## _wp15282

## Source details

**Canonical URL:** [_wp15282](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15282.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15282.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15282.pdf.json)

---

### Scope and objective
- Studies the potential economic impact of Chile’s 2014–15 structural reform agenda using the IMF multi-country dynamic general equilibrium model (GIMF).
- Reform agenda components: changes to the tax regime, boosts to infrastructure (primarily energy, transportation, and telecommunication), and reforms aimed at improving human capital (comprehensive education reform).
- Uses a wide range of shocks to account for uncertainty and trade-offs between short and long run.

### Model and calibration (GIMF features and Chile-specific calibration)
- Model features:
  - optimizing behavior by households and firms; tradable and nontradable goods sectors; nominal and real rigidities;
  - liquidity-constrained households and optimizing-saving households; macro-financial linkages;
  - non-Ricardian features yield relatively large fiscal multipliers.
- Country setup (three-country GIMF): Chile (domestic), Emerging Asia (EA), Rest of the World (RW).
  - Chile GDP share of world GDP: 0.6 percent.
  - EA: 17.4 percent.
  - RW: 82.0 percent.
- Key steady-state aggregates and shares (Chile, unless otherwise noted):
  - Imports of final goods: 25 percent of GDP.
  - Imports of investment goods: 7.6 percent of GDP.
  - Exports concentrated in intermediate goods: 15 percent of GDP; final goods: 8 percent of GDP.
  - Private consumption: 63 percent of GDP.
  - Total investment: 22 percent of GDP.
  - Government net debt-to-GDP ratios: Chile = -5 percent (average level in 2013 and 2014); EA = 33 percent; RW = 50 percent.
  - Public spending: Chile = 14 percent of GDP; EA = 18 percent; RW = 20 percent.
  - Tax rates (Chile, average 2014): labor income tax = 27.7 percent; capital income tax = 12.9 percent; consumption tax = 17.1 percent.
- Main calibration choices:
  - Real adjustment costs to imported final goods: Chile = 1.5; EA and RW = 1.
  - Share of tradable goods on real income: Chile = 50 percent; RW = 40 percent.
  - CES parameter ξ_Z = 1 (collapses to Cobb-Douglas).
  - Capital factor share α: tradable sector = 0.52 percent; nontradable sector = 0.38 percent (Chile).
  - Time preference β = 0.966 (implies capital-to-GDP ratio of 179).
  - Share of liquidity-constrained agents ψ = 0.5 (Chile) versus 0.3 (RW).
  - Fiscal rule: d_Y = 0.25; d_B = 0.
  - Monetary policy (Chile): coefficient on i_{t−1} = 0.3; coefficient on deviation from inflation target = 1.5; coefficients on output gap and exchange rate = 0.

### Chile’s identified structural gaps (summary diagnostics)
- Electricity:
  - Installed generating capacity per worker about 55 percent below OECD average.
  - Two main grids:
    - Northern Interconnected System (SING): about 20 percent of total installed capacity; relies mostly on thermal generation and suffers from overcapacity.
    - Central Interconnected System (SIC): accounts for about 70 percent of generating capacity and supplies 93 percent of the population; hydro-dominated and subject to rationing in dry years.
  - Distance and lack of interconnection reduce efficiency and increase costs; electricity costs among the highest in Latin America and among OECD countries.
- Transportation:
  - km of roads per worker: Chile is 67 percent below OECD average.
  - Ranks well behind OECD peers on other transportation measures (e.g., km of railroads).
  - Road infrastructure has not kept pace with rapid motorization and freight traffic growth.
- Telecommunication:
  - Broadly in line with OECD average in number of landlines and cell phones per worker; above Latin American countries.
- Composite infrastructure gap:
  - log Z_t = 0.4 log Υ + 0.4 log Ω + 0.3 log Θ
  - Chile’s 10-year average infrastructure capital about 50 percent below the OECD average, but above the Latin American average.
- Human capital gap:
  - Primary and secondary students: 47 percent municipal schools; 47 percent private subsidized schools; 6 percent non-subsidized schools.
  - Tertiary education: 80 percent of students enrolled in private universities or technical institutions.
  - Average years of schooling of Chile’s labor force about 12 percent below OECD level (as of 2010).
  - Based on 2012 PISA scores, Chile ranks 15 percent below the OECD average.
  - Education system highly segregated; debate exists on voucher system effects.

