## Appendix I: Details on the Pension System (wpiea2020153-print-pdf)

## Source details

**Canonical URL:** [Appendix I: Details on the Pension System (wpiea2020153-print-pdf)](https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020153-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2020/english/wpiea2020153-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2020/english/wpiea2020153-print-pdf.pdf.json)

---

### I. Introduction and model extensions
- Purpose: Document changes made to the Global Integrated Monetary and Fiscal model (GIMF) to incorporate demographic features so the IMF can analyze macroeconomic implications of demographic developments for bilateral and multilateral surveillance.
- Key model extensions introduced:
  - Time-varying birth and survival rates.
  - Age-decreasing hours supply to account for retirement.
  - A pay-as-you-go public pension system to capture fiscal implications of aging more fully.
- Rationale for simplified demographic structure:
  - Three parsimonious demographic variables capture essential dynamics:
    - the total population growth rate,
    - the average age,
    - the share of the working-age population in the total population.
  - Trade-off: enables multi-regional general equilibrium analysis at the cost of precise inter-generational redistributive detail.

### Summary of GIMF features relevant for demographic extension
- Model type and frictions:
  - Multicountry DSGE with optimizing households and firms, full intertemporal stock-flow accounting.
  - Frictions: sticky prices and wages, real adjustment costs, liquidity-constrained households, finite planning horizons.
  - Finite planning horizons enable well-defined steady states and meaningful transitions.
- Asset markets and financial accelerator:
  - Government debt: nominal, non-contingent, one-period domestic currency bonds held domestically.
  - International assets: nominal, non-contingent, one-period U.S. dollar bonds.
  - Financial accelerator: external finance premium rises with leverage; non-linearities produce steep premium increases for large negative net worth shocks.
- Household types:
  - Overlapping-generation households (OLG) with a 20-year planning horizon.
  - Liquidity-constrained households (LIQ) with marginal propensity to consume equal to unity.
- Fiscal and monetary policy:
  - Fiscal instruments include government consumption, investment, transfers, taxes, and tariffs.
  - Fiscal rule ensures long-run sustainability: government debt-to-GDP and deficit-to-GDP converge to targets.
  - Monetary policy: inflation forecast-based interest rate rule.

### Introducing demographic features: definitions and key algebra
- Time-varying demographic primitives introduced (notation preserved):
  - birth rates b_t (푏푏푡),
  - survival probabilities θ_t (휃휃푡),
  - age-dependent decay in hours endowment 0 < χ^h_t < 1 (0 < 휒휒ℎ𝑡 < 1).
- Population and age dynamics (notation preserved):
  - Total population evolution: N_t = n_t N_{t-1}, where n_t = b_t + θ_t.
  - Average age evolution: A_t n_t = b_t + (A_{t-1}+1) θ_t.
  - Working-age share s_t: n_t s_t = b_t + θ_t χ^h_t s_{t-1}.
- Lifetime utility and labor endowment:
  - Survival probability discounting introduced via Θ terms.
  - Age-dependent hours endowment enters via Χ terms.
  - Per-capita labor endowment l̄_t evolves as: n_t l̄_t = b_t + θ_t χ_a χ^h_{t+1} l̄_{t-1} (notation preserved).
- Human and pension wealth:
  - Human wealth per unit of labor endowment defined by a forward-looking recursive equation; per-capita human wealth is l̄_t times that per-unit term.
  - Pension wealth computed as difference of two composite forward- and backward-looking recursive terms; per-capita pension wealth e_w̄_t given by e̅_{t,1} eew_{t,1} − e̅_{t,2} eew_{t,2}.
  - Aggregate consumption identity in per-capita terms: c_t^{OLG} Θ_t = f_w̃_t + hw̃_t + ew̃_t (notation preserved).

