## cr18280

## Source details

**Canonical URL:** [cr18280](https://www.imf.org/-/media/files/publications/cr/2018/cr18280.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2018/cr18280.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2018/cr18280.pdf.json)

---

### WAGES AND COMPETITIVENESS IN NORWAY — Executive findings
- Wage growth was high during the 15 years before the 2014–16 oil downturn, substantially outpacing productivity growth and wages in trade partners.
- Norway avoided a large deterioration in aggregate competitiveness because of sizable terms of trade gains in oil (and oil-related industries), metals, and fisheries.
- Despite strong institutions to manage oil revenues, parts of the non-oil economy suffered during the oil boom and show signs of weakened competitiveness.
- Policy recommendations:
  - Continue the wage moderation started during the oil downturn.
  - Use the current economic upturn to start gradually tightening fiscal policy.

### Background: One country, two (interlinked) economies
- Oil sector size and spillovers:
  - Petroleum production represented 1/8 of output and 1/4 of exports in 2017.
  - Direct employment in the oil sector is 2 percent of total employment.
  - An estimated further 8 percent of employment indirectly depends on the oil sector.
  - The oil services industry accounts for 1/3 of mainland (non-oil) exports in 2017.
- Oil production trajectory:
  - Production reached a first peak by the mid-2000s and is expected to reach that peak once more as a large field comes onstream by the early 2020s.
- Economic structure and reallocation:
  - Rapid growth of oil and oil-related industries led to reallocation of resources toward these sectors, implying the non-oil economy grew less than peers even as mainland real GDP only increased in line with peers.

### The literature on Dutch Disease in Norway — transmission and mitigation
- Dutch Disease channels (Corden and Neary, 1982):
  - Spending effect: resource boom → higher demand for nontradables → higher nontradable prices and wages → tradable sector competitiveness declines.
  - Resource allocation effect: labor and capital reallocated to resource sector → higher remuneration draws labor → higher economy-wide labor costs and nontradable prices.
- Norway’s mitigating factors:
  - Strong institutions and policies: sovereign wealth fund policy saves oil revenues abroad; expected real returns revised from 4 percent to 3 percent in 2017 and only expected returns are injected gradually, delinking spending from contemporaneous oil revenues.
  - Development of related industries: expansion of oil services offset lower growth in other tradables; oil services created a large export sector and knowledge spillovers.
  - Migration: inward migration during booms buffered labor reallocation effects and softened wage pressures.
- Remaining vulnerabilities:
  - Evidence of resource movement effects: stagnation of non-oil exports and contraction of manufacturing; lack of high-tech manufacturing compared with Sweden and Finland.
  - Loose fiscal policy can exacerbate the spending channel; even the tightened fiscal rule (2017) that benchmarks spending at 3 percent of the sovereign wealth fund implies non-oil deficits of some 8 percent of mainland GDP.

### Wage and competitiveness developments (last two decades)
- Nominal wages and productivity:
  - Since 1995, nominal manufacturing wages in Norway rose by 160 percent, compared to less than 100 percent in other Nordics and less than 80 percent in Germany.
  - Since 1995, productivity in manufacturing in Norway grew by 50 percent, while other Nordic peers more than doubled (i.e., >100 percent).
  - In services, productivity increased somewhat more than in trading partners, but not enough to offset higher wage increases.
- Collective bargaining and pattern bargaining:
  - Norway’s manufacturing sector leads wage negotiations; the manufacturing wage target is applied economy-wide under the “pattern bargaining” process.
  - Manufacturing’s high wage increases—above 4 percent during 2001–13—were transmitted to follower sectors, contributing to broad-based wage growth.
- Unit labor costs and real exchange rate:
  - Non-agricultural ULC increased by more than 120 percent since 1995, compared to less than 40 percent in Nordic peers.
  - In manufacturing, ULCs increased by 70 percent since 1995.
  - The ULC-based REER in 2013 was some 70 percent more appreciated than in 1995; CPI-based REER remained roughly at or slightly below its 1995 level.
  - Only half of the ULC-based REER appreciation has been reversed during the recent oil downturn via krone depreciation.
- Sectoral outcomes:
  - Non-oil manufacturing’s value added has approximately halved in terms of mainland GDP since the late 1990s.
  - Oil-related manufacturing retained its share of value added through 2014, then was impacted by declines in global oil investment but is recovering.
- Labor market segmentation in some non-tradable sectors:
  - High wage costs have led to inward labor migration into non-tradable sectors (retail, restaurants, construction); migrants often accept lower wages and are less likely to be covered by collective agreements.
  - Union coverage in certain sectors has been decreasing rapidly.
- Recent moderation:
  - Wage growth moderated during 2014–17: manufacturing wage growth averaged less than 2 percent; other sectors reduced average wage growth from above 4 percent to 2.3 percent.
  - Reforms to the collective bargaining system effective 2014 likely contributed to this moderation.