### 2015–18 main reforms and measures (selected items)
- Energy:
  - Objectives: reduce electricity marginal costs and increase generating capacity.
  - Measures: improve connectivity between SING and SIC; boost incentives for renewable sources (target: renewables represent 45 percent of electricity generation capacity over the next decade); ease permit regulatory practices.
  - Ministry of Energy data on projects already approved estimate electricity generation capacity to increase by more than 40 percent by 2020.
- Transportation and telecommunication:
  - Strengthen urban/intercity connectivity and port infrastructure (including construction of a large central port) via investment and expedited PPP procedures.
  - Telecommunication: boost internet access, data transmission, fiber-optic national network coverage; establish a Fondo de Desarollo de Telecommunicaciones.
- Education (Ley de Inclusion, passed January 2015 for primary and secondary):
  - i) abolition of co-payments; ii) repeal of for-profit subsidized schools (by 2018); iii) end of early-education selection entry-tests.
  - Private co-payments replaced by public resources to expand lower-income families’ school choice.
  - Tertiary education program: free tertiary education for students in the bottom 60 percent of income distribution attending universities (public: 16; private: 9) belonging to the Consejo de Rectores or equivalent professional institutions.
  - Agenda contemplates greater spending on public education, including teacher formation, school infrastructure, and child care facilities.
- Income tax (September 2014 tax reform):
  - Designed to increase structural revenue by 3 percent of GDP to help finance the education reform.
  - SII preliminary estimates: effective marginal tax rate on capital income increases by 3 percentage points by 2018, from an initial steady-state level of about 13 percent.
  - New law gradually increases corporate tax rates and offers firms a choice between an integrated tax system and a semi-integrated regime.
  - In GIMF, higher capital taxation reduces return on capital, inducing less investment and lowering private consumption.

### Simulation design for infrastructure and human capital shocks
- Production with infrastructure: F(Ξ,Z,T;K,L) = (Ξ_t Z_t^φ)(K_t^α (T_t L_t)^{1−α}) with φ = 0.1.
- Human capital (labor-augmenting productivity): T_t(s) = κ^{1−μ} s^{1−μ} with κ = 0.32 and μ = 0.58.
- Infrastructure shocks calibrated by comparing Chile to OECD via extended production function; education shocks calibrated via labor-augmented productivity comparison.
- Government spending adjusts to stabilize debt and honor a budget target; government spending includes consumption and investment, with government investment augmenting public infrastructure capital.

### Tax shocks — description and calibration
- 2014 tax reform simulated as increases in taxes on capital income and consumption to finance structural reforms.
- SII estimates: effective marginal tax rate on capital income increases by 3 percentage points by 2018, from about 13 percent.
- Reform also extends the VAT base (including real estate) and increases excise taxes on non-primary goods.
- Modeled as increases in lump-sum taxes and consumption tax rates so overall increase in fiscal revenues from the full 2014 tax reform is 3 percent of GDP.
- Government investment and infrastructure-related measures can be calibrated as TFP shocks because they incentivize private investment.

### Scenario design: three key dimensions
- Effectiveness (percent of gap closed):
  - Infrastructure gaps closed: 20, 50, 80 (Low, Medium, High).
  - Human capital gaps closed: 20, 50, 80 (Low, Medium, High).
- Speed (years to close gaps):
  - Infrastructure: 15, 10, 5 (Low, Medium, High).
  - Human capital: 20, 15, 10 (Low, Medium, High).
  - Energy exception: assume 40 percent of energy generation gaps are closed within the first 5 years (implying almost 30 percent of overall infrastructure gap closed in first 5 years).
  - Education gap closure begins only after the first 5 years and is closed linearly thereafter.
- Credibility (delay before agents internalize policies):
  - Immediate (immediate), internalized after 2 years (2), and internalized after 4 years (4).
- Simulations consider independent combinations of credibility, effectiveness, and speed.

### Results — overview and sensitivity
- Net impact of reforms varies widely across scenarios.
- Tax reform exerts negative effects on investment and consumption in both short and long run; infrastructure and human capital reforms yield gradually building benefits.
- Long run: net impact mainly depends on effectiveness.
- Short and medium run: credibility and speed of implementation are more important.
- Implementation risk: IMF surveys show structural reforms are implemented in only about 50 percent of cases.

### Selected scenario outcomes (highlights)
- Medium effectiveness scenario (50 percent of both gaps closed; infrastructure 10 years; human capital 15 years):
  - Low credibility vs high credibility: level of real GDP after 5 years is 2-3 percentage points lower in low credibility scenario.
- Under high credibility, speed matters:
  - Low speed (infrastructure 15 years, human capital 20 years) vs high speed (infrastructure 5 years, human capital 10 years): real GDP is 3 percentage point lower after 5 years under low speed.
- Worst-case scenario (low effectiveness, low credibility, low speed):
  - Real GDP immediately falls relative to no-reform baseline.
  - By 2020 real GDP is -0.6 percent below baseline and only marginally above by 2025 and in the long run.
- Most optimistic scenario:
  - GDP immediately increases by 1.3 percent and is about 15 percent above baseline by 2025.

