## _wp13181 - Executive Summary

## Source details

**Canonical URL:** [_wp13181 - Executive Summary](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13181.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13181.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13181.pdf.json)

---

### Key stylized facts: pre-crisis boom and post-crisis collapse
- Construction shares (in percent of GDP) varied greatly across many European countries before and during the recent economic crisis.
- During 2000-08, the real estate boom in Ireland, Spain, and Cyprus was synonymous with a construction boom that boosted growth; emerging Europe experienced similar overheating in the Baltic countries and, to a lesser extent, Croatia.
- Since the crisis, most countries experienced a collapse in construction activities and a sharp decline in construction shares.
- Europe exhibited one of the largest variations in construction shares globally; Spain, Lithuania, and Iceland had pre-crisis construction shares over 10 percent of GDP.
- Some countries experienced declines of around 5 percentage points of GDP in construction shares from 2007-2011 (Ireland, Iceland, and Latvia).
- The construction industry in 2011 counted for 30.7% of industrial employment and 7% of Europe’s total employment (European Construction Industry Federation, 2012).

### Norms, fundamentals, and empirical approach
- Actual construction shares revolve around a norm determined by country-specific fundamentals in an auto-regressive, error-correcting process.
- Fundamental drivers included in the norm specifications:
  - Geography: population density; tourism intensity (tourism expenditure as percent of GDP).
  - Demographics: population growth, dependency ratio, share of urban population.
  - Economic conditions: income per capita, excess rent inflation (relative to CPI), credit conditions (interest rate or private sector credit growth), stock market performance (average daily return annualized, stock market volatility), unemployment rate.
  - Control: government capital expenditure (pct of GDP).
- Dataset: annual data for over 23 advanced economies and 25 emerging economies, spanning mid-1990 to 2011; estimations performed separately for advanced and emerging groups.
- Modeling framework:
  - Norm equation estimated via panel GLS to generate an error-correction (EC) term.
  - Short-run dynamics modeled as an AR error-correction process.
  - Dynamic estimation via Arellano and Bond (1991) dynamic panel method.

### Main empirical findings on fundamentals (norm equation results)
- Population density raises construction share in both advanced and emerging economies.
- Tourism expenditure raises construction share for both types of economies.
- Advanced economies: higher dependency ratio negatively affects construction share.
- Emerging economies: population growth is a significant positive explanatory variable.
- Economic-condition effects (both groups):
  - Higher per capita income contributes to higher construction shares.
  - Higher excess rent inflation (relative to CPI) contributes to higher construction shares.
  - More favorable credit conditions (lower interest rates or higher credit growth) contribute to higher construction shares.
  - A booming stock market (higher average daily return) and lower stock market volatility generally contribute to higher construction shares.
- Government capital expenditure (pct of GDP) is a significant positive control variable.
- Global real GDP growth or a country’s own GDP growth are generally not significant in the norm equation specifications presented.

### Adjustment dynamics and pro-cyclicality
- The change in construction share is well captured by a simple AR(1) error-correction process.
- Coefficients and interpretation:
  - Coefficients of the error-correction terms for both advanced and emerging economies are remarkably similar at 0.3, implying that for each period about a third of the previous period’s gap between the construction share norm and actual level is corrected in the current period.
  - AR(1) coefficients are reported as 0.46 and 0.3 respectively for the two types of economies.
- The adjustment pace depends on the size of the deviation from the norm and the previous period’s change in construction shares themselves.
- The gradual auto-regressive adjustment process explains the pro-cyclicality of construction shares: booms generate overshooting above the norm; recessions reverse the process and can produce undershooting.

