## _cr07256

## Source details

**Canonical URL:** [_cr07256](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2007/_cr07256.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2007/_cr07256.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2007/_cr07256.pdf.json)

---

### I. MEDIUM-TERM GROWTH AND PRODUCTIVITY IN ESTONIA: MACRO SUMMARY
- Between 1995 and 2005, Estonia’s real GDP per capita rose by an average of 6½ percent a year.
- In 2005 Estonia’s GDP per capita (at current prices and purchasing parity standard, PPS) was 50.9 percent of the EU-15 average.
- Estonia’s capital-labor ratio is only about one-third of that of the EU-15.
- The TFP gap relative to the EU-15 was 40 percent in 2005.
- Output per capita identity used:
  - Y/Pop = (N/Pop) × (E/N) × (L/E) × (Y/L)
  - Where Y is GDP at constant 1995 prices; Pop is population; N/Pop is working-age population to total population (inverse dependency ratio); E/N is the employment rate; L/E is average hours worked per employee; Y/L is labor productivity (output per hour worked).
- Labor productivity decomposition under Cobb-Douglas:
  - Y/L = A × (K/L)^α, with α set to 0.35. A represents TFP (residual).
- Key macro findings (1995–2005 and subperiods):
  - Demographics contributed positively but modestly: annual increase in GDP per capita averaging some 0.3 percent.
  - Labor utilization fell sharply during the 1990s and reversed after 2000; employment rate and hours worked declined in the 1990s but recovered with stronger job creation after 2000.
  - Labor productivity was the main driver of catch-up:
    - Estonia’s labor productivity growth was more than four times that in EU-15 countries during the mid-1990s and about 75 percent higher than in the NMS.
    - TFP growth was the most important component of labor productivity growth, with capital deepening also important.
  - Contributions eased since 2000 but remained high relative to EU-15.
  - A lower recent rise in capital intensity may reflect slower substitution of capital for labor, improved labor market flexibility, and a shift toward less capital-intensive sectors (services), potentially absorbing lower-skilled workers and contributing to a slowdown in measured TFP growth.

### II. MICRO DATA, METHODOLOGY, AND SAMPLE COVERAGE
- Micro data source and coverage:
  - Estonian Business Registry, covering the period 1997-2004 (sample uses data from 1997 due to missing information).
  - Firm-level data from all economic sectors; absence of size thresholds; transactions data available.
  - Because of missing information, only data for about 40 percent of the firms in the registry are used.
  - Value added of enterprises in the sample accounted for 45 percent of aggregate value added in 1997 and increased to 60 percent by 2004.
  - Employment coverage improved over time and was above 50 percent of the macro level in 2004 for all sectors except agriculture and public services.
  - Sample does not include data for the two major banks in Estonia; financial sector results could be biased.
- Deflation and measurement:
  - Value added and intermediate inputs deflated by respective deflators from the Statistical Office of Estonia.
  - Capital deflated by the gross capital formation price index.
  - Results should be interpreted with caution due to data coverage, measurement, and conceptual problems.
- Stylized facts on firm dynamics and structure:
  - Estonian entry and exit rates are fairly high by international standards, although entry rates have declined over time.
  - Firm size is very small: about 70 percent of firms are micro enterprises (less than 10 employees) and firm size is getting smaller across all sectors.
  - Most firms belong to the services sector, where smaller firms dominate.
- TFP estimation methodologies:
  - Method 1: Industry shares (IS)
    - TFP computed as log Yit − αl log Lit − αk log Kit, with αk = 1 − αl under constant returns to scale.
  - Method 2: Levinsohn-Petrin (LP)
    - log Yit = αl log Lit + αk log Kit + ωit + εit; uses intermediate inputs as proxy for productivity (Levinsohn and Petrin, 2003).
    - Allows testing (and rejection) of constant returns to scale because αk and αl need not sum to one.
- Aggregation and decomposition of aggregate productivity:
  - TFPt = Σi sit * tfpit, where sit is output share of firm i in period t.
  - Productivity growth decomposed into: within effect; between effect; covariance (cross) term; entry and exit effects (following Foster, Haltiwanger, and Krizan (1998)).

### III. MICRO RESULTS: MAGNITUDES, SECTORAL PATTERNS, AND FIRM DYNAMICS
- Aggregate and method comparison:
  - Aggregated micro results are consistent with macro data: labor productivity has been the main driver of growth at the enterprise level, with TFP accounting for a large share of its variation.
  - Divergence between productivity growth and value-added growth in 2000, possibly due to improved firm coverage that year.
  - LP tracks developments in labor productivity more closely than IS.
  - Constant returns to scale is rejected for most sectors; IS likely overestimates capital contribution; LP gives a better picture.
- Average TFP growth (1997–2004):
  - IS measure: 0.4 percent.
  - LP measure: 2.6 percent.
  - Macro estimate: 3.1 percent.
- Sectoral contributions:
  - Agriculture and public services made negative contributions to TFP growth during the period.
  - Trade and business services made large positive contributions.
  - Industry made positive but smaller contributions.
  - Ranking differences:
    - IS: business services at the top.
    - LP: trade at the top.
  - Business services performance driven by remarkable TFP growth in financial intermediation, associated with entry of foreign institutions.
  - Within manufacturing, low-tech and, surprisingly, high-tech firms exerted a drag on productivity growth.
- Role of firm turnover and reallocation:
  - Firm turnover and reallocation have been key factors driving TFP dynamics.
  - IS estimates: contributions to TFP growth from net entrants exceeded those from continuing firms for 1998-2004 and 2001-2004.
  - LP estimates: net entrants exceed continuing firms only after 2000; on average, continuing firms made larger contributions during 1998-2004 per LP (interpret cautiously).
  - Entering firms on average have higher productivity than incumbents had; exiting firms have lower productivity than continuing firms.
  - Within and between components of continuing firms have been consistently negative but offset by a large and positive covariance component — aggregate productivity growth boosted by reallocation of output across continuing firms with rising productivity.
  - Within each industry, output reallocation (covariance) has been the main engine of TFP growth; the between component is negative in many industries.
  - Firm turnover (entry/exit) important mainly for construction and financial services.

### IV. CAVEATS, LIMITATIONS, AND DATA NOTES
- Deflators and price effects:
  - Deflating output and inputs by industry-level price index implicitly assumes perfect competition; under imperfect competition, firm-level price deviations enter the error term and produce omitted price-variable bias, potentially important in financial sector.
- Selection and measurement biases:
  - Estimates may suffer from selection bias generated by relationship between unobserved productivity and shutdown decisions.
  - Covariance effect may be overestimated because measurement errors in output yield a positive covariance between productivity changes and changes in shares and a spuriously low within-firm effect.
- Data quality and scope:
  - Data quality improves after 2000; estimates for period prior to 2000 may be biased.
  - Net entry effects depend on the horizon over which productivity growth is measured; annual studies tend to find smaller net entry contributions than multi-year horizons.
  - Industry-level entry/exit definitions may treat mergers and sector shifts as exits from one industry even when the firm continues operating in another.