### Median scenario — assumptions and quantitative path (selected variables; percent deviation from baseline)
- Median assumptions:
  - Infrastructure: 50 percent of gap closed in 10 years.
  - Human capital: 50 percent of gap closed in 15 years.
  - Public fully internalizes effectiveness after 2 years.
- Median scenario outcomes (percent deviation from baseline):
  - GDP: 2015 = -0.3; 2016 = -0.4; 2020 = 1.9; 2030 = 7.5; SS = 8.0.
  - Consumption: 2015 = -0.6; 2016 = -0.9; 2020 = 0.3; 2030 = 3.1; SS = 4.4.
  - Investment: 2015 = -1.6; 2016 = -1.7; 2020 = 3.2; 2030 = 7.3; SS = 5.8.
  - Government spending: 2015 = -0.2; 2016 = 0.2; 2020 = 1.4; 2030 = 1.3; SS = 1.3.
  - Export: 2015 = 0.2; 2016 = 0.2; 2020 = 0.9; 2030 = 7.8; SS = 8.2.
  - Import: 2015 = -0.6; 2016 = -0.7; 2020 = 2.7; 2030 = 2.3; SS = 2.4.
  - REER (+=Deprec): 2015 = 0.1; 2016 = 0.1; 2020 = 0.4; 2030 = 2.6; SS = 2.6.
- Median scenario headline results:
  - Increase of real GDP by about 6 percent in 2025 and 8 percent in the long run, with a small negative impact during the first two years.
  - Private investment declines by about 2 percent in the first two years, reflecting higher taxes on capital income.
  - Short-term impact on real GDP is small: real GDP only slightly below no-reform baseline by 0.3 percent in 2015 and about 0.4 percent in 2016.
- Mechanisms limiting short-term negative impact of higher taxation:
  - Weaker private demand reduces labor demand, lowering marginal cost and domestic prices.
  - Fall in inflation leads monetary authority to reduce nominal policy rate, lowering real interest rate and cost of capital.
  - Lower real interest rate leads to REER depreciation, boosting net exports.
  - Higher tax revenue is gradually used to finance higher government spending equivalent to 3 percent of GDP, supporting aggregate demand.

### Additional median-scenario projection by 2020
- By 2020, when positive effects of structural reforms are fully internalized:
  - Real GDP about 2 percent higher than no-reform baseline.
  - Investment and consumption would be 3 and 0.4 percent higher than the baseline case by 2020, respectively.
  - Exports are 1 percent higher as the REER further depreciates with reforms increasingly bearing fruits.
  - (Source text contains an incomplete sentence: "with TFP growth increasing by  percent.")

### Conclusions — key takeaways
- The 2015–18 reform agenda has potential to increase long-run productivity and output, but a smooth transition depends on credibility, effectiveness, and speed of implementation.
- Chile has accumulated infrastructure and human capital gaps relative to OECD averages; closing these gaps can sharply improve GDP level.
- Negative impact on GDP from higher capital income taxes is likely to be minor and soon offset by positive effects of structural reforms on productivity in many scenarios.
- Badly designed reforms that remove only a very small fraction of gaps, implement slowly, and have little credibility can greatly reduce positive GDP impact.
- Implementation risk is high: structural reforms are implemented in only about 50 percent of cases (IMF, 2015b), and education reform impacts are particularly uncertain due to implementation lags and uncertainty about effects on education quality and human capital accumulation.

*Source: _wp15282 - IMF Working Paper sections 1, 3.1–4.2, and 5 (provided content).*

### 1. Introduction ........................................................................................................

### _wp15282 - 1. Introduction

### Scope and objective
- The paper studies the potential economic impact of Chile’s 2014–15 structural reform agenda using the IMF multi-country dynamic general equilibrium model (GIMF).
- The agenda includes changes to the tax regime, boosts to infrastructure (primarily energy, transportation, and telecommunication), and reforms aimed at improving human capital (comprehensive education reform).