### Government budget constraint and pensions
- Per-capita transfers include pensions via:
  - τ̌^P_t = (e̅_{t,1} − e̅_{t,2}) w̃^p_t.
- Budget financing in simulations reported:
  - Lump-sum taxes adjust endogenously to finance pension payments (alternative instruments are allowed by the model).
- Pension system determinants and notation:
  - t is current period (year); a is age measured as periods since newcomer.
  - W_{t−a} is the wage level when the household was a newcomer.
  - h is the hours endowment of a newcomer; h_{a,t} is hours endowment at period t for the cohort.
  - “hours in retirement” = h − h_{a,t}.
- Pension formula components:
  - Cohort-specific replacement rate decided at entry: τ_{s p_{t−a}}.
  - Current replacement rate applied to all retirees: τ_{h p_{t}}.
  - Replacement wage growth factor (indexation) π_{t p} defined by:
    - log(π_{t p}) = (1−χ_{t}) log(π_{t c}) + χ_{t} log(π_{t w}), where χ_{t} is policy-determined weight.
  - General pension level formula (notation preserved):
    - P_{a,t} = τ_{h p_{t}} τ_{s p_{t−a}} (h_{t−a} − h_{a,t}) ∏_{k=0}^{a−1} π_{t−k p} W_{t−a}.
  - Full wage indexation case (π_{t p} = π_{t w}):
    - P_{a,t} = τ_{h p_{t}} τ_{s p_{t−a}} (h_{t−a} − h_{a,t}) W_{t}.

### Multi-country demographic features and calibration
- Country/region specific and time-varying demographic variables:
  - population growth factor n_{t}^r,
  - planning horizon via survival probability,
  - hours endowment decline factor χ^h_t.
- Long-run restriction for steady state:
  - Demographic variables must stabilize; population growth rates must stabilize to a common value across regions.
- World aggregates (notation preserved):
  - World population growth factor: n_t ≡ Σ_{r∈R} s_{t-1}^{we,r} n_t^r.
  - Share in world population evolves: s_{t}^{we,r} = (n_t^r / n_t) s_{t-1}^{we,r}.
- Calibration choices:
  - Model covers working-age population and retirees; excludes young population below entry-age (country-specific threshold 15 or 20 years).
  - Survival probability set so steady-state planning horizon equals 20 years.
  - Hours endowment decline factor calibrated to reproduce observed working-age population shares.

### Properties of demographic shocks: simulation experiments (three illustrative scenarios)
- Overview:
  - Demographic transition represented by joint changes in: birth rate (population growth), survival probability (average age), and speed of labor endowment decline (working-age share).
  - Experiments conducted in a three-country model; graphical results summarize transitional dynamics.
- Experiment 1 — Decline in population growth rate:
  - Shock: birth rate falls by 0.25 percentage points over 10 years (temporary shock).
  - Outcomes (exact phrasing preserved where numeric):
    - Population size falls by about 2½ percent after 10 years.
    - Share of working-age population declines by about half a percent after 10 years.
    - Per-capita GDP declines by about ½ a percent at its trough.
    - Investment drops immediately; reduced capital accumulation lowers potential output.
    - Demand declines more than supply → increase in national saving → trade surplus and temporary accumulation of foreign assets.
    - Central bank lowers nominal rate; small short-lived depreciation in real exchange rate, later real appreciation.
    - Net effect on rest of world roughly zero (global interest rate falls stimulating investment vs. negative terms-of-trade via domestic currency appreciation).
- Experiment 2 — Increased life expectancy with unchanged population growth:
  - Shock: permanent increase in survival probability by one percentage point, birth rate reduced so population growth unchanged.
  - Outcomes:
    - Share of working-age population stabilizes about 4 percent below initial level.
    - Output declines by about 6 percent in the long run.
    - Each household accumulates more assets over lifecycle → strong improvement in trade balance peaking at almost 2 percent after about 7 years.
    - Short-run: lower demand → lower inflation and nominal interest rate; real exchange rate depreciates initially, appreciates in long run.
    - Spillovers to rest of world larger and positive in long run; global real interest rate declines.
- Experiment 3 — Earlier retirement (accelerated decline in hours endowment):
  - Shock: permanent raise in decay rate of labor endowment in age by one percentage point.
  - Outcomes:
    - Adjustment closely resembles increased life expectancy shock (with reduced birth rate).
    - Accelerated decline in per-capita labor endowment akin to higher average age.
    - Incentive to accumulate assets strengthened due to longer expected retirement.

