## cr18131

## Source details

**Canonical URL:** [cr18131](https://www.imf.org/-/media/files/publications/cr/2018/cr18131.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2018/cr18131.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2018/cr18131.pdf.json)

---

### Netherlands’ Wage Bargaining System — key findings
- Unemployment rate: fell from close to 8 percent in 2014 to 4.4 percent in last quarter of 2017.
- Nominal wage growth: around one percent; real wage growth: close to zero.
- Negotiated wages:
  - Since the crisis and up to 2017: growing in the range of 1 to 2 percent.
  - Projected: 2.2 percent in 2018 and 3.2 percent in 2019, compared with 1.5 percent in 2017.
- Wage drift:
  - Exceptionally low in recent years; turned negative in 2014 and 2016.
- Total labor cost share:
  - Including self-employed income, total labor cost share has shown a modest decline since the global financial crisis.
- Labor market slack indicators:
  - Despite unemployment falling below 5 percent, involuntary part-time employment and long-term unemployment remain elevated.
- Comparative wages:
  - Post-crisis wages have grown more slowly than most euro area countries and much more slowly than Germany.

### Institutional structure and stylized drivers
- Bargaining structure:
  - Collective bargaining predominantly at sectoral level with scope for firm/local flexibility; sectoral agreements can be made generally binding when covering a “substantial proportion” (normally 55 percent or more).
- Institutional facts:
  - Major unions: CNV, FNV, VCP.
  - Major employers’ organization: VNO-NCW (represents more than 80 percent of all medium-sized companies and nearly all large corporate institutions).
  - System penetration and stability: high penetration (> 80 percent including extension); union density currently 24 percent.
- Trends affecting bargaining:
  - Rising share of self-employed and temporary contract workers.
  - Increasing company-level agreements and individualized pay.
- Supply-side and structural drivers:
  - Trend labor productivity declined from above 2 percent to around 1 percent after the crisis.
  - Inflation remains subdued; projected to rise only gradually to the 2 percent target.
  - International spillovers: Dutch wages track trading partners, notably Germany.

### Model, data, and measurement choices
- Sample period: 1995Q1 to 2017Q1.
- Primary dependent variable: labor compensation per hour from national accounts.
- Real wages deflated by GDP deflator to align with real hourly labor productivity.
- Nominal wage measures: total labor compensation (includes employers’ social security contributions) and wages and salaries (robustness).
- Error-correction specification:
  - Difference operator covers 4 quarters.
  - Error correction term uses real wages and trend productivity 4 quarters earlier.
  - ECM includes domestic cyclical factors (D), foreign cyclical factors (F), structural factors (S), and interactions of structural variables with unemployment.

### Estimation results and interpretation
- Cointegration:
  - Coefficients on the lagged level of the real wage are statistically significant in all variants, indicating cointegration between real wages and productivity.
  - Rate of error correction: ranges from -0.5 to -0.7.
- Short-run drivers (models 1–6):
  - Significant variables: unemployment gap, expected inflation, labor productivity growth.
  - Involuntary part-time employment not significant in models 1–3.
- Foreign spillovers (models 6–9):
  - Foreign wages, especially German wages, have a significant impact on Dutch compensation growth.
- Structural change (models 10–14):
  - Rising share of temporary employment has a significant negative impact on wage growth.
  - Rising share of self-employed employment has a significant negative impact on wage growth.
  - Both temporary and self-employed shares reduce responsiveness of wage growth to the unemployment gap.
- Long-run elasticities (long-run regression, regression 7):
  - "1 percent increase in productivity is associated with 0.84 percent increase in the compensation, and the coefficient is significantly less than 1."
  - With temporary employment accounted for, compensation growth aligns more closely with trend productivity.
- Model fit:
  - Domestic wage curve model R-squared: 80 percent.
  - Adding spillovers and structural changes contributes an additional 10 percent to model fit.
  - ECM explains more than 80 percent of total variation of actual wage development.
- Decomposition of recent wage moderation (model 14):
  - Main factors: sluggish productivity growth; lower expected inflation; moderate wage growth in the euro area (including Germany); rising share of temporary workers; negative contribution from the error correction term reflecting structurally lower real wages driven by rising temporary work.

### Policy implications (wages)
- Key drivers to address: slower productivity growth; lower expected inflation; remaining slack; foreign wage spillovers; rising labor market duality/flexibility.
- Policy suggestion:
  - Reforms to harmonize employment contracts to increase flexibility while allowing greater bargaining power for more flexible employees to permit both greater flexibility and higher wages.

---

### Household income, financial net, and housing market — key facts
- Real personal disposable income (RPDI): grew by 2.0 percent per year on average in the 1990s.
- Post-GFC: sluggish RPDI growth and sharply rising unemployment contributed to housing downturn during 2007—13.
- Recent period: favorable economic and labor market trends plus rapid accumulation of financial net wealth exerted upward pressure on housing demand.
- Mortgage rates: declined substantially since 2000 to historically low levels.
- Demographics:
  - Annual population growth averaged about 0.5 percent from 1990–2016.
  - Urbanization averaged 1.1 percent over 2010–15.
  - Four major cities: average annual population growth rate of 1.2 percent during 2007–16.
  - Statistics Netherlands projects number of households will grow by some 640,000 to 8.4 million (8 percent) by 2030.
- Housing supply:
  - New-building permits stagnated since the GFC; housing completion in four major cities fell to a record low in 2014.
  - Long-run price elasticity of new housing supply estimated at about 0.2 (OECD average 0.6).

### Tax, rental regulation, and distributional effects
- Mortgage interest deductibility (MID):
  - Authorities reducing maximum tax rate for MID by 0.5 percentage points annually from 52 percent in 2013, to 38 percent in 2041 (50 percent in 2017).
  - Tax relief for housing financing remains among the most generous in the OECD.
- Rental market:
  - Rental market size: about 40 percent of total dwelling stock.
  - Social rental housing: 30 percent of total dwelling stock.
  - Rent regulation: Netherlands among the third most stringent in the OECD.
  - About 18 percent of social rentals estimated to be occupied by households earning too much relative to their rent in 2015.
- Implications:
  - Generous MID and light capital gains taxation encourage higher house and land prices, excessive leverage, and benefits concentrated among higher-income earners.
  - Strict rent control and social housing subsidies cause “locked-in” effects, long waiting lists, contraction of private rental stock (< 10 percent since the 1970s), and reduced labor mobility.

---

### Cross-country housing valuation model — specification and main estimates
- Long-run inverted demand determinants: real per capita household disposable income y; real after-tax mortgage rate morr; real per capita household net financial wealth w; housing stock per capita s.
- Augmentations: morr^2; OECD tax relief index interaction tr * y; rent control interaction rc * s; supply-elasticity interactions sr * (demand variables).
- Estimation sample: 20 advanced OECD countries, 1991:Q3–2016:Q4; country fixed effects; robust standard errors clustered at country level.
- Main coefficient ranges and interpretations:
  - Income (y, log): ~1.533–1.652 — a 1 percent increase in per capita disposable income raises long-run equilibrium house prices by 1.5–1.7 percent.
  - Mortgage rate (morr): coefficients about -1.776 to -2.759 — a one percentage point increase in real mortgage rate reduces real house prices by about 1.8–2.8 percent.
  - Mortgage rate squared (morr^2): small positive coefficients (non-linearity).
  - Household net financial wealth (w): coefficients around 0.020–0.056.
  - Housing stock per capita (s): coefficients around -1.070 to -1.322 — a one percent increase in housing stock associated with about a 1.3 percent reduction in prices.
  - Tax relief interaction (tr * y): positive ~0.351–0.487 — in countries with more generous tax relief, income shocks translate into larger price impacts (for the Netherlands tax relief implies about 0.5 percent higher house price from a one percent income increase).
  - Rent control interaction (rc * s): positive (1.156 in one specification; 0.436 in another) — rent control reduces the dampening effect of increased housing stock on prices.
- Model fit and tests:
  - Observations: 2042.
  - Adjusted R-squared: 0.853–0.867.
  - Panel cointegration test (Kao): t-Statistics -3.806, Prob. 0.0001.
  - Panel unit root test on residuals (Levin, Lin & chu): t-Statistics -2.705, Prob. 0.003.

### Dutch house-price valuation and scenarios
- Based on column (5) estimates, average house prices in the Netherlands in 2016:Q4 are about 5 percent above estimated equilibrium (ε = P - P*).
- Real mortgage rates are below their 1990–average by about 2 percent and below their 2000–average by about ½ percent.
  - If real mortgage rates unwind to their 1990–average (up by about 2 percent), equilibrium house prices would lower by up to about 5 percent.
  - If unwinding to the 2000–average (up by about ½ percent), equilibrium house prices would lower by about 2 percent.
- Implication: house prices could be up to 10 percent (or up to 7 percent) overvalued depending on mortgage rate normalization extent.

