Housing Affordability of Residents in Malaysia

Transcription

1 Housing Affordability of Residents in Malaysia Monica Selvaraja UCSI University, Malaysia Frankie Goh Song Peng UCSI University, Malaysia Vahid Biglari University of Newcastle (UoN), Singapore Joanne Tan Poh Lee UCSI University, Malaysia Lee Kah Mun (Carmen) UCSI University, Malaysia Abstract Housing affordability has often been discussed among researchers and policy makers as it has become one of the major concern among countries who recognize its significant role to the nation s economy and social wellbeing. This research attempts to understand housing affordability by using a proxy ratio comprising housing price and household income. The research, examines the significance of relationship between four macroeconomic factors including GDP, inflation rate, interest rate and exchange rate with housing affordability. The findings show that there is significant negative relationship between GDP and exchange rate towards housing affordability while inflation rate showed significant positive relationship with housing affordability. The findings benefit current and potential homeowners, researchers and policymakers in Malaysia and neighbouring countries. Keywords: Housing affordability, GDP, Malaysia, Macroeconomic Variables, Interest rate JEL Classification : A1, E00, D

2 1. Introduction Journal of Emerging Issues in Economics, Finance and Banking (JEIEFB) Real estate industry is an expanding industry. Malaysia property market is incrementing since 2001 to The total number of property sales increased from 242,600 units to 430,400 units from 2000 to 2011 (International Monetary Fund 2016) 1. There are 4 major sectors in Malaysia s property market which are residential, commercial, industrial, and agricultural. According to the report of Rating Agency of Malaysia (2007)2, the residential market contributes the largest portion to the property market by providing more than 60% of the total transactions. According to New Strait Times (2016), most Malaysians are still looking for buying a house rather than renting one. Furthermore, most of the residents in Malaysia originate from traditional social norms. They feel owning a house is more secure, thus they feel protected. According to a PropertyGuru Malaysia s Property Market Sentiment Survey Report3, 36% of Malaysian prefer to own a new house compare to 17% who wants to buy from the secondary market due to property developers are offering attractive sales packages for new houses (PropertyGuru, 2016). The real estate market in Malaysia especially residential property market is growing rapidly since 1950 due to the population increase and other minor activities (Bujang et al., 2015). The growing of residential property market in Malaysia is effected by various factors such as inflation, exchange rate, demographic, disposable income, level of education and household size. Housing industry plays an eminent role in the country s economy in terms of employment, capital market, consumption and financial wealth thus stimulating the business cycle (Hashim, 2010). Since the housing industry is highly correlated to major dimensions of Malaysia s economy, the health of housing industry reflects the purchasing power of the residents. Therefore, it is necessary for policymakers to constantly monitor the condition of housing market. If the housing prices are skyrocketing to unaffordable prices this will cause oversupply problems which sooner or later will lead to housing bubble. Housing affordability can be measured by dividing the median house price with median household income. It is important to investigate the factors that will affect the affordability of housing. The aim of this research is to address how inflation rate, exchange rate, interest rate and gross domestic product (GDP) affects the housing affordability of the residents in Malaysia. 1 International Monetary Funds. (2016). Global House Price Index. Retrieved from 2 DEPARTMENT, S. (2016). Population and Demography. Retrieved from Statistic Department: NWJwRWVSZklWdzQ4TlhUUT09 3 PROPERTYGURU Malaysia Property Market Sentiment Survey Report 2512

3 Malaysia s economy is expanding and cities are growing rapidly. The property prices has jumped to three times or more as compared to the original prices during past few decades. It is getting harder for the younger generations to purchase a house in the city area. Housing affordability is not only a major problem for individual households to worry but, it will affect a nation s economy. Gabriel et al. (2005) found that high housing costs also may affect firms in rapid developing states facing wages pressures and unable to attract key workers to build the firm s competitive advantage. These firms may not be able to compete with their competitors and will be forced to shut down, with this, the unemployment rates will increase as GDP will decrease, and thus will result to drastic drop of country s economy. Nevertheless, the low level of housing affordability also leads to a significant social problem. When the costs of construction are high, developers will build more high-end houses to cover their costs or even gain a good profit. Thus, the rental fees of houses will increase too, low-income families will face higher risk of not being able to sustain their tenancy and finally become homeless. When the volume of homeless families or individuals increase, the criminal rates will increase simultaneously. Macroeconomic factors that originate from government policies can significantly affect housing affordability. In this paper, we aim to find macroeconomic factors that are behind housing affordability. Thus, the objective of this study is to look into the housing affordability of residents in Malaysia with the macroeconomic variables. Thus, we are going to examine the relationship of macroeconomic variables such as inflation, interest rate, exchange rate and gross domestic product with housing affordability. The next section reviews Dependent Variable (DV) and Independent Variables (IV), the theoretical review, conceptual framework, and hypotheses development. Section three shows methods of measuring the variables. Section four covers the findings and outcome of this study. Section five concludes with the support of appropriate recommendations for the future studies. 1. Literature review 2.1 Review of Theory Income Effect Maslow s Hierarchy of Needs explains that people are motivated to meet certain needs and some needs has more priority over others (Maslow, 1943). He categorized human needs into 5 distinct levels physiological, safety, belonging, esteem and self-actualisation needs (McLeod, 2016). The purchase of housing for own consumption can be categorized as fulfilling one s physiological need as it provides shelter. As such, housing as a whole; regardless of type and price range; can be categorized as normal goods in which income effect will show positive relationship between income and quantity of housing demanded. As income rise drives demand in housing which in turn pushes the housing price upwards. Should income increase alone, the 2513

