## _wp0673 - 3. Sensitivity Analysis

## Source details

**Canonical URL:** [_wp0673 - 3. Sensitivity Analysis](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0673.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0673.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0673.pdf.json)

---

### Introduction and core tradeoff
- Sudden Stop phenomenon: sudden reversals of capital inflows and current account deficits; collapses in output and private absorption; large relative price corrections in domestic goods prices and asset prices.
- Two competing hypotheses:
  - Globalization hazard: world capital markets imperfect; an IFO offering ex-ante price guarantees could prevent Sudden Stops by credibly supporting asset prices.
  - International moral hazard: ex-ante guarantees raise foreign investors’ demand for emerging-market assets, transferring downside risk to the IFO and possibly inducing excessive indebtedness and worse Sudden Stops.
- Objective: study the globalization hazard–moral hazard tradeoff quantitatively in a DSGE model (Mendoza and Smith (2006) augmented with an IFO offering ex-ante asset-price guarantees).

### Model setup and frictions
- Agents and markets:
  - SOE with representative household and firm producing tradable output exp(εt)F(k,n).
  - Households trade equity (α) and one-period foreign bonds (b); equity price qt.
  - Representative foreign securities firm trades SOE equity.
  - IFO guarantees a minimum liquidation price (put-like); finances guarantees with lump-sum tax on foreign investors’ profits.
- Financial frictions:
  - Margin constraint: bt+1 ≥ -καt+1 q t k, with 0≤κ≤1 and short-selling constraint αt+1 ≥ χ, -∞<χ<1.
  - Trading costs: quadratic trading costs (parameter a) and recurrent trading cost θ.
- Preferences and production:
  - GHH preferences: utility over c - h(n); Cobb-Douglas production with labor share γ.
- IFO guarantee specifics:
  - Guaranteed return R̃t = (dt + q̃t)/q̃t-1 where q̃t is guaranteed price.
  - Lump-sum tax Tt* = max{0, k(q̃t - qt)} to cover executed guarantees.

### Equilibrium and pricing implications
- Competitive equilibrium: prices {wt, dt, qt} and allocations {ct, nt, αt, bt, Tt*} satisfy optimization, IFO budget and market clearing αt + αt* = 1.
- Key analytic insights:
  - Margin calls lower domestic asset prices directly (forced sales) and indirectly (increase covariance between marginal utility and returns).
  - Foreign demand is finite-elastic due to trading costs; fundamentals price qf defined as discounted dividends.
  - Foreign traders’ demand increases with percent deviation (qf − qt)/qt (elasticity 1/a) and with expected present value of excess prices from guarantees (international moral hazard).
  - Ex-ante guarantees can either prevent Sudden Stops by supporting foreign demand or create persistent overvaluation if set too high.

### Quantitative solution method and calibration
- State variables: α, b, ε. Grids: Α={α1<...<αNA} with α1=χ; Β={b1<...<bNB}. Productivity realizations E={εL<εH}.
- Algorithm: modified Mendoza–Smith quasi-planning with conjectured G(α,b,ε) for expected present discounted value of “excess prices”.
- Taxes implied by internalizing foreign demand are negligible in simulations:
  - Maximum taxes in absolute value range between 0.08 (0.4) and 0.2 (0.8) percent when a =0.2 (2).
  - Average tax in absolute value is 0.03 (0.3) percent in simulations with a =0.2 (2).
- Calibration (deterministic steady state and parameters):
  - σ =2.
  - γ =0.65.
  - Real interest rate R set to 6.5 percent annually, implying R=1.0651/4.
  - Labor disutility coefficient δ =2.
  - Mexican average GDP shares: sc =0.684, si =0.19, sg =0.092, snx =0.034.
  - Exogenous absorption = 0.282 (28.2 percent of steady-state GDP).
  - Deterministic steady-state values:
    - Depreciation rate (annual) = 7.75 percent.
    - Capital-output ratio sk ≈ 2.5.
    - Fundamentals equity price qf = 2.19.
    - Steady-state capital stock k = 79.
    - Time-preference elasticity β =0.0118.
  - Portfolio indeterminacy: any (α,b) with α qf k + b = s and non-binding constraints; debt portfolios require α > 0.9; low κ required for margin constraint to bind.