### V. MEDIUM-TERM GROWTH SCENARIOS (ILLUSTRATIVE)
- Purpose: illustrate how changes in TFP, capital accumulation, labor utilization, and demographics affect medium-term GDP per capita paths.
- Common assumptions (Table 2):
  - Population (P): -0.27 (U.N. (medium variant)).
  - Inverse dependency ratio (N/P): -0.19 (U.N. (medium variant)).
  - Employment rate (E/N): 0.9 (based on average employment growth since 1998 (0.4 percent); implies a sharp fall in unemployment to 2 percent unless participation rates rise above 73 percent).
  - Average hours worked per employee (L/E): -0.25.
  - Capital-labor ratio: 2.52.
  - Capital stock growth assumption: increased by 7½ percent per annum over the next 10 years.
- Scenario 1: Historical TFP patterns (based on IS and LP estimates for 2001-04)
  - Average performance (TFP 1/GDP): Min 3.7; Max 3.7; Min 6.1; Max 6.1
  - Best performance: Min 10.4; Max 11.2; Min 12.8; Max 13.7
  - Worst performance: Min -2.0; Max 0.5; Min 0.4; Max 3.0
  - Comment: Even if Estonia were to maintain the average TFP growth over the recent past, real GDP growth would slow to about 6 percent.
- Scenario 2: Decline in entry and exit
  - Assumption: a 50 percent decline in entry and exit rates.
  - Results (TFP 1/GDP): Min 1.1; Max 2.7; Min 3.5; Max 5.2
  - Comment: TFP growth is dampened by the reduction of firm entry and exit, resulting in GDP growth rates below 6 percent.
- Scenario 3: No further output reallocation
  - Assumption: sectoral productivity equal to the average productivity over 2001-04 and industries' output shares fixed at the 2004 level.
  - Results (TFP 1/GDP): Min 1.6; Max 3.5; Min 4.1; Max 6.0
  - Comment: By keeping output shares fixed, this scenario rules out the reallocation of resources toward sectors with faster productivity growth, reducing GDP growth to 4-6 percent.
- Key implication: A reduction of firm turnover or resource reallocation could have a significant negative impact on GDP growth.

### VI. AGING, FISCAL SUSTAINABILITY, AND POLICY IMPLICATIONS
- Executive summary of fiscal outlook:
  - Demographic changes will age Estonia’s population over the next 40-50 years, expanding age-related expenditures on pensions and health care and, together with planned tax cuts, creating a fiscal sustainability gap.
  - In a continuation-of-current-policies scenario (including continued ad-hoc pension benefit increases) age-related expenditures on pensions and health care will gradually increase by about 6½ percentage points of GDP by 2050.
  - Existing legislation will reduce the personal and corporate income tax rates from 24 in 2005 to 20 percent in 2009, potentially weakening the fiscal balance by 1-1½ percentage points of GDP.
  - The combined baseline effect is a decline in the fiscal balance by more than 8 percentage points of GDP over the next 45 years; in NPV terms this is equivalent to a permanent fiscal deterioration of about 6½ percentage points of GDP.
  - Stabilizing the government’s net financial position long-term requires a fiscal adjustment of about 2½ percentage points of GDP over the next 8-10 years if combined with an indexation rule that links pension benefits to wage growth and ends ad-hoc increases. Delaying the adjustment by 20-25 years will double the needed adjustment.
- Demographic and labor supply projections:
  - United Nations projections: life-expectancy will increase by 7-8 years; the most populated five-year cohort will shift from 15-19 to 60-64.
  - Projected decline in number of working-age people (15 to 64) by about 27 percent.
  - Projected increase in number of elderly people by 30 percent.
  - Labor supply projected to decrease by 25 percent from 2005 to 2050.
  - Labor force participation ratio projected to increase by only 3 percentage points over the period.
- Macro assumptions and average annual growth rates (percent):
  - Structural unemployment rate (level): 5.5 for 2005-10, 5.5 for 2010-20, 5.5 for 2020-30, 5.5 for 2030-50, 5.5 for 2005-50.
  - Labor force: 0.5 for 2005-10, 0.0 for 2010-20, -0.9 for 2020-30, -0.9 for 2030-50, -0.5 for 2005-50.
  - Structural employment: 0.5 for 2005-10, 0.0 for 2010-20, -0.9 for 2020-30, -0.9 for 2030-50, -0.5 for 2005-50.
  - Productivity (potential): 6.9 for 2005-10, 5.1 for 2010-20, 3.4 for 2020-30, 2.5 for 2030-50, 3.8 for 2005-50.
  - Potential GDP: 7.5 for 2005-10, 5.1 for 2010-20, 2.5 for 2020-30, 1.6 for 2030-50, 3.2 for 2005-50.
  - Inflation: 4.6 for 2005-10, 2.7 for 2010-20, 2.2 for 2020-30, 2.0 for 2030-50, 2.5 for 2005-50.
  - Wage rate: 11.8 for 2005-10, 7.9 for 2010-20, 5.8 for 2020-30, 4.5 for 2030-50, 6.3 for 2005-50.
  - Growth outlook: Real GDP growth projected to decline to just above 1½ percent from almost 8 percent during 2005-10.
  - Growth-adjusted interest rate on government debt set equal to 1 percent on average over the projection period.
- Quantified baseline fiscal impacts (changes relative to 2005, percentage points of GDP):
  - Health care expenditures: 0.0 (2005), 0.1 (2010), 0.5 (2020), 1.0 (2030), 2.2 (2050), 1.7 (NPV terms).
    - demographic effect: 0.0 (2005), -0.1 (2010), 0.0 (2020), 0.2 (2030), 0.9 (2050), 0.8 (NPV terms).
    - per capita GDP effect: 0.0 (2005), 0.2 (2010), 0.5 (2020), 0.8 (2030), 1.3 (2050), 0.9 (NPV terms).
  - Pension expenditures: 0.0 (2005), 0.1 (2010), 0.9 (2020), 2.1 (2030), 4.5 (2050), 3.4 (NPV terms).
    - demographic effect: 0.0 (2005), -0.1 (2010), 0.6 (2020), 1.7 (2030), 4.4 (2050), 3.6 (NPV terms).
    - retirement age (women) effect: 0.0 (2005), -0.3 (2010), -0.9 (2020), -1.1 (2030), -1.3 (2050), -1.1 (NPV terms).
    - benefit ratio effect: 0.0 (2005), 0.0 (2010), 0.4 (2020), 0.8 (2030), 1.2 (2050), 1.0 (NPV terms), 0.9 (NPV terms).
  - Taxes: 0.0 (2005), -1.3 (2010), -1.3 (2020), -1.4 (2030), -1.5 (2050), -1.4 (NPV terms).
    - personal and corporate income tax cuts: 0.0 (2005), -1.2 (2010), -1.2 (2020), -1.2 (2030), -1.2 (2050), -1.1 (NPV terms).
    - payroll tax: 0.0 (2005), -0.1 (2010), -0.2 (2020), -0.2 (2030), -0.4 (2050), -0.3 (NPV terms).
  - Fiscal balance effect: 0.0 (2005), 1.5 (2010), 2.7 (2020), 4.4 (2030), 8.2 (2050), 6.6 (NPV terms).
- Pension system and indexation:
  - Three pillars: PAYG first pillar; fully funded compulsory second pillar for new entrants and persons born after 1983 (voluntary for others); voluntary private third pillar.
  - Social contribution rate: 20 percent for persons not in second pillar, 16 percent for persons that have joined.
  - Retirement age: 63 for men and 60 for women, with women’s retirement age rising to 63 in 2016.
  - First pillar indexation rule: equal weights to inflation and growth in social tax revenue (wage growth).
  - Second pillar contribution rate: 4 percent, with an additional 2 percent paid by employees from gross wages.
  - Baseline assumes continued ad-hoc pension increases, with growth in the pension benefit ratio of 30 percent over 25 years.
  - Baseline pension expenditures increase by 4½ percentage points of GDP between 2005 and 2050; alternative constant pension benefit ratio → increase by about 2 percentage points of GDP; strict adherence to current indexation rule → pension expenditures would decline relative to GDP and first-pillar benefit ratio would drop by more than 50 percent.
- Health care expenditure drivers:
  - Projected health care spending increase: more than 2 percentage points of GDP by 2050, driven by rising share of elderly and increase in per capita GDP.
  - Box 2 cross-country relationship (2004, 25 EU countries):
    - Health care spending to GDP = 4.357 + 0.078*per capita GDP in million euros (R-square = 0.498; t-values: 11.4 and 4.5).
    - Projection implication: health care spending projected to increase by about 1½ percentage points of GDP over the next 45 years over and above demographics-induced spending increase.
- Fiscal model simulations and policy recommendations:
  - IMF’s Global Fiscal Model (GFM) simulations show without fiscal adjustment government initially accumulates assets but later enters explosive debt accumulation.
  - Stabilizing net financial position long-term requires fiscal adjustment of about 2½ percentage points of GDP over the next 8-10 years combined with indexation tying pension benefits to wage growth and ending ad-hoc increases.
  - Early adjustment implies long-term accumulation of government assets, requires less adjustment overall, and yields a larger impact on output compared with a delayed response.
  - Delaying adjustment by 20-25 years doubles the needed adjustment.
  - Recommended policy package:
    - Adopt an indexation rule that avoids excessive ad-hoc pension hikes but does not erode the benefit ratio.
    - Implement fiscal consolidation of about 2½ percentage points of GDP over 8-10 years combined with pension indexation to wage growth.
    - Consider pre-funding strategies to reduce required future adjustment and improve long-term outcomes.
    - Target average primary fiscal surpluses of about 3 percent of GDP to 2015.
  - Rationale: preserves inter-generational equity, reduces long-term adjustment costs, leverages current fiscal position (almost no debt and a significant fiscal surplus).

