## sipea2025100

## Source details

**Canonical URL:** [sipea2025100](https://www.imf.org/-/media/files/publications/selected-issues-papers/2025/english/sipea2025100.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/selected-issues-papers/2025/english/sipea2025100.pdf.md)
- [Structured JSON version](/-/media/files/publications/selected-issues-papers/2025/english/sipea2025100.pdf.json)

---

### A. Introduction
- Iceland: small, advanced economy with a floating exchange rate and significant trade openness.
- Central Bank of Iceland (CBI) inflation-targeting framework with mandate to maintain inflation at a target rate of 2.5 percent.
- Since the inflation target was set in 2001, inflation has averaged 4.9 percent.
- Introduces IceQMod, an Iceland-specific Quarterly Projection Model (QPM):
  - QPM: medium-sized semi-structural model used for forecasting and monetary policy analysis.
  - Calibrated to Icelandic data; tailored to capture essential features of the Icelandic economy.
- Objective: assess implications of a change in methodology for calculating imputed rent (owner-occupied housing services) in the CPI, moving from “user cost” to “rental equivalence” effective June 2024.
- Imputed rent sub-component weight in CPI: 20 percent.
- Under the prior “user cost” approach, imputed rent contributed an average 1.8 percentage point to headline CPI inflation between 2001 and mid-2024.

### B. Overview of IceQMod
- Model character:
  - Semi-structural gap model with New Keynesian short-run dynamics (sticky wages and prices).
  - State-space representation allowing Kalman filter decomposition into gap and trend components.
- Iceland-specific features:
  - Export equation reflects dominance of fisheries, aluminum, and tourism.
  - Import equation captures low dependence on energy imports.
  - Trade equations capture openness of the economy.
  - Population growth (substantial over past decade) influences real estate prices.
  - Highly elastic labor supply due to access for citizens of EU countries and procyclical immigration.
  - Separate Phillips curves for housing and non-housing inflation: housing component has different response to demand and exchange rate shocks and a higher trend inflation rate.
  - UIP condition adapted for strong connection to euro-denominated financial markets and high capital mobility.
- Appendix contains detailed description of IceQMod.

### C. Measuring imputed rent: methods and synthetic CPI construction
- Three broad approaches to estimating imputed rent (Consumer Price Index Manual, 2004):
  - User cost:
    - Includes mortgage cost, taxes, maintenance costs, depreciation based on current market value minus expected capital gain.
    - House prices and mortgage payments tend to dominate estimates.
    - Iceland used a version of this approach until May 2024.
  - Rental equivalence:
    - Estimates the rent homeowners would pay if they rented their homes.
    - Suitable where active rental market properties can proxy owner-occupied homes.
    - Iceland adopted this methodology from June 2024.
  - Acquisitions:
    - Looks at cost of purchasing and owning a dwelling, including renovations, home insurance, maintenance, and transfer costs.
- Theoretical note:
  - In frictionless, perfect-data setting, user cost and rental equivalence should yield the same imputed rent (National Academy of Sciences, Engineering, and Medicine, 2022).
- Synthetic historical CPI construction (second-best approach):
  - Expanded HMS rental database only recently established; cannot directly replace historical user-cost imputed rent.
  - Synthetic CPI created by assuming imputed rent mirrors existing actual rent sub-component.
  - Caveats:
    - Actual rents database includes private rents, social housing, and student rentals; rental equivalence uses private rentals only.
    - Owner-occupied property types may differ from rental stock; housing mix differences may bias synthetic measure.
- Empirical comparisons (2002–2024):
  - Synthetic CPI tends to follow published CPI but with significant divergences in specific periods (2005–06, 2008, 2015–19, 2022–24).
  - Example divergence: 2015–2019 synthetic inflation ≈ 1.2 percent versus published inflation 2.2 percent.
  - Averages and volatility (Text Table 1):
    - Since 2002 Published CPI Average: 4.9
    - Since 2002 IMF Synthetic CPI Average: 4.7
    - Standard deviation Published CPI: 3.3
    - Standard deviation IMF Synthetic CPI: 4.1
  - Synthetic measure is on average 0.2 percentage points lower than published CPI over 2002–2024.
  - Synthetic CPI is more volatile than published CPI, despite rental inflation being more stable than house prices.
  - Explanation: house prices in Iceland are negatively correlated with CPI excluding housing; house price inflation tends to move with a stronger exchange rate while CPI excluding housing moves opposite, producing offsetting effects on headline CPI.