### Calibration of financial frictions and Nearly Frictionless Economy (NFE)
- Calibrated frictions:
  - a = 0.2
  - θ = 0.001
  - κ = 0.03
- NFE steady-state outcomes:
  - α = 0.931
  - b = -4.825 (implies debt-GDP ratio ≈ 0.62)
  - Foreign traders hold stationary equity at price q = qf /(1+aθ), yielding Rq > R.
- Key relation: η/λ = (Rq - R)/(Rq - Rκ).
- With binding margin constraint, b = -κ α qk and α solves the steady-state expression.

### Stochastic process, grids and baseline experiments
- Productivity shock process: two-point symmetric Markov with targets:
  - Standard deviation = 2.64 percent
  - First-order autocorrelation = 0.683
  - Resulting parameters: σε = 1.79 percent; ρε = 0.683
  - Two points: εL = -εH = 0.0179
  - Long-run probability of each state = ½
  - Transition probabilities: staying = 0.8415; shifting = 0.1585
- State grids:
  - Equity: 78 nodes spanning [χ, 1] with χ = 0.84.
  - Bonds: 120 nodes spanning [-5.2, 25.7].
- Four simulations:
  1. NFE (nearly frictionless economy)
  2. BMR (binding margin requirements) with κ = 0.005
  3. NSCG (non-state-contingent guarantee)
  4. SCG (state-contingent guarantee applied only in Sudden Stop region)
- Sudden Stop region: high-debt region where margin constraint binds (first 25 nodes of B grid and all 72 nodes of A grid).

### Main quantitative findings on guarantees and sensitivity
- Guarantees and execution frequency:
  - Guarantees set ½ to 1 percent above fundamentals price contain Sudden Stop effects and virtually eliminate probability of margin calls.
  - Non-state-contingent guarantee at that level: long-run probability of execution ≈ 1/3; predicts persistent overvaluation.
  - Same guarantee but state-contingent (applied only in high-debt episodes): long-run probability < 1/100 and equally effective at containing Sudden Stops without persistent overvaluation.
- Welfare and distributional implications by foreign demand elasticity:
  - High elasticity of foreign demand:
    - Guarantees improve domestic welfare from initial Sudden Stop state.
    - Long-run welfare changes negligible.
    - Value of foreign traders’ firms falls slightly in Sudden Stop state but long-run average rises sharply.
  - Low elasticity:
    - Higher guarantees needed to prevent Sudden Stops.
    - Moral hazard reduces domestic welfare gains at Sudden Stop states and can produce average welfare losses in stochastic steady state.
    - Guarantees can yield short-term macro/welfare improvements at expense of long-term welfare loss.

### Scenarios, mechanisms, and conditional dynamics
- Partial-equilibrium intuition:
  - Margin call shifts domestic demand down; price falls to B (Sudden Stop). IFO guarantee above B shifts foreign demand out and can restore price to fundamentals if set correctly.
  - Guarantee too low: no effect; too high: overpricing.
- Sudden Stop drivers:
  - Leverage ratio = debt / market value of equity.
  - Equity market liquidity = α - χ.
- Conditional impact examples (one-standard-deviation negative productivity shock):
  - High leverage initial state (α = 0.938, b = -4.68):
    - c drops by 1.5 percent more than in NFE.
    - q drops by 0.4 percent more than in NFE.
    - ca/y rises by about 1 percentage point of GDP.
  - Low leverage initial state (same α, b = -3.38):
    - c and ca/y responses in BMR nearly identical to NFE.
    - q still falls by about same amount as in high leverage scenario.
  - Persistence: Sudden Stops short-lived in baseline calibration; BMR responses converge to NFE in about 4 quarters.