### VII. DATA, CALIBRATION, AND MODEL PARAMETERS (SELECTED NUMBERS)
- Micro sample summary statistics (means and standard deviations):
  - Value added: Mean 3.7; Std. Dev. 27.6
  - Output: Mean 11.6; Std. Dev. 69.2
  - Employment: Mean 17.0; Std. Dev. 92.3
  - Capital: Mean 3.8; Std. Dev. 87.5
  - Intermediate inputs: Mean 8.0; Std. Dev. 51.7
  - Staff costs: Mean 1.4; Std. Dev. 8.7
- GFM initial steady-state ratios (Estonia, percent of GDP):
  - Consumption 48.6; Government consumption 20.4; Investment 31.0; Imports 51.0; Capital income 57.1; Labor income 42.9; Gross debt 4.6.
- Tax rates and revenue (percent of GDP):
  - Payroll Taxes (Worker and Employer) 10.5; On Personal Income 5.4; On Corporate Income 1.3; On Consumption (VAT and excises) 11.7.
- Behavioral parameters (selected):
  - Frisch elasticity = -0.10; Elasticity of substitution between labor and capital = -0.75; Elasticity of intertemporal substitution = 0.40; Wedge between rate of time preference and yield on government bonds = 15 percent (planning horizon of 7 years); Fraction of consumers liquidity-constrained = 0.70 (these households consume 40 percent of aggregate consumption); Markup over marginal cost: tradables = 17 percent; nontradables = 25 percent.
- Stylized sustainability indicator example:
  - Initial primary deficit = 1 percent of GDP; Initial debt stock = 90 percent of GDP; Growth-adjusted interest rate = 2 percent; Aging costs increasing linearly by 5 percent of GDP during 2005–50.
  - An increase in initial debt stock or NPV of aging costs by 1 percentage point of GDP reduces the indicator by 0.02 percentage point of GDP.
  - Indicator adjustment decomposition: Balance component 1 percentage point of GDP; Debt component 1.8 percentage points of GDP; Aging component 3.3 percentage points of GDP; Indicator total: -6.1 (percent of GDP).

### VIII. CONCLUSIONS AND KEY IMPLICATIONS
- Estonia’s exceptional GDP per capita catch-up since the mid-1990s was driven mainly by labor productivity growth, particularly TFP improvements, complemented by capital deepening.
- Important remaining gaps: low capital-labor ratio (about one-third of EU-15) and substantial TFP gap (40 percent in 2005), implying catch-up requires both capital accumulation and further TFP convergence.
- Micro-level analysis is essential to disentangle within-industry efficiency gains from compositional shifts across sectors and to assess roles of firm restructuring and reallocation in sustaining future TFP growth.
- Recovery in labor utilization since 2000 has reinforced growth, but continued improvements in employment and hours worked, alongside capital deepening and TFP improvements, will be central to sustaining medium-term growth.
- Risks: Entry and exit rates have fallen over time and there is a limit to continuous reallocation of resources; productivity growth might decline, dampening GDP growth.
- Policy priorities to sustain rapid income convergence and mitigate fiscal pressures:
  - Ensure continuous creative destruction and reinvention to move toward (and eventually shift) the technology frontier.
  - Promote human capital development.
  - Increase R&D investment.
  - Provide public goods with positive production externalities.
  - Ensure efficient and flexible markets.
  - Raise employment rates and accelerate capital accumulation.
  - Implement fiscal consolidation of about 2½ percentage points of GDP over 8-10 years combined with pension indexation to wage growth and end ad-hoc pension increases.

*Prepared by Marialuz Moreno-Badia (EUR) and Michael Skaarup (FAD); Approved by European Department; July 11, 2007.*

### 2007. The views expressed in this document are those of the staff team and do not necessarily reflect

### I. MEDIUM-TERM GROWTH AND PRODUCTIVITY IN ESTONIA: A MICRO PERSPECTIVE

### A. Introduction
- Between 1995 and 2005, Estonia’s real GDP per capita rose by an average of 6½ percent a year.
- In 2005 Estonia’s GDP per capita (at current prices and purchasing parity standard, PPS) was 50.9 percent of the EU-15 average.
- Estonia’s capital-labor ratio is only about one-third of that of the EU-15.
- The TFP gap relative to the EU-15 was 40 percent in 2005.
- Paper objective and approach:
  - Establish stylized macro facts using growth-accounting.
  - Examine determinants of TFP using micro data, exploiting better micro coverage (including capital stock), sectoral decomposition, firm dynamics, and semiparametric Levinsohn and Petrin (2003) estimation.
  - Analyze sectoral contributions to productivity growth and present medium-term growth scenarios.