### D. Counterfactual experiment: monetary policy with rental-equivalence imputed rent
- Experiment setup:
  - Replace actual CPI in IceQMod with synthetic CPI (rental-equivalence proxy).
  - Keep model parameters unchanged.
  - Make policy interest rate unobservable and estimate it via the Kalman filter.
- Key findings:
  - Policy rate would have been slightly lower on average under synthetic inflation and more volatile.
  - Significant differences between actual CBI policy rate and model-implied policy rate based on synthetic CPI during periods of robust house price inflation: 2005–08 and 2013–21.
  - In those periods, synthetic CPI would have implied a more accommodative monetary policy stance than actual decisions, suggesting the Monetary Policy Committee may have been effectively “leaning against the wind” on asset prices.
  - Likely counterfactual: house price inflation would have been higher if policy had followed synthetic CPI; central bank could have used macroprudential tools to stabilize housing market.

### E. Policy implications and dynamics going forward
- Measurement change implications:
  - For an unchanged inflation target, a methodological change that reduces average inflation allows the central bank to maintain a looser nominal policy stance for a period.
  - The methodology change entails an increase in the equilibrium nominal interest rate.
    - Footnote example: UK switch from RPIX to CPI lowered average CPI by 0.8 percentage points; Chancellor revised target from 2.5 percent to 2.0 percent; decision not to fully revise resulted in de facto 0.3 percentage point rise in level of inflation consistent with target and equilibrium nominal interest rate rose by 0.3 percentage points (Nickell, 2003).
  - Equilibrium real interest rate is unchanged; no bearing on real economy from the nominal-equilibrium change alone.
  - Because revised methodology excludes house prices:
    - Monetary policy could be looser during periods of growing house prices and tighter during periods of weakness, assuming unchanged monetary policy reaction function.
- Macroprudential policy:
  - Increased need to adopt macroprudential measures to manage excessive house price movements, since monetary policy may no longer be implicitly targeting asset prices via headline CPI.

### F. Conclusions (summarized)
- Measurement effects:
  - Synthetic CPI is on average 0.2 percentage points lower than published CPI over 2002–2024.
  - Synthetic inflation displays higher standard deviation: 4.1 versus published CPI standard deviation 3.3.
- Monetary policy effects:
  - Central bank could keep monetary policy looser for a period if average CPI inflation is lower.
  - Policy rate behavior: lower during high house price inflation episodes and higher during low house price inflation episodes if reaction function unchanged.
- Recommendation:
  - Greater reliance on macroprudential measures may be required to manage house price fluctuations now that headline CPI excludes direct house price effects.