4 housing affordability ratio decreases, the housing price is more affordable to the people. However, when housing price follows the upward movement, the housing affordability ratio may increase, decrease or remain the same depending on the percentage of change on both income and housing price. According to Flavin and Yamashita, people s demand for residential housing is primarily determined by the price of such property, income growth rate, household formation rate as well as population growth (Flavin and Yamashita, 2002). Cost of housing, interest rates, availability of credit as well as social trends influence the demand of housing (Piscetek, 2013). San Ong (2013) provides empirical results that the gross domestic product (GDP), population and rate of property gain tax are the key determinants of housing prices. Green and Lee (2016) recognized that higher population growth and urbanization will lead to higher demand in housing. Low interest rate, credit availability at ease and low property price lower the overall cost of housing, which will stimulate the demand of housing. At the same time, conducive economic environment would stimulate demand for housing be it for consumption or investment purposes (Beltratti and Morana, 2010). Housing supply is affected by a few factors namely land availability, building costs, regulatory barriers, and the time required for planning, development, and building. Land availability has a positive relationship to housing supply, while the remaining factors that serve as barrier to housing supply possess negative relationships. The supply curve is usually steeper than the demand curve in the short-run, this is because the supply of housing is relatively inelastic due to time lags in construction and land release (Grimes et al., 2013). An intervention of factors influencing supply as a whole will cause the supply curve to shift to the left or to the right. For instance, an increase in land release will lead to a rightward shift in the supply curve, resulting in an increase in supply for housing. In the short run, a higher housing demand, creating an upward pressure on housing price. As the housing supply is not able to quickly respond to increases in demand, the equilibrium price becomes higher. This results in a higher housing price. In the long-run, housing supply will respond to cope with the demand, as such, the supply curve will shift rightwards, lowering the equilibrium price thus achieving a lower housing price. The term affordability has been used too frequently addressing disparate issues such as the distribution of housing prices, income distribution, capacity and accessibility of households to credit, housing quality, government policies on housing, other factors affecting housing supply and demand (Gabriel et al., 2005). MacLennan and Williams have done a great job in defining 2514

5 affordability which explains as such it is concerned with securing some given standard of housing (or different standards) at a price or rent which does not impose, in the eyes of some third party (usually government), an unreasonable burden on household incomes (Maclennan and Williams, 1990). On the other hand, Whitehead defines affordability as the relationship between housing expenditure and household income, and that they seek to establish a standard in respect of which the amount of income spent on housing is deemed unaffordable. The standard can be defined in terms of absolute residual income once housing costs have been met, or as a ratio measure specifying the acceptable proportion to be spent on housing (Whitehead, 1991). The straightforward calculation of housing costs and income ratios; which has been commonly adopted in the U.S.; shall continue to be viewed as an appropriate first step in deriving the cost component of housing affordability, before considering household composition and spatial variation. Housing affordability measured through housing expenditure-to-income ratio; has been deemed as an appropriate indicator of a population s ability to pay for housing (Hulchanski, 1995). It was mentioned that the real estate market is highly correlated with GDP as well as other macroeconomic variables. Valadez found that there is a relationship between U.S. Housing Price Index and Real GDP whereby a quarterly change in the former may yield a quarterly change in the latter (Valadez, 2011). Case et al. are on the same standpoint, stating that global property markets are largely correlated with GDP effects. It was shown that international housing price co-movements are at least partially explained by common exposure to global business cycles (Case et al., 2000). Ong has mentioned in his research that the house price index in Malaysia has witnessed significant growth in the past decade due to economic development measured in real GDP rate (Ong and Chang, 2013). When real GDP rate increases, people are optimist about the economic condition, thus increase in housing demand and housing price as well (Piazzesi and Schneider, 2009) However, another study showed that countries that experience sharp decline in GDP growth had shown the strongest increase in house prices (Tumbarello and Wang, 2010). Zhu s country specific research which attempted to explain the relationship between GDP and house prices found that both components are negatively related in Korea and Singapore but unrelated in Hong Kong (Zhu, 2006). A research in University Putra Malaysia has found that variations in the GDP are significantly related to the number of terraced, semi-detached and long houses built in Sarawak (Ong and Chang, 2013). 2515