### B. What Do The Macro Data Tell Us? (Growth-accounting framework and international comparisons)
- Output per capita identity decomposes GDP per capita into demographics, labor utilization, and labor productivity:
  - Y/Pop = (N/Pop) × (E/N) × (L/E) × (Y/L)
  - Where Y is GDP at constant 1995 prices; Pop is population; N/Pop is working-age population to total population (inverse dependency ratio); E/N is the employment rate; L/E is average hours worked per employee; Y/L is labor productivity (output per hour worked).
- Labor productivity decomposition under Cobb-Douglas:
  - Y/L = A × (K/L)^α, with α set to 0.35.
  - A represents TFP, computed as a residual and encompassing technological progress, human capital, institutional quality, measurement errors, omitted variables, and model misspecification.
- Key macro findings (1995–2005 and subperiods):
  - Demographics contributed positively but modestly (annual increase in GDP per capita averaging some 0.3 percent), reflecting an increase in inverse dependency ratio.
  - Labor utilization fell sharply during the 1990s and reversed after 2000; employment rate and hours worked declined in the 1990s but recovered with stronger job creation after 2000.
  - Labor productivity was the main driver of catch-up:
    - Estonia’s labor productivity growth was more than four times that in EU-15 countries during the mid-1990s and about 75 percent higher than in the NMS.
    - TFP growth was the most important component of labor productivity growth, with capital deepening also important.
  - Contributions eased since 2000 but remained high relative to EU-15.
  - A lower recent rise in capital intensity may reflect slower substitution of capital for labor, improved labor market flexibility, and a shift toward less capital-intensive sectors (services), potentially absorbing lower-skilled workers and contributing to a slowdown in measured TFP growth.

### C. What Is The Micro Story? (data and approach — summary of intent and advantages)
- Rationale for micro approach:
  - Better data coverage for capital and other firm-level variables enables cross-checking macro findings.
  - Micro data allow sectoral decomposition of productivity growth.
  - Micro data reveal firm dynamics (restructuring, reallocation) critical for understanding productivity differences across sectors and over time.
- Empirical contributions highlighted:
  - Analysis of productivity growth from macro and micro perspectives.
  - Use of a more recent and broader business-sector dataset.
  - Matching sectoral firm distribution with macro data.
  - Improved TFP estimation using Levinsohn and Petrin (2003) semiparametric method.
  - Analysis of sectoral contributions to productivity growth.

### D. Growth Scenarios (structure and purpose)
- Purpose: present scenario analysis of the impact of changes in specific growth factors on medium-term growth.
- Scenarios outlined in paper (Tables referenced):
  - Table 2. Growth Scenarios—Assumptions
  - Table 3. Total Factor Productivity and GDP Growth: Scenarios
- Scenario methodology: use macro accounting and micro insights on TFP and capital-labor dynamics to simulate medium-term paths for GDP per capita under alternative assumptions on TFP growth, capital accumulation, labor utilization, and demographics.
- (Detailed scenario assumptions and quantified projections are presented in Table 2 and Table 3 of the source document.)

### E. Conclusions (key implications and interpretation)
- Estonia’s exceptional GDP per capita catch-up since the mid-1990s was driven mainly by labor productivity growth, in particular TFP improvements, complemented by capital deepening.
- Important remaining gaps: low capital-labor ratio (about one-third of EU-15) and substantial TFP gap (40 percent in 2005), implying catch-up requires both capital accumulation and further TFP convergence.
- Micro-level analysis is essential to disentangle within-industry efficiency gains from compositional shifts across sectors and to assess the roles of firm restructuring and reallocation in sustaining future TFP growth.
- The recovery in labor utilization since 2000 has reinforced growth, but continued improvements in employment and hours worked, alongside capital deepening and TFP improvements, will be central to sustaining medium-term growth.

*Prepared by Marialuz Moreno-Badia (EUR) and Michael Skaarup (FAD); Approved by European Department; July 11, 2007.*

### 11. Our micro data come from the Estonian

### 11. Our micro data come from the Estonian

### Data and coverage
- Micro data source: Estonian Business Registry, covering the period 1997-2004.
- Database features:
  - Firm-level data from all economic sectors.
  - Absence of size thresholds.
  - Availability of transactions data (e.g., mergers).
  - Detailed information on balance sheets, income, and costs.
- Sample restrictions and coverage:
  - Because of missing information, only data for about 40 percent of the firms in the registry are used.
  - The number of business entities in the registry between 1997 and 2004 more than doubled, and coverage improved over time.
  - Value added of enterprises in the sample accounted for 45 percent of aggregate value added in 1997 and increased to 60 percent by 2004.
  - Employment coverage improved over time and was above 50 percent of the macro level in 2004 for all sectors except agriculture and public services.
  - Improvement in data quality may be related to the introduction in 2000 of fines penalizing firms that do not submit income or balance sheet statements; growth rates in 2000 may be biased because of the improved coverage in that year.
- Deflation and measurement:
  - Value added and intermediate inputs are deflated by the respective deflators of the system of national accounts provided by the Statistical Office of Estonia.
  - Capital is deflated by the gross capital formation price index.
  - Results should be interpreted with caution due to data coverage, measurement, and conceptual problems.
- Additional sample notes:
  - The sample does not include data for the two major banks in Estonia; results for the financial sector could be biased.
  - Comparability between micro and macro data is limited owing to methodological inconsistencies; value added at the macro level is a broader concept.
  - According to the Statistical Office of Estonia, all enterprises registered in Estonia accounted for about 70 percent of aggregate value added in 2005.

### Stylized facts on firm dynamics and structure
- Entry, exit, and size:
  - Estonian entry and exit rates are fairly high by international standards, although entry rates have declined over time.
  - High firm turnover may relate to restructuring during transition from large-scale production to smaller units.
  - Firm size is very small: about 70 percent of firms are micro enterprises (less than 10 employees) and firm size is getting smaller across all sectors.
  - The recent increase in the number of enterprises is explained mainly by the birth of micro enterprises.
- Sectoral shifts:
  - Agriculture contracted while construction expanded during the sample period.
  - Most firms belong to the services sector, where smaller firms dominate.

### Methodology
- Two alternative methodologies to estimate TFP are used to test sensitivity:
  - Method 1: Industry shares (IS).
    - Assumes a Cobb-Douglas production function.
    - TFP computed as log Yit − αl log Lit − αk log Kit, where αl and αk are industry cost shares measured at the two-digit industry level and αk = 1 − αl under constant returns to scale.
  - Method 2: Levinsohn-Petrin (LP).
    - Assumes a Cobb-Douglas production function with log Yit = αl log Lit + αk log Kit + ωit + εit, where ω is unobserved TFP and ε is an error term uncorrelated with input choices.
    - Uses intermediate inputs as a proxy for productivity following Levinsohn and Petrin (2003) to address input–unobservable correlation.
    - Allows testing (and rejection) of constant returns to scale because αk and αl need not sum to one.
- Aggregation and decomposition:
  - Aggregate productivity TFPt is the weighted average of firm productivities: TFPt = Σi sit * tfpit, where sit is output share of firm i in period t.
  - Following Foster, Haltiwanger, and Krizan (1998), productivity growth is decomposed into:
    - Within effect: productivity growth within continuing firms keeping market shares fixed.
    - Between effect: productivity growth due to changing output shares.
    - Covariance (cross) term: positive when output shares increase for continuing firms with rising productivity.
    - Entry and exit effects: entering firms contribute positively if they have higher productivity than aggregate productivity the year before; exiting firms contribute positively only if they had lower-than-average productivity the previous year.