### Fiscal policy and competitiveness implications
- Sovereign wealth fund and fiscal rule:
  - The sovereign wealth fund has grown above 300 percent of mainland GDP.
  - The 2017 tightening of the fiscal rule sets a long-run benchmark of spending 3 percent of the sovereign wealth fund.
  - Even at the 3 percent spending benchmark, non-oil deficits are expected to be some 8 percent of mainland GDP.
- Policy guidance:
  - Given the current economic up-cycle, a gradual fiscal tightening is appropriate to reduce indirect competitiveness pressures associated with expanded non-oil fiscal deficits.
  - Continued wage moderation would help restore competitiveness in exposed non-oil tradable sectors and facilitate the transition as oil production declines.

### Box 1. The 2013 Amendments to the Wage Setting Agreement — Background and Key changes
- Background:
  - In 2013, a commission examined wage formation since the introduction of the fiscal rule and the monetary policy inflation target and recommended that wage moderation would be needed in the long run.
  - The timing coincided with the 2014 sharp decline in oil prices, providing an opportunity to test the framework.
  - The framework has delivered wage moderation during the last few years.
- Key changes implemented starting in 2014:
  - 1) Setting a wage increase for all workers. Before 2014 the agreement only set blue-collar workers’ wages.
  - 2) Giving the NHO (main employers’ confederation) and LO (confederation of Trade Unions) the task to set a benchmark for wage growth. This reduced uncertainty and disputes at firm and sector level.
    - Before 2014, the wage leading agreement was set by the Federation of Norwegian industries (Norsk Industri) and the metal workers.
    - As many blue-collar workers received additional wage increases at the firm level, the scheme was fostering higher wage increases.
  - 3) More focus on benchmarking competitiveness and wage growth against trading partners.
- Effects and interpretation:
  - The amendments broadened the economy‑wide benchmark and reduced firm- and sector-level disputes over wage setting.
  - The reform reduced firm-level top-ups for blue-collar workers that previously fostered higher wage increases.
  - Implementation coincided with the 2014 oil price shock and observed wage moderation.

### House Prices and Labor Mobility in Norway — Major findings
- Nationwide and regional price movements:
  - Nationwide house prices were 55 percent higher than in 2010 (as of May 2018).
  - Annual average of real house prices in 2017 was 10 percent higher than the average observed during 2016.
  - Real house prices in Oslo now stand 60 above their 2010 level—compared to 35 percent for the whole of Norway.
  - Oslo experienced nominal house price declines of 10.5 percent in 2017; national house prices rose by 7.5 percent during January to May of 2018 (seasonally-adjusted basis).
- Regional differentials can limit regional labor mobility and slow income and productivity convergence (citing Ganong and Shoag, 2015; Hsieh and Moretti, 2017).

### Regional House Price Developments and Macroprudential Measures
- Recent dynamics:
  - Oslo: real appreciation of over 20 percent in 2016; declined by 11 percent between March and December 2017; by May 2018 above average 2016 level.
  - Oil regions: house prices still below levels observed before the 2014 oil price bust.
  - Rest of country: average real annual house price growth between 2013 and 2017 was 2.5 percent.
- Mortgage regulation changes (January 2017):
  - a debt-to-income (DTI) limit of five;
  - tightened conditions for applying an amortization requirement;
  - a lower limit for the maximum percentage of new mortgage lending in Oslo to deviate from one or more regulatory requirements.
- Macroprudential impact:
  - Regulations, especially the DTI limit, have been more binding in Oslo than in the rest of the country.
  - Staff analysis suggests macroprudential tools targeted at the housing market in Norway have contributed to improving the composition of household credit, and have had a dampening impact on growth in household credit and house prices.
- Contributing factors to regional trends:
  - Population growth outpacing residential construction in Oslo; supply response sluggish until recently.
  - Oil shock: 2014 oil price decline hit Rogaland and Hordaland harder, translating into slower or declining house prices in those areas.
  - Low interest rates: gradual reduction of the Norges Bank’s policy rate since late 2014 and low global rates reduced mortgage rates nationally.
  - Preferential property taxes: property tax rates differ across municipalities; Oslo introduced property tax in 2016 levied on a relatively small share of properties. Maximum property tax rate is 0.7 percent of a property’s value.
- Conclusion:
  - Differences in supply and demand factors across regions may not be large enough to fully explain recent house price dynamics; rapid growth in some areas raises the question whether fundamentals alone account for divergences.