### Country-level patterns and deviations from model predictions
- Advanced Europe:
  - Large overshooting at boom peak: Spain’s actual share exceeded its norm by close to six percentage points of GDP; Ireland’s construction share peaked in 2006 exceeding its norm by about 1.7 percentage points of GDP.
  - Model predictions of cumulative adjustment since 2008 are close to actual changes (difference less than 2 percentage points of GDP) for Austria, Norway, Sweden, Belgium, Italy, France, UK, Iceland, and France.
  - Large model–actual gaps:
    - Spain: actual decline in construction share has been more benign than predicted by the model.
    - Ireland and Cyprus: actual declines much larger than model predictions.
    - Netherlands and Switzerland: model predicted increases but actual change was very small (Switzerland) or close to zero (Netherlands).
    - Greece: actual decline larger and sharper than model predictions.
  - Medium-term deviations (2011 vs medium-term norm):
    - Above predicted medium-term norm (implying likely downward adjustment): Finland, Spain, and the United Kingdom.
    - Below predicted medium-term norm (room to increase): Ireland, Greece, the Netherlands, and Germany.
- CESEE (Central, Eastern, and Southeastern Europe):
  - Pre-crisis overshooting: Croatia, Bulgaria, Estonia, Latvia, Lithuania, and Slovak Republic exceeded their norms before the crisis; Croatia’s peak gap close to 3 percentage points of GDP.
  - Pace of post-2008 adjustment varied:
    - Lithuania and Latvia: actual declines more severe than projected.
    - Ukraine, Hungary, Serbia, and to a lesser extent Slovenia, Czech Republic, and Turkey: actual decline was not severe and contradicted predicted increases.
    - Estonia, Poland, Bulgaria: adjustment close to model predictions.
    - Croatia, Romania, Slovak Republic: actual changes smaller than or opposite to model predictions.
  - Medium-term deviations (2011 vs medium-term norm):
    - Below medium-term norm (room to increase): Ukraine, Serbia, Latvia, Lithuania, and Hungary.
    - Above medium-term norm (possible downward adjustment ahead): Slovak Republic and Romania, and to a lesser extent Croatia, Bulgaria, and Estonia.

### Aggregate and policy-relevant implications
- The improvement in construction shares, or lack thereof, will have serious implications for the speed of recovery in economic activity and for employment.
- Evidence of overshooting during the 2000–07 booms is clear in many countries (Spain, Ireland, the Baltic countries, Croatia, Romania).
- Since the crisis, reversal occurred but with non-uniform pace across countries; some countries face further adjustment risk before aligning with medium-term norms.
- Specific medium-term risks identified:
  - Construction shares likely to recover over the medium term in Greece, Iceland, and Ireland (advanced Europe), and Latvia, Lithuania, Hungary, and Ukraine (emerging Europe) when economic conditions normalize.
  - Construction shares could decline further in Spain, the United Kingdom, Romania, and the Slovak Republic.
  - For Spain in particular, the large gap between actual and medium-term norm indicates a potential decline in construction activity that would weigh heavily on economic activity and employment.

### Technical notes and robustness
- Modeling choices: two-step Engel and Granger procedure to estimate norm and generate EC term; dynamic estimation via Arellano-Bond for short-run dynamics.
- Panel GLS used to estimate norm equation to control for cross-country heterogeneity; tests reject structural homogeneity across advanced and emerging groups.
- Limitations:
  - Data availability constraints (e.g., housing price series, stock of business/housing structures) required proxies (excess rent inflation for housing price dynamics; tourism expenditure for tourism infrastructure intensity).
  - National accounts construction definitions may vary across countries and over time but panel methods mitigate heterogeneity concerns.