### Annex I. IceQMod core structure (select equations and parameters)
- Real sector (IS curve):
  - Y_t_gap = α1⋅Y_t-1_gap − α2⋅MCI_t + α3⋅Y_t^F_gap + α4⋅FI_t + α5⋅TP_t_gap + ε_t^Y
  - Monetary Conditions Index (MCI):
    - MCI_t = w⋅RRG_t + (1−w)⋅Ẑ_t
- Prices and inflation:
  - Headline CPI composition:
    - ΔP_t = δ⋅ΔP_t^XH + (1−δ)⋅ΔP_t^H + ψ_t_CCC
  - Non-housing CPI inflation:
    - π_t^XH = β1⋅π_t−1^XH + (1−β1−β3)⋅π_t+1^XH + β2⋅RRM_XH + β3⋅π_t^imp,xh + ε_t^XH
  - Steady-state inflation assumptions:
    - Housing prices inflation steady state: 3.25
    - Non-housing steady state: 2.25 percent
- Housing inflation construct:
  - π_t^H = c_1 ⋅ π_{t−1}^H + (1−c_1) ⋅ π_{t−1}^H + c_2 ⋅ RM_M^H_t + c_3 ⋅ RP̄_t + ε_t^{SHOCK,H}
  - Definitions and behavioral assumption:
    - π_t^H – housing inflation
    - RM_M^H_t – real marginal costs for housing
    - ΔRP_t – trend of real housing price (notationally indicated in source)
    - ε_t^{SHOCK,H} – shock
    - Housing inflation is assumed to be procyclical to the monetary policy stance using real interest rate gap (housing costs increase with higher short-term interest rate as it translates into higher mortgage interest rates).
- Monetary policy (Taylor-type rule):
  - RN_t = γ_1 ⋅ RN_{t−1} + (1−γ_1) ⋅ RRR_t + π^{target}_t + γ_2 ⋅ (π_{t+3} − π^{target}_t) + γ_3 ⋅ Ỹ_t + ε_t^{RSHOCK}
  - Definitions:
    - RN_t – nominal interest rate
    - RN_{t−1} – lagged nominal interest rate
    - RRR_t – neutral real interest rate
    - π_{t+3} – expected annual inflation
    - Ỹ_t – output gap
    - ε_t^{RSHOCK} – shock
- Exchange rate and UIP:
  - S_t = S_t^e + (R^F_t − R_t + CR_PM_t^4) + ε_t^s
  - S_t^e = ω S_{t+1} + (1−ω)(S_{t−1} + 2/4∙∆S̄_t)
  - Definitions:
    - S_t – nominal exchange rate
    - S_t^e – expected exchange rate
    - R^F_t – foreign nominal interest rate
    - R_t – domestic nominal interest rate
    - PRPM_t – risk premium (notation in source)
    - ∆S̄_t – nominal exchange rate trend
    - ε_t^s – shock
  - Key bilateral exchange rate used in UIP: ISK/EUR.
- Fiscal policy and fiscal impulse:
  - FI_t = ε_t^{CPPF,SPEC_RS} + PDISCRETIONARY_t
  - PRIMARY_F_t = PRIMARY_F_t^{STRUCTURAL} + PRIMARY_F_t^{CYCLICAL} + PRIMARY_F_t^{ONE-OFF}
  - Definitions:
    - FI_t – fiscal impulse
    - PRIMARY_F_t – primary balance
    - PRIMARY_F_t^{STRUCTURAL} – structural component
    - PRIMARY_F_t^{CYCLICAL} – cyclical component
    - PRIMARY_F_t^{ONE-OFF} – one-off discretionary component
- Labor market:
  - UNEMP_GAP_t = c_1 ⋅ UNEMP_GAP_{t−1} − c_2 ⋅ Ỹ_t + ε_t^{UNEMP_SHOCK}
  - UNEMP_TR_t = c_1 ⋅ UNEMP_TR_{t−1} + (1−c_1) ⋅ UNEMP_TR_{ss} + ε_t^{NAIRU_SHOCK}
  - Definitions:
    - UNEMP_GAP_t – unemployment gap
    - Ỹ_t – output gap
    - ε_t^{UNEMP_SHOCK} – unemployment shock
    - UNEMP_TR_t – unemployment trend (NAIRU)
    - UNEMP_TR_{ss} – steady-state rate of unemployment
    - ε_t^{NAIRU_SHOCK} – shock
- External economy and partners:
  - Partner economies modeled via equations for output gap, inflation, and exchange rates.
  - Key partners: European Union, United States, United Kingdom.
  - Together these partners account for approximately 85 percent of total exports.
- Calibration, priors, and estimation:
  - Model is partially calibrated; most parameters estimated using a Bayesian approach.
  - Calibrated steady states and shares/weights:
    - Potential output growth steady state is set at 2.7 percent.
    - Iceland’s inflation target per Central bank of Iceland policy is set at 2.5 percent.
    - Real exchange rate depreciation is set to 0 percent.
    - Real neutral interest rate is set to 2.7 percent.
    - Key economic partners have inflation target in 2 percent.
    - The real neutral interest rate for key bilateral partner, EA, is set to 0.5 percent.
  - Other parameters, including standard deviations, are estimated using data from the 2001 to 2024 period.
- Selected estimated parameter values (priors / posterior modes as reported):
  - IS curve (Eq. 1) parameters:
    - α_1 = 0.485
    - α_3 = 0.225
    - α_3 (second listing) = 0.70
    - α_4 = 0.45
    - α_5 = 0.025
  - Headline CPI (Eq. 2) and non-housing inflation (Eq. 3) parameters:
    - δ = 0.8
    - β_1 = 0.55
    - β_2 = 0.285
    - β_3 = 0.04
  - Interest rate rule (Eq. 4) parameters:
    - γ_1 = 0.675
    - γ_2 = 1.275
    - γ_3 = 0.25
- Model properties, shocks, and impulse responses:
  - Analyses performed: Impulse response functions (IRFs), forecast error variance decomposition, and in-sample historical simulations.
  - Shock experiments described:
    - Demand shock represented by fiscal impulse (one time increase in structural deficit).
    - Interest rate shock (one time increase in key monetary policy rate).
  - Qualitative findings:
    - Results highlight important role of policy regime for both fiscal and monetary policies.
    - The stabilizing character of the fiscal policy: the fiscal balance returns to its long-term value; a fiscal impulse produces contraction the next period after the assumed increase in deficit.