### Results
- Aggregate and measure comparison:
  - Aggregated micro results are consistent with macro data: labor productivity has been the main driver of growth at the enterprise level, with TFP accounting for a large share of its variation.
  - There is a divergence between productivity growth and value-added growth in 2000, possibly due to improved firm coverage that year.
  - Of the two TFP estimates, LP tracks developments in labor productivity more closely than IS.
  - Since the assumption of constant returns to scale is rejected for most sectors, IS probably overestimates the contribution of capital to growth; LP estimates give a better picture.
- Magnitudes and sectoral contributions:
  - Average TFP growth between 1997 and 2004 ranged from 0.4 percent (IS measure) to 2.6 percent (LP measure), compared with a macro estimate of 3.1 percent.
  - Both methods indicate agriculture and public services made negative contributions to TFP growth during the period, more than offset by large contributions from trade and business services.
  - Industrial sector made positive but smaller contributions.
  - Ranking differences:
    - Industry shares (IS): business services at the top.
    - Levinsohn-Petrin (LP): trade at the top.
  - The performance of business services is driven by remarkable TFP growth in financial intermediation, associated with entry of foreign institutions.
  - Within manufacturing, low-tech and, surprisingly, high-tech firms exerted a drag on productivity growth.
- Role of firm turnover and reallocation:
  - Firm turnover and reallocation have been key factors driving TFP dynamics.
  - According to IS estimates, contributions to TFP growth from net entrants exceeded those from continuing firms for periods 1998-2004 and 2001-2004.
  - LP estimates show the same result but only after 2000; on average, continuing firms made larger contributions to growth during 1998-2004 per LP (interpret cautiously).
  - Positive contribution of firm turnover is due to both entering and exiting firms: entering firms on average have higher productivity than incumbents had; exiting firms have lower productivity than continuing firms.
  - Decomposition shows within and between components of continuing firms have been consistently negative but offset by a large and positive covariance component — aggregate productivity growth boosted by reallocation of output across continuing firms with rising productivity.
  - Within each industry, output reallocation (covariance) has been the main engine of TFP growth; the between component is negative in many industries.
  - Within component indicates firms with larger output shares have experienced negative productivity growth.
  - Firm turnover (entry/exit) important mainly for construction and financial services.

### Caveats and limitations
- Deflators and price effects:
  - By deflating output and inputs by an industry-level price index, the analysis implicitly assumes perfect competition; under imperfect competition, firm-level price deviations will end up in the error term and cause omitted price-variable bias, potentially important in the financial sector where price-setting power may exist.
- Selection and measurement biases:
  - Estimates may suffer from selection bias generated by the relationship between unobserved productivity and shutdown decisions.
  - The covariance effect may be overestimated because measurement errors in output yield a positive covariance between productivity changes and changes in shares and a spuriously low within-firm effect.
- Data quality and scope:
  - Data quality improves after 2000; estimates for the period prior to 2000 may be biased.
  - Net entry effects depend on the horizon over which productivity growth is measured; high-frequency (annual) studies tend to find a smaller contribution of net entry compared with studies using 2-, 3-, or 5-year horizons.
  - Industry-level entry/exit definitions may treat mergers and sector shifts as exits from one industry even when the firm continues operating in another.

### Growth scenarios (illustrative)
- Three illustrative scenarios considered for medium-term growth prospects:
  - Scenario 1: Three different TFP growth patterns based on historical data.
  - Scenario 2: Implication of a 50 percent decline in firm entry and exit.
  - Scenario 3: Impact of no further reallocation of output across sectors.
- All scenarios are purely illustrative and not projections of medium-term growth.

*Italic: Source — Excerpt from IMF chapter: micro data and productivity analysis for Estonia (Estonian Business Registry, 1997-2004).*

### 22. The three scenarios share some common assumptions (Table 2). Population and

### _cr07256 - 22. The three scenarios share some common assumptions (Table 2). Population and

### Common assumptions (Table 2)
- Population (P): -0.27 (U.N. (medium variant)).
- Inverse dependency ratio (N/P): -0.19 (U.N. (medium variant)).
- Employment rate (E/N): 0.9
  - Comment: Based on the average employment growth since 1998 (0.4 percent).
  - This assumption implies a sharp fall in unemployment (to 2 percent) unless participation rates rise above 73 percent.
- Average hours worked per employee (L/E): -0.25
  - Comment: Based on income effects and convergence toward Europe.
  - Notes: A further compositional shift of employment toward services or part-time employment could cause a larger decline in hours worked than assumed here.
- Capital-labor ratio: 2.52
  - Comment: Based on average increase in capital stock since 1995.
- Capital stock growth assumption: increased by 7½ percent per annum over the next 10 years.
  - Comment: This is the estimated annual increase in the net capital stock over the period 1998-2005. There is scope for faster capital accumulation, given the large capital-labor ratio gap with the EU-15.

Sources cited for assumptions: United Nations; and IMF staff calculations.

### Scenario analysis and projections (Table 3)
- Overview:
  - Maintaining the strong growth performance of recent years is not guaranteed.
  - Three scenarios explore impacts of alternative TFP histories, reduced firm entry/exit, and no further output reallocation.
- Scenario 1: Historical TFP patterns (based on Industry shares and Levinsohn-Petrin estimates for the period 2001-04)
  - Average performance (TFP 1/GDP): Min 3.7; Max 3.7; Min 6.1; Max 6.1
  - Best performance: Min 10.4; Max 11.2; Min 12.8; Max 13.7
  - Worst performance: Min -2.0; Max 0.5; Min 0.4; Max 3.0
  - Comment: Even if Estonia were to maintain the average TFP growth over the recent past, real GDP growth would slow to about 6 percent.
- Scenario 2: Decline in entry and exit 2/
  - Assumption: a 50 percent decline in entry and exit rates.
  - Results (TFP 1/GDP): Min 1.1; Max 2.7; Min 3.5; Max 5.2
  - Comment: TFP growth is dampened by the reduction of firm entry and exit, resulting in GDP growth rates below 6 percent.
- Scenario 3: No further output reallocation 3/
  - Assumption: sectoral productivity equal to the average productivity over the period 2001-04 and industries' output shares fixed at the 2004 level.
  - Results (TFP 1/GDP): Min 1.6; Max 3.5; Min 4.1; Max 6.0
  - Comment: By keeping output shares fixed, this scenario rules out the reallocation of resources toward sectors with faster productivity growth, reducing GDP growth to 4-6 percent.
- Key implication from scenarios:
  - A reduction of firm turnover or resource reallocation could have a significant negative impact on GDP growth.

### Findings on productivity dynamics and growth potential
- Labor productivity increases have underpinned Estonia’s positive performance during the last decade.
- Most convergence with the EU-15 achieved during this period stems from closing the gap in labor productivity and, specifically, TFP.
- Remarkable productivity gains occurred in the business services and trade sectors.
- Productivity dynamics have been dominated by entry and exit:
  - New firms have introduced technologies and innovation, boosting productivity and displacing inefficient firms.
  - Reallocation of output toward firms with faster productivity growth has been an important determinant of aggregate and within-sector productivity.
- Estonia still faces a large TFP gap with respect to the EU-15, underscoring substantial growth potential.

