RDP 2004-08: Housing Construction Cycles and Interest Rates Appendix B: Econometric Results

Laura Berger-Thomson and Luci Ellis

October 2004

In the panel showing results for the dynamics, the numerical entries denote the sum of the coefficients on current values and lags of that variable. Where more than one lag of a variable is included, the sign of the first coefficient (current-dated or shortest lag) is shown as a subscript on the summation. The orders of the lags used are shown in parentheses alongside the reported coefficient sum. The significance levels of the first coefficient in the dynamics and of the long-run coefficients at the 10, 5 or 1 per cent levels are indicated by superscripted *, ** or †.

To conserve space and make interpretation easier, we report individual coefficients only for the long-run relationships in each equation. We report the sum of the coefficients on dynamic terms for each variable, with the range of contemporaneous or lagged variables indicated in parentheses alongside this summation. Sequences of lagged explanators can sometimes include coefficients with different signs, or where intermediate lags are not significant. We therefore also report the sign and significance of the estimated coefficient on the lowest lag of each variable as subscripts and superscripts to the sum of the coefficients.

Table B1: System Results – Australia
Equation
Dependent variable		 Supply mark-up
Δsp		 Total supply
Δhc		 Quantity demand
Δs		 House prices
Δhp
Long-run variables (t – 1-dated)
Starts (s) 0.02†   –   −0.14†   –  
Construction (hc) –   −0.17†   –   –  
Costs (cc) 0.07†   –   –   –  
Housing prices (hp) 0.02**   0.14**   –   −0.12†  
Structure prices (sp) −0.11†   −0.09   –   –  
Interest rates (i) –   −0.00   –   –  
Income (y) –   –   –   0.16†  
CPI (p) –   −0.05   –   0.012†  
Vacancy rate (v) –   –   −3.42†   −1.56**  
Household growth (hh) –   –   0.04   –  
Dynamics
Δs –   –   0.51†+ (4–6) –  
Δhc –   0.17** (1) –   –  
Δcc −0.33†− (4–5) –   –   –  
Δhp 0.19†+ (2–3) 0.19**+ (1–6) −0.46** (6) −0.24† (1)
Δsp 0.52† (2–3) 0.15† (2–6) −0.31† (1−4) 1.61† (0)
Interest rates 0.09†− (2–3) −0.21**− (2–5) −0.68† (1) −0.42† (1)
Δy –   –   −10.48†− (1–5) −0.74† (5)
Δp –   −0.60* (6) −2.43† (2) 0.55** (3)
Δv –   –   −23.28**− (1–4) 8.25† (3)
Δhh –   –   −0.12** (3) –  
GST dummy 0.08† (0) −0.13*+ (4–1) −0.11†+ (4–0) −0.30† (0)
Post-GST dummy 0.01**   –   –   –  
Constant −0.31†   0.52*   1.20†   −0.89†  
R-bar squared 0.83   0.70   0.71   0.54  
Breusch-Godfrey LM 4.24   −0.59   −0.12   3.95  
Jarque-Bera 87.61   96.15   0.54   0.36  
Notes: Estimated 1975:Q4–2003:Q3. The GST dummy takes the value 1 for the 2000:Q3. The time indicators for the GST dummy refer to leads. The post-GST dummy takes the value 1 for every observation from 2000:Q3. Instruments for endogenous variables included lags of endogenous variables, construction costs and permanent income.
Table B2: System Results – United Kingdom
Equation
Dependent variable		 Supply mark-up
Δsp		 Total supply
Δhc		 Quantity demand
Δs		 House prices
Δhp
Long-run variables (t – 1-dated)
Starts (s) 0.02   –   −0.45†   –  
Construction (hc) –   −0.54†   –   –  
Costs (cc) 0.24†   –   –   –  
Housing prices (hp) 0.14†   −0.10   –   −0.02*  
Structure prices (sp) −0.37†   0.73†   –   –  
Interest rates (i) –   −0.11†   –   –  
CPI (p) –   −0.74†   –   0.02*  
Vacancy rate (v) –   –   −4.90†   0.13  
Household growth (hh)I –   –   −0.15†   –  
Household growth (hh)II –   –   0.07   –  
Dynamics
Δs I 0.09** (4) –   –   0.17†+ (0–4)
Δs II –   –   –   0.17†+ (0–4)
Δhc –   0.92†+ (3–6) –   –  
Δcc I 0.43** (5) –   –   –  
Δhp I −0.03**− (1–6) 1.58**+ (2–6) −1.94† (4) 0.60†+ (1–3)
Δhp II −0.29** (2) 1.58**+ (2–6) –   0.60† (1–3)
Δsp I 0.37**+ (2–4) −3.59†− (1–4) 0.49** (4) –  
Δsp II 0.41† (4) −2.64†− (1–3) –   –  
Interest rates I –   −0.81† (2) −2.35† (2) −0.23† (2)
Interest rates II –   −0.40** (2) 0.62* (5) −0.23† (2)
Δy I –   –   1.244†+ (2–3) 0.93**+ (0–4)
Δy II –   –   10.00**+ (0–3) 0.93**+ (0–4)
Δp I –   0.99†+ (1–6) 1.79** (3) 0.39†+ (0–1)
Δp II –   1.29† (1) –   0.39†+ (0–1)
1979 dummy –   –   −0.31†   –  
Constant I −1.18†   13.73†   7.03†   0.10**  
Constant II −1.18†   13.73†   4.11†   0.10**  
R-bar squared 0.59   0.43   0.36   0.64  
Breusch-Godfrey LM 4.81   1.12   4.98   −3.87  
Jarque-Bera 22.02   625.84   10.42   9.33  