### Estimation of Regional House Price Overvaluation — Method and national implications
- Methodological approach:
  - Two-stage approach to address short regional time series: first estimate national equilibrium house prices; second-stage regional regression specified in deviations from national-level averages.
  - National equilibrium estimated using approach in Geng (2018) based on a panel of 20 advanced economies over 1990:Q3–2016:Q4.
  - Two model specifications used:
    - Model 1: core fundamental factors—real household income and wealth per capita, building stock, and interest rate.
    - Model 2: larger set including policy measures and interactions (per capita income, net financial wealth, mortgage rate interacted with country-specific elasticity of housing supply s; rent control index * stock of buildings pc; tax relief index * real income pc).
- National-level model implications:
  - Both models imply a real overvaluation of house prices of 10–20 percent in periods prior to the Nordic banking crisis and before the global financial crisis.
  - For 2017:
    - Model 1 suggests a real overvaluation of national house prices of 19 percent.
    - Model 2 suggests that prices were broadly in line with fundamentals in 2017 on average.
- Selected coefficient estimates (Model 1 and Model 2):
  - real income pc, log: 1.63*** (model 1); 1.53*** (model 2)
  - real net financial wealth pc, log: 0.08*** (model 1); 0.06** (model 2)
  - stock of buildings per capita (in %): -1.08*** (model 1); -1.32*** (model 2)
  - mortgage rate: -2.76*** (model 1); -1.78*** (model 2)
  - (mortgage rate)^2: 0.08** (model 1); 0.06* (model 2)
  - tax relief index * real income pc: 0.49*** (model 2)
  - rent control index * stock of buildings pc: 0.44*** (model 2)
  - mortgage rate * s: 1.13*** (model 2)
  - real net financial wealth pc * s: -0.06* (model 2)
  - Observations: 2,080 (both models)
  - R-squared: 0.781 (both models)
  - Number of countries: 20 (both models)
  - Country fixed-effects: YES (both models)
  - Robust standard errors: YES (both models)
  - Significance notation: *** p<0.01, ** p<0.05, * p<0.1

### Regional estimation details and 2017 results (Model 2 preferred)
- Two-stage regional procedure:
  - Stage one: national equilibrium from panel regressions (Model 2 preferred; Model 1 residuals non-stationary).
  - Stage two: regional regression of deviations from national equilibrium using county-level variables; fitted regional deviations added to national equilibrium price.
- Data and estimation:
  - Annual data for 19 Norwegian counties between 2005–2016.
  - Regional variables: registered unemployment, population aged 20–50 years, residential building stock, real income, public housing, and property taxes.
  - Panel regression with county fixed effects; Discoll-Kraay standard errors to correct for serial correlation and cross-section dependence.
- 2017 regional overvaluation results (Model 2):
  - Oslo: estimated house price overvaluation of 11 percent in 2017 when applying national equilibrium prices from model 2.
  - Oslo overall overvaluation estimated at around 10–20 percent in 2017 (given higher weight on Model 2 and size of confidence bands).
  - Oil-dependent regions: estimated to be somewhat undervalued; no overvaluation in the oil regions.
  - Rest of Norway (non-oil, non-Oslo): mild overvaluation of 5–10 percent in 2017.
- Model 1 contrasts:
  - Model 1 implies much higher overvaluations across all regions in 2017: around 30 percent in Oslo, and 15–20 percent in the rest of the country.
  - Model 1 implies an average overvaluation of 19 percent at the national level.
- Confidence bands for Oslo overvaluation in 2017:
  - Model 2: 95 percent confidence band is 6–18 percent.
  - Model 1: 95 percent confidence band is 29–39 percent.