### Conclusions and applications
- Demographic features (time-varying birth rates, survival rates, and age-dependent labor endowments) have been incorporated into GIMF and coded.
- The model illustrates adjustment properties to changes in the three demographic levers and can be used to examine macroeconomic and fiscal implications of demographic change.
- Application example: this GIMF version has been applied in multilateral surveillance (e.g., Fernandez-Corugedo and Colacelli (2018) for Japan and policy offsets to demographic headwinds).

*Source: Appendix I: Details on the Pension System, wpiea2020153-print-pdf*

### Appendix I: Details on th e Pen sion Sy stem ........................................................................21

### Appendix I: Details on th e Pen sion Sy stem

### I. INTRODUCTION
- Purpose: Document changes made to the Global Integrated Monetary and Fiscal model (GIMF) to incorporate demographic features so the IMF can analyze macroeconomic implications of demographic developments for bilateral and multilateral surveillance.
- Context: The original GIMF model is documented by Kumhof and others (2010) and Anderson and others (2013), and extended by Carton and others (2017). The original assumes a Blanchard-Yaari overlapping-generation household sector with constant population growth rate and constant probability to survive (Yaari, 1965; Blanchard, 1985).

### Key model extensions introduced
- Time-varying birth and survival rates.
- Age-decreasing hours supply to account for retirement.
- A pay-as-you-go public pension system to capture fiscal implications of aging more fully.

### Rationale for a simplified demographic structure
- A fully-fledged demographic structure would track year-by-year total population growth and the age pyramid (repartition by age).
- The paper introduces a parsimonious demographic structure that captures dynamics in three variables:
  - the total population growth rate,
  - the average age,
  - the share of the working-age population in the total population.
- These three variables are identified as key to explaining:
  - the growth rate of labor supply,
  - the aggregate saving rate,
  - the fiscal implications of a public pay-as-you-go pension system.
- Trade-off: The simplified structure does not allow precise analysis of inter-generational redistributive effects of macroeconomic shocks or policy reforms, but it facilitates multi-regional frameworks to assess global general equilibrium implications of country and regional differences in demographic trends.

### Illustrative uses and simulated outcomes
- The extended model is used to estimate macroeconomic impacts of demographic change.
- Simulations demonstrate that changes in population and labor force have important implications for:
  - growth,
  - savings and investment,
  - real interest rates,
  - fiscal policy.

### Paper organization (as stated)
- Section II: Overview of GIMF.
- Section III: Documentation of changes made to incorporate demographic features.
- Section IV: Simulations that illustrate model adjustment properties to shocks embodying demographic change.
- Section V: Conclusion.

*Source: Appendix I: Details on th e Pen sion Sy stem, wpiea2020153-print-pdf*

### Appendix I provides details of the algebraic derivation associated with the introduction of

### wpiea2020153-print-pdf - Appendix I provides details of the algebraic derivation associated with the introduction of

### Summary of the Global Integrated Monetary and Fiscal Model (GIMF)
- GIMF is a multicountry DSGE model with optimizing households and firms, and full intertemporal stock-flow accounting.
- Frictions: sticky prices and wages, real adjustment costs, liquidity-constrained households, finite planning horizons.
- Finite planning horizons allow well-defined steady states with countries as long-run debtors or creditors and meaningful transitions.
- Non-Ricardian features produce non-neutrality of spending-based and revenue-based fiscal measures; sustained government deficits crowd out private investment and net foreign assets in the long run.
- Asset markets are incomplete:
  - Government debt: nominal, non-contingent, one-period bonds in domestic currency and held domestically.
  - International assets: nominal, non-contingent, one-period bonds denominated in U.S. dollars issued by the U.S. government and private agents.
  - Firms are domestically owned; equity is not traded domestically; households receive lump-sum dividends.
- Financial sector includes a procyclical financial accelerator (external finance premium rises with leverage).
- Model version used comprises 3 regions with different calibrations for exposition and to analyze policy spillovers.