### Model and approach
- Uses GIMF, which features:
  - optimizing behavior by households and firms;
  - tradable and nontradable goods sectors;
  - nominal and real rigidities;
  - liquidity-constrained households and optimizing-saving households;
  - macro-financial linkages.
- The model’s non-Ricardian features make fiscal measures non-neutral, yielding relatively large fiscal multipliers.
- The model is calibrated to the Chilean economy and simulates the 2014 tax reform, the education reform, and the infrastructure plan.
- A wide range of shocks is calibrated to account for uncertainty and trade-offs between short and long run.

### Key summary findings (from simulations)
- The agenda has the potential to increase Chile’s long-run productivity and output, but a smooth transition depends on three dimensions:
  - credibility of the reforms;
  - effectiveness in closing structural gaps;
  - speed of implementation.
- The impact of the reforms in the medium run lies within a wide range of -0.2 and 6 percent of GDP, determined by the combination of the three key dimensions.
- The median results point to:
  - an increase of GDP by 2 percent in 2020;
  - an increase of GDP by 8 percent in the long run.
- The net outcome of the government reform agenda is positive in most simulation scenarios, especially in the long run, but short- and medium-run outcomes are sensitive to reform design, implementation speed, and credibility.
- Badly designed reforms that remove only a very small fraction of the existing structural gaps, are implemented slowly, and lack credibility can greatly reduce the positive impact on GDP.

### Uncertainty and empirical context
- Recent IMF surveys show that structural reforms are implemented in only about 50 percent of cases (IMF, 2015b), underscoring implementation risk.
- Education reform impacts are particularly uncertain due to implementation lags and uncertainty about the effect of specific policies on education quality and human capital accumulation.
- The Chilean education reform includes measures aimed at reducing segregation, improving school infrastructure and teacher curricula, and providing free access to tertiary education; however, whether segregation and lower outcomes are exclusively due to the voucher system remains an open question.

---

### Chile’s identified structural gaps (summary statistics and diagnostics)
- Electricity:
  - Electricity potential in Chile, as measured by installed generating capacity per worker, is about 55 percent below OECD average.
  - The electricity distribution is organized around two main grids:
    - Northern Interconnected System (SING): serves mining regions and accounts for about 20 percent of total installed capacity; relies mostly on thermal generation and suffers from overcapacity.
    - Central Interconnected System (SIC): supplies energy to the central part of the country and to most of the population (93 percent), accounting for about 70 percent of the generating capacity; hydro-dominated and subject to rationing in dry years.
  - Distance and lack of interconnection between the two main grids reduces efficiency and increases costs; electricity costs are among the highest in Latin America and among OECD countries.
- Transportation:
  - In terms of km of roads per worker, Chile is 67 percent below OECD average.
  - Chile ranks well behind OECD peers on other transportation measures, such as km of railroads.
  - Road infrastructure has not kept pace with rapid motorization and freight traffic growth.
- Telecommunication:
  - Chile is broadly in line with OECD average in terms of telecommunication infrastructure (number of landlines and cell phones per worker) and is largely above Latin American countries.
- Composite infrastructure gap:
  - Using a principal-component approach combining electricity (Υ), transportation (Ω), and telecommunication (Θ), the composite indicator (Z) is:
    - log Zt = 0.4 log Υ + 0.4 log Ω + 0.3 log Θ
  - Based on this indicator, Chile’s 10-year average infrastructure capital is about 50 percent below the OECD average, but above the average for Latin American countries.

### Human capital gap (summary statistics)
- Education system composition:
  - Primary and secondary students: 47 percent attend municipal schools, 47 percent attend private subsidized schools, and 6 percent attend non-subsidized schools.
  - Tertiary education: 80 percent of students are enrolled in private universities or technical institutions.
  - The system is highly segregated, with distribution of students strongly dependent on socioeconomic status.
- Quantified gaps:
  - Average years of schooling of Chile’s labor force was about 12 percent below OECD level (as of 2010, Barro and Lee (2014)).
  - Based on 2012 PISA scores, Chile ranks 15 percent below the OECD average.
- Education quality has not clearly improved since the introduction of the voucher system; debate exists on whether vouchers caused increased segregation.