### Appendix highlights (sample, data sources, selected estimation results)
- Sample:
  - Advanced economies sample includes: Australia; Austria; Belgium; Canada; Cyprus; Denmark; Finland; France; Germany; Greece; Iceland; Ireland; Italy; Japan; Korea, Rep.; Netherlands; Norway; Portugal; Spain; Sweden; Switzerland; United Kingdom; United States; Turkey; Ukraine.
  - Emerging economies sample includes: Argentina; Bulgaria; Chile; China; Colombia; Croatia; Czech Republic*; Estonia*; Hungary; Indonesia; Kazakhstan; Latvia; Lithuania; Mexico; Philippines; Poland; Romania; Saudi Arabia; Serbia; Slovak Republic*; Slovenia*; South Africa; Thailand.
  - Note: "* Czech Republic, Slovak Republic, Estonia, and Slovenia ... are classified in this paper with the rest of the CESEE countries because for the majority of the period of the investigation, they are classified as emerging economies."
- Data series and sources (selected): Population — IFS; Share of urban population — OECD; Dependency ratio — World Development Index; GDP — WEO; Construction (Value added) — Haver Analytics, OECD, Eurostat; Interest rate — IFS; Private Sector Credit — IFS and Haver Analytics; CPI: Rent for housing — Haver Analytics; Tourism expenditure — World Development Index, World Bank; Income per capita (in US dollars) — WEO; Government capital expenditure — AMECO, World Development Index; Stock market index — Bloomberg, World Development Index.
- Selected estimation magnitudes (Advanced Economies, GLS estimates, coefficient (standard error)):
  - Dependency ratio (pct): -0.0980 (0.00944)***.
  - Density (Population over area, in logs): 0.503 (0.0256)***.
  - Excess rent inflation (relative to CPI): 0.0617 (0.0176)***.
  - Tourism expenditure (share of GDP): 0.295 (0.0151)***.
  - Income per capita: 0.00643 (0.00151)***.
  - Unemployment rate (pct): -0.168 (0.0115)***.
  - Interest rate (pct): -0.0342 (0.00991)***.
  - Stock market index, avg daily return annualized (pct): -0.00145 (0.000512)***.
  - Stock market volatility (mean adjusted): -0.883 (0.353)**.
  - Government capital expenditure (pct of GDP): 0.440 (0.0224)***.
  - Constant: 14.05 (0.471)***.
- Selected estimation magnitudes (Emerging Economies, GLS estimates, coefficient (standard error)):
  - Density (Population over area, in logs): -0.171 (0.0624)***.
  - Population growth (log difference): -1.250 (0.0554)***.
  - Excess rent inflation (relative to CPI): 0.0691 (0.00352)***.
  - Tourism expenditure (share of GDP): -0.216 (0.0279)***.
  - Real credit growth (log difference): 0.00575 (0.00196)***.
  - Stock market index, avg daily return annualized (pct): -0.00179 (0.000273)***.
  - Income per capita: 0.101 (0.00872)***.
  - Government capital expenditure (pct of GDP): 0.0722 (0.00785)***.
  - Constant: 3.714 (0.713)***.
- Structural difference tests (including real credit, Table A6a):
  - LR/Chi-test evidence rejects parameter constancy across advanced vs emerging groups; Type of Economy Dummy: -8.389 (1.407)***.
  - Selected coefficients (coefficient (standard error)) include: Excess rent inflation 0.0327 (0.00775)***; Tourism expenditure 0.207 (0.0133)***; Unemployment rate -0.157 (0.00801)***; Stock market volatility 1.242 (0.384)***.
- Structural difference tests (including interest rate, Table A6b):
  - Chi-test: 370.0; Prob>Chi: 20.
  - Selected coefficients (coefficient (standard error)) include: Interest rate (pct) -0.0297 (0.0108)*** in (1) and 0.0635 (0.0157)*** in (2); Excess rent inflation 0.0458 (0.0106)***; Unemployment rate -0.121 (0.0131)***; Stock market volatility 1.217 (0.592)**.
- Estimation notes: Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. Estimations using generalized least square (GLS) with no country dummies unless otherwise noted.

*Source: _wp13181 - EXECUTIVE SUMMARY*

### Executive Summary ......................................................................................................

### _wp13181 - Executive Summary

### I. Document structure and major sections
- Executive Summary ...................................................................................................................4
- I. Introduction ............................................................................................................................5
- II. Literature Review ................................................................................................................11
- III. Data Set and Main Methodology .......................................................................................12
- IV. What Do the Results Reveal? ............................................................................................16
  - A. Advanced Europe ....................................................................................................16
  - B. Central, Eastern, and Southeastern Europe .............................................................17 
- V. Conclusion ..........................................................................................................................18

### II. Tables included
- Table 1. Advanced economies: Norm Equation Estimation Results .................................................20
- Table 2. Emerging and Developing Economies: Norm Equation Estimation Results .......................21
- Table 3. Advanced Economies: Dynamic Equation EstimationRresults ...........................................25
- Table 4. Emerging and Developing Economies: Dynamic Equation Estimation Results .................25