6 Inflation is referred to the substantial and persistent movement in general level of prices of which the rate is derived from changes in a country s Price Index (Makinen, 2003). A moderate level of inflation stirs economy growth. An inflation rate of 2 or 3% stimulates purchases and borrowing due to lower level of interest rate. Tsatsaronis and Zhu (2004) shows that inflation contribute to 66-90% of the total price variation. They explained that inflation might have both positive and negative effect on housing price. The former is elaborated with the reason of residential real estate as better investment vehicle to hedge against inflation risk over bonds and equities. Therefore, a higher inflation will lead to higher demand for housing, thus increasing the price of such. The negative relationship is formed when a higher inflation increases the real value of mortgage repayment, decreasing the demand for housing, hence negatively impact house prices (Debelle, 2004). Rogers (2001) states that there was a negative relationship between inflation rate and housing price in European countries between 1999 and Brunnermeier and Julliard s (2008) findings are in line with Roger s. They explained it as people s wrong understanding on the nature of real and nominal interest rates, assuming that a decrease in inflation is due to a decline in the real interest rate which consequently underestimate the real cost of future mortgage payments. As a result, they cause an upward pressure on housing prices when inflation decreases (Brunnermeier and Julliard, 2008). Feldstein (1983) research showed that when there is an increase in the expected rate of inflation, there will be an immediate increase in the relative price of land or house. However, Teck-Hong (2010) showed that inflation is not a significant determinant of housing price. However, he concludes that it may be too shallow to reject inflation rate as an important influencing factor based on one set of sample. Higher interest rates could potentially lead to repercussions for homeowners and the economy where a significant decline in house prices observed. This phenomenon will be more significant if house prices are already viewed as overvalued. A higher interest rates indicates a higher mortgage repayments where the house purchaser can no longer borrow more. This will dampen the housing demand, bringing down the housing prices (Gan and Hill, 2009). According to (Tumbarello and Wang, 2010) in the Australian market, a decline in mortgage interest rates in 2008/2009 has increased residents ability to pay for housing expenditure. Other researchers such as Craig and Hua (2011) have also noted that the decline of interest rate will bring forth the increase in housing prices. However, Campbell and Mankiw (1989) showed that movements observed in consumption cannot be explained as a rational response to movements in real interest rates. Having said that, 2516

7 the justification on changes in housing demand due to interest rate may not stand. Khan et al. (2014) believed that there is an indirect positive relationship between interest rate and household income. Real effective exchange rate is commonly used in researches and journals. It is defined as weighted average of nominal exchange rates that have been adjusted for relative price differential between the domestic and foreign countries. Since it has taken price differential and inflation into consideration, it is believed to be a better indicator of the currency competitiveness of the country. A rise in the level of index indicates appreciation of currency and vice-versa (Kumar, 2012). Glindro et al. (2011) concluded in their research that real effective exchange rate appreciation is expected to be positively related to property market prices. An exchange rate appreciation is usually associated with housing booms in Asia where there is substantial demand from foreign direct investment (FDI). More investment is pooled in as the currency strengthens, investors are expecting better returns. This is supported by Meidani s finding with her fellow researchers too (Meidani et al., 2011). On the other hand, according to Abelson et al, the demand for housing is boosted at weak exchange rate. It is believed that foreign investors are able to convert more currency into that country at this point to invest in the relatively safe property market. Investors common principle of buy low sell high is applied here as they await the time where the currency strengthens in future (Abelson and Chung, 2004). Mallick and Mahalik (2015) mentioned that there is no significant relationship between real effective exchange rate and housing price, but FDI will influence housing price positively. However, we understand that exchange rate has direct impact on FDI, hence, we have reason to believe that real effective exchange rate does influence housing price, either positively or negatively. Looking at the other element of housing affordability, exchange rate does affect household income. (Zakaria, 2013) found significant relationship between exchange rate volatility and trade. From Garcia s (1999) model, it is observed that more inequality will depreciate the real exchange rate through the effect of Samuelson-Balassa. Recent finding in the University of Texas has made a statement that exchange rates are related to the domestic pay distribution because of the impact on export and non-export sectors upon currency value fluctuations. Domestic currency revenue in export-oriented sectors increases when its currency depreciates and vice versa. Currency depreciations are accompanied by a rise in inequality levels whilst currency appreciations are associated with falling pay inequality (Rossi and Galbraith, 2016). 2517