### 2015–18 main reforms and measures (selected items)
- Energy:
  - Objectives include reducing electricity marginal costs and increasing generating capacity.
  - Measures include improving connectivity between SING and SIC; boosting incentives for renewable sources (target: renewables represent 45 percent of electricity generation capacity over the next decade); easing permit regulatory practices to facilitate private involvement and competitive tendering.
  - Ministry of Energy data on projects already approved estimate electricity generation capacity to increase by more than 40 percent by 2020.
- Transportation and telecommunication:
  - Strengthen urban/intercity connectivity and port infrastructure (including construction of a large central port) via direct investment and incentivized private sector involvement through expedited PPP procedures.
  - Telecommunication aims to boost internet access, data transmission, and fiber-optic national network coverage, including establishing a Fondo de Desarollo de Telecommunicaciones.
- Education (Ley de Inclusion, passed January 2015 for primary and secondary):
  - i) abolition of co-payments that allowed families to send children to more expensive schools;
  - ii) repeal of for-profit subsidized schools (by 2018);
  - iii) end of early-education selection entry-tests.
  - Private co-payments will be replaced by public resources to expand lower-income families’ school choice.
  - Tertiary education program: ensure free tertiary education for students in the bottom 60 percent of the income distribution attending universities (public: 16; private: 9) belonging to the Consejo de Rectores or professional institutions meeting similar standards.
  - The agenda contemplates greater spending on public education, including teacher formation, school infrastructure, and child care facilities.
- Income tax (September 2014 tax reform):
  - The reform was designed to increase structural revenue by 3 percent of GDP and help finance the education reform.
  - Preliminary estimates by the Chilean Internal Revenue Service (SII) suggest the effective marginal tax rate on capital income would increase by 3 percentage points by 2018.
  - The new law gradually increases corporate tax rates and offers firms a choice between:
    - an integrated tax system (less generous than the old system because dividends are now taxed when accrued), and
    - a semi-integrated regime (dividends taxed only when distributed but at a higher rate and shareholders can only partially deduct that tax).
  - In GIMF, higher capital taxation reduces return on capital, inducing less investment and lowering private consumption as household income falls; thus the tax reform is expected to reduce output mainly by reducing investment.

### Literature and positioning
- Related to macroeconomic impact assessments of structural reforms (Bouis and Duval, 2011; Barnes et al., 2013; Lusinyan and Muir, 2013; Gomes et al., 2013).
- Draws on infrastructure-to-output contributions (Calderon and Serven, 2004; Calderon, Moral-Benito and Serven, 2014) and human capital effects on growth (Bills and Klenow, 2000; Psacharopoulos, 2004).
- Tax impacts on investment and dynamics are informed by Hassett and Hubbard (2002), Chang (1995), Santoro and Wei (2011).

*Source: _wp15282 - 1. Introduction (IMF PDF content unit)._*

### 3.1    The model

### 3.1    The model

### Households
- Two household types per country: optimizing agents with finite time horizon and liquidity-constrained agents, in measure (1−ψ) and ψ respectively.
- Optimizing agents are overlapping generations (OLG) with finite planning horizons (myopia governed by θ), face constant probability of death, and choose consumption, savings, and labor supply to maximize lifetime utility as in equation (2).
- Utility features:
  - Habit persistence on previous aggregate consumption with intensity ν (equation (2)).
  - CES variety preferences for consumption with elasticity of substitution σR (equation (3)).
  - Leisure term (S_t − l_a,t) and preference weight η.
- Budget constraint for OLG households is given by equation (4) and includes:
  - Retailer price P^R_t, consumption tax τ_c,t, lump-sum taxes τ^ls,OLG_a,t, and transfers Υ^OLG_a,t.
  - Holdings of domestic government bonds B^G_a,t, bank bonds B^N_a,t and B^T_a,t, and foreign bonds F_a,t converted at nominal exchange rate ϕ_t.
  - Domestic bond interest rates include a domestic risk-premium ξ_t,b (ξ_t,b < 0 indicates private sector faces larger marginal funding rate than public sector); foreign bonds include foreign exchange risk-premium ξ_t,f.
- Liquidity-constrained households:
  - Do not access credit markets and consume after-tax income each period.
  - Utility similar to OLG (equation (5)) with variety preferences as in equation (3).
  - Budget constraint provided in equation (6).
  - Labor supply for liquidity-constrained agents can be treated as exogenous; then consumption equals after-tax income.
- Money:
  - Model can yield a money demand function via separable preference for money, but with monetary policy as an interest-rate rule the money-demand equation is redundant (note in text).

### Production
- Multi-layer production structure with competitive input markets and monopolistically competitive output markets in each layer (except distribution).
- Manufacturing firms (continuum indexed z∈[0,1]) produce tradable and nontradable goods using capital from capital goods producers and labor from unions in CES form (equation (7)):
  - Country-specific TFP A_t, capital K_t(z), labor L_t(z), capital share α, substitution elasticity ξ_Z, labor-augmenting productivity T_t. Trend productivity g = T_t / T_{t−1}.
- Adjustment features:
  - Investment adjustment costs G_I,t(I_t,I_{t−1}) introduce a marginal q that depends on r_t, u_t, MPK, τ_K,t and adjustment costs (equation (8)), with dq/dτ_K < 0.
- Distribution chain:
  - Manufacturing → intermediate goods → several distributor layers → final consumption and investment goods.
  - Distributors face nominal rigidities; import volume changes face quadratic adjustment costs.
  - Retailers sell to households in monopolistic competition with real price rigidities.
- Capital goods producers supply capital subject to investment adjustment costs; the financial accelerator (Bernanke, Gertler and Gilchrist, 1999) operates at this layer.