*Source: IMF staff paper "Assessing the Monetary Policy Implications of Changes in Owner-Occupied Housing Cost Measurement: Insights from IceQMod" (May xx, 2025).*

### References ____________________________________________________________________________ 12

### References ____________________________________________________________________________ 12

### A. Introduction
- Iceland: small, advanced economy with a floating exchange rate and significant trade openness.
- Central Bank of Iceland (CBI) inflation-targeting framework with mandate to maintain inflation at a target rate of 2.5 percent.
- Since the inflation target was set in 2001, inflation has averaged 4.9 percent.
- Introduces IceQMod, an Iceland-specific Quarterly Projection Model (QPM):
  - QPM: medium-sized semi-structural model used for forecasting and monetary policy analysis.
  - Calibrated to Icelandic data; tailored to capture essential features of the Icelandic economy.
- Objective: assess implications of a change in methodology for calculating imputed rent (owner-occupied housing services) in the CPI, moving from “user cost” to “rental equivalence” effective June 2024.
- Imputed rent sub-component weight in CPI: 20 percent.
- Under the prior “user cost” approach, imputed rent contributed an average 1.8 percentage point to headline CPI inflation between 2001 and mid-2024.

### B. Overview of IceQMod
- Model character:
  - Semi-structural gap model with New Keynesian short-run dynamics (sticky wages and prices).
  - State-space representation allowing Kalman filter decomposition into gap and trend components.
- Key Iceland-specific features captured:
  - Export equation reflects dominance of fisheries, aluminum, and tourism.
  - Import equation captures low dependence on energy imports.
  - Trade equations capture openness of the economy.
  - Population growth (substantial over past decade) influences real estate prices.
  - Highly elastic labor supply due to access for citizens of EU countries and procyclical immigration.
  - Separate Phillips curves for housing and non-housing inflation: housing component has different response to demand and exchange rate shocks and a higher trend inflation rate.
  - UIP condition adapted for strong connection to euro-denominated financial markets and high capital mobility.
- Appendix contains detailed description of IceQMod.