8 There are several factors that affect the demand and supply of housing. However, our emphasis is on the macroeconomic factors which is being categorised as a factor influencing the housing demand. As we explore through the literatures, it is undeniable that the macroeconomic factors identified as our explanatory variables in this research have significant relationship with housing price. At the same time, a handful of researches have also mentioned their impact on household income. With that, we are interested to look into the net effect of each variables on housing affordability instead of respective elements separately. Household income is sometimes being understood as an indicator of borrowing capacity of a household which is referred to accessibility to borrowings or credit. This in turn has a positive influence on housing demand. The higher the household income, the higher the borrowing capacity, the lower the housing affordability ratio, the higher the affordability among households. Overall, literature reveals that there is established relationship between GDP and housing price. There is mixed findings on the relationship between inflation and interest rate and housing price. There has been mixed finding on the relationship between exchange rate and housing demand. The conceptual framework of this study is shown in Figure 1. Figure 1: Conceptual framework Since GDP comprises household consumption, changes in household consumption directly affects a nation s GDP result. Implying the Income Effect, a higher income will lead to a higher demand on goods and services, hence, increasing the demand for housing which in turn will impact the housing price as well. Keynes Multiplier Effect explains that a higher demand for goods leads to higher employment rate at the supplier side, which will result in a higher income to pay wages that can be used for consumption. In the short run, GDP s effect on housing price should be more significant than its effect on household income, decreasing the housing affordability. In the long run, the housing affordability might increase as household income 2518

9 picks up its momentum. Therefore, GDP is believed to possess a net negative relationship with housing affordability since it affects both the numerator and denominator of the function. Inflation rate is believed to represent the purchasing power of a nation which can be translated into buying demand. As inflation rate increases, the purchasing power decreases, the real value of mortgage repayment increases, decreasing the demand for housing, hence negatively impact house prices (Debelle, 2004). We believe that Debelle s finding will have higher likelihood of occurrence in Malaysia instead of Debelle (2004), which stated that housing price will increase as demand decrease. Therefore, we believe that inflation have a net positive relationship with housing affordability since we did not come across any empirical evidence that shows significant relationship between inflation and household income. Interest rate affects the borrowing and saving pattern of a nation. A higher interest rates also means a higher mortgage repayments where the house purchaser can no longer borrow more. According to Gan and Hill (2009), this will exert downward pressure to housing prices as demand lowers. On the other hand, saving rate would increases as savers believe that real rate of return on deposit would be positive and shall increase household income in future (Khan et al., 2014). With that, we believe that there will be a net positive relationship between interest rate and housing affordability regardless if it is in view of short run or long run. This is because interest rate influences housing price negatively but household income positively. Exchange rate plays an important role in ensuring a healthy economic growth be it from the perspective of international trade or financial sector. Glindro et al. (2011) had found that an exchange rate appreciation is usually associated with housing booms in Asia where there is substantial demand from foreign direct investment (FDI). We are in line with this perspective since Malaysia is part of the emerging market where FDI plays an important role. This is finding is supported with recent evidence of MYR2,255millions of FDI outflow in 3rd quarter of 2016 and a dip in housing price as Ringgit depreciates against U.S. Dollar at a 18 years low since Asian Financial Crisis (Malaysia Foreign Direct Investment, n.d.). At the same time, Rossi and Galbraith s (2016) research had showed that domestic currency revenue in export-oriented sectors increases when its currency depreciates. Since Malaysia has a positive net export amount, we believe that the same relationship applies (OEC-Malaysia, n.d.). With that, we believe that there will be a net negative relationship between exchange rate and housing affordability regardless if it is in view of short run or long run. This is because exchange rate influences housing price positively but household income negatively. 2.4 Hypothesis Development Based on the explanation earlier, we propose hypothesis for respective variables as follows: H1: There is negative relationship between GDP and housing affordability H2: There is positive relationship between inflation rate and housing affordability 2519

10 H3: There is positive relationship between interest rate and housing affordability H4: There is negative relationship between exchange rate and housing affordability 3. Methodology A total of 44 observations are collected by using eleven years quarterly data (from 2005Q1 to 2015Q4). This study uses the quantitative approach to conduct a non-experimental study the relationship between the determined variables. This study is using correlational research design which is the most relevant design to find out the correlation and covariation among the variables, as in what degree independent and dependent variables would relate to one another. Five dependent and independent variables data are collected with quarterly data from year 2005 first quarter to year 2015 fourth quarter, meaning that 44 observations are collected in this study. Variables Inflation Rate (CPI) Exchange Rate Interest Rate GDP Growth Rate Sources Bank Negara Malaysia Investing.com BaseRate.my The World Bank Housing affordability Dependent variable Is calculated by using formula of: Median house price/median household income Also known as Price-Income Ratio This formula is also used in the researches of Thalmann (2003), and Gan and Hill (2009) All house price is a proxy of median house price, it represents average house price of all types of residential property, data collected from JPPH. Median household income is collected from NAPIC, however it only shows median household income of year 2004, 2007, 2009, 2012, and In order to align with other variables, we derive the quarterly data by averaging the growth rate per quarter. Gross Domestic Product (GDP) One of the independent variables as it also reflects the economy of a country. According to the research of Wang and Murie (2011) and Zandi et al. (2015)), housing affordability is positively correlated to GDP growth rate. 2520