### Valuation caveats and vulnerabilities
- Valuation estimates indicative, subject to uncertainty.
- Investment purchases by high-income households and low interest rates raise equilibrium prices and can bias up long-run estimates.
- Macro-financial vulnerabilities:
  - Overvalued house prices and elevated household debt pose vulnerabilities.
  - A large correction driven by slower real income growth, sentiment reversal, or interest rate hikes could weaken household balance sheets, depress private demand, and affect corporate and bank earnings.

### Housing policy measures implemented and recommended reforms
- Measures implemented:
  - Additional bank capital buffers; LTV and DSTI caps since 2013.
  - Gradual reduction of LTV limit to 100 percent by 2018.
  - Tax exemption for gifts for housing down payments or mortgage repayments.
  - Allowing MID only for new fully amortizing loans.
  - Gradual reduction of maximum MID tax rate from 52 percent in 2013 to 38 percent in 2042 in steps of ½ percent per year.
- Recommended further reforms:
  - Accelerate phasing down of MID toward neutrality relative to other assets; implement reductions now given low interest rate environment.
  - Improve housing supply responsiveness in large cities by streamlining building aesthetics criteria, zoning plans, and permitting processes; address impediments to urban redevelopment; improve public transportation.
  - Phase out rent control and reform social housing:
    - Gradually raise rents on regulated housing toward market rates.
    - Protect vulnerable households with targeted housing allowances.
    - Expected outcomes: more efficient use of stock, larger private rental market, improved mobility.
  - Tighten macroprudential measures:
    - Gradually lower maximum LTV by at least 1 percentage point per year to no more than 90 percent by 2028 (FSC recommendation) and to 80 percent thereafter.
    - Introduce prudential ceilings on DSTI caps by income category that could not be relaxed during strong growth.
  - Consider temporarily allowing partial use of pension savings for housing down payments (examples: US 401K, Switzerland, Canada, Singapore).

---

### Health care system — design, outcomes, and bargaining dynamics
- Public spending on health and long-term care have increased to about 6 and 4 percent of GDP, respectively.
- 2006 Health Insurance Act (Zvw): managed competition, mandatory insurance for a legally defined basic package.
- Funding architecture:
  - Income-related contribution: 6.9 percent of income before taxes and social premiums for workers in regular employment (paid by employer to tax office and allocated to insurers via risk equalization fund).
  - Average direct health insurance premium: 110 euros per month on average.
  - Deductible: 385 euros per year for basic benefits.
  - Self-employed contribution: 5.654 percent of their income.
  - About 85 percent take supplemental insurance; about 40 percent receive income-related health care allowances.
- Supply-side bargaining and market design:
  - Insurers and providers bargain on premia and prices; patients can switch insurers once a year; insurers must charge community-based premia.
  - “Segment B” negotiable hospital spending raised to 70 percent of hospital budgets; “segment A” (30 percent) set nationally by NZa.
  - Stakeholder agreements set yearly ceilings for health care expenditure growth; breach triggers prorated across-the-board savings.
- Outcomes and concerns:
  - Coverage improved (uninsured down from about 200,000 to negligible).
  - Productivity gains: diminishing lengths of stay; expansion of ZBCs for routine care.
  - Premium differentiation increased; average premium decreased by 2 percent in 2006–2008.
  - Insurers improved solvency ratios to about 160 percent.
  - Evidence of hospital mergers, potential increases in hospital bargaining power.
  - Overall health spending envelope continued to increase due to volume growth and possible upcoding.
- Preliminary empirical findings (regional panel 2012–2015):
  - For annual spending ~€1,190 on average:
    - A one notch increase in insurer concentration (HHI) associated with €244 increase in total health expenditure (coefficient 244.046**, standard error (86.717)).
    - Coefficient on average hospital expenditures: 177.420** (standard error (61.178)).
    - Ratio 65 year + / total population (percent): 212.723*** (standard error (24.952)).
  - Observations: 36; R-squared: 0.939, 0.628, 0.795.
  - Caution: low number of observations and likely omitted variable bias (notably hospital sector concentration).
- Policy implications (health):
  - Continue vigilance from regulatory and monitoring agencies.
  - Monitor concentration trends in insurers and hospitals to preserve competition and limit upward pressure on prices.

---

### Climate commitments, carbon pricing, and fiscal implications
- International and national targets:
  - EU pledge: reduce GHGs by 40 percent relative to 1990 levels by 2030.
  - Netherlands target: increase GHG reduction target for 2030 to 49 percent below 1990 levels.
- Model framework and scenarios:
  - BAU projections account for previously implemented mitigation policies but not planned mitigation policies.
  - Envisioned policies reduce nationwide CO2 emissions by an estimated 26 percent below BAU in 2030; BAU emissions already 10 percent below 2005 levels.
  - Emission reductions by component (percent of BAU in 2030):
    - National policies: 9 percent.
    - ETS: 6 percent.
    - Energy efficiency policies for the ETS sector: 6 percent.
    - Energy efficiency policies for the non-ETS sector: 4 percent.
    - Vehicle emissions rate standard: 2 percent.
- Revenue impacts of envisioned policies:
  - Envisioned policies reduce revenue by 0.25 percent of GDP in 2030.
  - ETS allowance auctions assumed to raise 0.15 percent of GDP (assuming half of allowances auctioned).
- Emissions-neutral revenue-raising scenarios (holding emissions fixed):
  - Replace vehicle emissions standards with higher road fuel taxes:
    - Raises 0.4 percent of GDP; requires fuel price increase of €0.60 per liter.
  - Extend uniform carbon tax to all non-ETS emissions (removing extra road fuel tax):
    - Raises 1.1 percent of GDP; requires €170 per tonne of CO2.
  - Fully auction ETS allowances:
    - Raises additional 0.15 percent of GDP.
  - Introduce CO2 surcharge for ETS sector (removing other CO2 policies):
    - Raises 0.2 percent of GDP; requires raising CO2 price by €52 per tonne.
  - Harmonizing prices across ETS and non-ETS (keeping emissions fixed) implies economy-wide price of €136 per tonne of CO2 and loses a modest amount of revenue.
- Costs and cost-effectiveness:
  - Aggregate cost of envisioned policies: around 0.6 percent of GDP in 2030.
    - Largest source: national targets for non-ETS: 0.35 percent of GDP.
    - Vehicle emissions standard: 0.15 percent of GDP.
    - Energy efficiency policies within ETS: 0.1 percent of GDP.
  - Netting domestic environmental benefits lowers costs to 0.45 percent of GDP.
  - Pricing reforms could lower economic costs to 0.4 percent of GDP, or costs net of environmental benefits to 0.1 percent of GDP.
- Instruments and implementation precedents:
  - UK: variable carbon tax on top of ETS where tax equals difference between target combined price and ETS price.
  - France: tax planned from €31 (2017) to €65 (2020) and €86 (2022).
  - Ireland: carbon tax at €20 per tonne.
  - A Dutch ETS surcharge would not directly reduce EU emissions unless allowances withdrawn from the market.
- Complementary measures:
  - Feebates for power sector (pivot point at industry average emission rate for revenue neutrality).
  - Fiscal incentives for CCS; fiscal analogs of regulations (out-of-compliance fees).
  - Infrastructure investments (grid, CCS pipelines).
- Near-term pricing credibility:
  - Priority on establishing robust near-term carbon prices to increase credibility; current global average CO2 price about €1 per tonne; low EU ETS price example €5 per tonne.