### Fiscal and monetary policy
- Fiscal policy objective: ensure a stable government debt-to-GDP ratio in the long run, excluding sovereign default.
- Fiscal rule (Chile) allows countercyclical response (equation (9)):
  - (ΣY)_t = (ΣY)_SS + d_B[(B/Y)_t − (B/Y)_SS] + d_Y(Y_t/Y_SS)
  - (ΣY)_t is budget surplus as share of GDP; d_B and d_Y govern reactions to debt deviations and the output gap.
- Fiscal instruments: labor income tax, consumption taxes, capital return taxes, lump-sum taxes; spending includes consumption, investment, and lump-sum transfers. Government issues debt.
- Monetary policy:
  - Central bank follows strict inflation targeting (single mandate) using a Taylor-type rule (equation (10)):
    - i_t depends on lagged rate i_{t−1} (intensity δ_i), weighted present and forward inflation π (weight δ_p), deviation from target π*, output gap (Y_t/Y_SS), and nominal exchange rate depreciation η relative to η* with intensity δ_η.
  - For Chile calibration: coefficients on i_{t−1} and inflation deviation set at 0.3 and 1.5 respectively; response coefficients to output gap and exchange rate set to 0 (single mandate).

### Calibration (section 3.2): model baseline and country shares
- Three-country GIMF: Chile (domestic), Emerging Asia (EA), Rest of the World (RW).
  - Chile GDP share of world GDP: 0.6 percent.
  - EA: 17.4 percent.
  - RW: 82.0 percent.
- Expenditure-to-GDP and steady-state aggregates (Table 4 summarized in text):
  - Chile high external dependence:
    - Imports of final goods: 25 percent of GDP (Chile) versus 12.6 (EA) and 4.4 (RW).
    - Imports of investment goods: 7.6 percent (Chile) versus 3.4 (EA) and 1.1 (RW).
  - Exports concentration in intermediate goods: 15 percent of GDP in intermediate goods versus 8 percent in final goods (Chile).
  - Private consumption: 63 percent of GDP.
  - Total investment: 22 percent of GDP.
  - Government net debt-to-GDP ratios:
    - Chile: -5 percent (average level in 2013 and 2014).
    - EA: 33 percent.
    - RW: 50 percent.
  - Public spending:
    - Chile: 14 percent of GDP.
    - EA: 18 percent.
    - RW: 20 percent.
  - Tax rates (Chile, average 2014):
    - Labor income tax: 27.7 percent.
    - Capital income tax: 12.9 percent.
    - Consumption tax: 17.1 percent.

### Main parameters (section 3.2.2)
- Many parameters follow Kumhof et al. (2010) (US calibration), with Chile-specific adjustments:
  - Real adjustment costs to imported final goods: Chile 1.5 versus 1 for EA and RW.
  - Share of tradable goods on real income: 50 percent (Chile) versus 40 percent (RW).
  - CES parameter ξ_Z set to 1 so equation (7) collapses to Cobb-Douglas (equation (11)).
  - Capital factor share α:
    - Tradable sector α set to 0.52 percent.
    - Nontradable sector α set to 0.38 percent.
    - RW: tradable 0.45 and nontradable 0.30.
  - Time preference β chosen to be consistent with capital-to-GDP ratio of 179, which implies β = 0.966.
  - Share of liquidity-constrained agents ψ = 0.5 (Chile) versus 0.3 (RW).
    - Motivation: capture limited credit access due to high income inequality and skewed collateral distribution.
- Fiscal rule calibration:
  - d_Y (reaction of fiscal surplus to output gap) set to 0.25 (as in Kumhof et al., 2010 for EA).
  - d_B (reaction to debt-to-GDP) set to 0.
- Monetary policy calibration for Chile:
  - Coefficient on i_{t−1}: 0.3.
  - Coefficient on deviation from inflation target: 1.5.
  - Coefficients on output gap and exchange rate: 0.