### C. Measuring imputed rent: methods and synthetic CPI construction
- Three broad approaches to estimating imputed rent (Consumer Price Index Manual, 2004):
  - User cost:
    - Includes mortgage cost, taxes, maintenance costs, depreciation based on current market value minus expected capital gain.
    - House prices and mortgage payments tend to dominate estimates.
    - Iceland used a version of this approach until May 2024.
  - Rental equivalence:
    - Estimates the rent homeowners would pay if they rented their homes.
    - Suitable where active rental market properties can proxy owner-occupied homes.
    - Iceland adopted this methodology from June 2024.
  - Acquisitions:
    - Looks at cost of purchasing and owning a dwelling, including renovations, home insurance, maintenance, and transfer costs.
- Theoretical note:
  - In frictionless, perfect-data setting, user cost and rental equivalence should yield the same imputed rent (National Academy of Sciences, Engineering, and Medicine, 2022).
- Synthetic historical CPI construction (second-best approach):
  - Expanded HMS rental database only recently established; cannot directly replace historical user-cost imputed rent.
  - Synthetic CPI created by assuming imputed rent mirrors existing actual rent sub-component.
  - Caveats:
    - Actual rents database includes private rents, social housing, and student rentals; rental equivalence uses private rentals only.
    - Owner-occupied property types may differ from rental stock; housing mix differences may bias synthetic measure.
- Empirical comparisons (2002–2024):
  - Synthetic CPI tends to follow published CPI but with significant divergences in specific periods (2005–06, 2008, 2015–19, 2022–24).
  - Example divergence: 2015–2019 synthetic inflation ≈ 1.2 percent versus published inflation 2.2 percent.
  - Averages and volatility (Text Table 1):
    - Since 2002 Published CPI Average: 4.9
    - Since 2002 IMF Synthetic CPI Average: 4.7
    - Standard deviation Published CPI: 3.3
    - Standard deviation IMF Synthetic CPI: 4.1
  - Synthetic measure is on average 0.2 percentage points lower than published CPI over 2002–2024.
  - Synthetic CPI is more volatile than published CPI, despite rental inflation being more stable than house prices.
  - Explanation: house prices in Iceland are negatively correlated with CPI excluding housing; house price inflation tends to move with a stronger exchange rate while CPI excluding housing moves opposite, producing offsetting effects on headline CPI.

### D. Counterfactual experiment: monetary policy with rental-equivalence imputed rent
- Experiment setup:
  - Replace actual CPI in IceQMod with synthetic CPI (rental-equivalence proxy).
  - Keep model parameters unchanged.
  - Make policy interest rate unobservable and estimate it via the Kalman filter.
- Key findings:
  - Policy rate would have been slightly lower on average under synthetic inflation and more volatile.
  - Significant differences between actual CBI policy rate and model-implied policy rate based on synthetic CPI during periods of robust house price inflation: 2005–08 and 2013–21.
  - In those periods, synthetic CPI would have implied a more accommodative monetary policy stance than actual decisions, suggesting the Monetary Policy Committee may have been effectively “leaning against the wind” on asset prices.
  - Likely counterfactual: house price inflation would have been higher if policy had followed synthetic CPI; central bank could have used macroprudential tools to stabilize housing market.

### E. Policy implications and dynamics going forward
- Measurement change implications:
  - For an unchanged inflation target, a methodological change that reduces average inflation allows the central bank to maintain a looser nominal policy stance for a period.
  - The methodology change entails an increase in the equilibrium nominal interest rate.
    - Footnote example: UK switch from RPIX to CPI lowered average CPI by 0.8 percentage points; Chancellor revised target from 2.5 percent to 2.0 percent; decision not to fully revise resulted in de facto 0.3 percentage point rise in level of inflation consistent with target and equilibrium nominal interest rate rose by 0.3 percentage points (Nickell, 2003).
  - Equilibrium real interest rate is unchanged; no bearing on real economy from the nominal-equilibrium change alone.
  - Because revised methodology excludes house prices:
    - Monetary policy could be looser during periods of growing house prices and tighter during periods of weakness, assuming unchanged monetary policy reaction function.
- Macroprudential policy:
  - Increased need to adopt macroprudential measures to manage excessive house price movements, since monetary policy may no longer be implicitly targeting asset prices via headline CPI.

### F. Conclusions (summarized)
- The change in imputed rent measurement (to rental equivalence) could produce:
  - Lower average inflation: synthetic CPI is on average 0.2 percentage points lower than published CPI over 2002–2024.
  - More volatile inflation: synthetic inflation displays higher standard deviation (4.1) versus published CPI (3.3).
- Monetary policy effects:
  - Central bank could keep monetary policy looser for a period if average CPI inflation is lower.
  - Policy rate behavior: lower during high house price inflation episodes and higher during low house price inflation episodes if reaction function unchanged.
- Recommendation:
  - Greater reliance on macroprudential measures may be required to manage house price fluctuations now that headline CPI excludes direct house price effects.