### Household sector: types, behavior, and implications
- Two household types:
  - Overlapping-generation households (OLG) with a 20-year planning horizon.
  - Liquidity-constrained households (LIQ) who do not save and have marginal propensity to consume of unity.
- Taxes: direct on labor income, indirect on consumption, and lump-sum taxes for all households.
- OLG saving instruments: domestic government bonds, international U.S. dollar bonds, fixed-term deposits.
- OLG consumption depends on financial wealth and present discounted value of after-tax wage, pension, and investment income.
- Key implications:
  - A short-run tax cut raises consumption for OLG; if matched by future tax increases leaving long-run debt unchanged, short-run consumption still rises (time profile tilt).
  - Permanent increases in government debt raise real interest rates and crowd out private capital.
  - Increases in the interest rate reduce consumption mainly via wealth effects; intertemporal substitution calibrated to empirical evidence.
  - The intertemporal elasticity of substitution determines magnitude of long-run crowding-out.

### Production sector: firms and adjustment frictions
- Firms produce tradable and nontradable intermediate goods; managed according to owners (finite-lived households) implying a substantial equity premium driven by impatience.
- Nominal rigidities in price setting and real adjustment costs in hiring and investment.
- Firms pay capital income taxes, wages, and dividends to OLG households.
- Investment cannot be fully financed by retained earnings; firms borrow from financial intermediaries and face bankruptcy/auditing costs.
- Markets are monopolistically competitive; prices contain a markup over marginal cost.
- Firms use public infrastructure (government capital stock) as an input—government capital augments productivity.

### Financial sector: assets, banks, and premia
- Financial assets menu limited: domestic one-period government bonds; one-period fixed-term deposits offered by banks.
- Ownership of firms, deposits, and domestic government debt are non-tradable internationally.
- OLG households may issue/purchase tradable U.S.-dollar obligations.
- Lending rates include an external financing premium that varies with leverage (financial accelerator); non-linearities cause steep premium increases for large negative net worth shocks.
- Uncovered interest parity may not hold due to country risk premiums, producing both short- and long-run interest rate deviations across regions.

### International dimensions and spillovers
- All bilateral trade flows and relative prices (including exchange rates) are explicitly modeled and calibrated to steady-state observed flows.
- International linkages driven by global saving and investment decisions, resulting from consumers’ finite horizons; yields uniquely defined current account balances and net foreign asset positions.
- Net foreign asset positions represented by nominal, non-contingent one-period U.S. dollar bonds.
- Spillovers determined mainly by uncovered interest parity, world real interest rate movements, and trade linkage magnitudes.

### Fiscal and monetary policy design in GIMF
- Fiscal instruments: government consumption, investment, lump-sum transfers (general or LIQ-targeted), taxes on labor and corporate income, consumption taxes, lump-sum taxes, and tariffs.
- Government investment augments public infrastructure that depreciates at a constant rate.
- Fiscal policy rule:
  - Ensures long-run sustainability: government debt-to-GDP and deficit-to-GDP converge to targets (precludes sovereign default).
  - Allows short-run countercyclical policy via automatic stabilizers.
  - Instruments: labor and capital income taxes (can be replaced by other instruments per region).
- Monetary policy: inflation forecast-based interest rate rule adjusting the gap between actual policy rate and long-run equilibrium to stabilize inflation.