### III. Figures included
- Figure 1. Europe: GDP Growth and Construction ................................................................................6
- Figure 2. Europe: Unemployment and Construction .............................................................................6
- Figure 3. Share of Construction in GDP, 1980–2011 ...........................................................................7
- Figure 4. Advanced Europe: Share of Construction in GDP, 1980–2011 ............................................7
- Figure 5. CESEE: Share of Construction in GDP, 1980–2011 .............................................................8
- Figure 6: Advanced Economies: Construction Share ............................................................................8
- Figure 7. Emerging Economies: Construction Share ............................................................................9
- Figure 8. Construction Share in Selected Advanced European Countries ..........................................10
- Figure 9. Selected Advanced European Countries: Construction Share .............................................22
- Figure 10. Selected CESEE European Countries: Construction Share ...............................................23
- Figure 11. Europe: Construction Share, Deviation from Norm in 2007 .............................................24
- Figure 12. Europe: Cumulative Adjustment in Construction Share ....................................................26
- Figure 13. Europe: Deviation from Medium-Term Fundamental in Construction Shares .................27

### IV. Appendices and supplementary tables
- Appendices
  - Table A1. List of Countries Included in the Sample ...............................................................28
  - Table A2. List of Data and Its Sources ....................................................................................29
  - Table A3. Heterogeneity test results ........................................................................................29
  - Table A4. Advanced Economies: Alternative Estimation for Construction Norm Equation ..30
  - Table A5. Emerging Economies: Alternative Estimation for Construction Norm Equation ..31
  - Table A6a. Structural Difference Test Results (Including Real Credit) ..................................32
  - Table A6b. Structural Difference Test Results (Including Interest Rate) ................................33
- References ................................................................................................................................34

*Source: _wp13181 - Executive Summary (PDF).*

### EXECUTIVE SUMMARY

### EXECUTIVE SUMMARY

### Key stylized facts: pre-crisis boom and post-crisis collapse
- Construction shares (in percent of GDP) varied greatly across many European countries before and during the recent economic crisis.
- During 2000-08, the real estate boom in Ireland, Spain, and Cyprus was synonymous with a construction boom that boosted growth; emerging Europe experienced similar overheating in the Baltic countries and, to a lesser extent, Croatia.
- Since the crisis, most countries experienced a collapse in construction activities and a sharp decline in construction shares.
- Europe exhibited one of the largest variations in construction shares globally; Spain, Lithuania, and Iceland had pre-crisis construction shares over 10 percent of GDP.
- Some countries experienced declines of around 5 percentage points of GDP in construction shares from 2007-2011 (Ireland, Iceland, and Latvia).
- The construction industry in 2011 counted for 30.7% of industrial employment and 7% of Europe’s total employment (European Construction Industry Federation, 2012).

### Norms, fundamentals, and empirical approach
- Actual construction shares revolve around a norm determined by country-specific fundamentals in an auto-regressive, error-correcting process.
- Fundamental drivers include:
  - Geography: population density; tourism intensity (tourism expenditure as percent of GDP).
  - Demographics: population growth, dependency ratio, share of urban population.
  - Economic conditions: income per capita, excess rent inflation (relative to CPI), credit conditions (interest rate or private sector credit growth), stock market performance (average daily return annualized, stock market volatility), unemployment rate.
  - Control: government capital expenditure (pct of GDP).
- Dataset: annual data for over 23 advanced economies and 25 emerging economies, spanning mid-1990 to 2011; estimations performed separately for advanced and emerging groups.
- Modeling framework:
  - Norm equation estimated via panel GLS to generate an error-correction (EC) term.
  - Short-run dynamics modeled as an AR error-correction process; estimation of dynamics via Arellano and Bond (1991) dynamic panel method.

