economic trend, make policy formulation in advance, and reasonably grasp the implementation time, so as to avoid the effect of the policy in the wrong time to achieve the ideal regulation effect or even cause adverse impact on the economy. For example, during the current epidemic period, the economy is stimulated by lowering interest rates, but we should also take into account the possibility of regret for rising housing prices in the future, and timely adopt some indicators such as money supply exchange rate for macrocontrol to ensure the stability of housing prices in the future.
3.2. Suggestions to Consumers
Rational view of real estate price changes and the use of information to accurately grasp the timing of house purchase: Housing price fluctuation is the result of a variety of economic factors. It is a complex system, but it is still in line with the role of the law of value. Its price still fluctuates around the value in the long run. Real estate prices are closely related to the implementation of national monetary policy, as buyers, whether for any purposes, such as to try to avoid overconfidence and survivor deviation error, to take the initiative to study the policy impact on the housing prices, combined with the macroeconomic form, local specific situation, and monetary authority decisions, to grasp the overall trend of price fluctuations, using monetary policy effective delay, to choose the optimal purchase time, and to reduce their purchase cost to realize the value of assets.
Establish the correct concept of housing purchase: Housing not only has the consumption attribute of self-living, but also has the investment attribute of investment products. At the same time, due to the influence of China’s idea of safe land relocation, it seems that every family must own their own housing; this traditional concept actually increases the housing demand and becomes the driving force for China to maintain high housing prices. But in other developed countries around the world, house rental is already very common. With the economic development of China, now new ideas, such as long-term rental apartments, gradually flourish in China. It is hoped that Chinese consumers can also accept the new concept of rental housing and relieve the pressure of domestic housing demand. At the same time, the relevant departments should also effectively protect the legitimate interests of the renters.
3.3. Advice to Real Estate Developers
At present, the real estate enterprises for the operation of real estate development cash flow mostly depend on bank loans, to prevent real estate excessive reliance on bank funds cause financial risks, should draw lessons from the successful experience at home and abroad, efforts to broaden the channels of investment and financing, the transformation of financing mode, more through the real estate trust business, money and capital markets, multi-channel direct financing mode to obtain funds. This not only has a positive significance to the security of developers’ funds, but also is conducive to promoting the implementation effect of China s monetary policy and reducing the abnormal fluctuations of housing prices. On the contrary, commercial banks should also strengthen the supervision and risk control system of the funds invested in the real estate field, and improve the loan authorization and approval system and the whole process of the loan management mode. At the same time, we will actively take innovative measures to implement differentiated mortgage business to control the direction of capital flow, and promote the central bank’s accurate measurement and release of the market money volume.
4. Conclusion
Through the cointegration test, it can be considered that the real estate price has a long-term stable relationship with the interest rate and the broad money supply. The authorities can control the real estate prices by reasonably adjusting the interest rates and the money supply to ensure the reasonable and stable operation of the real estate market. At the same time, although the monetary policy has a certain effect on regulating the housing price, the impact of the monetary policy is limited, and the effect also varies with different monetary policies. Therefore, we should not rely too much on the adjustment of monetary policy but reasonably choose the types of monetary policy.
Through the pulse response and the variance decomposition, the magnitude of the impact effect of different monetary policy variables on real estate prices and the contribution to the change in housing prices can be observed. Interest rate policy has a lag effect on real estate prices. In the short term, housing price is not sensitive to interest rate, and there is a lag period of 5–7 periods when using interest rate to adjust housing rate. However, in the medium and long term, the 5-year loan interest rate still has a negative impact on the housing price and lasts for a long time, and the regulation effect is gradually obvious over time. For the money supply, it can be known that one unit of new interest rate can quickly act on the housing price, exerting a gradually increasing positive impact on the housing price. Although there is a two-way effect of money supply on housing price from the perspective of the mechanism of action, it can be seen that the positive effect of money supply on housing price is greater than the reverse effect.
Therefore, the authorities to adopt expansionary monetary policy to increase the amount of money in circulation can immediately play a positive effect on housing prices. In the short term, the control of money supply can achieve better effect of housing price than the interest rate. From the contribution of various monetary policy factors to the fluctuation of housing price, in the long run, the impact of money supply changes on the housing price is greater than the interest rate, so the overall ability of money supply is greater than the ability of interest rate to control the housing price. The central bank can use the money supply to regulate the property market to achieve better results.
Data Availability
The dataset can be accessed upon request.
Conflicts of Interest
The author declares no conflicts of interest.