---

### Road transport, externalities, and charging reforms
- External cost estimates (2013):
  - Gasoline total: €1.58 per liter (Dutch); €1.10 per liter (IMF).
    - Congestion: €0.88 / €0.85 per liter.
    - Accidents: €0.44 / €0.17 per liter.
    - Global warming: €0.21 / €0.08 per liter.
    - Local air pollution: €0.05 / €0.01 per liter.
  - Diesel total (LDVs, HGVs, buses): about €1.12 per liter in both studies.
    - Congestion: €0.59 / €0.77 per liter.
    - Accidents: €0.15 / €0.13 per liter.
    - Global warming: €0.24 / €0.09 per liter.
    - Air pollution: €0.14 / €0.13 per liter.
    - Road damage: €0.01 / €0.02 per liter.
- Current excises vs. efficient taxes:
  - Current gasoline excise: €0.77 per liter.
    - Efficient tax (Dutch): €1.07 per liter.
    - Efficient tax (IMF): €0.70 per liter.
  - Current diesel excise: €0.49 per liter.
    - Efficient tax (Dutch): €0.83 per liter.
    - Efficient tax (IMF): €0.77 per liter.
  - Note: diesel faces higher annual road taxes and vehicle registration fees.
- Congestion pricing and km-based charging:
  - Efficient approach: peak-period km charges varying by time and location; national system could require GPS and independent billing.
  - Local urban systems can capture most gains if comprehensive across urban centers.
- HGV charges and design:
  - Road damage charges vary with axle weight; efficient illustrative tolls:
    - €0.45 per vehicle km in rural areas.
    - €2.60 per vehicle km for peak urban driving.
  - Surrounding-country charge: roughly €0.15 per HGV km; EU cap: €0.40 per km.
  - Transitional considerations: phase in charges and coordinate with other countries; lower diesel fuel taxes to offset truck burdens while preserving feebates for low-carbon vehicles.
- Vehicle registration fees and feebates:
  - Current registration: five CO2 brackets; fixed charges €356 (lowest) to €12,593 (highest); variable charges multiply gram/km differences by escalating per-gram rates.
  - Problems: raises less revenue as vehicles become cleaner; creates bunching at bracket thresholds.
  - Proposed reform: ad valorem tax on vehicle sales paired with continuous feebate (fees/rebates proportional to difference from pivot point gram/km).
  - Feebate for HGVs:
    - Apply to in-use fleet scaled to annual km and integrated into HGV charging system.
    - Consider varying pivot point by truck class.
- PAYD insurance and accident risk:
  - PAYD insurance: payments proportional to km driven and scaled by driver/vehicle rating factors.
  - Distributional effects: low-km drivers pay less; may need tax incentives to initiate transition.
- Key numeric references preserved exactly:
  - Gasoline/diesel stockholding fee: €0.008 per liter.
  - Second-best efficient gasoline tax around €1 per liter; diesel around €0.8 per liter.
  - HGV toll illustrative values: €0.45 (rural) to €2.60 (peak urban).
  - Current surrounding-country HGV charge: roughly €0.15 per HGV km; EU cap: €0.40 per km.
  - Current registration fixed charges: €356 and €12,593.
  - Example vehicle tax comparison: zero-emission vehicle fixed fee €356; vehicle with 100 gram/km pays €2,355 (fixed €2,077 and variable €278); feebate rate examples: €20 per gram/km or €40 per gram/km.

_Italic: IMF Country Report cr18131 (excerpt)._

### 1. Netherlands’ Wage Bargaining System ____________________________________________________ 6

### Netherlands’ Wage Bargaining System

### Introduction
- Wage growth has been subdued in the Netherlands despite tighter labor market conditions.
- Unemployment rate: fell from close to 8 percent in 2014 to 4.4 percent in last quarter of 2017.
- Nominal wage growth: around one percent; real wage growth: close to zero.
- Observed phenomenon: apparent flattening of the wage Phillips curve (falling unemployment fails to generate higher wage growth).
- Comparative context: many advanced economies show a disconnect between unemployment and nominal wage growth; wage growth in the Netherlands is among the weakest in advanced EU countries.

### Key empirical findings on wages and bargaining
- Negotiated wages:
  - Since the crisis and up to 2017: growing in the range of 1 to 2 percent (slower than 2000–2007 when close to 3 percent).
  - Projected: 2.2 percent in 2018 and 3.2 percent in 2019, compared with 1.5 percent in 2017.
- Wage drift:
  - Exceptionally low in recent years; turned negative in 2014 and 2016.
  - Low wage drift persists despite high and rising firm profitability.
  - Possible contributor: compositional changes (higher share of low-skilled workers), but puzzle remains.
- Total labor cost share:
  - Including self-employed income, total labor cost share has shown a modest decline since the global financial crisis.
  - Rising share of self-employed has made self-employed income an important part of total labor income; without self-employed income the employees’ total labor cost share would have shown a declining trend after the crisis.
- Labor market slack indicators:
  - Even with unemployment falling from above 8 percent to less than 5 percent, involuntary part-time employment and long-term unemployment remain elevated, suggesting remaining slack.
- Comparative wages:
  - Netherlands’ wage growth has moved closely with, or more slowly than, Germany and other Euro Area countries; post-crisis wages have grown more slowly than most euro area countries and much more slowly than Germany.

### Background of the Netherlands’ wage bargaining system
- Bargaining structure:
  - Collective bargaining predominantly at sectoral level, with scope for firm/local flexibility.
  - Sectoral agreements can set minimum/standard terms (a “wage floor”) which employers can complement or deviate from upwards at firm level; also allow “à la carte” trade-offs between wages and working conditions.
  - Sectoral bargaining is relatively independent across sectors but wages tend not to deviate much from those set in the export-oriented industry sector.
- Institutional facts:
  - Three major unions: Christelijk Nationaal Vakverbond (CNV), Federatie Nederlandse Vakbeweging (FNV), Federation of Managerial and Professional Staff Unions (VCP).
  - Major employers’ organization: Confederation of Netherlands Industry and Employers (VNO-NCW) covering almost all sectors; represents more than 80 percent of all medium-sized companies and nearly all large corporate institutions.
  - Collective agreements are legally binding for union members; employers must offer same terms to non-union members in practice; parties can request government to make agreements generally binding if they cover a “substantial proportion” (normally 55 percent or more).
  - Penetration and stability: system characterized by high penetration (> 80 percent including extension) and relative stability; union density currently 24 percent.
  - Trends: decreasing membership rates (fewer young sign-ups), increasing share of self-employed not covered by collective agreements, rising company-level agreements and individualized pay.
- 2013 statistics on agreements:
  - 182 agreements covering normal pay and conditions issues signed at industry level.
  - 519 company collective agreements covering together 5.9 million employees.
  - 10 percent of all wage agreements were covered through company collective agreements (company agreements predominate in the largest companies such as Philips, DSM and Shell).

### Stylized facts and drivers of wage moderation
- Productivity and inflation:
  - Trend labor productivity declined from above 2 percent to around 1 percent after the crisis.
  - Inflation remains subdued; projected to rise only gradually to the 2 percent target.
- Labor market flexibility and composition:
  - Higher flexibility has altered the relationship between wages and unemployment: more workers become self-employed to avoid high pension and social security contributions; flexible arrangements and temporary contracts have increased.
  - Share of self-employed and temporary contract workers rising faster than in other EU countries; may have weighed on employees’ wage bargaining power.
  - Labor unions’ emphasis on making contracts more permanent and strengthening employment protection in bargaining may also slow wage growth.
- Non-traditional slack indicators:
  - Despite falling headline unemployment, involuntary part-time employment and long-term unemployment remain elevated.
- International spillovers:
  - Wages in the Netherlands track trading partners (e.g., Germany); sectoral leadership by the industry sector induces sensitivity to foreign wages.

### Model specification and estimation approach
- Base model:
  - Use a wage curve model based on Blanchard (1997) to analyze drivers of moderate wage growth.
  - Wage curve in equilibrium: the level of real wages is negatively related to aggregate unemployment (U); augmented with labor productivity (TLP) over time.
  - Equation shown in source: loglogRWTLPUαβ=−
- Error-correction model (ECM):
  - An error-correction specification is used to estimate the wage curve (Sargan 1964; Zhang 2017 extensions).
  - ECM includes an error correction term depending on lagged levels of real wages and trend labor productivity; when real wages deviate from long-run equilibrium this term impacts nominal wage growth to restore equilibrium.
  - Full specification as given in source:
    - 041,,2 ,,3,, 
      1 11
      loglog(1)
      m mm
      ttm  timm timm timti
      K KK
      dWdWDFSUnemGapδλλλλ
      −− − −−
      ===
      =++++ +
      ∑∑∑

      1424
      (    loglog)
      ttt
      RWTLP
      φφ η
      −−
      +− +
  - Notes on specification:
    - Difference operator covers 4 quarters (reflecting common annual wage adjustment).
    - Error correction term uses real wages and trend productivity 4 quarters earlier.
    - Specification controls for lagged dependent variable, 4 log t dW − , to capture inertia or base effects.
    - A statistically significant negative coefficient on lagged real wages 4t RW − would indicate cointegration between real wages and trend productivity.
    - Parameter 1 φ is expected to be approximately equal to 2 φ such that real wages grow broadly at the pace of trend productivity in equilibrium.
- Variables included:
  - Nominal wages determined by:
    - Domestic cyclical factors (D): inflation expectations, unemployment gap, additional labor market slack indicators (e.g., involuntary part-time employment).
    - Foreign cyclical factors (F): German wage growth, foreign labor market slack (unemployment rate and involuntary part-time employment in the Euro Area).
    - Structural factors (S): share of temporary and self-employed workers.
  - Interaction terms: structural variables interacted with unemployment to examine effects of structural changes on the Phillips curve.

