WHAT DOES THE HOUSE PRICE-TO-

Transcription

1 WHAT DOES THE HOUSE PRICE-TO- INCOME RATIO TELL US ABOUT THE HOUSING AFFORDABILITY: A THEORY AND INTERNATIONAL EVIDENCE (THIS VERSION: AUG 2016) Charles Ka Yui LEUNG City University of Hong Kong Edward TANG Hong Kong Shue Yan University

2 INTRODUCTION: BIG PICTURE QN: (especially after 2008 global financial crisis): 1. What is the normal asset price? What is the normal relationship between the asset price and economic fundamentals? 2. Do we have any implementable way to measure the deviations of the actual asset price from the normal counterpart? Can we establish an early warning system for the asset price? 2

3 FROM DEMOGRAPHIA (2016) 3

4 IMF uses Price-to-income ratio (PIR) to detect possible housing bubble formation. Source: IMF Global Housing Watch 4

5 INTRODUCTION Even though PIR is widely accepted as a measure of housing affordability, formal modeling of PIR is relatively rare. We construct a simple DSGE model and derive the PIR out of that model. If agents are rational and foresee these movements, they would adjust their consumptionsaving decision, labor supply decision, as well as home purchase decision accordingly. Thus, we should expect the PIR to be somehow related to the movement of economic fundamentals, e.g. the output dynamics. 5

6 INTRODUCTION Oikainen (2009), Stadelmann (2010), Agnello and Schuknecht (2011), among others Aggregate output, labor wage and monetary policy are determinants of housing price Our paper attempts to relate PIR and the economic dynamics in a simple DSGE model when all these variables are determined endogenously. 6

7 A BENCHMARK MODEL Based on Greenwood and Hercowitz (1991), Benassy (1995) and Leung (2007, 2014). Time is discrete and the horizon is infinite. Population is assumed to be constant. Goods in the economy Non-durable consumption goods: C t Business (or physical) capital: K t Residential (or housing) capital: H t Representative agent Labor hours: L t Real cash balance: M t / P t Government prints nominal money: M t 7

8 MODEL (CONTINUED) Future stock of capital (business and housing) depends on current level of stock (after depreciation) and new investment. Firms employ labor and business capital from competitive input markets to produce output. Productivity shock is a AR(1) process. Government prints money at a rate / All agents in the model economy maximize their utility or profit. Goods market and housing market clear. 8

9 A BENCHMARK MODEL The representative household in the model maximizes the expected intertemporal utility: with a simple utility functional form 9

10 A BENCHMARK MODEL 10

11 A BENCHMARK MODEL 11

12 A BENCHMARK MODEL 12

13 A BENCHMARK MODEL 13

14 SOME BASIC RESULTS 14

15 A BENCHMARK MODEL 15

16 A BENCHMARK MODEL 16

17 MAIN RESULT (FLEXIBLE PRICE CASE) 17

18 CASE II: SHORT-TERM RIGID WAGES 18

19 SHORT-TERM RIGID WAGES (CONTIN.) 19

20 SHORT-TERM RIGID WAGES (CONTIN.) 20

21 MONETARY POLICY 21

22 A MODIFIED MONEY SUPPLY RULE 22

23 BASIC RESULT: MONETARY POLICY 23

24 MAIN RESULTS: SHORT-TERM RIGID WAGES 24

25 MAIN RESULTS: SHORT-TERM RIGID WAGES (II) 25

26 EMPIRICAL EVIDENCE Sampling period: 1997Q Q4 Countries: Australia, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Netherlands, New Zealand, Norway, Spain, Sweden, UK, US. Data Real GDP and wage index: OECD Statistics Housing price index: Bank of International Settlements 26

27 EMPIRICAL EVIDENCE Equation (30): 27

28 EMPIRICAL EVIDENCE Equation (30): 28

29 EMPIRICAL EVIDENCE Equation (29): Instruments Lags of dependent variable included Lags 2 to 3 Lags 2 to 4 Lags 2 to 5 w t p ht *** *** *** y t *** *** *** GMM weights AB-n-step AB-n-step AB-n-step J-statistics Prob.(J-statistics)

30 EMPIRICAL EVIDENCE One additional question: Are wage-to-house price ratio and real output having a long-run relationship? Yes, using panel cointegrating regressions. 30

31 EMPIRICAL EVIDENCE Panel unit root tests IPS W-statistic ADF-Fisher Chi-Square PP-Fisher Chi-square w t p h, t Δ(w t p h, t ) y t Δ(y t ) *** *** *** *** *** *** Notes: H 0 : Each country follows an individual unit root process. H 1 : At least one country s process is trend stationary. Exogenous variables: individual effects, individual linear trends. ***denotes 1% significance level. 31

32 EMPIRICAL EVIDENCE Panel cointegrating regressions Dependent variable: w t+1 p h, t+1 Fully-modified OLS Dynamic OLS y t *** *** ***denotes 1% significance level. 32

33 AN INTERPRETATION: Consider the economy is initially in the steady state (or, on a balanced growth path). Now a positive productivity shock. Output yt Typically, in the short run, the house price increases more than the wage, and hence wt pt Over time, however, 2 effects. Capital goods are accumulated wt Housing units are built as well pt Hence, wt pt Output and RIP can have a positive relationship in the long run. 33

34 34

35 Sep-97 Sep-98 Sep-99 Sep-00 Sep-01 Sep-02 Sep-03 Sep-04 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-10 Sep-11 Sep-12 Sep-13 Sep-14 wt -pht Japan Actual Fitted Sep-97 Sep-98 Sep-99 Sep-00 Sep-01 Sep-02 Sep-03 Sep-04 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-10 Sep-11 Sep-12 Sep-13 Sep-14 wt -pht Netherlands Actual Fitted Sep-97 Sep-98 Sep-99 Sep-00 Sep-01 Sep-02 Sep-03 Sep-04 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-10 Sep-11 Sep-12 Sep-13 Sep-14 wt -pht New Zealand Actual Fitted Sep-97 Sep-98 Sep-99 Sep-00 Sep-01 Sep-02 Sep-03 Sep-04 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-10 Sep-11 Sep-12 Sep-13 Sep-14 wt -pht Norway Actual Fitted

36 Sep-97 Sep-98 Sep-99 Sep-00 Sep-01 Sep-02 Sep-03 Sep-04 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-10 Sep-11 Sep-12 Sep-13 Sep-14 wt -pht Spain Actual Fitted Sep-97 Sep-98 Sep-99 Sep-00 Sep-01 Sep-02 Sep-03 Sep-04 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-10 Sep-11 Sep-12 Sep-13 Sep-14 wt -pht Sweden Actual Fitted Sep-97 Sep-98 Sep-99 Sep-00 Sep-01 Sep-02 Sep-03 Sep-04 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-10 Sep-11 Sep-12 Sep-13 Sep-14 wt -pht UK Actual Fitted Sep-97 Sep-98 Sep-99 Sep-00 Sep-01 Sep-02 Sep-03 Sep-04 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-10 Sep-11 Sep-12 Sep-13 Sep-14 wt -pht US Actual Fitted

37 CONCLUDING REMARKS Housing price-to-income ratio (PIR) is widely used in the media and policy discussion. This paper constructs a simple DSGE model, which studies the endogenous dynamics of the PIR. The volatility of PIR is positively and significantly correlated to the volatility of real GDP. PIR is positively correlated to its own lag and previous real GDP. PIR and real GDP has a long run relationship. 37

38 THANKS!! 38