### NSCG vs SCG quantitative effects (baseline guarantee = 0.5 percentage point above qf(εL) = 2.185; guaranteed price = 2.196)
- NSCG outcomes:
  - Long-run probability of binding margin constraints falls to 0.001%.
  - Long-run average of G = 0.01 (about ½ of a percent above mean equity price).
  - Mean equity price increases by 0.053 percent.
  - Consumption coefficient of variation falls by about 0.2 percentage points.
  - Equity price at date 0 ≈ 2/3 of a percentage point higher than in BMR in both high and low leverage states.
  - Guarantees executed often in both high- and low-leverage states; distort asset prices outside Sudden Stop region.
- SCG outcomes:
  - Long-run distributions less skewed than NSCG.
  - Long-run probability of binding margin constraints ≈ 0.001%.
  - Date-0 consumption and current account responses nearly identical to NSCG; recovery in asset prices slightly smaller.
  - SCG converges faster to NFE dynamics after date 0.
  - SCG yields higher asset prices mainly within Sudden Stop region, inducing smaller distortions outside Sudden Stop region relative to NSCG.

### Selected long-run business cycle moments (from Table 1)
- Panel I. NFE Economy highlights:
  - GDP: mean = 7.833; std. dev. = 2.644; first-order auto-correlation = 0.683
  - Consumption: mean = 5.366; std. dev. = 2.185; first-order auto-correlation = 0.770
  - equity price: mean = 2.187; std. dev. = 0.121; first-order auto-correlation = 0.606
  - foreign debt-GDP ratio: mean = 0.177; std. dev. = 56.694; first-order auto-correlation = 0.997
- Panel II. BMR Economy:
  - Probability of binding margin constraints = 3.973%
  - foreign debt-GDP ratio: mean = 0.499; std. dev. = 37.964
- Panel III. NSCG Economy:
  - Probability of binding margin constraints = 0.001%
  - equity price: mean = 2.198; std. dev. = 0.099
  - foreign debt-GDP ratio: mean = 1.336; std. dev. = 41.053
- Panel IV. SCG Economy:
  - Probability of binding margin constraints = 0.001%
  - equity price: mean = 2.188; std. dev. = 0.121
  - foreign debt-GDP ratio: mean = 1.114; std. dev. = 34.141

### Normative welfare results (compensating variation in date-0 consumption)
- Mean welfare costs (ergodic distribution) of deviating from NFE:
  - Average welfare losses in BMR, NSCG and SCG relative to NFE < 0.07 percent in c0.
- Welfare in Sudden Stop (high-leverage) state (short-run and long-run):
  - Short-run (date-0) welfare cost in c0:
    - BMR: 1.4 percent
    - NSCG: 0.9 percent
    - SCG: 1.0 percent
  - Long-run welfare gains (percent cut in c0 equating expected lifetime utility with NFE):
    - BMR: about 1 percent
    - NSCG: 1.9 percent
    - SCG: 1.3 percent
  - Total welfare:
    - NSCG: total welfare gain of 1 percent (domestic agents obtain total welfare gains)
    - SCG: implied positive total welfare but smaller than NSCG