### Main empirical findings on fundamentals (norm equation results)
- Population density raises construction share in both advanced and emerging economies.
- Tourism expenditure raises construction share for both types of economies.
- Advanced economies: higher dependency ratio negatively affects construction share.
- Emerging economies: population growth is a significant positive explanatory variable.
- Economic-condition effects (both groups):
  - Higher per capita income contributes to higher construction shares.
  - Higher excess rent inflation (relative to CPI) contributes to higher construction shares.
  - More favorable credit conditions (lower interest rates or higher credit growth) contribute to higher construction shares.
  - A booming stock market (higher average daily return) and lower stock market volatility generally contribute to higher construction shares.
- Government capital expenditure (pct of GDP) is a significant positive control variable.
- Global real GDP growth or a country’s own GDP growth are generally not significant in the norm equation specifications presented.

### Adjustment dynamics and pro-cyclicality
- The change in construction share is well captured by a simple AR(1) error-correction process.
- Coefficients and interpretation:
  - Coefficients of the error-correction terms for both advanced and emerging economies are remarkably similar at 0.3, implying that for each period about a third of the previous period’s gap between the construction share norm and actual level is corrected in the current period.
  - AR(1) coefficients are positive and reported as 0.46 and 0.3 respectively for the two types of economies.
- The adjustment pace depends on the size of the deviation from the norm and the previous period’s change in construction shares themselves.
- The gradual auto-regressive adjustment process explains the pro-cyclicality of construction shares: booms generate overshooting above the norm; recessions reverse the process and can produce undershooting.

### Country-level patterns and deviations from model predictions
- Advanced Europe:
  - Large overshooting at boom peak: Spain’s actual share exceeded its norm by close to six percentage points of GDP; Ireland’s construction share peaked in 2006 exceeding its norm by about 1.7 percentage points of GDP.
  - Model predictions of cumulative adjustment since 2008 are close to actual changes (difference less than 2 percentage points of GDP) for Austria, Norway, Sweden, Belgium, Italy, France, UK, Iceland, and France.
  - Large model–actual gaps:
    - Spain: actual decline in construction share has been more benign than predicted by the model.
    - Ireland and Cyprus: actual declines much larger than model predictions.
    - Netherlands and Switzerland: model predicted increases but actual change was very small (Switzerland) or close to zero (Netherlands).
    - Greece: actual decline larger and sharper than model predictions.
  - Medium-term deviations (2011 vs medium-term norm): Finland, Spain, and the United Kingdom—actual 2011 construction shares were above the predicted medium-term norm (implying likely downward adjustment); Ireland, Greece, the Netherlands, and Germany—2011 levels may be below medium-term norm (room to increase).
- CESEE (Central, Eastern, and Southeastern Europe):
  - Pre-crisis overshooting: Croatia, Bulgaria, Estonia, Latvia, Lithuania, and Slovak Republic exceeded their norms before the crisis; Croatia’s peak gap close to 3 percentage points of GDP.
  - Pace of post-2008 adjustment varied:
    - Lithuania and Latvia: actual declines more severe than projected.
    - Ukraine, Hungary, Serbia, and to a lesser extent Slovenia, Czech Republic, and Turkey: actual decline was not severe and contradicted predicted increases.
    - Estonia, Poland, Bulgaria: adjustment close to model predictions.
    - Croatia, Romania, Slovak Republic: actual changes smaller than or opposite to model predictions.
  - Medium-term deviations (2011 vs medium-term norm): Ukraine, Serbia, Latvia, Lithuania, and Hungary—2011 levels below medium-term norm (room to increase); Slovak Republic and Romania, and to a lesser extent Croatia, Bulgaria, and Estonia—2011 levels above medium-term norm (possible downward adjustment ahead).

### Aggregate and policy-relevant implications
- The improvement in construction shares, or lack thereof, will have serious implications for the speed of recovery in economic activity and for employment.
- Evidence of overshooting during the 2000–07 booms is clear in many countries (Spain, Ireland, the Baltic countries, Croatia, Romania).
- Since the crisis, reversal occurred but with non-uniform pace across countries; some countries face further adjustment risk before aligning with medium-term norms.
- Specific medium-term risks:
  - Construction shares likely to recover over the medium term in Greece, Iceland, and Ireland (advanced Europe), and Latvia, Lithuania, Hungary, and Ukraine (emerging Europe) when economic conditions normalize.
  - Construction shares could decline further in Spain, the United Kingdom, Romania, and the Slovak Republic.
  - For Spain in particular, the large gap between actual and medium-term norm indicates a potential decline in construction activity that would weigh heavily on economic activity and employment.