### Simulations: approach and assumptions (section 4 introduction)
- Structural reform impacts on productivity and long-run GDP are highly uncertain (implementation probability ~50 percent in IMF surveys cited).
- Policy experiments simulated with a three-country GIMF.
- Reforms imply macroeconomic and public finance changes during transition and in new steady state; e.g., a tax reform that raises capital taxes increases income-tax revenues and is offset by adjustments in other fiscal instruments.
- Government spending is assumed to adjust to honor a budget target and stabilize debt (motivations: Chile 2015–18 agenda includes spending increases financed by tax reform; focus on macroeconomic impacts rather than welfare/redistribution).

### Infrastructure and human capital shocks (section 4.1)
- Productivity decomposition and infrastructure inclusion:
  - Extended production function (equation (12)): F(Ξ,Z,T;K,L) = (Ξ_t Z_t^φ)(K_t^α (T_t L_t)^{1−α})
  - Parameter φ calibrated by Calderon, Moral-Benito and Serven (2014) at φ = 0.1.
  - Infrastructure shocks are calibrated by comparing Chile to OECD via relation in equation (13) using the extended production function; shock is calibrated as a TFP shock.
- Human capital specification:
  - Labor-augmenting productivity T_t(s) specified as (equation (14)): T_t(s) = κ^{1−μ} s^{1−μ}
  - Parameter values: κ = 0.32 and μ = 0.58 (Bills and Klenow, 2000 estimates).
  - Calibration of education shocks: reforms that improve school quality close part of the gap with OECD and the implied productivity shock is calculated via relation in equation (15) comparing Chile and OECD labor-augmented productivity.
- Simulation practice:
  - A wide range of shocks are calibrated for infrastructure and education reforms to reflect uncertainty in implementation and impacts on productivity and GDP.
  - All policy simulations assume government spending adjusts to stabilize debt and honor a budget target.

*Source: IMF Working Paper section 3.1–4.1 (The model; Calibration; Simulations).*

### 4.2    Tax shocks

### _wp15282 - 4.2    Tax shocks

### 4.2 Tax shocks — description
- The 2014 tax reform is simulated as an increase in taxes on capital income and consumption to help finance structural reforms.
- Chilean Internal Revenue Service (SII) preliminary estimates: the effective marginal tax rate on capital income increases by 3 percentage points by 2018, from an initial steady-state level of about 13 percent.
- In the model (GIMF):
  - Higher capital taxation reduces the return on capital, inducing firms to invest less and weakening private consumption as household income falls.
  - The 2014 tax reform also increases taxes on consumption by extending the VAT tax base (including on real estate) and increasing excise taxes on non-primary goods (such as tobacco and alcohol).
  - This is modeled as an increase in lump-sum taxes and in consumption tax rates so that the overall increase in fiscal revenues from the full 2014 tax reform is 3 percent of GDP.

### 4.2 Tax shocks — calibrated modeling notes
- Government spending in the model is both consumption and investment; government investment augments publicly provided infrastructure capital used in final output production.
- Infrastructure-related measures are calibrated as a TFP shock because their nature is mostly related to product market reform incentivizing private investment, especially in energy and transportation sectors.

---

### 4.3 Dimensions affecting the outcome of reforms — scenario design
- Simulated reforms vary across three key dimensions: effectiveness, speed, and credibility.
- Effectiveness (percent of gap closed):
  - Infrastructure gaps closed: 20, 50, 80 (Low, Medium, High).
  - Human capital gaps closed: 20, 50, 80 (Low, Medium, High).
- Speed (years to close gaps):
  - Infrastructure: 15, 10, 5 (Low, Medium, High).
  - Human capital: 20, 15, 10 (Low, Medium, High).
  - Note: Within each time span, gaps are closed linearly with the exception of energy. Given capacity built up in 2015, assume 40 percent of energy generation gaps are closed within the first 5 years (implying almost 30 percent of overall infrastructure gap closed in first 5 years). Education gap closure begins only after the first 5 years and is closed linearly thereafter.
- Credibility (delay before agents internalize policies):
  - Credibility scenarios: policies immediately effective (immediate), internalized after 2 years (2), and internalized after 4 years (4) for both infrastructure and human capital.
- Simulation coverage:
  - Simulations consider independent combinations of credibility, effectiveness, and speed, recognizing possible interactions between dimensions.

---

### 4.4 Results — overview
- Net impact of reforms shows large variation across scenarios.
- Tax reform has negative effects on investment and consumption in short and long run; infrastructure and human capital reforms yield benefits that build up gradually.
- Long run: net impact mainly depends on effectiveness.
- Short and medium run: credibility and speed of implementation are more important.