### Distributional accounting and foreign traders’ returns (selected Table 2 figures)
- Long-run averages (Domestic agents / Foreign traders):
  - Domestic agents: Welfare cost (long-run average, compensating variation) — NFE: 0.017; BMR: 0.062; NSCG: 0.057.
  - Foreign traders: Present value of traders' returns (long-run average) — NFE: 17.920; BMR: 20.519; NSCG: 27.350; SCG: 25.529.
    - Percent change w.r.t. NFE (long-run average) — BMR: 14.499; NSCG: 52.621; SCG: 42.460.
- High leverage Sudden Stop state (date-0 and present values):
  - Domestic agents: Welfare cost (compensating variation) — NFE: 0.367; BMR: -1.036; NSCG: -0.317.
    - Short-run cost (date-0) — NFE: 1.376; BMR: 0.862; NSCG: 0.978.
    - Long-run cost (difference) — NFE: -1.010; BMR: -1.899; NSCG: -1.295.
  - Foreign traders (high-leverage present value): NFE: 10.931; BMR: 10.940; NSCG: 10.858; SCG: 10.901.
    - Date-0 returns: NFE: -0.192; BMR: -3.770; NSCG: -3.796; SCG: -3.791.
    - Value of trades in high-leverage BMR/NSCG/SCG ≈ 3.9 (vs NFE 0.359), driving date-0 decline in net returns.
- Interpretation:
  - Long-run increases in foreign traders’ firm value arise mainly from increased average equity holdings and higher dividend earnings.
  - Conditional on Sudden Stops, foreign traders’ payoffs fall in price-guarantee economies; guarantees reduce downside risk and raise long-run average returns for foreign traders while increasing moral hazard.

### Sensitivity analysis highlights (Table 3)
- Parameters varied: larger/more persistent productivity shocks (σε =0.024, ρε =0.8), higher guaranteed price (1 percent above qf(εL)), higher θ =0.01, higher a =2. Column (I) = baseline.
- BMR robustness:
  - Probability of binding margin constraint:
    - rises to 4.2 percent when ρε increases.
    - falls to 2.3 percent when σε increases.
  - Higher θ =0.01: margin-call probability = 16.740 percent.
  - Higher a =2: margin-call probability = 19.444 percent.
- NSCG / SCG with parameter changes:
  - Price guarantees virtually eliminate long-run probability of binding margin constraints across panels.
  - Long-run probability of executing guarantees:
    - NSCG: ranges 29–34 percent (falls to 15 percent when σε higher).
    - SCG: below 1 percent in baseline; rises to 6.063 and 3.245 percent with θ =0.01 and a =2 respectively.
  - Raising guaranteed price to 1 percent above qf(εL):
    - further damps Sudden Stops; e.g., fall in consumption in high-leverage Sudden Stop: NSCG -0.5 percent; SCG -0.6 percent.
    - Domestic welfare gain in high-leverage Sudden Stop increases to 2.4 percent (NSCG Column IV) and 0.9 percent (SCG Column IV).
  - High per-trade costs (a =2):
    - baseline guarantee 0.5 percent above qf(εL) insufficient; domestic agents suffer substantial welfare loss in high-leverage Sudden Stop: NSCG 1.506 percent; SCG 2.596 percent.
    - Long-run average welfare costs increase about 5 (NSCG) and 2 (SCG) times relative to a =0.2 baseline.
- Key sensitivity conclusions:
  - Price guarantees always reduce long-run probability of binding margin constraints.
  - State-contingent guarantees greatly reduce frequency of IFO trades relative to NSCG.
  - Per-trade cost parameter a (price elasticity of world demand) is crucial: higher a magnifies moral hazard and can reverse long-run welfare gains from guarantees.
  - Higher guaranteed prices and state-contingent design mitigate Sudden Stops but increase moral hazard; optimal design must weigh trading costs and demand elasticity.

### Policy implications and recommendations
- Ex-ante price guarantees effective when:
  - Guaranteed prices set moderately above fundamentals (baseline: 0.5 to 1 percent above qf(εL)).
  - World demand elasticity for SOE equity is sufficiently high (per-trade costs implying elasticity around 5 in calibrated cases).
- State-contingent guarantees preferred when feasible:
  - Equally effective at reducing margin-call probability.
  - Much lower long-run probability of IFO intervention (reducing persistent overvaluation and trade frequency).
- Design tradeoffs:
  - Increasing guaranteed prices weakens Sudden Stops but exacerbates international moral hazard, raising long-run average welfare costs if trading frictions are large.
  - When per-trade costs a are large (low demand elasticity), higher guarantees can improve Sudden Stop outcomes but may yield long-run welfare losses unless guarantees are optimally adjusted.
- Practical challenge for IFOs:
  - Need a quantitative asset-price model that explains Sudden Stops to set sustainable external-debt levels, credible guaranteed prices, and evaluate moral hazard vs globalization hazard tradeoffs.