### Selected regional regression coefficient estimates (percentage-deviation specifications)
- Representative coefficients (significance shown):
  - population 20-50 years old: 0.4694***, 0.4060***, 0.4263***, 0.3911***, 0.3384***, 0.3562***
  - registered unemployment rate: -0.0673, -0.0814**, -0.0744*, -0.0633, -0.0759**, -0.0696*
  - median household real income per person: 3.3254***, 3.0113***, 3.4119***, 2.8259***, 2.5543***, 2.9057***
  - local property tax rate: -0.0033, -0.0095
  - public housing per 1000 inhabitants: 0.0701, 0.0681
  - stock of dwellings per person: -0.0039, -0.0214***, -0.0032, -0.0187***
  - (stock of dwellings per person)^2: 0.0001***, 0.0001***
- Estimation details:
  - Observations: 228
  - Number of counties: 19
  - Discoll-Kraay standard errors: YES
  - County fixed effects: YES
  - Significance legend: *** p<0.01, ** p<0.05, * p<0.1

### Internal migration analysis: regional house price differentials and labor mobility
- Context:
  - Large house price differentials can affect internal migration and hence the adjustment of employment, income, and productivity across regions.
  - Internal mobility in Norway is relatively high: annual regional migration flows have oscillated around 2.5 percent of the total population in recent years (compared to close to 1 percent of national populations in EU-15 countries on average).
  - Evidence of increasing outflows of prime-age cohorts (30–49-year-olds) from Oslo, often moving to surrounding regions for larger and more affordable dwellings.
- Migration regression setup:
  - Dependent variable: total net internal migration of persons aged 30–49 years (in percent of a county’s population of the same age cohort).
  - Explanatory variables (deviations from cross-county averages): unemployment rate, real labor compensation per person, real house price, total population. All dependent variables are lagged.
  - Expected sign: coefficient on house prices expected to be negative if higher relative real house prices reduce net migration to a county.
- Main quantitative results:
  - Real house price per sq. meter, t-1 coefficients across specifications: -0.03***, -0.03***, -0.007**, -0.006*, -0.006**, -0.007*, -0.006* (significant and negative across specifications).
  - Other sample/model details: Observations 209 (187 in one specification), Number of counties 19 (17 excluding Oslo and Akershus), R-squared values reported (e.g., 0.494, 0.675), county fixed effects included in several specifications, robust standard errors used.
  - Magnitude interpretation: for Oslo, a 25 percent increase of house prices above national average increases the net outflow of the prime age cohort by almost 10 percent.
  - Compensation and income generally not significant—likely reflecting very small variability over time and across regions in Norway.
  - The coefficient on house prices is not significant when using an aggregate house price index (suggesting segmentation of the housing market within the prime age cohort).

### Conclusions and implications
- Evidence indicates regional house price differences in Norway exceed levels implied by variation in fundamentals.
- Large deviations of house prices above equilibrium increase vulnerability to significant corrections.
- 2017 assessment:
  - Oslo real house prices exceeded equilibrium levels by around 10–20 percent (Model 2 weighted result).
  - Oil regions: house prices aligned with fundamentals (no overvaluation).
  - Non-oil, non-Oslo Norway: house price overvaluation around 5–10 percent.
- Labor mobility implications:
  - Statistically significant, quantitatively moderate negative effect of higher regional house prices on net migration of prime-age cohorts.
  - Continued divergence in house prices across regions can potentially weaken income and productivity convergence across regions going forward.

*This summary is based on material in cr18280 (Norway) as provided in the source content.*

### References _____________________________________________________________________________ 14

### cr18280 - References _____________________________________________________________________________ 14

### WAGES AND COMPETITIVENESS IN NORWAY — Executive findings
- Wage growth was high during the 15 years before the 2014–16 oil downturn, substantially outpacing productivity growth and wages in trade partners.
- Norway avoided a large deterioration in aggregate competitiveness because of sizable terms of trade gains in oil (and oil-related industries), metals, and fisheries.
- Despite strong institutions to manage oil revenues, parts of the non-oil economy suffered during the oil boom and show signs of weakened competitiveness.
- Policy recommendations highlighted:
  - Continue the wage moderation started during the oil downturn.
  - Use the current economic upturn to start gradually tightening fiscal policy.

### Background: One country, two (interlinked) economies
- Oil sector size and spillovers:
  - Petroleum production represented 1/8 of output and 1/4 of exports in 2017.
  - Direct employment in the oil sector is 2 percent of total employment.
  - An estimated further 8 percent of employment indirectly depends on the oil sector.
  - The oil services industry accounts for 1/3 of mainland (non-oil) exports in 2017.
- Oil production trajectory:
  - Production reached a first peak by the mid-2000s and is expected to reach that peak once more as a large field comes onstream by the early 2020s.
- Economic structure and reallocation:
  - Rapid growth of oil and oil-related industries led to reallocation of resources toward these sectors, implying the non-oil economy grew less than peers even as mainland real GDP only increased in line with peers.