### Technical notes and robustness
- Modeling choices: two-step Engel and Granger procedure to estimate norm and generate EC term; dynamic estimation via Arellano-Bond for short-run dynamics.
- Panel GLS used to estimate norm equation to control for cross-country heterogeneity; tests reject structural homogeneity across advanced and emerging groups.
- Limitations: data availability constraints (e.g., housing price series, stock of business/housing structures) required proxies (excess rent inflation for housing price dynamics; tourism expenditure for tourism infrastructure intensity); national accounts construction definitions may vary across countries and over time but panel methods mitigate heterogeneity concerns.

*Source: _wp13181 - EXECUTIVE SUMMARY*

### Appendix Tables

### Appendix Tables

### List of countries in the sample
- Advanced economies: Australia; Austria; Belgium; Canada; Cyprus; Denmark; Finland; France; Germany; Greece; Iceland; Ireland; Italy; Japan; Korea, Rep.; Netherlands; Norway; Portugal; Spain; Sweden; Switzerland; United Kingdom; United States; Turkey; Ukraine
- Emerging Economies: Argentina; Bulgaria; Chile; China; Colombia; Croatia; Czech Republic*; Estonia*; Hungary; Indonesia; Kazakhstan; Latvia; Lithuania; Mexico; Philippines; Poland; Romania; Saudi Arabia; Serbia; Slovak Republic*; Slovenia*; South Africa; Thailand
- Note: "* Czech Republic, Slovak Republic, Estonia, and Slovenia which have attained advanced economy status (e.g. in the IMF’s WEO classification) are classified in this paper with the rest of the CESEE countries because for the majority of the period of the investigation, they are classified as emerging economies."

### Data series and sources (Table A2 excerpts)
- Population — IFS
- Share of urban population — OECD
- Ratio of age 25-49 — OECD
- Dependency ratio — World Development Index
- Area — CIA World Fact Book
- GDP — WEO
- Construction (Value added) — Have Analytics, OECD, Eurostat
- Interest rate — IFS (lending rate, various definitions), World Development Index
- Private Sector Credit (Nominal) — IFS and Haver Analytics (various definitions)
- Unemployment rate — WEO, Haver Analytics
- Global GDP growth — WEO
- CPI: Rent for housing — Haver Analytics
- CPI — Haver Analytics
- Tourism expenditure — World Development Index, World Bank
- Income per capita (in US dollars) — WEO
- Government capital expenditure — AMECO (Europeans Commission), World Development Index
- Stock market index — Bloomberg, World Development Index
- Stock market index volatility (mean adjusted) — Calculated (based on daily stock market index data)

### Heterogeneity test results (Table A3)
- Specification (1) in Table 1 (Advanced Economies) vs Specification (1) in Table 2 (Emerging and Developing Economies)
- Null hypothesis: No heterogeneity in panel (both columns)
- Likelihood-ratio (LR) χ2(22) and χ2(24)
- LR statistic: 462.33162.71
- Prob > χ2: 00