### Data and measurement choices
- Nominal wages:
  - Primary measure: total labor compensation from national accounts as a ratio to hours worked.
  - Robustness: also test wages and salaries from national accounts as a ratio to hours worked.
  - Total compensation includes employers’ social security contributions; differs from wages and salaries measure by inclusion of employers’ contributions.
  - Labor Cost Index (LCI) wages and salary measures: based on a more stable basket of jobs, include bonuses and benefits (e.g., car, health care, sick leave); cover the business sector only.
  - National accounts measures reflect structural changes in composition of the labor force and align with labor productivity measures that are averages across changing job composition.
- Real wages:
  - Nominal wage indicator is deflated by the GDP deflator (not a consumer price measure) to ensure consistency with real hourly labor productivity and to reflect firms’ capacity to pay depending on the price of output.

*Source: cr18131 - 1. Netherlands’ Wage Bargaining System*

### 16.      The estimation results indicate that wage  formation has been influenced by both

### 16.      The estimation results indicate that wage  formation has been influenced by both

### Estimation framework and specification
- Sample period: 1995Q1 to 2017Q1.
- Dependent variable: labor compensation per hour from national accounts.
- Model: Error Correction Model (ECM) variants reported in Table 1 (Models 1–14).
- Cointegration: coefficients on the lagged level of the real wage are statistically significant in all variants, indicating cointegration between real wages and productivity.
- Rate of error correction: ranges from -0.5 to -0.7.

### Short-run drivers (models 1–6)
- Models 1–3 (domestic variables only):
  - Significant variables include: unemployment gap, expected inflation, and labor productivity growth.
  - Involuntary part-time (PT) employment is not significant.
- Models 4–5 (add euro area labor market slack):
  - Changes in the EA unemployment gap are statistically significant.
  - Domestic labor slack indicators remain statistically significant.
- Models 6–9 (add German or EA compensation/wage growth):
  - Foreign wages have a significant impact on Dutch compensation growth.
  - Significance is higher for German wages than for euro area wages.
  - In Models 7–9, coefficients on expected inflation, euro area unemployment changes, and domestic unemployment gap remain statistically significant.

### Structural change in employment (models 10–14)
- Adding structural changes in employment (temporary and self-employment shares):
  - The rising share of temporary employment has a significant negative impact on wage growth.
  - The rising share of self-employed employment has a significant negative impact on wage growth.
  - Both temporary and self-employed shares reduce the responsiveness of wage growth to the unemployment gap.
  - When temporary and self-employment are included simultaneously, the effect of the self-employed share becomes less important.

### Long-run relationship and elasticities
- Long-run regression (regression 7):
  - "1 percent increase in productivity is associated with 0.84 percent increase in the compensation, and the coefficient is significantly less than 1."
  - Expressed as: log0.84 log0.10RWTLPU=−
- With labor market structure changes included:
  - Compensation growth aligns more closely with trend productivity when temporary employment is accounted for.
  - Suggested relationship (as given): log1.00.41.08 log(0.17   0.90.79) *RWtempselfTLPtempselfU=−  + +−−  −
  - Interpretation: rising shares of temporary and self-employed workers may have lowered the responsiveness of the real wage to trend productivity and unemployment over the long run, possibly due to reduced bargaining power of more flexible employees.

### Model fit and contribution of factors
- Domestic wage curve model (includes unemployment, productivity, expected inflation) R-squared: 80 percent.
- Adding spillover effects from euro area labor market conditions, German wage growth, and structural changes contributes an additional 10 percent to model fit.
- Overall, the ECM explains more than 80 percent of the total variation of actual wage development.

### Decomposition of recent wage moderation
- Main factors associated with recent moderation in compensation growth (from most complete short-run regression, model 14):
  - Sluggish productivity growth.
  - Lower expected inflation.
  - Moderate wage growth in the euro area (including Germany).
  - Changes in form of employment (rising share of temporary workers) — negative contribution from the error correction term reflects structurally lower real wages driven by a rising share of temporary workers.

### Conclusions and policy implications
- Key drivers of wage moderation in the Netherlands:
  - Slower productivity growth.
  - Lower expected inflation.
  - Remaining slack in the labor market.
  - Foreign wage growth spillovers (notably from Germany and the euro area) due to competitiveness considerations.
  - Rising labor market duality/flexibility (higher shares of temporary and self-employed workers) may have contributed to stagnant wage growth.
- Policy suggestion:
  - Reforms to harmonize labor market employment contracts to increase flexibility while allowing greater bargaining power for more flexible employees could permit both greater flexibility and higher wages.

### Outlook and projections
- Projected negotiated wage increases:
  - Negotiated wages are projected to increase from 1.5 percent last year to 2.2 and 3.2 percent in 2018 and 2019 respectively.
- Uncertainties noted:
  - Projection of wage drift remains unclear given possibilities of compositional changes.
  - Uncertain net effect of proposed coalition agreement measures to balance employment protection between temporary and permanent workers; further research contingent on outcomes of social partner discussions.

*Italic: IMF staff calculations and analysis as presented in the source content.*

### 7.      Household income and financial net

### 7.      Household income and financial net

### Role of household income, employment, and financial net wealth
- Real personal disposable income (RPDI) grew by 2.0 percent per year on average in the 1990s.
- Post-GFC: sluggish RPDI growth and sharply rising unemployment contributed to the housing downturn during 2007—13.
- In more recent years, favorable economic and labor market trends, combined with a rapid accumulation of financial net wealth of households, exerted renewed upward pressure on housing demand.

### Interest rates and housing investment returns
- Mortgage rates have gone down substantially since 2000 to historically low levels in recent years.
- Declining long-term bond yields and the slide of policy rates made housing investment returns increasingly attractive after the GFC, stimulating purchases for investment purposes by wealthier households and further driving up prices.

### Demographics and urbanization
- Annual population growth averaged about 0.5 percent from 1990–2016.
- Urbanization averaged 1.1 percent over 2010–15, exerting additional pressure on demand in main urban areas.
- The four major cities: average annual population growth rate of 1.2 percent during 2007–16.
- Statistics Netherlands projects the number of households will continue to grow by some 640,000 to 8.4 million (8 percent) by 2030.

### Supply factors
- Issuance of new-building permits has stagnated since the GFC; housing supply has lagged expected household growth in most provinces (Economic Institute for Construction and Housing, 2016).
- Housing completion in the four major cities fell to a record low in 2014 despite a post-GFC surge in population.
- Even as residential investment rebounded in response to higher prices, housing completion remained well below estimated household formation, contributing to fast price increases.

### Institutional and structural constraints on supply
- Using 1989–2016 data and the OECD methodology (Caldera Sanchez et. al., 2011), The Netherlands has the second lowest price responsiveness of housing supply among OECD countries:
  - Long-run price elasticity of new housing supply estimated at about 0.2 compared to the OECD average of 0.6.
- Causes of sluggish supply response:
  - Natural constraints (topography).
  - Man-made constraints (stringent local land-use regulations, cumbersome building permitting process, restrictive zoning codes, building aesthetics criteria).
  - Capacity constraint of the construction sector following the onset of the GFC.
- Implication: In markets with inelastic long-run supply curves, given increases in long-run demand lead to greater price increases than in markets with elastic supply (Anundsen et al., 2016).

### Tax incentives, user cost, and distributional effects
- Interest on mortgages is fully tax deductible, which effectively reduces debt service costs and incentivizes higher borrowing and purchase of more expensive houses.
- Authorities have started gradually reducing the maximum tax rate that mortgage interest can be deducted against by 0.5 percentage points annually from 52 percent in 2013, to 38 percent in 2041 (50 percent in 2017).
- The tax relief for housing financing in the Netherlands remains one of the most generous in the OECD and leads to higher house and land prices.
- Capital gains tax is one of the lightest in the European Union; recurrent tax revenue from immovable properties is low compared with the OECD average.
- Economic implications:
  - Favorable tax treatment may crowd out capital from more productive uses, resulting in efficiency losses and housing demand distortions by reducing the user cost of owner-occupied housing and encouraging excessive leverage.
  - Benefits tend to favor higher-income earners (tax savings from mortgage interest deductibility tend to be larger when income are higher).