### The literature on Dutch Disease in Norway — transmission and mitigation
- Dutch Disease channels (Corden and Neary, 1982):
  - Spending effect: resource boom → higher demand for nontradables → higher nontradable prices and wages → tradable sector competitiveness declines.
  - Resource allocation effect: labor and capital reallocated to resource sector → higher remuneration draws labor → higher economy-wide labor costs and nontradable prices.
- Norway’s mitigating factors identified in the literature:
  - Strong institutions and policies: sovereign wealth fund policy saves oil revenues abroad; expected real returns revised from 4 percent to 3 percent in 2017 and only expected returns are injected gradually, delinking spending from contemporaneous oil revenues.
  - Development of related industries: expansion of oil services offset lower growth in other tradables; oil services created a large export sector and knowledge spillovers.
  - Migration: inward migration during booms buffered labor reallocation effects and softened wage pressures.
- Remaining vulnerabilities:
  - Evidence of resource movement effects: stagnation of non-oil exports and contraction of manufacturing; lack of high-tech manufacturing compared with Sweden and Finland.
  - Loose fiscal policy can exacerbate the spending channel; even the tightened fiscal rule (2017) that benchmarks spending at 3 percent of the sovereign wealth fund implies non-oil deficits of some 8 percent of mainland GDP.

### Wage and competitiveness developments (last two decades)
- Nominal wages and productivity:
  - Since 1995, nominal manufacturing wages in Norway rose by 160 percent, compared to less than 100 percent in other Nordics and less than 80 percent in Germany.
  - Since 1995, productivity in manufacturing in Norway grew by 50 percent, while other Nordic peers more than doubled (i.e., >100 percent).
  - In services, productivity increased somewhat more than in trading partners, but not enough to offset higher wage increases.
- Collective bargaining and pattern bargaining:
  - Norway’s manufacturing sector leads wage negotiations; the manufacturing wage target is applied economy-wide under the “pattern bargaining” process.
  - Manufacturing’s high wage increases—above 4 percent during 2001–13—were transmitted to follower sectors, contributing to broad-based wage growth.
- Unit labor costs and real exchange rate:
  - Non-agricultural ULC increased by more than 120 percent since 1995, compared to less than 40 percent in Nordic peers.
  - In manufacturing, ULCs increased by 70 percent since 1995.
  - The ULC-based REER in 2013 was some 70 percent more appreciated than in 1995; CPI-based REER remained roughly at or slightly below its 1995 level.
  - Only half of the ULC-based REER appreciation has been reversed during the recent oil downturn via krone depreciation.
- Sectoral outcomes:
  - Non-oil manufacturing’s value added has approximately halved in terms of mainland GDP since the late 1990s.
  - Oil-related manufacturing retained its share of value added through 2014, then was impacted by declines in global oil investment but is recovering.
- Labor market segmentation in some non-tradable sectors:
  - High wage costs have led to inward labor migration into non-tradable sectors (retail, restaurants, construction); migrants often accept lower wages and are less likely to be covered by collective agreements.
  - Union coverage in certain sectors has been decreasing rapidly.
- Recent moderation:
  - Wage growth moderated during 2014–17: manufacturing wage growth averaged less than 2 percent; other sectors reduced average wage growth from above 4 percent to 2.3 percent.
  - Reforms to the collective bargaining system effective 2014 likely contributed to this moderation.

### Fiscal policy and competitiveness implications
- Sovereign wealth fund and fiscal rule:
  - The sovereign wealth fund has grown above 300 percent of mainland GDP.
  - The 2017 tightening of the fiscal rule sets a long-run benchmark of spending 3 percent of the sovereign wealth fund.
  - Even at the 3 percent spending benchmark, non-oil deficits are expected to be some 8 percent of mainland GDP.
- Policy guidance from the analysis:
  - Given the current economic up-cycle, a gradual fiscal tightening is appropriate to reduce indirect competitiveness pressures associated with expanded non-oil fiscal deficits.
  - Continued wage moderation would help restore competitiveness in exposed non-oil tradable sectors and facilitate the transition as oil production declines.

