Literature Review

國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

16

bubbles. The following two parts then move to the framework and data section. The framework draws from Stiglitz’s (1990) theory on asset bubbles and applies the direct capitalization approach through weighted average cost of capital (WACC) to identify real estate bubbles in the four countries. In the empirical part, Vector Autoregression (VAR) and Vector Error Correction (VEC) models are set up and impulse response analysis applied to investigate the relationship between the monetary policy of the ECB and property bubbles in the four countries. Finally, the findings from the analysis are discussed and a summary of the central arguments of this essay presented.

2.2. Literature Review

2.2.1. Real Estate Bubbles

The definition of a bubble is simple. A bubble describes the situation where the market price is higher than the fundamental value or not justified by fundamental factors (Stiglitz, 1990). Although there is common consent about the definition of a bubble, the measurement of the fundamental value is a difficult task. In the literature, there are several approaches on how to determine the fundamental value of real estate.

The fundamental value is derived in one approach from an equilibrium model and usually contains a number of variables such as income, employment, construction cost and interest rates. For example, Hui and Yue (2006) applies a comparative study on housing price bubbles in Hong Kong, Beijing and Shanghai and uses disposable income, the stock of vacant new dwellings and local GDP as market fundamentals.

Mikhed and Zemcik (2009) investigate whether the recently high and rapidly decreasing US house prices have been justified by fundamental factors. In their structural model of the housing market, personal income, population, house rent, stock market wealth, building costs, and mortgage rate are used as fundamentals.

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

17

Another approach is the user cost framework developed by Poterba (1984). The key concept behind this framework is that individuals base their decision of owning or renting a house on the relative cost. Thereby the cost of owning a house is calculated by adjusting the house price by the user cost of housing⁶. Under the long-run housing market equilibrium, the cost of owning a house equals the cost of renting. Individuals adjust its consumption preference until the equilibrium is reached. For instance, if the cost of owning a house is high relative to the cost of renting it, house prices must fall or rents must rise to calibrate the market equilibrium. Related to the user cost approach of Poterba (1984) is the present-value approach. This approach focuses on the present value of the expected stream of future cash-flow. The idea is that the market is in equilibrium when the housing price equals the present value of the future cash-flow. There are several studies applying different variations of this approach.

Many studies (Bjoerklund & Soederberg, 1999; Chan, Lee, & Woo, 2001; J. D.

Hamilton, 1985; Hatzvi & Otto, 2008; Smith & Smith, 2006; Xiao & Tan, 2007) model the fundamental value as the sum of the expected rental income discounted at a constant rate of return. Unlike the equilibrium model outlined above, this approach does not rely on other macroeconomic variables and thus does not need a lot of data.

Mikhed and Zemcik (2009) also focuses on the rent as determinant of the cash-flow associated with owning a house. In their analysis, however, the rent-to-price ratio is used to detect the discrepancy between fundamental and market value. Smith and Smith (2006) acknowledge that rental savings are the central factor determining the fundamental value of an owner-occupied home but they emphasize that other factors such as transaction costs, down payment, insurance, maintenance costs, property taxes,

6 Following Poterba (1984), the user cost of housing includes the foregone interest that the homeowner could have earned by investing in an alternative risk-free asset known as “opportunity cost”, the cost of property taxes, tax deductibility of mortgage interest and property taxes, maintenance cost, expected capital gain or loss and an additional risk premium to compensate homeowners of the higher risk of owning versus renting.

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

18

mortgage payments, tax savings, and the proceeds if the home is sold at some point have an impact on the cash flow. Black et. al. present (2006) another variation of the present-value approach which is based on the present value of real disposable income to determine the fundamental value of housing.

Another representation of the present-value approach is the state-space model which can be used to empirically estimate the size of housing bubbles. This time series model includes one or more unobservable (state) variables and represents its dynamics in a state equation. The model further consists of the observation equation that captures the relationship between the observed variables and unobserved variables. In a recent study, Teng et. al. (2013) build on the state-space model of Alessandri (2006) on stock market and apply it to the context of real estate markets in Taipei and Hong Kong. In their model, the state equation captures the movement of the bubble and allows the bubble price to move at time varying rates determined by the bubble price and risk-free rate in the previous period. The observation equation captures the market price and contains the rent, risk-free interest rate and the bubble as unobservable component. The combination of these two equations form the state-space model which is used to separate the deviation of the observed market price from the fundamental price into measurement error and the bubble price, whose evolution is driven by lagged bubble price and interest rate. The underlying concept of Teng et.

al.’s (2013) state-space model is the same as the simple present-value framework.

This state-space model is just another more sophisticated representation of the present-value approach which allows separating the deviation of the observed market price from the fundamental price into measurement error generated by a white noise process and the bubble price which is driven by lagged bubble price and interest rate.

Instead of the state-space representation of the present-value approach, the following

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

19

analysis applies a simple present-value approach that takes, through the weighted average cost of capital, financial leverage into account. This approach is used because it is widely applied by practitioners in real estate markets and can be readily understood by a broad audience.

2.2.2. Monetary Policy and Property Bubbles

Several papers analyze the relationship between monetary policy variables and housing prices. For instance, Ahearne et. al. (2005) examine the rise and fall of real house prices since the 1970 in 18 major industrialized countries including Ireland and Spain. The analysis shows that real house prices are typically preceded by a period of easing monetary policy. Another recent study (Adams & Füss, 2010) applies panel cointegration analysis to a sample of 15 OECD countries, including Ireland and Spain, over the period of 30 years to study the macroeconomic determinants of international housing markets. The analysis shows that besides the economic activity and construction cost, the long-term interest rate also has a significant inverse impact on housing prices. As for Ireland and Spain, this inverse relationship is significant at the 1% level.

There are also several studies in the literature specifically looking at the relationship between monetary policy variables and housing bubbles. In this regard, Tsai and Peng (2011) analyze house prices in four cities in Taiwan. The empirical result of the panel unit root and cointegration test shows that bubble-like behavior of house prices in Taiwan after 1999 was primarily related to the mortgage rates. The study concludes that expansionary monetary policy, which leads to speculations and lower mortgage rates, is the key driver for housing bubbles. Another study (Agnello & Schuknecht, 2011) looks into the determinants of housing market booms and busts in eighteen industrialized countries, including Ireland and Spain, from 1980 to 2007. The

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

20

estimates from the multinomial probit model indicate that domestic credit and interest rates have a significant impact on the probability of booms and busts. The evidence indicates that regulatory policies which slow down money and credit growth reduce boom probabilities.

Confrey and Gerald (2010) show with the example of Ireland and Spain that the introduction of the single monetary policy under the ECB and its monetary policy was a main factor causing real estate bubbles in the booming periods. Specifically, the analysis points out that the regime change brought about a substantial reduction in the real cost of capital for households in many Eurozone countries. In the case of Ireland and Spain, the reduction was particularly strong which in turn supported households in these countries to finance their investment in the housing market. The regime change not only lowered the real cost of capital but also made it much easier for the domestic financial system in Ireland and Spain to fund the dramatic investment surge of households.

In sum, monetary policy variables play a crucial role in the rise of real estate prices and the formation of bubbles. This study contributes to this literature by analyzing the relationship between the bubble formation in Greece, Ireland, Portugal and Spain and the monetary policy of the ECB, the top authority controlling money supply and key interest rates in the European monetary union. This essay further sheds light on the reasons of diverging bubble formation across these countries.