### 4.4.1 Dimensions: credibility, effectiveness and speed — selected outcomes
- Medium effectiveness scenario: 50 percent of both infrastructure and human capital gaps closed at medium speed (infrastructure: 10 years; human capital: 15 years).
  - With low credibility vs high credibility: level of real GDP after 5 years is 2-3 percentage points lower in low credibility scenario.
- Medium effectiveness under high credibility:
  - With low speed (infrastructure 15 years, human capital 20 years) vs high speed (infrastructure 5 years, human capital 10 years): real GDP is 3 percentage point lower after 5 years under low speed.
- Worst-case scenario (low effectiveness, low credibility, low speed):
  - Real GDP immediately falls relative to no-reform baseline.
  - By 2020 real GDP is -0.6 percent below baseline and only marginally above by 2025 and in the long run.
- Most optimistic scenario:
  - GDP immediately increases by 1.3 percent and is about 15 percent above baseline by 2025.

### 4.4.2 Median scenario — quantitative path and mechanisms
- Median scenario assumptions:
  - Infrastructure: 50 percent of gap closed in 10 years.
  - Human capital: 50 percent of gap closed in 15 years.
  - Public fully internalizes effectiveness after 2 years.
- Median scenario outcomes (Table 7: % deviation from baseline):
  - GDP: 2015 = -0.3; 2016 = -0.4; 2020 = 1.9; 2030 = 7.5; SS = 8.0.
  - Consumption: 2015 = -0.6; 2016 = -0.9; 2020 = 0.3; 2030 = 3.1; SS = 4.4.
  - Investment: 2015 = -1.6; 2016 = -1.7; 2020 = 3.2; 2030 = 7.3; SS = 5.8.
  - Government spending: 2015 = -0.2; 2016 = 0.2; 2020 = 1.4; 2030 = 1.3; SS = 1.3.
  - Export: 2015 = 0.2; 2016 = 0.2; 2020 = 0.9; 2030 = 7.8; SS = 8.2.
  - Import: 2015 = -0.6; 2016 = -0.7; 2020 = 2.7; 2030 = 2.3; SS = 2.4.
  - REER (+=Deprec): 2015 = 0.1; 2016 = 0.1; 2020 = 0.4; 2030 = 2.6; SS = 2.6.
- Median scenario headline results:
  - Increase of real GDP by about 6 percent in 2025 and 8 percent in the long run, with a small negative impact during the first two years.
  - Private investment declines by about 2 percent in the first two years, reflecting higher taxes on capital income.
  - Short-term impact on real GDP is small: real GDP only slightly below no-reform baseline level by 0.3 percent in 2015 and about 0.4 percent in 2016.
- Model mechanisms that limit short-term negative impact of higher taxation:
  - Weaker private demand reduces labor demand, lowering marginal cost and domestic prices.
  - Fall in inflation leads monetary authority to reduce nominal policy rate, lowering real interest rate and the cost of capital, offsetting initial higher taxation impact.
  - Lower real interest rate leads to REER depreciation, boosting net exports.
  - Higher tax revenue is gradually used to finance higher government spending equivalent to 3 percent of GDP, supporting aggregate demand (consistent with government plan to use higher structural revenue to finance higher education spending).

### 4.4.2 Median scenario — additional projection by 2020
- By 2020, when positive effects of structural reforms are fully internalized:
  - Real GDP about 2 percent higher than no-reform baseline.
  - Investment and consumption would be 3 and 0.4 percent higher than the baseline case by 2020, respectively.
  - Exports are 1 percent higher as the REER further depreciates with reforms increasingly bearing fruits.
  - (The source text contains an incomplete sentence: "with TFP growth increasing by  percent.")

---

### 5 Conclusions — key takeaways
- The paper uses the IMF dynamic general equilibrium model (GIMF) to study the potential economic impact of the 2015-18 structural reforms agenda in Chile.
- Three key areas affect outcomes: credibility, effectiveness, and speed of implementation.
- The agenda has potential to increase long-run GDP, but design, credibility, and speed are crucial for a smooth transition to higher productive potential.
- Chile has accumulated gaps in infrastructure and human capital relative to average OECD economies; closing these gaps can sharply improve GDP level.
- Despite uncertainty, model simulations show:
  - Negative impact on GDP from higher taxes on capital income is likely to be minor and soon offset by positive effects of structural reforms on productivity.
  - Badly designed reforms that remove only a very small fraction of the gaps, implement slowly, and have little credibility can greatly reduce positive GDP impact.

*Source: _wp15282 - 4.2    Tax shocks (IMF working paper content provided).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15282.pdf_