_This summary is based on material in cr18280 (Norway) as provided in the source content._

### Box 1. The 2013 Amendments to the Wage Setting Agreement

### Box 1. The 2013 Amendments to the Wage Setting Agreement

### Background
- In 2013, Norway appointed a commission to examine wage formation experiences since the introduction of the fiscal rule and the monetary policy inflation target.
- The committee recommendations highlighted that wage moderation would be needed in the long run.
- The timing of the recommendations and their implementation was optimal, as in 2014 the sharp decline in oil prices provided an opportunity to test the framework.
- The framework has delivered wage moderation during the last few years.

### Key changes implemented starting in 2014
- 1) Setting a wage increase for all workers. Before 2014 the agreement only set blue-collar workers’ wages. This resulted in white-collar workers’ wages growing above blue-collar workers’.
- 2) Giving the NHO (main employers’ confederation) and LO (confederation of Trade Unions) the task to set a benchmark for wage growth. This reduced uncertainty and disputes at firm and sector level.
  - Before 2014, the wage leading agreement was set by the Federation of Norwegian industries (Norsk Industri) and the metal workers.
  - As many blue-collar workers received additional wage increases at the firm level, the scheme was fostering higher wage increases.
- 3) More focus on benchmarking competitiveness and wage growth against trading partners.

### Effects and interpretation
- The amendments broadened the economy‑wide benchmark and reduced firm- and sector-level disputes over wage setting.
- The reform reduced a mechanism that previously fostered higher wage increases through firm-level top-ups for blue-collar workers.
- Implementation coincided with a large external shock (sharp decline in oil prices in 2014), providing a real-world test that coincided with observed wage moderation.
- The box concludes: “The framework has delivered wage moderation during the last few years.”

*Source: Norwegian Government (2013).*

### References

### cr18280 - References

### References
- Brede, M. and C. Henn (2018), “Finland’s Public Sector Balance Sheet: A Novel Approach to the Analysis of Public Finance,” IMF Working Paper 18/78.
- Cabezon, E. and C. Henn (2018), “Counting the Oil Money and the Elderly: Norway’s Public Sector Balance Sheet,” IMF Working Paper, forthcoming.
- Norwegian Government (2015), “Fiscal policy in an oil economy,” Official Norwegian Reports NOU 2015:9.
- Norwegian Government (2017). “Long-term Perspectives on the Norwegian Economy—A Summary of Main Points,” Meld. St. 29 (2016–2017), Report to the Storting (white paper).

### House Prices and Labor Mobility in Norway: A Regional Perspective — Major Findings (Section A: Introduction)
- Nationwide house prices were 55 percent higher than in 2010 (as of May 2018).
- Annual average of real house prices in 2017 was nonetheless 10 percent higher than the average observed during 2016 (reasons enumerated in source).
- Real house prices in Oslo now stand 60 above their 2010 level—compared to 35 percent for the whole of Norway.
- Oslo experienced nominal house price declines of 10.5 percent in 2017; national house prices rose by 7.5 percent during January to May of 2018 (seasonally-adjusted basis).
- Large regional differentials in house prices can limit regional labor mobility and slow income and productivity convergence (citing Ganong and Shoag, 2015; Hsieh and Moretti, 2017).

### Regional House Price Developments (Section B)
- Recent dynamics:
  - Oslo: real appreciation of over 20 percent in 2016; declined by 11 percent between March and December 2017; by May 2018 above average 2016 level.
  - Oil regions: house prices still below levels observed before the 2014 oil price bust.
  - Rest of country: average real annual house price growth between 2013 and 2017 was 2.5 percent.
- Mortgage regulation changes (January 2017) included:
  - a debt-to-income (DTI) limit of five;
  - tightened conditions for applying an amortization requirement;
  - a lower limit for the maximum percentage of new mortgage lending in Oslo to deviate from one or more of regulatory requirements.
- Macroprudential impact:
  - Regulations, especially the DTI limit, have been more binding in Oslo than in the rest of the country.
  - Staff analysis suggests macroprudential tools targeted at the housing market in Norway have contributed to improving the composition of household credit, and have had a dampening impact on growth in household credit and house prices.
- Factors contributing to differential regional price trends:
  - Population growth outpacing residential construction in Oslo; supply response sluggish until recently.
  - Oil shock: 2014 oil price decline hit Rogaland and Hordaland harder, translating into slower or declining house prices in those areas.
  - Low interest rates: gradual reduction of the Norges Bank’s policy rate since late 2014 and low global rates reduced mortgage rates nationally.
  - Preferential property taxes: property tax rates differ across municipalities; Oslo introduced property tax in 2016 levied on a relatively small share of properties. Maximum property tax rate is 0.7 percent of a property’s value.
- Conclusion in Section B:
  - Differences in supply and demand factors across regions may not be large enough to fully explain recent house price dynamics; rapid growth in some areas raises the question whether fundamentals alone account for divergences.