11 Interest rate Journal of Emerging Issues in Economics, Finance and Banking (JEIEFB) Base lending rate is used in this paper due to it is the main interest rate affects ability of residents in making monthly mortgage payment. Zandi et al. (2015) showed that home owners tend to focus on the change of interest rate because they have a direct impact on the housing prices and Base Lending Rate (BLR) showed positive and significant correlation towards housing price. Consumer Price Index (CPI) One of the independent variables as it affects the purchasing power of Malaysian Is calculated by using CPI data with the unit of Index (2000=100) There is no journal to support that inflation rate can cause changes in housing affordability but in the research of Tsatsaronis and Zhu (2004) proved that inflation rate and housing price is positive correlated but the impact of inflation on housing price is greater than that of housing prices on inflation. Exchange rate One of the independent variables because it is highly correlated to a nation s economy. Data used in this research is U.S. Dollar (USD) to Malaysian Ringgit (MYR). The study of Yun Joe Wong et al. (2003) have proved that there is a positive relationship between exchange rate and housing prices. However, the research of Xu and Chen (2012) showed that housing prices are affect exchange rate significant than exchange rate affect housing price. Abelson et al. (2005), Mallick and Mahalik (2015)found out that exchange rate is negatively affecting the house price. In research, target population is defined as an identified group of people projected to be the respondent of a research (Sekaran and Bougie, 2011). Our target country of research is Malaysia. According to data from the Statistics Department, Malaysia currently has 31.7 millions living on the sovereign land of 329,613 square kilometres (DEPARTMENT, 2016). There is a total of millions of employed person in Malaysia, which is about 44.95% of the total population (TRADINGECONOMICS, 2016). These are the individuals who contribute income to their households which almost always the key decision maker in housing purchase. We would like to build a model to predict the factors that influences the housing affordability phenomena in this country. Housing affordability is dependent variable while the independent variables in this research are inflation rate, exchange rate, interest rate and GDP. The period of research is quarterly data of 10 years which falls between first quarter of 2006 to fourth quarter of The total number 2521

12 of observations (n) is 40. All the obtained data will run through the EViews 8 software to generate results that show evidence and further explanation. The data analysis involves two parts Inferential Statistics and Diagnostic Tests. Inferential statistic computation is to derive the first hand result of regression while 5 diagnostic tests will be conducted after that to determine if there is any underlying problems within the regression model that affect the results from the first part. 4. Analysis and Findings Equation consisting of all dependent and independent variables is computed: HA= β0 +β1gdp+β2inflation+β3blr+β4forex+μ (1) Whereas: HA = Housing affordability GDP = Gross Domestic Product INFLATION = Inflation Rate BLR = Interest Rate FOREX = Exchange Rate The above equation is refined by transforming Housing Affordability and Interest Rate data into logarithmic form, and creating a lagged variable from the transformed Interest Rate data. The equation is as follows: lnha= β0 +β1gdp+β2inflation+β3lnblr(- 1)+β4FOREX+μ The equation should fulfill the following 7 conditions to be deemed as statistically acceptable regression model: (1) Regression line must be fitted to data. (2) At least two independent variables should be significant to explain dependent variable. (3) Joint independent variables should be able to significantly explain dependent variable. (4) R-square value should be high. (5) The residuals (μ) have no sign of autocorrelation. (6) The residuals (μ) have no sign of heteroscedasticity. (7) The residuals (μ) are normally distributed. (2) The data collected for respective variables are being imported to EViews and a residual plot result was being generated to identify outliers. Then, in order to improve the result of the regression model, the most prominent data that lied out of the limited lines were removed. On top of that, adjustment has been made to lag the logarithmic version of interest rate, after that, 34 quarters was left as observation of this research. 2522

13 Jarque-Bera (JB) test was used to test for normality of error term. Since JB test result is ; which is smaller than the critical value (20.9); H0 is not rejected and the error term are said to be normally distributed. Table 1 shows the correlation coefficient between independent variables which is generated from EViews. Table 1: Correlation MAtrix Correlation (r) GDP INFLATION LAGLNBLR FOREX GDP * INFLATION * LAGLNBLR * FOREX * *Note: Correlation between two identical variables result in , indicating a perfect correlation. The correlation between all independent variables falls into the category of weak correlation with the exception for the pair GDP & LAGLNBLR which holds a strong positive correlation of This might affect the accuracy of our model as it indicates the possibility of multicollinearity problem. As such, we proceed by computing Variance Inflation Factors (VIF) using the formula: VIF = 1 1 R 2 Table 2: Variance Inflation Factor Regression on respective independent variables GDP =β 0 +β 1INFLATION+β 2lnBLR(-1)+β 3FOREX+μ VIF= R2= INFLATION =β 0 +β 1GDP+β 2lnBLR(-1)+β 3FOREX+μ R2= lnblr(-1)=β 0 +β 1GDP+β 2INFLATION+β 3FOREX+μ R2= FOREX =β 0 +β 1GDP+β 2INFLATION+β 3lnBLR(-1) +μ R2= VIF= VIF= 1 ( ) = ( ) = ( ) = VIF= ( ) = As shown in Table 2, all the VIF results computed above falls into the category of moderate correlation, therefore, the model is considered free from multicollinearity problem. 1 Table 3 shows the result of White test, which is used to see whether the variance of the errors in a regression model is constant. Table 3: White test Heteroscedasticity Obs*R-squared Prob. Chi-square Dependent variable Resid^2 2523