Notes from source footnotes:
- 100 percent deduction for all pre-2013 loans and for post-2013 fully amortizing loans (within 30 years). Imputed rent from home ownership is taxed at a low level much smaller than the mortgage interest deductibility (MID).
- The recently released coalition agreement proposes a more rapid phase-out in steps of 3 percentage points annually until the basic rate of 37 percent is reached in 2023, subject to parliamentary approval.
- The recurrent property tax in the Netherlands is levied at the local level and varies by region, ranging from 0.1–0.3 percent of property value.

### Rental market regulation and effects on owner-occupied demand
- Dutch rental market size: about 40 percent of total dwelling stock (about OECD average).
- Social rental housing accounts for 30 percent of total dwelling stock (comparators: France 19 percent, UK 15 percent, Germany 5 percent).
- Rent regulation:
  - Both private and social rental housing subject to strict rent regulation—the third most stringent in the OECD.
  - Social rental housing receives large direct/indirect public subsidies.
  - Allocation issues: about 18 percent of social rentals were estimated to be occupied by households earning too much relative to their rent in 2015 (Ministry of Economic Affairs, 2016).
- Consequences of strict rent control and subsidies:
  - “Locked-in” effects, long waiting lists, hindered efficient use of existing housing stock.
  - Crowding out of public and private investment in unregulated rental dwellings; private rental sector contracted to less than 10 percent of housing stock since the 1970s (with slight recent recovery).
  - Supply shortage of unregulated rental housing, especially in large cities, limits mobility to areas with greatest job availability and forces some households into owner-occupation, increasing excess demand for owner-occupied housing and high household debt.

### Cross-country housing valuation model: specification and data
- Long-run equilibrium specification: log-linear inverted demand function with determinants including:
  - Real per capita household disposable income y.
  - Real after-tax interest rate for mortgage borrowing morr.
  - Real per capita household net financial wealth w.
  - Housing stock per capita s.
- Augmented model additions:
  - Square term of real mortgage rate to capture non-linear relationship (morr^2).
  - OECD tax relief index (updated) and interaction term tr * y to capture larger tax savings for higher incomes.
  - Interaction term rc * s where rc is OECD rent control index (rescaled to 0–1) to test rent control effects.
  - Interaction terms with demeaned long-run supply elasticities sr * (demand variables) to allow differential impacts by supply elasticity.
- Estimation sample: 20 advanced OECD countries over 1991:Q3–2016:Q4.
- Country fixed effects used; robust standard errors clustered at country level.

### Estimation results (summary of main coefficients and statistical findings)
- Income (y, log): coefficients across models ~1.533–1.652; a one percent increase in per capita disposable income raises long-run equilibrium house prices by a cross-country average of 1.5–1.7 percent, consistent with housing as a luxury good.
- Mortgage rate (morr, percent): coefficients range about -1.776 to -2.759; a one percentage point increase in the real mortgage rate reduces real house prices by a cross-country average of about 1.8–2.8 percent.
- Mortgage rate squared (morr^2, percent): positive small coefficients (evidence of non-linearity); significance varies across specifications.
- Household net financial wealth (w, log): small positive impact (coefficients around 0.020–0.056).
- Housing stock per capita (s, percent): coefficients around -1.070 to -1.322; a one percent increase in housing stock relative to population associated with about a 1.3 percent reduction in house prices.
- Tax relief interaction (tr * y): positive coefficient ~0.351–0.487; in countries with more generous tax relief, positive income shocks translate into greater price impacts. Example: for the Netherlands, tax relief implies about 0.5 percent higher house price from a one percent increase in real per capita disposable income (i.e., 2.0–2.2 percent rise rather than 1.5–1.7 percent).
- Rent control interaction (rc * s): positive coefficient (1.156 in one specification; 0.436 in another) indicating rent control reduces the dampening effect of increased housing stock on prices. For the Netherlands rent control leads to 0.3 percentage points less decrease in real house prices for one percent increase in housing stock per capita (i.e., 1.0 percent fall rather than 1.3 percent).
- Supply-elasticity interactions (sr * ...): some interactions significant, indicating demand shocks have amplified impacts in more inelastic markets.
- Model fit and tests:
  - Observations: 2042 (across specifications).
  - Adjusted R-squared: 0.853–0.867 across models.
  - Number of countries: 20.
  - Panel cointegration test (Kao): t-Statistics -3.806, Prob. 0.0001.
  - Panel unit root test on residuals (Levin, Lin & chu): t-Statistics -2.705, Prob. 0.003.

### Model-based assessment of Dutch house price valuation
- Based on column (5) estimates, the average house prices in the Netherlands in 2016:Q4 are found to be about 5 percent above the estimated equilibrium value implied by fundamentals (error term ε = P - P*).
- However, real mortgage rates are below their 5 percent average since 1990 by about 2 percent (or below their 3½ percent average since 2000 by about ½ percent) and are likely to unwind (at least partially) over time.
  - If real mortgage rates unwind to their 1990–average (up by about 2 percent), this would lower housing prices by up to about 5 percent in equilibrium.
  - If unwinding is to the 2000–average (up by about ½ percent), this would lower housing prices by about 2 percent in equilibrium.
- Implication: house prices could be up to 10 percent (or up to 7 percent) overvalued depending on the extent of mortgage rate normalization.

_Italic: Source: cr18131 - 7.      Household income and financial net_

### 17.      The implied valuations from this exercise should be interpreted with caution. The

### 17.      The implied valuations from this exercise should be interpreted with caution.

### Valuation caveats and housing market uncertainties
- Estimated equilibrium price levels are subject to uncertainties and should be interpreted as indicative of potential valuation gaps rather than precise measures.
- Factors complicating valuation estimates:
  - Purchases for investment purposes by high-income households—as housing investment returns exceed long-term bond yields.
  - Low interest rates have driven up equilibrium house prices which mitigates overvaluation concerns, but:
    - They do not rule out that demand is excessive.
    - Demand could fall sharply as interest rates normalize.
  - These factors could bias up the estimates of long-run equilibrium prices.

### Conclusions on drivers of house-price increases (points 18–19)
- Institutional and structural factors have significantly contributed to high and rapidly rising house prices in the Netherlands, beyond conventional demand and supply factors.
- Key contributing factors:
  - Large direct and indirect subsidies for social housing and a highly regulated rental market likely skew housing needs and use.
  - A well-developed mortgage market combined with large tax preferences for owner-occupied housing and mortgage debt has further fueled demand for homeownership and household debt.
  - Sluggish housing supply response exacerbated demand pressures by failing to cushion the impact of rising demand.
- Macro-financial vulnerabilities:
  - Overvalued house prices and elevated household debt pose vulnerabilities to financial and macroeconomic stability.
  - The recent house-price cycle left the Netherlands with elevated household debt and a significant share of underwater mortgages.
  - A large correction of house prices—driven by slower real income growth, a reverse in sentiment, or interest rate hikes—could weaken household balance sheets, depress private demand, and adversely affect corporate and bank earnings.

### Policies implemented to date (point 20)
- Authorities introduced measures targeting owner-occupied housing and strengthening resilience of banks and households:
  - Additional bank capital buffer requirements in line with Basel III/CRD IV.
  - Introduction of LTV and debt service-to-income (DSTI) caps since 2013.
  - Gradual reduction of LTV limit for mortgages to 100 percent by 2018.
  - A tax exemption for gifts used for housing down payments or mortgage repayments.
  - Allowing MID only for new fully amortizing loans.
  - Gradual reduction of the maximum tax rate allowed for MID from 52 percent in 2013 to 38 percent in 2042 in steps of ½ percent per year.

### Further reforms recommended (point 21)
- Rationale: Comprehensive reforms are needed to address housing market risks, enhance macro-financial resilience, reduce boom-bust cycles, and improve labor mobility.
- Recommended policy actions:
  - Reduce generous tax preferences for owner-occupied housing and mortgage debt to prevent demand distortions and excessive leverage:
    - Accelerate phasing down of MID to ultimately bring it to a neutral level relative to the taxation of other assets.
    - Given the current low interest rate environment limiting the effective benefit of MID, implement the reduction now.
    - Note: The recently released coalition agreement proposes a much more rapid phase-out in steps of 3 percentage points annually until the basic rate of 37 percent is reached.
  - Improve housing supply responsiveness in large cities by:
    - Streamlining and relaxing stringent building aesthetics criteria, restrictive zoning plans, and cumbersome building permission processes.
    - Addressing impediments to urban redevelopment and improving public transportation to relieve demand pressures in major centers.
  - Phase out rent control and reform social housing to enhance flexibility:
    - Gradually raise rents on regulated rental housing to be aligned with market rates.
    - Protect vulnerable households through targeted housing allowances.
    - Expected benefits: more efficient use of existing housing stock, a larger private rental market, and improved mobility across housing types and locations.
  - Tighten macroprudential measures to contain household financial vulnerabilities:
    - Gradually lower the maximum limit on LTV ratios by at least 1 percentage point per year to no more than  90 percent by 2028 (as recommended by the Financial Stability Committee (FSC)) and to 80 percent thereafter.
    - Introduce prudential ceilings on DSTI caps by income category that could not be relaxed during periods of strong growth.
  - Consider temporarily allowing partial use of pension savings for housing purchases to ease liquidity constraints for first-time home buyers (e.g., meeting part of the down payment), thereby reducing debt burdens and total savings needs of home purchasers:
    - Comparable arrangements: US (401K), Switzerland, Canada, Singapore.