### Estimation of Regional House Price Overvaluation (Section C)
- Methodological approach:
  - Two-stage approach to address short regional time series: first estimate national equilibrium house prices; second-stage regional regression specified in deviations from national-level averages.
  - National equilibrium estimated using approach in Geng (2018) based on a panel of 20 advanced economies over 1990:Q3–2016:Q4.
  - Two model specifications used:
    - Model 1: core fundamental factors—real household income and wealth per capita, building stock, and interest rate.
    - Model 2: larger set including policy measures and interactions (per capita income, net financial wealth, mortgage rate interacted with country-specific elasticity of housing supply s; rent control index * stock of buildings pc; tax relief index * real income pc).
- National-level model implications:
  - Both models imply a real overvaluation of house prices of 10–20 percent in periods prior to the Nordic banking crisis and before the global financial crisis.
  - For 2017:
    - Model 1 suggests a real overvaluation of national house prices of 19 percent.
    - Model 2 suggests that prices were broadly in line with fundamentals in 2017 on average.
- Selected coefficient estimates from Box 2 (Geng (2018) specifications used for Norway):
  - real income pc, log: 1.63*** (model 1); 1.53*** (model 2)
  - real net financial wealth pc, log: 0.08*** (model 1); 0.06** (model 2)
  - stock of buildings per capita (in %): -1.08*** (model 1); -1.32*** (model 2)
  - mortgage rate: -2.76*** (model 1); -1.78*** (model 2)
  - (mortgage rate)^2: 0.08** (model 1); 0.06* (model 2)
  - tax relief index * real income pc: 0.49*** (model 2)
  - rent control index * stock of buildings pc: 0.44*** (model 2)
  - mortgage rate * s: 1.13*** (model 2)
  - real net financial wealth pc * s: -0.06* (model 2)
  - Observations: 2,080 (both models)
  - R-squared: 0.781 (both models)
  - Number of countries: 20 (both models)
  - Country fixed-effects: YES (both models)
  - Robust standard errors: YES (both models)
  - Significance notation: *** p<0.01, ** p<0.05, * p<0.1
- Model-implied historical overvaluation series summarized graphically (1990–2016) with both models shown.

*Source: cr18280 - References*

### Box 2. Estimating National Equilibrium House Prices in Norway (concluded)

### Box 2. Estimating National Equilibrium House Prices in Norway (concluded)

### Model specification, estimation approach, and key assumptions
- Two-stage approach:
  - Stage one: estimate national equilibrium house prices using panel regressions (two model specifications: Model 1 and Model 2). Model 2 is the preferred specification; Model 1 residuals (overvaluation time series) are non-stationary over the sample period.
  - Stage two: estimate regional equilibrium house prices by regressing deviations of regional house prices from the national equilibrium price on deviations of regional fundamentals from national averages; fitted values are added to national equilibrium prices.
- Formal regional regression specification (variables defined as percentage deviations from national averages):
  - Regional price deviation regressed on vector of regional explanatory variables in percentage deviations.
  - Regional equilibrium house price obtained by adding fitted regional deviations to national equilibrium price.
- Data and estimation:
  - Annual data for 19 Norwegian counties between 2005–2016.
  - Regional house price data from Real Estate Norway.
  - Explanatory variables: registered unemployment, population aged 20–50 years, residential building stock, real income, public housing, and property taxes from Statistics Norway.
  - Panel regression with county fixed effects.
  - Discoll-Kraay standard errors used to correct for serial correlation and cross-section dependence of error terms.
- Important implicit assumptions and limitations:
  - Variables used in the national-level regression and not available regionally (e.g., mortgage rate or financial wealth) do not vary significantly across regions; otherwise omitted variable bias may occur.
  - In absence of regional CPI, national price indicators used to obtain real values—regional differences in prices of the same goods may exist.
  - Results should be interpreted with caution due to these assumptions.