### Overall conclusions
- Ex-ante price guarantees can substantially reduce Sudden Stop real effects and nearly eliminate long-run probability of margin calls under reasonable calibration (guarantees ½ to 1 percent above low-productivity fundamentals price).
- Guarantees create international moral hazard that raises foreign demand and long-run average value of foreign traders’ firms, benefiting foreign traders on average but potentially hurting them conditional on Sudden Stops.
- Welfare implications hinge critically on price elasticity of world demand for the SOE’s equity (linked to per-trade cost parameter a):
  - High elasticity (example elasticity of 5): guarantees yield domestic welfare gains at Sudden Stops and negligible long-run welfare change; foreign traders much better off on average.
  - Low elasticity (example elasticity of 1/2): guarantees may reduce Sudden Stop severity but moral hazard can dominate, producing long-run welfare losses unless guarantees are optimally adjusted.
- Recommendation: design guarantees that are state-contingent and sized conditional on credible asset-price models; adjust guarantee levels upward when trading frictions are higher, but account for increased moral hazard.

*Source: _wp0673 - 3. Sensitivity Analysis (IMF working paper content provided).*

### References..............................................................................................................

### _wp0673 - References................................................................................................................................39

### Figures
- 1. Equilbrium in the Asset Market ...........................................................................................34
- 2. Ergodic Distributions of Domestic Equity and Bond Holdings...........................................35
- 3. Consumption and Current Account-GDP Ratio Impact Effects of a Negative 
         Productivity Shock in the Sudden Stop Region of Equity & Bonds...............................36
- 4. Conditional Responses to a Negative, One-Standard-Deviation Productivity Shock .........37
- 5. Equity Pricing Function in the Low Productivity State .......................................................38

### Tables
- 1. Long Run Business Cycle Moments....................................................................................19
- 2. Payoffs of Domestic Agents and Foreign Traders in Baseline Simulations ........................24

*_wp0673 - References.............................................................................................................._*

### 3. Sensitivity Analysis ................................................................................................

### 3. Sensitivity Analysis

### Introduction: tradeoff framed
- Sudden Stop phenomenon characterized by: sudden reversals of capital inflows and current account deficits; collapses in output and private absorption; and large relative price corrections in domestic goods prices and asset prices.
- Two competing hypotheses:
  - Globalization hazard: world capital markets are imperfect and prone to contagion/overreaction; an IFO offering ex-ante price guarantees (Calvo (2002) proposal) could prevent Sudden Stops by credibly supporting asset prices.
  - International moral hazard: ex-ante guarantees raise foreign investors’ demand for emerging-market assets (downside risk transferred to the IFO), possibly inducing excessive indebtedness and worsening Sudden Stops (Lerrick and Meltzer, 2003; Meltzer Commission view).
- Paper objective: study the globalization hazard–moral hazard tradeoff quantitatively in a dynamic, stochastic general equilibrium (DSGE) model based on Mendoza and Smith (2006) augmented with an IFO offering ex-ante asset-price guarantees.