### Annex note (graphs)
- Note included in source: Blue lines represent actual house prices; red lines refer to estimated long-run equilibrium prices; and green lines on the RHS axis tell valuation gap in percent.

### Health care system summary (selected points from subsequent section)
- Overview and fiscal challenge:
  - The Dutch health care system delivers good outcomes but at rapidly increasing public costs.
  - Public spending on health and long-term care have increased faster than GDP to reach about 6 and 4 percent of GDP, respectively.
  - National projections: if expenditures continue to rise linearly without policy response, total health care cost would reach 31 percent of GDP by 2040, and the average household would have to contribute about half of its income for its funding.
- 2006 reform and “regulated competition”:
  - The 2006 reform (Health Insurance Act, Zvw) introduced “managed competition” and mandatory insurance of the whole population by private health insurers for a legally defined set of basic health care services.
  - Funding architecture now comprises:
    - An income related contribution of 6.9 percent of income before taxes and social premiums for workers in regular employment (paid by the employer to the tax office and allocated to insurers via a risk equalization fund).
    - Direct payments by individuals to their chosen health insurers (together with the income contribution, each contributing about half to the notional health care budget).
  - Other financial parameters and coverage:
    - Average direct health insurance premium: 110 euros per month on average (selected financial indicators table).
    - Deductible currently set at a minimal level of 385 euros per year for basic benefits coverage.
    - Self-employed workers pay a contribution representing 5.654 percent of their income.
    - About 85 percent of Dutch households also take supplemental insurance; about 40 percent of households receive income-related health care allowances; government covers health care provision for children.

*IMF Country Report cr18131 (excerpt).*

### 6.      On the supply side, price competition is    expected to result from bargaining

### 6.      On the supply side, price competition is    expected to result from bargaining

### Supply-side design and bargaining mechanisms
- Insurers and health care providers are meant to jointly determine health care premia and prices of health care services through bargaining mechanisms.
- Patient mobility and enrollment rules:
  - Patients can switch among health insurers once a year without restriction.
  - Insurers are not allowed to refuse enrollment or discriminate among applicants based on risk.
  - Some ex ante equalization mechanisms offset structural discrepancies across categories of patients.
  - Insurers must charge the same “community-based” premia to all of their insured population.
  - Insurers are strictly bound by the content of the basic health care package, including guarantees in terms of geographic access to care.
- Competition dimensions available to insurers:
  - Relative combination of premia and deductibles.
  - Mix of in kind or reimbursement policies.
  - Tailored policies such as selective or group contracting.
- Insurers use market power to bargain on purchase of medical goods and services from health care providers.
- Hospital price segmentation:
  - “Segment B” (negotiable) spending envelope has been gradually raised to 70 percent of hospital budgets.
  - The remaining 30 percent of hospital prices (“segment A”), generally pertaining to research and complex care, are set nationally by the Dutch Healthcare Authority (NZa).
  - Negotiations conducted on the basis of a standardized system of coding for inpatient, outpatient and specialist costs, simplified in 2012.

### Policy adjustments strengthening competition and fiscal anchors
- Key 2012 reforms and tools:
  - Shift from ex post risk compensation to an ex ante risk equalization mechanism in 2012, increasing pressure on insurers to negotiate costs and differentiate premia.
  - Introduction of “stakeholder agreements” per sector (primary care, hospitals, mental healthcare) among providers, insurers and patient associations under government aegis.
  - Stakeholder agreements consist of yearly ceilings for health care expenditure growth; breach can trigger across-the-board savings initiated by government, prorated to market shares of insurers and hospitals.
  - The macro-budget tool provides an anchor for centralized negotiations among stakeholders, organizing supply-side competition.

### Impacts on health care markets and delivery
- Coverage and service delivery outcomes:
  - Drop in number of uninsured persons from about 200,000 people to negligible amounts.
  - Productivity gains in hospitals evidenced by diminishing lengths of stay.
  - Improvements in client service: lengthening of opening hours, facilities to prevent unnecessary emergency care use, reduction in waiting lists.
  - Expansion of service supply via multiple independent treatment centers (ZBCs) providing routine care at lower costs (ophthalmology, orthopedics, etc.).
- Insurance market developments:
  - Administrative costs of health insurers trended downward after an initial peak following the reform.
  - A 2 percent decrease in the average premium was triggered by competition for market shares in 2006–2008.
  - Increasing premium differentiation: up to about 30 percent difference between the highest and lowest nominal premium in 2014.
  - Increasing uptake of higher deductibles in exchange for lower monthly payments.
  - Insurers improved solvency ratios to about 160 percent, in compliance with Solvency II requirements, after initial financial losses and windfall profits on basic health insurance.
  - Annual switching behavior: proportion of people switching insurers settled slightly above 6 percent, i.e. about 1.1 million people, after an initial spike to 17 percent at reform enactment.
  - Degree of competitiveness remains limited; persistent barriers to entry noted.
- Provider-side price and volume dynamics:
  - Real prices for “segment B” services have declined or increased more slowly than non-negotiated prices, with substantial variation across provider types (university hospitals, general hospitals, etc.).
  - Overall health spending envelope continued to increase due to a pickup in volume of care and possible ‘upcoding’ to preserve income amid lower prices.
  - Evidence of substantial hospital mergers in recent years, potentially increasing hospital bargaining power against insurers.
  - Administrative costs may have increased for providers due to higher number of contractual arrangements and information requirements from insurers and supervisors.
  - Price of pharmaceutical products underwent a sharp decrease since 2012 due to insurers reimbursing only lowest price generic drugs.

### Preliminary empirical findings (2012–2015)
- Methodology note:
  - Degree of competition on the insurance market assessed by Herfindahl-Hirschman index (HHI) of market shares.
  - Analysis uses regional data covering post-reform years (2012–2015).
  - Cautions: low number of observations, limited explanatory power, likely omitted variable bias (notably hospital sector concentration).
- Key econometric results (Fixed effects panel estimations, robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1):
  - For annual spending on health care totaling about €1,190 on average over 2012–2015:
    - A one notch increase in the level of insurer concentration (HHI) is associated with a €244 increase in total health expenditure (coefficient 244.046**, standard error (86.717)).
    - Coefficient on average hospital expenditures: 177.420** (standard error (61.178)).
    - Coefficient on average GP expenditures: 28.420 (standard error (17.918)) — not significant.
  - Disposable income per capita (in €): coefficients 0.067**, -0.004, -0.003 (standard errors (0.030), (0.018), (0.007)) — muted or not significant.
  - Ratio 65 year + / total population (in percent): 212.723***, 22.911*, -3.724 (standard errors (24.952), (12.361), (5.212)) — ageing ratio contributes positively to spending.
  - Time dummy for 2014: -90.864***, 26.460***, 16.093*** (standard errors (12.669), (4.367), (2.314)) — 2014 dummy negatively impacts total expenditure but positively impacts hospital and GP spending.
  - Constant: -3,131.711***, 343.013, 155.178 (standard errors (525.888), (395.814), (166.491)).
  - Observations: 36 for each regression.
  - R-squared: 0.939, 0.628, 0.795 (Adjusted R2 reported as 0.768 for each column).
- Interpretation:
  - Preliminary estimations suggest more concentration among insurers is positively related to average health care price developments, likely driven by hospital spending.
  - Results warrant caution due to data limitations and potential omitted variables (notably hospital market concentration).

### Conclusion and policy implications
- Evidence points to significant efficiency gains in the health care sector (coverage gains, productivity improvements, service expansion), but the effect on overall price developments remains uncertain.
- Ongoing trends:
  - Continued pressure towards concentration in both the health insurance and hospital sectors.
  - Complex interactions between insurers and hospitals complicate assessment of relative bargaining power and ultimate impact on costs and quality.
- Risk and regulatory recommendation:
  - Concern exists that excessive cost-saving measures may risk lower quality of care in the future.
  - Calls for continued vigilance from regulatory and monitoring agencies within the new institutional framework.