### Annex I. IceQMod core structure (select equations and parameters)
- Real sector (IS curve):
  - Y_t_gap = α1⋅Y_t-1_gap − α2⋅MCI_t + α3⋅Y_t^F_gap + α4⋅FI_t + α5⋅TP_t_gap + ε_t^Y
  - Monetary Conditions Index (MCI): weighted combination of real interest rate gap and real exchange rate gap:
    - MCI_t = w⋅RRG_t + (1−w)⋅Ẑ_t
- Prices and inflation:
  - Headline CPI P_t comprises housing P_t^H and non-housing P_t^XH components:
    - ΔP_t = δ⋅ΔP_t^XH + (1−δ)⋅ΔP_t^H + ψ_t_CCC
  - Non-housing CPI inflation:
    - π_t^XH = β1⋅π_t−1^XH + (1−β1−β3)⋅π_t+1^XH + β2⋅RRM_XH + β3⋅π_t^imp,xh + ε_t^XH
  - Steady-state inflation assumptions:
    - Housing prices inflation steady state: 3.25
    - Non-housing steady state: 2.25 percent
    - These constitute total inflation at (text truncated in source).

*Source: IMF staff paper "Assessing the Monetary Policy Implications of Changes in Owner-Occupied Housing Cost Measurement: Insights from IceQMod" (May xx, 2025).*

### 2.5 percent as per central bank inflation target. Given the housing inflation indicator construct it is

### sipea2025100 - 2.5 percent as per central bank inflation target. Given the housing inflation indicator construct it is

### Housing inflation construct
- Housing inflation equation (notation preserved):
  - π_t^H = c_1 ⋅ π_{t−1}^H + (1−c_1) ⋅ π_{t−1}^H + c_2 ⋅ RM_M^H_t + c_3 ⋅ RP̄_t + ε_t^{SHOCK,H}
- Definitions:
  - π_t^H – housing inflation
  - RM_M^H_t – real marginal costs for housing
  - ΔRP_t – trend of real housing price (notationally indicated in source)
  - ε_t^{SHOCK,H} – shock
- Behavioral assumption:
  - Housing inflation is assumed to be procyclical to the monetary policy stance using real interest rate gap (housing costs increase with higher short-term interest rate as it translates into higher mortgage interest rates).

### Monetary policy (Taylor-type rule)
- Interest rate rule (notation preserved):
  - RN_t = γ_1 ⋅ RN_{t−1} + (1−γ_1) ⋅ RRR_t + π^{target}_t + γ_2 ⋅ (π_{t+3} − π^{target}_t) + γ_3 ⋅ Ỹ_t + ε_t^{RSHOCK}
- Definitions:
  - RN_t – nominal interest rate
  - RN_{t−1} – lagged nominal interest rate
  - RRR_t – neutral real interest rate
  - π_{t+3} – expected annual inflation
  - Ỹ_t – output gap
  - ε_t^{RSHOCK} – shock

### Exchange rate and UIP
- UIP condition (notation preserved):
  - S_t = S_t^e + (R^F_t − R_t + CR_PM_t^4) + ε_t^s
- Expected exchange rate:
  - S_t^e = ω S_{t+1} + (1−ω)(S_{t−1} + 2/4∙∆S̄_t)
- Definitions:
  - S_t – nominal exchange rate
  - S_t^e – expected exchange rate
  - R^F_t – foreign nominal interest rate
  - R_t – domestic nominal interest rate
  - PRPM_t – risk premium (notation in source)
  - ∆S̄_t – nominal exchange rate trend
  - ε_t^s – shock
- Key bilateral exchange rate used in UIP: ISK/EUR given importance of EU.