14 Method Least square Variable Coefficient Std. Error T-statistic Prob. C GDP^ INFLATION^ LagLnBLR^ Forex^ R-squared F-statistic Adjusted R-squared Prob. (F-statistic) Since the p-value of Chi-square is which is higher than 0.05, H0 is not rejected, thus no sign of heteroscedasticity. Using Durbin-Watson (DW) d statistic at 5% significance points of d L and d U, k=4, n=34, the value of d U,d L,4 d U,4 d L are obtained from DW critical values table. Critical values are shown in figure 2. Figure 2: Durbin Watson critical values table 0 d L = d U = d U 4 d L 4 =2.272 =2.792 A parameter is being constructed as follows: Table 4: EViews result, Durbin Watson Statistic Dependent variable LnHA Included observations 34 after adjustments Variable Coefficient Std. Error t-statistic Prob. C GDP Inflation LagLnBLR Forex R-squared F-statistic Adjusted R-squared Prob. (F-statistic) Durbin-Watson Stat Table 4 shows Durbin-Watson statistic, d = Since d = dl (1.208), H0 is rejected, positive serial correlation is detected. Breusch-Godfrey Serial Correlation LM Test is also conducted (not shown here) and has confirmed the same problem. In order to resolve this issue, we proceed to include one lagged 2524

15 period of Housing Affordability data into (2 and re-run the same test using the estimated (3 as follows: lnha= β 0 +β 1GDP+β 2INFLATION+β 3lnBLR(-1)+β 4FOREX+β 5lnHA(- 1)+μ Table 5: Breusch-Godfrey Serial Correlation LM Test (3) F-statistic Probability F Observer R squared Probability Chi Square Dependent variable RESID Included observations 34 Variable Coefficient Std. Error t-statistic Prob. C GDP Inflation LagLnBLR Forex LagLnHA RESID(-1) RESID(-2) R-squared F-statistic Adjusted R-squared Prob. (F-statistic) Durbin-Watson Stat The new LM Test result shown in Table 5 indicates that autocorrelation problem has been resolved, where : Dustin-watson stat, d = d U (1.728) Prob Chi-square= > level of significance (0.05) Therefore, H0 is not rejected, no positive serial correlation is detected. Ramsey RESET test is being run through EViews and the result is shown in Table 6. Table 6: Ramsey RESET test result F-statistic Probability F T-statistic Probability Chi Square Dependent variable LnHA Included observations 34 Variable Coefficient Std. Error t-statistic Prob. C GDP Inflation LagLnBLR Forex Fitted^ R-squared F-statistic Adjusted R-squared Prob. (F-statistic) Durbin-Watson Stat Since the p-value of F-statistic is which is higher than 0.05, H0 is not rejected, thus the model is correctly specified. 2525

16 Table 7 shows the result of Augmented Dickey-Fuller (ADF) Unit Root test which was used to check whether a unit root is present in a time series sample. Table 7: Result of Augmented Dickey-Fuller (ADF) Unit Root test Augmented Dickey Fuller (ADF) Level Variable Constant Without Trend Constant With Trend None LnHA GDP * (1) (1) (1) INFLATION ** (4) (2) (4) LaglnBLR (1) (1) (3) Forex (1) First Difference LnHA ** ** ** GDP * ** INFLATION ** ** ** (3) (3) (3) LaglnBLR * ** (2) (2) (2) Forex ** ** ** Second Difference LnHA ** ** ** GDP ** ** ** INFLATION ** ** ** (4) (4) (4) LaglnBLR ** ** ** Forex ** ** ** Note: (i) ** and * denotes significant at 1% and 5% significance level respectively. (ii) The figure in parenthesis ( ) represents the optimum lag length selected based on Schwarz Info Criterion. According to the result in Table 7, the ADF test are unable to reject H0 of variables Housing Affordability (LNHA), Interest Rate (LAGLNBLR) and Exchange Rate (FOREX) at level form. The p-value of these three variables are more than It indicates that these variables are not stationary and have unit root. On the other hand, H₀ of variables GDP (GDP) and Inflation Rate (INFLATION) are being rejected as their p-values are less than 0.05 while their t-stat results fall on 5% and 1% level of significance respectively. This indicates that these two variables are stationary and do not have unit root at level form. 2526