### Advanced Economies — Alternative estimation for construction norm equation (Table A4)
- Estimation methods reported: GLS, Fixed effect, Random Effect
- Sample size: Observations 273; Number of countries 23
- Key estimated coefficients (coefficient (standard error)):
  - Dependency ratio (pct): GLS -0.0980 (0.00944)***; Fixed effect -0.118 (0.0717); Random Effect -0.134 (0.0558)**
  - Density (Population over area, in logs): GLS 0.503 (0.0256)***; Fixed effect -0.880 (6.461); Random Effect -0.640 (0.427)
  - Excess rent inflation (relative to CPI): GLS 0.0617 (0.0176)***; Fixed effect 0.0411 (0.00757)***; Random Effect 0.0418 (0.00862)***
  - Tourism expenditure (share of GDP): GLS 0.295 (0.0151)***; Fixed effect 0.0349 (0.220); Random Effect 0.0884 (0.169)
  - Income per capita: GLS 0.00643 (0.00151)***; Fixed effect 0.0332 (0.0151)**; Random Effect 0.0299 (0.00927)***
  - Unemployment rate (pct): GLS -0.168 (0.0115)***; Fixed effect -0.268 (0.0343)***; Random Effect -0.261 (0.0492)***
  - Interest rate (pct): GLS -0.0342 (0.00991)***; Fixed effect 0.0141 (0.0586); Random Effect 0.00501 (0.0377)
  - Stock market index, avg daily return annualized (pct): GLS -0.00145 (0.000512)***; Fixed effect 0.00343 (0.00106)***; Random Effect 0.00310 (0.00112)***
  - Stock market volatility (mean adjusted): GLS -0.883 (0.353)**; Fixed effect 0.00659 (0.333); Random Effect 0.135 (0.313)
  - Government capital expenditure (pct of GDP): GLS 0.440 (0.0224)***; Fixed effect 0.342 (0.210); Random Effect 0.345 (0.188)*
  - Constant: GLS 14.05 (0.471)***; Fixed effect 3.221 (58.12); Random Effect 6.098 (3.730)
- Estimation notes: Data source Haver Analytics. Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. Estimations using generalized least square (GLS) with no country dummies.

### Emerging Economies — Alternative estimation for construction norm equation (Table A5)
- Estimation methods reported: GLS, Fixed effect, Random Effect
- Sample size: Observations 215; Number of countries 25
- Key estimated coefficients (coefficient (standard error)):
  - Density (Population over area, in logs): GLS -0.171 (0.0624)***; Fixed effect 1.004 (8.906); Random Effect -0.0193 (0.361)
  - Population growth (log difference): GLS -1.250 (0.0554)***; Fixed effect -0.0668 (0.498); Random Effect -0.235 (0.274)
  - Excess rent inflation (relative to CPI): GLS 0.0691 (0.00352)***; Fixed effect 0.0336 (0.0133)**; Random Effect 0.0338 (0.0136)**
  - Tourism expenditure (share of GDP): GLS -0.216 (0.0279)***; Fixed effect -0.0987 (0.131); Random Effect -0.109 (0.0920)
  - Real credit growth (log difference): GLS 0.00575 (0.00196)***; Fixed effect 0.00999 (0.00725); Random Effect 0.0109 (0.00622)*
  - Stock market index, avg daily return annualized (pct): GLS -0.00179 (0.000273)***; Fixed effect -0.00163 (0.000811)*; Random Effect -0.00173 (0.000849)**
  - Income per capita: GLS 0.101 (0.00872)***; Fixed effect 0.127 (0.0376)***; Random Effect 0.129 (0.0328)***
  - Government capital expenditure (pct of GDP): GLS 0.0722 (0.00785)***; Fixed effect 0.147 (0.115); Random Effect 0.106 (0.0761)
  - Constant: GLS 3.714 (0.713)***; Fixed effect 14.14 (85.67); Random Effect 4.465 (3.877)
- Estimation notes: Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. Estimations using generalized least square (GLS) with no country dummies.