### Introducing demographic features: key model changes
- Time-varying demographic variables introduced: birth rates (푏푏푡), survival probabilities (휃휃푡), and age-dependent decay in hours endowment (0 < 휒휒ℎ𝑡 < 1).
- Total population evolution: N_t = n_t N_{t-1}, where n_t = b_t + θ_t.
- Average age evolution: A_t n_t = b_t + (A_{t-1}+1) θ_t.
- Working-age share s_t evolves: n_t s_t = b_t + θ_t χ^h_t s_{t-1}; χ^h_t captures retirement as gradual decline in hours endowment.
- Lifetime expected utility adjusted to include:
  - Survival probability discounting (Θ terms).
  - Age-dependent hours endowment via Χ terms.
- Budget constraint modified to include public pensions: pension receipt appears as last term scaled by (1 − Χ_{...}^h) and an effective replacement wage W_t^p based on birth-period nominal net wage (indexation allowed).
- Per-capita labor endowment l̄_t becomes demographic-dependent and evolves as:
  - n_t l̄_t = b_t + θ_t χ_a χ^h_{t+1} l̄_{t-1} (notation preserved as in source).
- Human wealth defined per unit of labor endowment with a forward-looking recursive equation; per-capita human wealth is l̄_t times that per-unit term.
- Pension wealth computed as difference of two composite forward- and backward-looking recursive terms; per-capita pension wealth e_w̄_t given by e̅_{t,1} eew_{t,1} − e̅_{t,2} eew_{t,2}.
- Aggregate consumption identity in per-capita terms replaced by:
  - c_t^{OLG} Θ_t = f_w̃_t + hw̃_t + ew̃_t (notation preserved).

### Government budget constraint and pensions
- Per-capita transfers now include pensions:
  - τ̌^P_t = (e̅_{t,1} − e̅_{t,2}) w̃^p_t.
- In simulations reported, lump-sum taxes adjust endogenously to finance pension payments (model allows other instruments).

### Multi-country features of demographic variables
- Country/region specific and time-varying: population growth factor n_{t}^r, planning horizon (via survival probability), hours endowment decline factor.
- Long-run restriction: these variables must stabilize; population growth rates must stabilize to a common value to ensure existence of a final steady state.
- World population growth factor defined as n_t ≡ Σ_{r∈R} s_{t-1}^{we,r} n_t^r.
- Share in world population evolves: s_{t}^{we,r} = (n_t^r / n_t) s_{t-1}^{we,r}.

### Calibrating demographic variables
- Model covers working-age population and retirees; excludes young population. Population above entry-age (country-specific threshold 15 or 20 years) used.
- Planning horizon calibration: survival probability set so steady-state planning horizon equals 20 years (informed by wealth effect estimates) despite average life expectancy being higher.
- Hours endowment decline factor calibrated to reproduce observed share of working-age population.