17 As we proceed with first differences of the same test, H 0 of all variables are rejected with their p-values being less than 0.05 and their t-stat results fall on 1% or 5% significant level. Therefore, we can conclude that all variables are stationary and do not have unit root at first differences. Similar to the results of first differences, a test on second differences allows us to reject H₀ of all variables as their p-values are less than 0.05 and their t-stat results fall on 1% significant level. We conclude that all variables are stationary and do not have unit root at second differences. Diagnostic tests have confirmed that the regression model based on Equation 4 fulfils most of the seven assumptions of an ideal regression model. We shall proceed with inferential analysis; including both F-statistic test and T-statistic test; using the same estimated equation as follows: lnha= β0 +β1gdp+β2inflation+β3lnblr(-1)+β4forex+μ (4) Table 8: Regression results Dependent variable LnHA Included observations 34 after adjustments Variable Coefficient Std. Error t-statistic Prob. C GDP INFLATION LagLnBLR Forex R-squared F-statistic Adjusted R-squared Prob. (F-statistic) Durbin-Watson Stat The estimated equation is entered through EViews to churn out the result as shown in table 8. Since F-stat = > and p-value = , we proceed to reject H 0. This indicates that there is relationship between dependent variable, Y and independent variable(s), X. Using degree of freedom: n-k= 30 at 5% significant level the t-critical value of is obtained from the t table. Comparing the t-statistical results of each independent variables to the t-critical value of 1.697, we proceed to reject the H 0 of variables GDP (GDP), Inflation Rate (INFLATION) and Exchange Rate (FOREX). This indicates that there is relationship between dependent variable, Y and independent variable, X respectively. The alternative method of comparing p-value has confirmed the same, which also indicates that there is no relationship between dependent variable, Y and lagged value of dependent variable Interest Rate (LagLnBLR). 2527

18 Based on the result above, R-Squared shows that % of variation in Housing Affordability can be explained by all four (4) independent variables (namely GDP, Inflation Rate, Interest Rate and Exchange Rate). Adjusted R-Squared is identified as %. Since both R-Squared and adjusted R- Squared figures are above 60%, we conclude that the data is closely fitted to the regression line. The coefficient of each independent variables indicates the mean change in respond to a unit change in the dependent variable, which is referred to Housing Affordability (LNHA) in our model. The following model is an extension of (4 that includes the coefficients obtained from EViews: lnha= gdp inflation lnblr(-1) forex (5) With that, we may proceed to explain the relationship between respective independent variable and Housing Affordability as follows: Holding all other variables constant, every increase in a unit of GDP growth rate will lead to an increase of in Housing Affordability ratio. The relationship between these two variables is expected to be negative where the higher the GDP growth rate, the lower the ability to own a house, translating to a higher Housing Affordability ratio. This matches our EViews result as the coefficient for GDP appear to be a positive value. Holding all other variables constant, every increase in a unit of Inflation Rate will lead to a decrease of in Housing Affordability ratio. The relationship between Housing Affordability and Inflation Rate is expected to be positive. The higher the Inflation Rate, the higher the ability to own a house, translating to a lower Housing Affordability ratio. This matches our EViews result as the coefficient for Inflation Rate is a negative value. Holding all other variables constant, every increase in a unit of Interest Rate will lead to an increase of in Housing Affordability ratio. Housing Affordability and Interest Rate is expected to have positive relationship between each other. A higher Interest Rate is said to increase Housing Affordability, which in turn lowers the housing expenditure-to-income ratio. However, this does not match our EViews result which shows its coefficient in a positive value. Holding all other variables constant, every increase in a unit of Exchange Rate will lead to an increase of in Housing Affordability ratio. 5. Discussion and Occlusion All results attained are desirable and positive which means that the regression model possess no major problem that may render it inadequate. Table 9: Summary of Inferential Statistic Inferential Statistics F-statistic Test Results There is relationship between variables 2528

19 T-test R-Squared & Adjusted R- Squared No relationship between Housing Affordability and Interest Rate. There is relationship between Housing Affordability and other independent variables. Data is closely fitted to the regression line The result in Table 8 showed that variable of GDP is significant at 5% level because the p-value is at which is less than Therefore, it is said that GDP possesses significant relationship with Housing Affordability ratio. This result is supported by researchers like, Tsatsaronis and Zhu (2004), (Valadez, 2011), (Case et al., 2000), Piazzesi and Schneider (2009) and (Tumbarello and Wang, 2010). Piazzesi and Schneider (2009) explained that when real GDP rate increases, people are optimist about the economic condition, thus increase in housing demand and housing price as well. GDP is a measurement of national income and often being used as proxy of income and standard of living in researches. Therefore, GDP and household income should possess a positive relationship too. Nonetheless, the stronger impact on housing price should justify the net positive impact of GDP to the Housing Affordability ratio. Previous researches examined impact of macroeconomic variables on housing price and income. Researchers like Tsatsaronis and Zhu (2004), Feldstein (1983), (Frappa and Mesonnier, 2010) and Debelle (2004) have all concluded their researches indicating that there is significant relationship between inflation rate and housing price. Debelle (2004) mentioned that the negative relationship is formed when a higher inflation increases the real value of mortgage repayment, decreasing the demand for housing, hence negatively impact house prices. Brunnermeier and Julliard (2008) also explained that this negative relationship is due to wrong understanding on the nature of real and nominal interest rates, assuming that a decrease in inflation is due to a decline in the real interest rate which consequently underestimate the real cost of future mortgage payments. As a result, it causes an upward pressure on housing prices when inflation decreases. Since we have not come across any resources that indicate the relationship between inflation rate and household income, we would believe that this has resulted in a net negative effect on housing affordability ratio. This research examined the effect of macroeconomic variables such as inflation rate, exchange rate, GDP, and interest rate on housing affordability. The result showed that variable of Interest Rate is insignificant at 5% level because the p-value is at which is more than Therefore, it is said that Interest Rate does not possess significant relationship with Housing Affordability ratio. Despite of this, referring to the coefficient of the regression model, the former is observed to be positively related to the latter which is not in consistence with our expectation earlier. 2529