### Structural difference tests (Table A6a — including real credit)
- Model columns (1) and (2); Observations 474; Number of countries 48
- Chi-test: 294.3; Prob>Chi: 20
- Reported notable coefficients (coefficient (standard error)), interaction (Dummy*) terms, and type dummy:
  - Density (Population over area, in logs): -0.118 (0.0277)*** in (1); -0.0495 (0.0968) in (2)
  - Population growth (log difference): -0.162 (0.0717)** in (1); -0.790 (0.113)*** in (2)
  - Excess rent inflation (relative to CPI): 0.0327 (0.00775)*** in (1); 0.0419 (0.0105)*** in (2)
  - Tourism expenditure (share of GDP): 0.207 (0.0133)*** in (1); -0.200 (0.0332)*** in (2)
  - Real credit growth (log difference): -0.00343 (0.00298) in (1); 0.00225 (0.00520) in (2)
  - Stock market index (in logs): 0.342 (0.0183)*** in (1); -0.0605 (0.0424) in (2)
  - Per capita Income: -0.0108 (0.00176)*** in (1); 0.0475 (0.0145)*** in (2)
  - Government capital expenditure (% of GDP): 0.0389 (0.0129)*** in (1); -0.0226 (0.0112)** in (2)
  - Unemployment rate (pct): -0.157 (0.00801)*** in (1); -0.109 (0.0142)*** in (2)
  - Stock market volatility (mean adjusted): 1.242 (0.384)*** in (1); -0.602 (0.817) in (2)
  - Selected Dummy* interaction terms with significance:
    - Dummy*Population growth (log difference): 1.537 (0.217)***
    - Dummy*Tourism expenditure (share of GDP): 0.465 (0.0456)***
    - Dummy*Stock market index (in logs): 0.606 (0.0732)***
    - Dummy*Per capita Income: -0.0540 (0.0152)***
    - Dummy*Government capital expenditure (% of GDP): 0.460 (0.0625)***
    - Dummy*Stock market volatility (mean adjusted): 2.819 (1.262)** 
  - Type of Economy Dummy: -8.389 (1.407)***
  - Constant: 2.422 (0.269)*** in (1); 7.737 (1.054)*** in (2)
- Note: Includes credit as explanatory variable. The Chi-test tests whether parameters are constant across countries of different types of economy (advanced or developing, emerging). Standard errors in parentheses. Estimations using generalized least square (GLS) with no country dummies.

### Structural difference tests (Table A6b — including interest rate)
- Model columns (1) and (2); Observations 508; Number of countries 48
- Chi-test: 370.0; Prob>Chi: 20
- Reported notable coefficients (coefficient (standard error)), interaction (Dummy*) terms, and type dummy:
  - Density (Population over area, in logs): -0.413 (0.0416)*** in (1); -0.215 (0.101)** in (2)
  - Population growth (log difference): -0.299 (0.0975)*** in (1); -0.914 (0.106)*** in (2)
  - Excess rent inflation (relative to CPI): 0.0458 (0.0106)*** in (1); 0.0431 (0.00997)*** in (2)
  - Tourism expenditure (share of GDP): 0.0731 (0.0189)*** in (1); -0.186 (0.0296)*** in (2)
  - Interest rate (pct): -0.0297 (0.0108)*** in (1); 0.0635 (0.0157)*** in (2)
  - Stock market index (in logs): 0.244 (0.0285)*** in (1); -0.0128 (0.0408) in (2)
  - Per capita Income: -0.0162 (0.00302)*** in (1); 0.0753 (0.0160)*** in (2)
  - Government capital expenditure (% of GDP): 0.00118 (0.0117) in (1); 0.00678 (0.0133) in (2)
  - Unemployment rate (pct): -0.121 (0.0131)*** in (1); -0.112 (0.0124)*** in (2)
  - Stock market volatility (mean adjusted): 1.217 (0.592)** in (1); -0.686 (0.891) in (2)
  - Selected Dummy* interaction terms with significance:
    - Dummy*Population growth (log difference): 1.465 (0.213)***
    - Dummy*Tourism expenditure (share of GDP): 0.512 (0.0416)***
    - Dummy*Interest rate (pct): -0.0920 (0.0285)***
    - Dummy*Stock market index (in logs): 0.595 (0.0722)***
    - Dummy*Per capita Income: -0.0779 (0.0167)***
    - Dummy*Government capital expenditure (% of GDP): 0.452 (0.0491)***
    - Dummy*Stock market volatility (mean adjusted): 2.800 (1.260)**
  - Type of Economy Dummy: -5.741 (1.499)***
  - Constant: 1.321 (0.472)*** in (1); 4.880 (1.166)*** in (2)
- Note: Includes credit as explanatory variable. The Chi-test tests whether parameters are constant across countries of different types of economy (advanced or developing, emerging). Standard errors in parentheses. Estimations using generalized least square (GLS) with no country dummies.

*Source: _wp13181 - Appendix Tables*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13181.pdf_