### Risks and policy implications
- Risks:
  - Entry and exit rates have fallen over time and there is a limit to continuous reallocation of resources.
  - Productivity growth might decline in the near future, dampening GDP growth.
- Policy implications and recommendations (as emphasized in the chapter):
  - Ensure continuous creative destruction and reinvention to move toward (and eventually shift) the technology frontier.
  - Policies that will be critical to foster innovation and rapid productivity growth include:
    - Promoting human capital development.
    - Increasing R&D investment.
    - Provision of public goods with positive production externalities.
    - Ensuring efficient and flexible markets.
  - In addition, higher employment rates and faster capital accumulation could also help maintain rapid income convergence.

### Data sources, definitions, and methodology (appendix highlights)
- Macro data: Annual Macroeconomic Database (AMECO) of the European Commission unless otherwise noted.
- Key definitions:
  - Real GDP: Real GDP at 1995 prices; converted into PPS for comparison.
  - Working population: Population aged between 15 and 64 years.
  - Capital stock: Sum of produced fixed assets; for new member states calculated using perpetual inventory model with initial capital-output ratios from Schadler and others (2006); gross fixed investment from AMECO; depreciation assumed to be 5 percent.
  - Hours worked: Number of hours worked per year and per person employed. Sources: OECD and the Total Economy Database of the Groningen Growth and Development Centre and the Conference Board.
- Micro data: Estonian Business Registry database (covers 1995-2004; sample uses data from 1997 due to missing information).
  - Sample construction steps: construct longitudinal panel using registration codes with corrections for transfers and acquisitions; exclude unrealistic observations (zero/negative value added, employment, capital, intermediate inputs); exclude firms with unclear industry information.
- TFP estimation: Levinsohn and Petrin methodology assuming a Cobb-Douglas production function; uses intermediate inputs to control for unobservables, polynomial approximation in K and M, a two-stage estimation, bootstrap with 50 replications to construct standard errors.
- Selected estimation results (Table A.5): estimated coefficients of log labor and log capital by industry and Wald test of constant returns to scale (5 percent significance). Examples include:
  - Agriculture, hunting, and related activities: Log Labor coeff 0.41 (Std. Err 0.03); Log Capital coeff 0.30 (Std. Err 0.04); Constant Returns to Scale: No.
  - Computer and related activities: Log Labor coeff 0.86 (Std. Err 0.05); Log Capital coeff 0.21 (Std. Err 0.05); Constant Returns to Scale: Yes.
  - Financial intermediation: Log Labor coeff 0.65 (Std. Err 0.10); Log Capital coeff 0.21 (Std. Err 0.14); Constant Returns to Scale: Yes.
- Summary statistics (Table A.1) for micro sample (means and standard deviations):
  - Value added: Mean 3.7; Std. Dev. 27.6
  - Output: Mean 11.6; Std. Dev. 69.2
  - Employment: Mean 17.0; Std. Dev. 92.3
  - Capital: Mean 3.8; Std. Dev. 87.5
  - Intermediate inputs: Mean 8.0; Std. Dev. 51.7
  - Staff costs: Mean 1.4; Std. Dev. 8.7

*Source: IMF staff calculations and Estonian Business Registry; United Nations; AMECO; and sources cited in the chapter.*

### 26.      The aging of Estonia’s population creates a fiscal sustainability gap that is

### _cr07256 - 26.      The aging of Estonia’s population creates a fiscal sustainability gap that is

### Executive summary
- Demographic changes (rising longevity and low fertility) will age Estonia’s population over the next 40-50 years, expanding age-related expenditures on pensions and health care and, together with planned tax cuts, creating a fiscal sustainability gap.
- In a continuation-of-current-policies scenario (including continued ad-hoc pension benefit increases) age-related expenditures on pensions and health care will gradually increase by about 6½ percentage points of GDP by 2050.
- Existing legislation will reduce the personal and corporate income tax rates from 24 in 2005 to 20 percent in 2009, which could weaken the fiscal balance by 1-1½ percentage points of GDP.
- The combined baseline effect is a decline in the fiscal balance by more than 8 percentage points of GDP over the next 45 years; in NPV terms this is equivalent to a permanent fiscal deterioration of about 6½ percentage points of GDP.
- Stabilizing the government’s net financial position in the long-term requires a fiscal adjustment of about 2½ percentage points of GDP over the next 8-10 years if combined with an indexation rule that links pension benefits to wage growth and ends ad-hoc increases. Delaying the adjustment by 20-25 years will double the needed adjustment.

### Demographic outlook and labor supply
- United Nations projections: life-expectancy will increase by 7-8 years; the most populated five-year cohort will shift from 15-19 to 60-64; projected rise in life-expectancy contrasts with previous 30 years where it remained roughly constant.
- Projected demographic effects on population composition:
  - Decline in the number of working-age people (15 to 64) by about 27 percent.
  - Increase in the number of elderly people by 30 percent.
- Labor supply projection:
  - Labor supply is projected to decrease by 25 percent from 2005 to 2050.
  - Labor force participation ratio projected to increase by only 3 percentage points over the period.
  - Labor force participation ratio projected using a “cohort” methodology.

### Macro assumptions and growth projections
- Key macro assumptions and average annual growth rates (percent):
  - Structural unemployment rate (level): 5.5 for 2005-10, 5.5 for 2010-20, 5.5 for 2020-30, 5.5 for 2030-50, 5.5 for 2005-50.
  - Labor force: 0.5 for 2005-10, 0.0 for 2010-20, -0.9 for 2020-30, -0.9 for 2030-50, -0.5 for 2005-50.
  - Structural employment: 0.5 for 2005-10, 0.0 for 2010-20, -0.9 for 2020-30, -0.9 for 2030-50, -0.5 for 2005-50.
  - Productivity (potential): 6.9 for 2005-10, 5.1 for 2010-20, 3.4 for 2020-30, 2.5 for 2030-50, 3.8 for 2005-50.
  - Potential GDP: 7.5 for 2005-10, 5.1 for 2010-20, 2.5 for 2020-30, 1.6 for 2030-50, 3.2 for 2005-50.
  - Inflation: 4.6 for 2005-10, 2.7 for 2010-20, 2.2 for 2020-30, 2.0 for 2030-50, 2.5 for 2005-50.
  - Wage rate: 11.8 for 2005-10, 7.9 for 2010-20, 5.8 for 2020-30, 4.5 for 2030-50, 6.3 for 2005-50.
- Growth outlook: Real GDP growth projected to decline to just above 1½ percent from almost 8 percent during 2005-10.
- The growth-adjusted interest rate on government debt (nominal interest rate minus nominal growth) set equal to 1 percent on average over the projection period.

### Baseline tax and expenditure assumptions
- Baseline tax assumptions:
  - Gradual reduction of the personal and corporate income tax rates to 20 percent (from 24 in 2005).
  - Other taxes left unchanged in the baseline.
  - The revenue from social contributions projected to decline by about ¼ percentage point of GDP due to wider coverage of the mandatory fully funded pension system.
- Baseline pension and health-care assumptions:
  - Health care spending evolution based on demographic projection and cross-country evidence on health spending rising with income.
  - Baseline for pensions assumes continued practice of ad-hoc increases in benefits, with size of increases gradually tapering off.
  - All other expenditure and revenue components assumed to remain unchanged as a share of GDP throughout the projection period.
- Resulting trajectory: Age-related expenditure trajectory mirrors the evolution of the old-age dependency ratio.