20 1) Interest rate might have laggard effect towards housing price. As a result, housing price will probably experience downward trend a few quarters after the interest rate adjustment takes place, thus, no direct correlation from a particular time period is observed. 2) Interest rate increment could be due to higher number of loan applications which indicate higher housing demand, where higher housing demand will contribute to higher housing price generally. However, the result shows that there is a higher chance of this result to occur randomly, which renders the relationship between these two variables insignificant. This is supported by Campbell and Mankiw s research which showed that expected real interest rates does not have significant relationship with expected changes in consumption (Campbell & Mankiw, 1989). In other words, consumers do not adjust their consumption growth in response to interest rates. As such, interest rate does not affect housing demand, neither will it affect housing price nor housing affordability ratio significantly. The significant positive relationship aligns with the findings of 3 other research papers namely from Glindro et al. (2011), and Rossi and Galbraith (2016). Glindro et al. (2011) concluded in their research that real effective exchange rate appreciation is expected to be positively related to property market prices. Without opposing what Rossi & Galbraith have showed in their research, exchange rate appreciation is believed to bring more positive impact to the numerator than denominator, which will result in a net positive effect on housing affordability ratio. This research has laid down a foundation stating that the independent variables should be taken into consideration while performing activities that is directly or indirectly related to the subject matter such as setting housing price, forecasting future housing demand, preparing policy to safeguard the welfare of low-middle income families, curbing poverty issues, or even preparing for a relevant research. The significant positive relationship between Housing Affordability ratio and GDP has shown that it is important to evaluate the economic condition of the country when measuring its residents ability to pay for housing. The greater the GDP growth, the more favourable the economic condition of the country, money supply increases in a growing economy. This will result in increase in housing price as well as household income but the latter is usually impacted after a lagged period. Such response will result in a net push on housing affordability ratio. Should the household income not pick up the momentum sooner, it will be more difficult for the residents to pay for housing. Therefore, it is important not to make assumption that a growing economy reflects a higher ability to afford housing. The significant negative relationship of Inflation rate with housing affordability ratio shows that the changes in price of goods and services in a country will have a positive direct impact 2530

21 to the nation s ability in paying for housing. This is likely to be explained by the perception of money illusion among the public where a higher inflation rate increases the real cost of mortgage payment. This will lower the demand for housing which will then bring down the housing price in general. As such, it is important to not generalise the movement on inflation with the price movement of housing within the country. This information is especially important for government and central bank, when they are dealing with inflation rate and poverty issue. An increase in inflation rate will decrease the purchasing power of the people, which may worsen poverty issue but is expected to improve housing affordability. This could be a contradicting situation which might eventually result in greater rich-poor gap. Despite general understanding and basis of numerous journals to support the significant relationship between interest rate and housing price, one still cannot assume that this will translate into the same effect on housing affordability. The process of transforming the housing price data into housing affordability might have exposed it to higher chance of random occurrence. It could also be due to geographical reason where Base lending Rate (BLR) does not relates significantly but other types of interest rate such as the financial institution lending rate might have a more significant relationship. This opens up other research opportunities for those who are interest to examine further on this area. With USD being one of the most important trading currency with Malaysia, it is within our expectation that the exchange rate will carry a significant relationship with housing affordability. However, with the influx of investment from the Chinese especially on real estate property, this keeps us wondering if the Chinese Yuan will likewise possess significant relationship to housing affordability. Looking at the result that shows net significant effect on housing affordability, it is not ideal to draw conclusion that the relationship is only with housing price, though many journals have examined the same concern. Exchange rate might somewhat affect one s income significantly especially when overseas investment is getting more accessible to the mass public. This research is prepared in hope to achieve the stated objectives however there are few unavoidable limitations occurred when conducting this research. Median household income data is needed to calculate housing affordability ratio however only incomplete household income data can be found in NAPIC website. There is only few years median household income data can be found where 2005, 2006, 2008, 2010, 2011, 2013 and 2015 data are missing. So the quarterly median household income data used in this paper is calculated by average out the growth rate and divides accordingly. Hence, the estimated 2531