### Fiscal policy and fiscal impulse
- Fiscal impulse definition (notation preserved):
  - FI_t = ε_t^{CPPF,SPEC_RS} + PDISCRETIONARY_t
- Primary balance decomposition (notation preserved):
  - PRIMARY_F_t = PRIMARY_F_t^{STRUCTURAL} + PRIMARY_F_t^{CYCLICAL} + PRIMARY_F_t^{ONE-OFF}
- Definitions:
  - FI_t – fiscal impulse; comprises shock to the structural balance (including permanent discretionary items) and one-off discretionary component
  - PRIMARY_F_t – primary balance
  - PRIMARY_F_t^{STRUCTURAL} – structural component
  - PRIMARY_F_t^{CYCLICAL} – cyclical component
  - PRIMARY_F_t^{ONE-OFF} – one-off discretionary component of the balance

### Labor market
- Unemployment gap equation (notation preserved):
  - UNEMP_GAP_t = c_1 ⋅ UNEMP_GAP_{t−1} − c_2 ⋅ Ỹ_t + ε_t^{UNEMP_SHOCK}
- NAIRU (unemployment trend) equation (notation preserved):
  - UNEMP_TR_t = c_1 ⋅ UNEMP_TR_{t−1} + (1−c_1) ⋅ UNEMP_TR_{ss} + ε_t^{NAIRU_SHOCK}
- Definitions:
  - UNEMP_GAP_t – unemployment gap
  - Ỹ_t – output gap
  - ε_t^{UNEMP_SHOCK} – unemployment shock
  - UNEMP_TR_t – unemployment trend (NAIRU)
  - UNEMP_TR_{ss} – steady-state rate of unemployment
  - ε_t^{NAIRU_SHOCK} – shock

### External economy and partners
- Representation:
  - Partner economies modeled via equations for output gap, inflation, and exchange rates.
- Key partners:
  - European Union is major trade partner and counterpart in UIP condition.
  - United States is significant partner in services, particularly tourism.
  - United Kingdom also a significant partner.
  - Together these partners account for approximately 85 percent of total exports.

### Calibration, priors, and estimation
- Calibration and estimation approach:
  - Model is partially calibrated; most parameters estimated using a Bayesian approach.
  - Calibrated steady states and shares/weights include potential output, inflation target, real exchange rate depreciation, real interest rate, housing/non-housing inflation shares, external trade, monetary conditions, and marginal costs.
  - Other parameters, including standard deviations, are estimated using data from the 2001 to 2024 period.
- Calibrated steady states and values (preserved exactly):
  - Potential output growth steady state is set at 2.7 percent.
  - Iceland’s inflation target per Central bank of Iceland policy is set at 2.5 percent.
  - Real exchange rate depreciation is set to 0 percent (historical average over last two decades).
  - Real neutral interest rate is set to 2.7 percent—to balance yield of financial assets with potential output growth.
  - Key economic partners have inflation target in 2 percent.
  - The real neutral interest rate for key bilateral partner, EA, is set to 0.5 percent.

### Selected estimated parameter values (priors / posterior modes as reported)
- IS curve (Eq. 1) parameters:
  - α_1 = 0.485
  - α_3 = 0.225
  - α_3 (second listing) = 0.70
  - α_4 = 0.45
  - α_5 = 0.025
- Headline CPI (Eq. 2) and non-housing inflation (Eq. 3) parameters:
  - δ = 0.8
  - β_1 = 0.55
  - β_2 = 0.285
  - β_3 = 0.04
- Interest rate rule (Eq. 4) parameters:
  - γ_1 = 0.675
  - γ_2 = 1.275
  - γ_3 = 0.25

### Model properties, shocks, and impulse responses
- Analyses performed:
  - Impulse response functions (IRFs), forecast error variance decomposition, and in-sample historical simulations.
- Shock experiments described:
  - Demand shock represented by fiscal impulse (one time increase in structural deficit).
  - Interest rate shock (one time increase in key monetary policy rate).
- Qualitative findings highlighted:
  - Results highlight important role of policy regime for both fiscal and monetary policies.
  - The stabilizing character of the fiscal policy: the fiscal balance returns to its long-term value; a fiscal impulse produces contraction the next period after the assumed increase in deficit.

*Source: sipea2025100 - 2.5 percent as per central bank inflation target. Given the housing inflation indicator construct it is*

---


_Source: https://www.imf.org/-/media/files/publications/selected-issues-papers/2025/english/sipea2025100.pdf_