### Quantified baseline fiscal impacts (compared to 2005)
- Table of changes (percentage points of GDP):
  - Health care expenditures: 0.0 (2005), 0.1 (2010), 0.5 (2020), 1.0 (2030), 2.2 (2050), 1.7 (NPV terms).
    - demographic effect: 0.0 (2005), -0.1 (2010), 0.0 (2020), 0.2 (2030), 0.9 (2050), 0.8 (NPV terms).
    - per capita GDP effect: 0.0 (2005), 0.2 (2010), 0.5 (2020), 0.8 (2030), 1.3 (2050), 0.9 (NPV terms).
  - Pension expenditures: 0.0 (2005), 0.1 (2010), 0.9 (2020), 2.1 (2030), 4.5 (2050), 3.4 (NPV terms).
    - demographic effect: 0.0 (2005), -0.1 (2010), 0.6 (2020), 1.7 (2030), 4.4 (2050), 3.6 (NPV terms).
    - retirement age (women) effect: 0.0 (2005), -0.3 (2010), -0.9 (2020), -1.1 (2030), -1.3 (2050), -1.1 (NPV terms).
    - benefit ratio effect: 0.0 (2005), 0.0 (2010), 0.4 (2020), 0.8 (2030), 1.2 (2050), 1.0 (NPV terms), 0.9 (NPV terms).
  - Taxes: 0.0 (2005), -1.3 (2010), -1.3 (2020), -1.4 (2030), -1.5 (2050), -1.4 (NPV terms).
    - personal and corporate income tax cuts: 0.0 (2005), -1.2 (2010), -1.2 (2020), -1.2 (2030), -1.2 (2050), -1.1 (NPV terms).
    - payroll tax: 0.0 (2005), -0.1 (2010), -0.2 (2020), -0.2 (2030), -0.4 (2050), -0.3 (NPV terms).
  - Fiscal balance effect: 0.0 (2005), 1.5 (2010), 2.7 (2020), 4.4 (2030), 8.2 (2050), 6.6 (NPV terms).

### Pension system, indexation and scenarios
- Structure:
  - Three pillars: first pillar PAYG public pension; fully funded compulsory second pillar for new entrants and persons born after 1983 (voluntary for others); voluntary private third pillar.
  - First pillar financed by social contributions and budgetary transfers; social contribution rate: 20 percent for persons not in second pillar, 16 percent for persons that have joined.
  - Retirement age: 63 for men and 60 for women, with women’s retirement age rising to 63 in 2016.
  - First pillar indexation rule: equal weights to inflation and growth in social tax revenue (wage growth).
  - Second pillar contribution rate: 4 percent, with an additional 2 percent paid by employees from gross wages.
- Observed practice and baseline assumption:
  - Over the past five years pension benefits have grown almost twice as fast as prescribed by the formal indexation rule; ad-hoc increases have raised pension benefits twice as fast as the rule suggests (and faster than wage growth).
  - Baseline assumes continued ad-hoc increases, with growth in the pension benefit ratio of 30 percent over 25 years.
- Scenario comparisons:
  - Baseline pension expenditures increase by 4½ percentage points of GDP between 2005 and 2050.
  - Alternative: constant pension benefit ratio → pension expenditures would increase by about 2 percentage points of GDP.
  - If the current indexation rule were followed strictly → pension expenditures would decline relative to GDP and the pension benefit ratio would drop by more than 50 percent (first-pillar benefit ratio), though second pillar maturation would partly mitigate income decline for future pensioners.

### Health care expenditure drivers and projections
- Projected health care spending increase: more than 2 percentage points of GDP by 2050, driven by rising share of elderly and increase in per capita GDP.
- Assumptions incorporate a “constant health” scenario: projected higher life-expectancy is not all spent in “bad health”; age distribution of expenditure shifts so that average health care cost at a given age (e.g., 60) will be lower in 2050 than now.
- Empirical regularity: share of health care spending in GDP depends on per capita GDP; OECD evidence indicates real per capita health care expenditures have grown 1½ percentage point faster per year than can be explained by demographics and per capita GDP growth combined.

### Fiscal sustainability and policy implications
- Simulations with the IMF’s Global Fiscal Model (GFM):
  - Without fiscal adjustment the government initially accumulates assets but later enters explosive debt accumulation.
  - Stabilizing the government’s net financial position long-term requires a fiscal adjustment of about 2½ percentage points of GDP over the next 8-10 years if combined with an indexation rule tying pension benefits to wage growth and ending ad-hoc increases.
  - An early adjustment implies long-term accumulation of government assets, requires less adjustment overall, and yields a larger impact on output compared with a delayed response.
  - Delaying adjustment by 20-25 years will double the needed adjustment.
- Rationale for early action:
  - Leaves more room for future possible additional adjustment needs given upside risks to spending pressures.
  - Addresses inter-generational equity by ensuring all generations share the adjustment burden.
- Policy choices highlighted:
  - Adopt an indexation rule that avoids excessive ad-hoc pension hikes but does not erode the benefit ratio — an intermediate course between the formal rule and current practice.
  - Implement fiscal consolidation of about 2½ percentage points of GDP over 8-10 years combined with reform of pension indexation to wage growth to secure long-term sustainability.

*Source: _cr07256 - 26.      The aging of Estonia’s population creates a fiscal sustainability gap that is*

### Box 2. Relationship Between Health Spending and Per Capita GDP

### Box 2. Relationship Between Health Spending and Per Capita GDP

### Estimated cross-country relationship (2004, 25 EU countries)
- Based on cross country 2004 data for 25 European Union countries the following relationship between health care spending (as a share of GDP) and per capita GDP can be estimated
  - Health care spending to GDP = 4.357 + 0.078*per capita GDP in million euros
    - (11.4)     (4.5)                                         R-square = 0.498
    - t-values in parenthesis below the parameters.
- The estimated relationship is used in this paper to project non-demographic growth in health care spending until 2050.

### Projection implication
- On that basis health care spending is projected to increase by about 1½ percentage points of GDP over the next 45 years over and above the demographics-induced spending increase.

### Figure: Health Care Spending and Per Capita GDP in EU countries, 2004
- Figure title: Health Care Spending and Per Capita GDP in EU countries, 2004.
- Country labels shown in the figure:
  - AUT
  - CZE
  - DEN
  - EST
  - FIN
  - FRA
  - GER
  - GRE
  - HUN
  - IRE
  - ITA
  - LAT
  - LIT
  - MLT
  - NLD
  - POL
  - PRT
  - SLK
  - SLO
  - SPA
  - SWE
  - UK
- Figure axes/labels as shown:
  - Horizontal axis: Nominal GDP per capita (1,000 Euros)
  - Vertical axis: Health care spending (Percent of GDP)

*Source: _cr07256 - Box 2. Relationship Between Health Spending and Per Capita GDP*

### 56.      Population aging will lead to a sizeable expansion in pension and health care

### Population aging will lead to a sizeable expansion in pension and health care costs in Estonia

### Main findings
- With a continuation of current policies and practices, age-related expenditures on pensions and health care will rise by 6½ percentage points of GDP between 2005 and 2050.
- This future fiscal weakening is compounded by planned reductions in personal and corporate income tax rates until 2009.
- Ensuring long-term fiscal sustainability requires a reduction in non-age related expenditures of about 2½ percentage points of GDP until 2015 if combined with a pension indexation rule linking benefits to wage growth.
- Targeting primary fiscal surpluses of about 3 percent of GDP on average to 2015 is commensurate with the above expenditure reduction.
- Delaying the expenditure reduction by 20-25 years will double the needed adjustment.
- Benefits of pre-funding the fiscal costs of aging:
  - Smaller fiscal adjustment the earlier it is implemented.
  - Long-term impact on economic activity is positive if implemented early; it may be negative if significantly delayed.
  - Early adjustment leaves more room for future unanticipated additional adjustment needs.
  - Addresses inter-generational equity by ensuring both current and later generations bear the burden.
  - The current fiscal position (almost no debt and a significant fiscal surplus) allows such a strategy without imposing an unmanageable fiscal burden.