### Regional equilibrium and overvaluation results (2017)
- Summary of 2017 results using Model 2 (preferred):
  - Oslo: estimated house price overvaluation of 11 percent in 2017 when applying national equilibrium prices from model 2.
  - Oslo overall overvaluation estimated at around 10–20 percent in 2017 (given higher weight on Model 2 and size of confidence bands).
  - Oil-dependent regions: estimated to be somewhat undervalued; no overvaluation in the oil regions.
  - Rest of Norway (non-oil, non-Oslo): mild overvaluation of 5–10 percent in 2017.
- Summary of 2017 results using Model 1:
  - Much higher overvaluations across all regions in 2017: around 30 percent in Oslo, and 15–20 percent in the rest of the country.
  - Model 1 implies an average overvaluation of 19 percent at the national level.
- Confidence bands for Oslo overvaluation in 2017:
  - Model 2: 95 percent confidence band is 6–18 percent.
  - Model 1: 95 percent confidence band is 29–39 percent.

### Selected estimated coefficients from regional equilibrium regressions (percentage-deviation specifications)
- Variables and representative coefficient estimates reported in Table 1 (specifications differ slightly across columns; significance levels shown):
  - population 20-50 years old: 0.4694***, 0.4060***, 0.4263***, 0.3911***, 0.3384***, 0.3562***
  - registered unemployment rate: -0.0673, -0.0814**, -0.0744*, -0.0633, -0.0759**, -0.0696*
  - median household real income per person: 3.3254***, 3.0113***, 3.4119***, 2.8259***, 2.5543***, 2.9057***
  - local property tax rate: -0.0033, -0.0095
  - public housing per 1000 inhabitants: 0.0701, 0.0681
  - stock of dwellings per person: -0.0039, -0.0214***, -0.0032, -0.0187***
  - (stock of dwellings per person)^2: 0.0001***, 0.0001***
- Estimation details from Table 1:
  - Observations: 228
  - Number of counties: 19
  - Discoll-Kraay standard errors: YES
  - County fixed effects: YES
  - Significance legend: *** p<0.01, ** p<0.05, * p<0.1

### Internal migration analysis: regional house price differentials and labor mobility
- Context:
  - Large house price differentials can affect internal migration and hence the adjustment of employment, income, and productivity across regions.
  - Internal mobility in Norway is relatively high by international standards: annual regional migration flows have oscillated around 2.5 percent of the total population in recent years (compared to close to 1 percent of national populations in EU-15 countries on average).
  - Evidence of increasing outflows of prime-age cohorts (30–49-year-olds) from Oslo, often moving to surrounding regions for larger and more affordable dwellings.
- Migration regression setup:
  - Dependent variable: total net internal migration of persons aged 30–49 years (in percent of a county’s population of the same age cohort).
  - Explanatory variables (deviations from cross-county averages): unemployment rate, real labor compensation per person, real house price, total population. All dependent variables are lagged.
  - Expected sign: coefficient on house prices expected to be negative if higher relative real house prices reduce net migration to a county.
- Main quantitative results (Table 2 and discussion):
  - Real house price per sq. meter, t-1 coefficients across specifications: -0.03***, -0.03***, -0.007**, -0.006*, -0.006**, -0.007*, -0.006* (significant and negative across specifications).
  - Other sample/model details in Table 2: Observations 209 (187 in one specification), Number of counties 19 (17 excluding Oslo and Akershus), R-squared values reported (e.g., 0.494, 0.675), county fixed effects included in several specifications, robust standard errors used.
  - Magnitude interpretation: for Oslo, a 25 percent increase of house prices above national average increases the net outflow of the prime age cohort by almost 10 percent.
  - Result is statistically significant though moderate in magnitude; robustness checks include excluding Oslo and Akershus and alternative price measures.
  - Notes: compensation and income generally not significant—likely reflecting very small variability over time and across regions in Norway. The coefficient on house prices is not significant when using an aggregate house price index (suggesting segmentation of the housing market within the prime age cohort).

### Conclusions and implications
- Evidence indicates regional house price differences in Norway exceed levels implied by variation in fundamentals.
- Large deviations of house prices above equilibrium increase vulnerability to significant corrections.
- 2017 assessment:
  - Oslo real house prices exceeded equilibrium levels by around 10–20 percent (Model 2 weighted result).
  - Oil regions: house prices aligned with fundamentals (no overvaluation).
  - Non-oil, non-Oslo Norway: house price overvaluation around 5–10 percent.
- Labor mobility implications:
  - Statistically significant, quantitatively moderate negative effect of higher regional house prices on net migration of prime-age cohorts.
  - Continued divergence in house prices across regions can potentially weaken income and productivity convergence across regions going forward.

*IMF staff summary based on Box 2, “Estimating National Equilibrium House Prices in Norway (concluded).”*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2018/cr18280.pdf_