*Source: CBS; Vektis; Dutch Healthcare Authority (NZa); and IMF staff calculations.*

### 1.      The new Dutch government fully embraced the Paris Climate Agreement and

### 1.      The new Dutch government fully embraced the Paris Climate Agreement and

### Climate commitments and currently envisioned policies
- EU pledge: reduce CO2 and other greenhouse gases (GHGs) by 40 percent relative to 1990 levels by 2030.
- Netherlands target: increase GHG reduction target for 2030 to 49 percent below 1990 levels.
- Existing policies designed to meet the EU pledge include:
  - EU Emissions Trading System (ETS) reducing power generation and large industrial emissions 43 percent below 2005 levels by 2030.
  - National-level targets for non-ETS emissions—for the Netherlands a 36 percent reduction below 2005 levels by 2030.
  - EU goals for energy efficiency (a 30 percent improvement by 2030) and renewables.
  - EU standards for vehicle CO2 emission rates.
- Government agreement measures mentioned:
  - introducing a minimum price for CO2 emissions from power generation on top of the ETS;
  - shifting taxes off electricity and onto gas generation;
  - phasing out coal plants and natural gas for new buildings by 2030;
  - subsidizing carbon capture and storage (CCS);
  - expanding offshore wind power.
- Transportation reforms being considered:
  - full penetration of electric vehicles into the new car fleet by 2030;
  - adoption of km-based taxation for HGVs;
  - stiffer penalties to deter dangerous driving;
  - infrastructure upgrades to alleviate traffic congestion.
- Revenue concern: progressive erosion of traditional LDV revenue sources—fuel taxes and CO2-related vehicle taxes—due to electric vehicle penetration prompts analysis of replacement revenue instruments.

### Analytical approach and model structure
- Analysis uses an IMF spreadsheet tool parameterized to the Netherlands.
- Model starts with use of fossil fuel products and other fuels in power generation, road transport, industry, and household/commercial sectors.
- BAU (business-as-usual) projections account for previously implemented mitigation policies (implicit in recent fuel use data) but not planned mitigation policies, with assumptions about:
  - GDP growth;
  - income elasticities;
  - autonomous rates of technological change;
  - future international energy prices;
  - price responsiveness of fossil fuels in different sectors.
- An “envisioned policy” reference case represents:
  - a simplified EU ETS;
  - regulations represented by implicit or “shadow” prices to meet energy efficiency, vehicle emission rates, and the Netherlands target for non-ETS emissions.
- Reforms considered replace regulatory approaches with pricing policies while preserving emission targets.
- Parameter sources: IMF (GDP growth and domestic environmental impacts); International Energy Agency (fuel use by sector); Dutch authorities (current fuel prices and taxes); empirical evidence/results from energy models (fuel price responsiveness and technological change rates).
- Assumed fuel price elasticity example: each 1 percent increase in fuel price reduces consumption of that fuel by 0.6 percent, with two-thirds from adoption of more fuel-efficient or cleaner technologies and one-third from reduced intensity of use.

### Impacts of currently envisioned policies (model findings)
- Collectively, currently envisioned policies reduce nationwide CO2 emissions in the Netherlands by an estimated 26 percent below BAU levels in 2030.
- BAU emissions are already 10 percent below 2005 levels.
- Emission reductions by policy component (percent of BAU in 2030):
  - National policies: 9 percent.
  - ETS: 6 percent.
  - Energy efficiency policies for the ETS sector: 6 percent.
  - Energy efficiency policies for the non-ETS sector: 4 percent.
  - Vehicle emissions rate standard: 2 percent.
- Revenue impact: envisioned policies reduce revenue by 0.25 percent of GDP in 2030.
  - ETS allowance auctions assumed to raise 0.15 percent of GDP, assuming half of allowances are auctioned.
- Reason for revenue loss: energy efficiency policies erode bases for fuel and electricity taxes, offsetting ETS auction revenue gains.

### Emissions-neutral reforms and revenue-raising scenarios
- Replacing vehicle emissions standards with higher road fuel taxes (holding road emissions fixed):
  - Raises 0.4 percent of GDP.
  - Requires a fuel price increase of €0.60 per liter.
- Extending a uniform carbon tax to all non-ETS emissions (keeping non-ETS emissions fixed and removing the extra road fuel tax):
  - Raises revenues of 1.1 percent of GDP.
  - Requires a price of €170 per tonne of CO2.
- Fully auctioning ETS allowances:
  - Would raise an additional 0.15 percent of GDP in revenue.
- Introducing a CO2 surcharge for the ETS sector (while removing other CO2-related policies):
  - Raises revenues of 0.2 percent of GDP.
  - Requires raising the CO2 price by €52 per tonne.
- Harmonizing prices across ETS and non-ETS sectors (keeping nationwide emissions fixed):
  - Loses a modest amount of revenue.
  - Implies an economy-wide price of €136 per ton of CO2.

### Economic costs, environmental benefits, and cost-effectiveness of pricing reforms
- Costs of envisioned policies (around 2030):
  - Aggregate cost: around 0.6 percent of GDP in 2030.
  - Largest source: national level targets for the non-ETS sector—0.35 percent of GDP (assumed to take the form of regulations reducing fossil fuel intensity for households, commerce, and small industry).
  - Vehicle emissions standard: about 0.15 percent of GDP.
  - Energy efficiency policies within the ETS sector: 0.1 percent of GDP.
  - ETS and (EU level) energy efficiency requirements for the ETS sector: around 0.03 percent of GDP each.
- Netting domestic environmental benefits (primarily reduced air pollution mortality) lowers costs moderately to 0.45 percent of GDP overall.
- Pricing reforms could lower economic costs to 0.4 percent of GDP, or costs net of environmental benefits to 0.1 percent of GDP.
- Largest cost savings arise from carbon taxes for the non-ETS sector.

### International precedents and implementation details for carbon pricing
- UK practice: variable carbon tax on top of the ETS emissions price for power generators where the tax rate equals any prevailing difference between a target for the combined tax/ETS price and the ETS price.
  - Example: tax (out to 2021) set equal to the difference between £18 (€20) per tonne and the EU ETS emissions price.
- French and Irish practices:
  - France: carbon tax slated to rise from €31 per tonne in 2017 to €65 in 2020 and €86 in 2022.
  - Ireland: carbon tax currently fixed at €20 per tonne.
- Notes on Dutch ETS surcharge:
  - A Dutch surcharge for ETS emissions would have no direct impact on EU-wide emissions (ETS cap fixes EU emissions), unless the Netherlands purchased ETS allowances and withdrew them from the market.
  - A Dutch tax might spur similar measures in other member states, increasing pressure for ETS reform.

### Complementary fiscal instruments and infrastructure to support deeper mitigation
- Feebates for the power sector:
  - Charges for emissions-intensive generators proportional to output times the difference between their emission rate and a pivot point emission rate.
  - Subsidies for non-emissions-intensive generators proportional to output times the difference between the pivot point and their emissions rate.
  - If pivot point set at industry average emission rate, feebate is revenue neutral and has only a modest impact on electricity prices.
  - Feebates are flexible, easily adjusted over time to decarbonize the power sector without raising tax burdens on the sector.
  - Feebates strike an efficient balance between gas and other emitting fuels like coal with CCS.
- Other complementary measures:
  - Fiscal incentives for adoption of CCS at industrial plants.
  - Fiscal analogs of regulations (e.g., progressively tightening natural gas standards for new buildings with the option to pay out-of-compliance fees).
  - Targeted infrastructure investments to enhance effectiveness of carbon pricing (modifications to the grid to accommodate more renewables; pipelines for CCS).
- Policy preference: targets for emissions prices generally preferred on economic grounds because they provide more certainty for investment and better accommodate uncertainties than quantitative emissions targets.

### Near-term pricing credibility
- Priority: establishing robust near-term carbon prices may be more urgent than fine-tuning distant emissions targets.
- Rationale:
  - Credibility for distant targets may be challenging if market participants expect possible scale-backs of distant targets.
  - Establishing more aggressive carbon prices in the near term (minimum prices for CO2 emissions, carbon surcharges, taxes) strengthens near-term mitigation and could enhance credibility of longer term targets.
- Context: current global average CO2 price about €1 per tonne (WBG 2017); low EU ETS price (example €5 per tonne) may reflect expectations of possible scaling back of climate goals.