### Policy recommendations and scenarios
- Implement expenditure reductions of about 2½ percentage points of GDP until 2015 combined with a pension indexation rule linking benefits to wage growth.
- Target average primary fiscal surpluses of about 3 percent of GDP to 2015.
- Consider pre-funding strategies to reduce the required future adjustment and improve long-term economic outcomes.
- Avoid delaying adjustment: a 20-25 year delay doubles the necessary fiscal correction.

### Methodology for projecting age-related expenditures
- Basic expenditure identity: E = b ⋅ R (benefit level b times number of recipients R).
- Extended formulation used in projections:
  - E = b ⋅ (R/P) ⋅ (P/ P1564) ⋅ (P1564/ L) ⋅ L
  - Where P reflects the potentially eligible population, P1564 is the working-age population, L is employment, and R/P is the coverage ratio.
- Benefit level linked to nominal GDP per worker as proxy for wage rate:
  - b = br ⋅ (Y / L)
  - br denotes the benefit ratio capturing indexation rules or ad-hoc benefit changes.
- Log-change equation for growth in age-related expenditures as a share of GDP (preserving source structure as a log-decomposition):
  - Δ log E = Δ log br + Δ log (R/P) + Δ log (P/ P1564) + Δ log (P1564/ L) + (Δ log Y − Δ log L)
- In Estonia projections:
  - Eligible population assumed to be persons older than 65 for both pension and health care expenditures.
  - Projected number of pension recipients accounts for changes in statutory retirement age.
  - Health care user index based on age distribution (five-year cohorts) of health care spending (includes entire population).
- The projections allow for benefit ratio increases for pensions due to ad-hoc decided benefit hikes and rising health care expenditures from non-demographic drivers.

### Labor force projection (cohort methodology)
- Cohort methodology allows changing participation ratios in line with observed exit and entry rates in five-year cohorts rather than assuming unchanged age-specific participation.
- Projected cohort effects for Estonia:
  - Female labor force participation projected to rise by about 6½ percentage points.
  - Male labor force participation projected to decline by 1½ percentage points.
  - By 2025 women will have higher labor force participation than men (figures referenced in source).

### The Global Fiscal Model (GFM): overview and role in simulations
- GFM is a four-country dynamic general equilibrium model based on the NOEM tradition, designed to examine fiscal policy issues, especially temporary or permanent changes in taxes or expenditures and gradual fiscal pressures such as aging.
- Non-Ricardian features introduced via three channels:
  - Households with finite horizons (temporary fiscal changes affect consumption).
  - Liquidity-constrained households that consume all disposable income each period.
  - Distortionary labor- and capital taxes affecting incentives.
- Other model features:
  - Consumption and production with constant elasticity of substitution functions; firms and workers have market power.
  - Traded and non-traded goods; two factors of production (capital and labor) mobile across sectors but not internationally.
  - Investment driven by Tobin’s Q with adjustment costs; wages and prices fully flexible (monetary policy ineffective).
  - Two financial assets: government debt (traded internationally) and equity (held domestically); international trade in government debt equalizes nominal interest rates over time.
- GFM used to demonstrate that early expenditure reductions and pension indexation can restore long-term sustainability with manageable short-term fiscal adjustments.

### GFM calibration and behavioral parameters (key values preserved)
- Key macroeconomic ratios in initial steady state (Estonia):
  - Consumption 48.6 (percent of GDP)
  - Government consumption 20.4 (percent of GDP)
  - Investment 31.0 (percent of GDP)
  - Imports 51.0 (percent of GDP)
  - Capital 57.1 (factor income percent of GDP)
  - Labor 42.9 (factor income percent of GDP)
  - Gross debt 4.6 (percent of GDP)
- Tax rates and revenue (as % of GDP):
  - Payroll Taxes (Worker and Employer) 10.5
  - On Personal Income 5.4
  - On Corporate Income 1.3
  - On Consumption (VAT and excises) 11.7
- Trade flow shares (percent of total exports):
  - Total exports 51.0 (to Euro area 40.0, to Nordic and Baltics 45.0, to Rest of the World 30.0)
  - Export shares to specific partners: to Estonia 1.2; to Euro area 17.9; to Nordic and Baltics 15.3; to Rest of the World 17.8
- Behavioral and structural parameters (selected):
  - Frisch elasticity (sensitivity of labor supply to real after-tax wage) = -0.10 (baseline)
  - Elasticity of substitution between labor and capital = -0.75
  - Elasticity of intertemporal substitution = 0.40
  - Wedge between rate of time preference and yield on government bonds = 15 percent (planning horizon of 7 years)
  - Fraction of consumers liquidity-constrained = 0.70 (these households consume 40 percent of aggregate consumption)
  - Markup over marginal cost: tradables sector = 17 percent; nontradables sector = 25 percent
- Additional parameters (from Table A.2 and calibration listing):
  - Planning horizon of consumers 7 years
  - Labor disutility parameters 0.90
  - Fraction of rule-of-thumb consumers 0.70
  - Intertemporal elasticity of substitution 0.40
  - Elasticity of substitution between capital and labor 0.75
  - Depreciation rate on capital (listed but value not specified in extracted text)
  - Capital adjustment cost parameters 1.0
  - Elasticity of substitution between varieties (Tradables sector) 16.7
  - Price markup over marginal cost (tradables) 1.06
  - Price markup over marginal cost (nontradables) 1.04
  - Capital share in production tradables sector 0.70
  - Capital share in production nontradables sector 0.70
  - Utility from real money balances 0.02
  - Price stickiness parameters 0
  - Home bias in government consumption yes
  - Home bias in private consumption no
  - Elasticity of substitution between traded and nontraded goods 0.40
  - Bias towards domestically produced tradable over nontradables 2.50

### Stylized example explaining the sustainability indicator (preserved numbers)
- Stylized economy assumptions:
  - Initial primary deficit = 1 percent of GDP
  - Initial debt stock = 90 percent of GDP
  - Growth-adjusted interest rate = 2 percent
  - Aging costs increasing linearly by 5 percent of GDP during 2005–50 and stay constant as a share of GDP beyond 2050
- With these parameters, an increase in the initial debt stock or the NPV of aging costs by 1 percentage point of GDP reduces the indicator by 0.02 percentage point of GDP.
- The indicator produces an adjustment of 6 percentage points of GDP to restore sustainability, with component contributions:
  - Balance component: 1 percentage point of GDP
  - Debt component: 1.8 percentage points of GDP
  - Aging component: 3.3 percentage points of GDP
  - Indicator total: -6.1 (percent of GDP)

*Source: IMF staff report chapter — “Population aging will lead to a sizeable expansion in pension and health care costs in Estonia.”*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2007/_cr07256.pdf_