Notes: Estimated 1971:Q4–2003:Q2. Instruments for endogenous variables included lags of endogenous variables, construction costs and permanent income. A structural break is allowed for in 1986:Q1. Coefficients that apply only to the first segment of the data are labelled I, while those applying only in the second segment are labelled II. Dummy for 1979:Q1 to capture significant fall in starts.
Table B3: System Results – United States
Equation
Dependent variable		 Supply mark-up
Δsp		 Total supply
Δhc		 Quantity demand
Δs		 House prices
Δhp
Long-run variables (t – 1-dated)
Starts (s) 0.00   –   −0.10†   –  
Construction (hc) –   −0.19†   –   –  
Costs (cc) 0.04   –   –   –  
Housing prices (hp) 0.04*   0.22   –   −0.07†  
Structure prices (sp) −0.06**   0.05   –   –  
Interest rates (i) –   −0.02†   –   –  
Income (y) –   –   –   0.09†  
CPI (p) –   −0.25**   –   0.07†  
Vacancy rate (v) –   –   −1.53   −0.11  
Household growth (hh) –   –   0.01   –  
Dynamics
Δs –   –   −0.12 (3) –  
Δce II −0.30* (0) –   –   –  
Δhc I –   0.13†+ (1–4) –   –  
Δhc II –   0.37†+ (1–2) –   –  
Δhp I 0.58†+ (2–3) 6.42†+ (2–6) −3.08† (0) 0.33† (5)
Δhp II –   –   −1.79* (4) –  
Δsp I –   −4.15− (3–6) –   0.867†+ (0–3)
Δsp II –   −1.56* (0) –   1.27†+ (2–5)
Interest rates (i) I –   −0.08†− (2–3) −1.18† (0) −0.08† (3)
Interest rates (i) II –   –   −1.46† (2) 0.14†− (5–6)
Δy II –   –   –   −0.69†+ (0–2)
Δp I –   −6.16†− (0–4) −1.17†− (1–4) –  
Δp II –   −6.16†− (0–4) –   −0.27†− (3–4)
Δv II –   –   –   0.79** (2)
Constant I −0.29**   3.25†   1.36†   −0.34†  
Constant II −0.29**   3.25†   1.36†   −0.36†  
R-bar squared 0.47   0.51   0.14   0.53  
Breusch-Godfrey LM 0.84   4.99   4.66   2.01  
Jarque-Bera 408.20   37.21   2.22   0.48  

Notes: Estimated 1967:Q3–2003:Q2. Instruments for endogenous variables included lags of endogenous variables, construction costs and permanent income. A structural break is allowed for in 1986:Q1. Coefficients that apply only to the first segment of the data are labelled I, while those applying only in the second segment are labelled II.
Table B4: System Results – Canada
Equation
Dependent variable		 Supply mark-up
Δsp		 Total supply
Δhc		 Quantity demand
Δs		 House prices
Δhp
Long-run variables (t – 1-dated)
Starts (s) −0.02†   –   −0.42†   –  
Construction (hc) –   −0.07**   –   –  
Costs (cc) 0.09†   −0.09   –   –  
Housing prices (hp) 0.14†   −0.18**   –   −0.06†  
Structure prices (sp) −0.27†   0.22   –   –  
Interest rates (i) –   0.00**   –   –  
Income (y) –   –   –   –  
CPI (p) –   0.03   –   0.06†  
Vacancy rate (v) –   –   −5.10†   −0.32  
Household growth (hh) –   –   0.25†   –  
Dynamics
Δs 0.02† (1) –   −0.16** (4) –  
Δhc –   −0.18†+ (1–4) –   –  
Δcc −0.47*+ (2–4) −1.56† (1) –   –  
Δhp 0.37†+ (0–3) 2.29†+ (0–5) 2.17†+ (1–2) 0.31† (4)
Δsp 0.37**+ (1–4) −2.21†− (1–4) −1.42* (4) –  
Interest rates −0.00**+ (1–2) −0.01† (0) −0.01† (0) −0.00 (3)
Δy –   –   4.00† (3) 2.22**+ (0–3)
Δp –   1.49† (2) 5.57† (2) –  
Δv –   –   35.59† (1) −10.09† (3)
GST dummy 0.04†   −0.06**   –   0.10†  
GST dummy (-1) –   –   –   0.09†  
Post-GST dummy −0.01**   –   –   –  
Constant 0.41†   1.50**   2.13†   0.05*  
R-bar squared 0.74   0.63   0.52   0.36  
Breusch-Godfrey LM 1.63   −0.37   −0.17   1.68  
Jarque-Bera 0.50   0.72   3.49   0.78  

Notes: Estimated 1976:Q4–2003:Q3. GST dummy takes the value 1 for the 1990:Q1. The post-GST dummy takes the value 1 for every observation from 1990:Q1. Instruments for endogenous variables included lags of endogenous variables, construction costs and permanent income.