### Road transportation: external costs, tax gaps, and policy implications
- External cost estimates for gasoline vehicles (2013):
  - Dutch source total: €1.58 per liter.
  - IMF source total: €1.10 per liter.
  - Component breakdowns (Dutch / IMF):
    - Congestion: €0.88 / €0.85 per liter.
    - Traffic accidents: €0.44 / €0.17 per liter.
    - Global warming: €0.21 / €0.08 per liter.
    - Local air pollution: €0.05 / €0.01 per liter.
  - Note: Dutch and IMF estimates assume CO2 damage values of €85 and €32 per tonne respectively.
- External cost estimates for diesel vehicles (averaged over LDVs, HGVs, buses, 2013):
  - Total: about €1.12 per liter in both studies.
  - Component breakdowns (Dutch / IMF):
    - Congestion: €0.59 / €0.77 per liter.
    - Accidents: €0.15 / €0.13 per liter.
    - Global warming: €0.24 / €0.09 per liter.
    - Air pollution: €0.14 / €0.13 per liter.
    - Road damage: €0.01 / €0.02 per liter.
- Current excises and efficient tax comparisons:
  - Current gasoline excise: €0.77 per liter.
    - Efficient tax implied by Dutch estimate: €1.07 per liter.
    - Efficient tax implied by IMF estimate: €0.70 per liter.
  - Current diesel excise: €0.49 per liter.
    - Efficient tax implied by Dutch estimate: €0.83 per liter.
    - Efficient tax implied by IMF estimate: €0.77 per liter.
  - Caveat: diesel vehicles face higher annual road taxes and vehicle registration fees than gasoline vehicles; cross-border refueling limits unilateral tax increases.
- Treatment of externalities that vary with driving:
  - For externalities that vary with changes in driving (congestion, accidents, road damage) but not fuel efficiency, the model multiplies these externalities by the fraction of tax-induced fuel reduction that comes from reduced driving (assumed to be 0.4 in Figures 1d and 1e).

### Congestion pricing and km-based charging
- Congestion is most efficiently addressed by peak period pricing of busy roads, administered nationally or locally.
- Severe congestion concentrated in the modest share of driving occurring in densely populated areas during peak periods.
- Effective charge design: charges for vehicle km driven on busy roads, progressively rising and falling over the course of the rush hour.
- National-level system design:
  - Record annual km driven by motorists and levy charges on each km varying by when and where driving occurs to reflect prevailing congestion costs.
  - Implementation could require GPS technology in all vehicles to inform motorists of route charges and transmit driving information to an independent billing agency.
- Local urban systems:
  - Could charge by km according to route within the network and time of day and can capture most gains from a national approach if applied comprehensively across urban centers.
  - Past local schemes have typically been limited to downtown area charges or individual highway tolls (e.g., London, Stockholm, Milan).
- Policy tradeoff:
  - Tension between keeping charging systems simple and easy to understand versus more finely tuned systems with rates varying by major road and time of day.
  - Recommendation: systems can start simple and be progressively refined as acceptability improves.

*International Monetary Fund — Selected Issues Paper excerpt*

### 14.      Accidents are more effectively reduced through distance-based charging related to

### 14. Accidents are more effectively reduced through distance-based charging related to accident risk, either through explicit taxes or pay-as-you-drive (PAYD) insurance.

### Distance-based charging and accident risk
- Principal efficient tax: levied on a km basis, with rates scaled to both driver risk (e.g., rating factors accounting for age, prior crash record) and vehicle risks (e.g., higher for larger vehicles posing greater risk to other road users).
- Fiscal instruments have not been introduced comprehensively; a promising alternative is voluntary transition from lump-sum annual insurance payments to PAYD insurance where payments are:
  - directly proportional to km driven, and
  - per km charges scaled by drivers’ rating factors.
- Distributional and behavioral implications:
  - No new tax burden on the average motorist; low-km drivers pay less under PAYD.
  - Increased rates for remaining drivers provide incentives to switch to km-based insurance.
  - Tax incentives may be needed to kick-start the transition because an individual insurer does not capture benefits to others from reduced multi-vehicle collision risk.

### HGV charges: road damage, congestion, accidents, and emissions
- Road damage should be incorporated into upcoming distance-based charges for HGVs; charges should ultimately account for other environmental costs and vary by location and time of day.
- Efficient design features:
  - Road damage charge would vary with axle weight (footnote: road damage increases exponentially with axle weight) and vulnerability of roads where driving occurs.
  - HGV charges should include components for congestion and accidents and, if levied in proportion to local air emission rates, provide incentives for abatement equipment and cleaner fuels.
- Illustrative efficient toll levels (from Figure 1f):
  - about €0.45 per vehicle km in rural areas (road damage significant, other environmental costs small).
  - about €2.60 per vehicle km for peak urban driving (primarily due to congestion; air pollution damages are also higher).
- Current policy constraints and transitional considerations:
  - Surrounding countries charge roughly €0.15 per HGV km and EU legislation caps charges at €0.40 per km.
  - Charges in the Netherlands would need to be phased in gradually and increased in coordination with other countries and revisions to maximum EU rates.
  - Diesel fuel taxes could be lowered to contain new tax burdens for trucks, with incentives for low-carbon vehicles preserved through feebates.

### Vehicle registration fees, feebates, and balancing environmental and fiscal objectives
- Current registration fee system for imported passenger vehicles:
  - Allocates new vehicles into one of five CO2 emission rate brackets, varying from below 73 gram/km to above 162 gram/km.
  - Imposes an escalating system of fixed charges: €356 for the lowest emission rate bracket and €12,593 for the highest bracket.
  - Variable charges equal the difference between the emission rate and the lower bound of the bracket, times a charge rising from €2 per gram/km for the lowest bracket to €458 per gram/km for the highest bracket.
- Identified problems with current system:
  - Raises less revenue the more successful it is in shifting people to lower emission vehicles.
  - Violates uniform incentive principle and creates bunching of vehicle demand at the top of the next lowest emission bracket.
- Proposed reform: combine an ad valorem tax on vehicle sales prices (set to meet fiscal objectives) with a continuous feebate:
  - Feebate applies fees or rebates in proportion to the difference between a vehicle’s CO2 gram/km and a common pivot point gram/km equal to the average across the imported vehicle fleet.
  - Benefits of feebate:
    - Provides uniform incremental incentive to reduce gram/km across all vehicles.
    - Rate can be chosen to maintain or strengthen existing incentives for low emission vehicles without eroding revenue base for ad valorem tax.
    - Provides strong incentives for electric vehicles without mandating market entry irrespective of future costs and acceptability.

### Feebate design for HGVs
- A feebate could reduce carbon intensity of HGVs if applied to the in-use fleet, scaled to annual km, and integrated into the upcoming charging system.
- Design specifics:
  - Fees and rebates for each vehicle scaled in proportion to its carbon emission rate (averaged over annual vehicle trips with and without freight) and multiplied by the truck’s annual km driven.
  - Administrative ease: fees/rebates could be integrated into prospective HGV charging system.
  - Consider varying pivot point by truck class (e.g., average among trucks within a given weight classification) to avoid overly penalizing large trucks that have scale economy advantages.
- Rationale against sales-only application:
  - Applying feebates to HGV sales only would be administratively challenging and less effective because manufacturers often build components rather than complete vehicles and used trucks (with long lifetimes) would not be covered.

### Example numeric references and illustrative comparisons (preserved exactly)
- Gasoline and diesel fuel are also subject to a modest stockholding fee of €0.008 per liter.
- Marginal congestion costs (UK, 2015) example: about 1 to 10 pence per vehicle km where volume-to-capacity ratio < 75 percent (91 percent of total traffic); about 80 to 170 pence per vehicle km where volume-to-capacity ratio approaches 100 percent (9 percent of traffic).
- Efficient excise tax levels (Figure captions):
  - Second-best efficient gasoline tax around €1 per liter.
  - Second-best efficient diesel tax around €0.8 per liter.
- HGV toll illustrative values: €0.45 per vehicle km (rural) to €2.60 per vehicle km (peak urban).
- Current surrounding-country HGV charge: roughly €0.15 per HGV km; EU cap: €0.40 per km.
- Current registration fixed charges: €356 (lowest bracket) and €12,593 (highest bracket).
- Example vehicle tax comparison:
  - Zero-emission vehicle currently pays a fixed fee of €356 while a vehicle with 100 gram/km pays €2,355 (fixed and variable fees of €2,077 and €278 respectively), or about €2,000 more.
  - Feebate examples: a feebate rate of €20 per gram/km would preserve the current difference in taxes between these vehicles (assuming pivot point is 100 gram/km); a feebate rate of €40 per gram/km would double the current tax difference.

*Source: IMF staff analysis in "KINGDOM OF THE NETHERLANDS—NETHERLANDS" chapter.*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2018/cr18131.pdf_