### Model setup and frictions
- Agents and markets:
  - Small open economy (SOE) with representative household and representative firm producing tradable output exp(εt)F(k,n).
  - Households trade equity (α) and one-period foreign bonds (b). Equity price qt.
  - Representative foreign securities firm specializes in trading SOE equity.
  - An IFO guarantees a minimum liquidation price (put-like guarantee) for foreign traders; it finances guarantees with a lump-sum tax on foreign investors’ profits.
- Key financial frictions (from Mendoza and Smith (2006)):
  - Margin constraint on foreign borrowing: bt+1 ≥ -καt+1 q t k, with 0≤κ≤1 and short-selling constraint αt+1 ≥ χ, -∞<χ<1.
  - Trading costs for foreign securities firm: quadratic trading costs (parameter a) and recurrent trading cost θ.
- Preferences and production:
  - Utility with endogenous time preference (impatience effect), composite good c - h(n) (GHH preferences).
  - Cobb-Douglas production with labor share γ.
- IFO guarantee specifics:
  - Guarantee implies a guaranteed return R̃t = (dt + q̃t)/q̃t-1 where q̃t is the guaranteed price.
  - IFO sets lump-sum tax Tt* = max{0, k(q̃t - qt)} to cover executed guarantees.

### Competitive equilibrium and pricing implications
- Equilibrium defined by prices {wt, dt, qt} and allocations {ct, nt, αt, bt, Tt*} satisfying households’, foreign traders’, firms’ optimization, IFO budget and market clearing αt + αt* = 1.
- Household first-order conditions yield:
  - Marginal utility condition λt = Uc(c,n).
  - Labor condition h′(nt) = wt.
  - Euler equations for equity and bonds lead to forward solution for equity prices and an expression for excess returns.
- Key analytic insights:
  - Margin calls lower domestic asset prices directly (forced sales) and indirectly (increase covariance between marginal utility and returns, tightening discounting of dividends).
  - Foreign demand less-than-infinitely-elastic because of trading costs; fundamentals price qf ≡ Σi=0∞ E[d t+i ]/(R−1) (discounted at world interest rate).
  - Foreign traders’ demand is increasing in the percent deviation (qf − qt)/qt (elasticity 1/a) and in expected present value of excess prices due to guarantees (international moral hazard effect).
  - Ex-ante guarantees can: (i) prevent Sudden Stops by shifting foreign demand up to support fundamentals; or (ii) create persistent overvaluation if set too high.

### Quantitative analysis: solution method and calibration
- Recursive equilibrium state variables: α, b, and productivity ε. State grids of NA nodes Α={α1<...<αNA} with α1=χ and NB nodes Β={b1<...<bNB}. Productivity realizations E={εL<εH}.
- Solution algorithm: modified Mendoza–Smith quasi-planning algorithm with a conjectured G(α,b,ε) returning expected present discounted value of “excess prices”; iterate Bellman problem and foreign demand implied pricing to convergence.
- Note on method: algorithm assumes emerging economy internalizes foreign demand function; implied proportional taxes/subsidies are negligible in simulations:
  - Maximum taxes in absolute value range between 0.08 (0.4) and 0.2 (0.8) percent when a =0.2 (2).
  - Average tax in absolute value is 0.03 (0.3) percent in simulations with a =0.2 (2).
- Functional forms and calibration (deterministic steady state):
  - Production: Cobb-Douglas, F(k,n) with γ the labor income share.
  - Period utility: u(c - h(n)) with CRRA parameter σ.
  - Time-preference function parameterized by β (elasticity of rate of time preference w.r.t. 1+c-h(n)).
  - Parameter values:
    - σ =2.
    - γ =0.65.
    - Real interest rate R set to 6.5 percent annually, implying R=1.0651/4.
    - Labor disutility coefficient δ =2.
    - Mexican average GDP shares used: sc =0.684, si =0.19, sg =0.092, snx =0.034.
    - Combined investment and government purchases treated as exogenous absorption 0.282 (28.2 percent of steady-state GDP).
  - Resulting deterministic steady-state values:
    - Depreciation rate (annual) = 7.75 percent.
    - Capital-output ratio sk ≈ 2.5.
    - Fundamentals (steady-state) equity price qf = 2.19.
    - Steady-state capital stock k = 79.
    - Time-preference elasticity solved from steady-state Euler condition yields β =0.0118.
  - Portfolio indeterminacy: any (α,b) consistent with α qf k + b = s and non-binding constraints; debt portfolios require α > 0.9; low κ required to make margin constraint bind given these steady-state portfolios.