### Properties of demographic shocks: simulation experiments
- Demographic transition represented by joint changes in three variables: birth rate (population growth), survival probability (average age/life expectancy), and speed of labor endowment decline (working-age share).
- Three illustrative simulations in a three-country model:
  1. Decline in population growth rate: birth rate falls by 0.25 percentage points over 10 years (temporary shock).
     - Population size falls by about 2½ percent after 10 years.
     - Share of working-age population declines by about half a percent after 10 years.
     - Per-capita GDP declines by about ½ a percent at its trough.
     - Output, consumption, investment, imports and exports adjust downwards at the twenty-year horizon roughly in line with labor supply decline.
     - Investment drops immediately; reduced capital accumulation lowers potential output.
     - Demand declines more than supply → increase in national saving → trade surplus and temporary accumulation of foreign assets.
     - Central bank lowers nominal rate; small short-lived depreciation in real exchange rate, later real appreciation.
     - Net effect on rest of world roughly zero (global interest rate falls stimulating investment vs. negative terms-of-trade via domestic currency appreciation).
  2. Increased life expectancy: permanent increase in survival probability by one percentage point, birth rate reduced so population growth unchanged.
     - Share of working-age population stabilizes about 4 percent below initial level.
     - Output declines by about 6 percent in the long run.
     - Each household accumulates more assets over lifecycle → strong improvement in trade balance peaking at almost 2 percent after about 7 years.
     - Short-run: lower demand → lower inflation and nominal interest rate; real exchange rate depreciates initially, appreciates in long run.
     - Spillovers to rest of world larger and positive in long run; global real interest rate declines.
  3. Earlier retirement: permanent raise in decay rate of labor endowment in age by one percentage point.
     - Adjustment closely resembles increased life expectancy shock (with reduced birth rate).
     - Accelerated decline in per-capita labor endowment akin to higher average age.
     - Incentive to accumulate assets strengthened due to longer expected retirement.
- Graphical results (Figures 1–3) illustrate transitional dynamics for each experiment (per-source figures include time paths for population size, real aggregates, working-age share, average age, real effective exchange rate, trade balance, inflation and interest rates, net foreign assets, and rest-of-world responses).

### Conclusions and applications
- Demographic features (time-varying birth rates, survival rates, and age-dependent labor endowments) have been incorporated into GIMF and coded.
- Model illustrates adjustment properties to changes in the three demographic levers and can be used to examine macroeconomic and fiscal implications of demographic change.
- This GIMF version has been applied in multilateral surveillance, e.g., Fernandez-Corugedo and Colacelli (2018) for Japan and policy offsets to demographic headwinds.

*Source: IMF Working Paper (GIMF documentation and demographic extension) — content extracted from the supplied PDF content unit.*

### APPENDIX I: DETAILS ON THE PENSION SYSTEM

### APPENDIX I: DETAILS ON THE PENSION SYSTEM

### Notation and household characteristics
- t is the current period (year) and a the age of the households (periods since he was a newcomer).
- 푊푊_{t−a} is the wage level when the household was a newcomer.
- ℎ is the hours endowment of a newcomer.
- ℎ_{a,t} is the hours endowment of the same household at period t.
- ℎ−ℎ_{a,t} is “hours in retirement”.

### Determinants of the pay-as-you-go public pension
- The level of pensions P_{a,t} depends on:
  - the wage level when the household was a newcomer 푊푊_{t−a};
  - the cohort-specific replacement rate, decided when the household was a newcomer 휏_{s p_{t−a}};
  - the current replacement rate (apply to all retirees) 휏_{h p_{t}};
  - the replacement wage growth factor due to indexation 휋_{t p}.

### Replacement wage growth factor (indexation)
- 휋_{t p} is a geometric average of CPI and wage inflation with a policy-determined weight 휒_{t}:
  - log(휋_{t p}) = (1−휒_{t}) log(휋_{t c}) + 휒_{t} log(휋_{t w})

### Pension level formula
- The level of pensions is given by:
  - P_{a,t} = 휏_{h p_{t}} 휏_{s p_{t−a}} (ℎ_{t−a} − ℎ_{a,t}) �∏_{k=0}^{a−1} 휋_{t−k p}� 푊푊_{t−a}

### Full wage indexation case
- Under full wage indexation 휋_{t p} = 휋_{t w}, the level of pensions simplifies to:
  - P_{a,t} = 휏_{h p_{t}} 휏_{s p_{t−a}} (ℎ_{t−a} − ℎ_{a,t}) 푊푊_{t}

### Policy variables and reform differentiation
- Two policy variables allow differentiation between types of pension reforms:
  - current replacement rate 휏_{h p_{t}} (affects all households the same way);
  - cohort-specific replacement rate 휏_{s p_{t−a}} (allows grandfathering, affecting only new generations).

*APPENDIX I: DETAILS ON THE PENSION SYSTEM (wpiea2020153-print-pdf)*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020153-print-pdf.pdf_