### Key quantitative findings on guarantees and sensitivity
- Guarantee levels and outcomes:
  - Guarantees set slightly above the model’s fundamentals price (by ½ to 1 percent) contain Sudden Stop effects and virtually eliminate the probability of margin calls in the stochastic steady state.
  - A non-state-contingent guarantee at that level is executed often, with a long-run probability of about 1/3, and predicts persistent overvaluation of asset prices above frictionless prices.
  - A guarantee set at the same level but offered only when external debt is high is executed much less often (long-run probability below 1/100) and is equally effective at containing Sudden Stops without inducing persistent overvaluation.
- Welfare and distributional implications (normative analysis):
  - When elasticity of foreign demand for domestic assets is high:
    - Guarantees improve domestic welfare measured from initial conditions at a Sudden Stop state.
    - Long-run welfare changes are negligible.
    - Value of foreign traders’ firms falls slightly in a Sudden Stop state but their long-run average rises sharply.
    - Net balance favors using price guarantees to contain globalization hazard.
  - When elasticity of foreign demand is low:
    - Higher guarantees are needed to prevent Sudden Stops.
    - Large moral hazard distortions reduce domestic welfare gains at Sudden Stop states and produce average welfare losses in the stochastic steady state.
    - Price guarantees can produce short-term macro/welfare improvements at the expense of long-term welfare loss.

### Scenarios and mechanism summaries
- Partial-equilibrium intuition (Figure 1 narrative):
  - HH = domestic agents’ demand curve; FF = foreign traders’ demand curve.
  - Margin call shifts HH → HH′; without guarantees price falls to B (Sudden Stop).
  - IFO guarantee set above B shifts FF → FF′; price moves to C = fundamentals if IFO sets guarantee at correct level.
  - If guarantee < price at B: no effect. If guarantee too high: overpricing exceeding initial underpricing.
- Mechanisms:
  - Direct effect of margin calls: fire-sale equity lowers qt and raises expected excess return.
  - Indirect effect: binding borrowing limit increases covariance between marginal utility and returns, increasing discounting of future dividends.
  - Trading costs make foreign demand finite-elastic; guarantees shift foreign demand outward (international moral hazard).

### Implementation-relevant sensitivity points
- Guarantee design matters:
  - Small state-contingent guarantees targeted to high-debt episodes can preserve the Sudden Stop–preventing benefits while avoiding frequent execution and persistent overvaluation.
  - Non-state-contingent guarantees, even if modestly above fundamentals, may be executed frequently (≈1/3 long-run probability) and generate persistent overvaluation.
- Parameter sensitivity:
  - Elasticity of foreign demand (related to trading cost parameter a) is central: higher elasticity (lower a) amplifies benefits of guarantees for domestic welfare in crisis states; lower elasticity increases moral hazard costs and long-run welfare losses.
  - The margin parameter κ governs how readily margin constraints bind; low κ makes margin calls more likely given calibrated steady-state portfolios.

*Source: _wp0673 - 3. Sensitivity Analysis (IMF working paper content provided).*

### 0.10 for the margin constraint to bind for at least some of the multiple steady-state pairs of

### _wp0673 - 0.10 for the margin constraint to bind for at least some of the multiple steady-state pairs of

### Calibration of financial frictions and the Nearly Frictionless Economy (NFE)
- Objective: choose financial frictions (a, θ, κ) so deterministic RBC-SOE allocations and prices are closely approximated by a deterministic steady state with negligible recurring trading costs and a margin constraint that is just slightly binding (the “nearly frictionless economy” (NFE)).
- Rationale: low κ values represent the fraction of domestic capital useful as collateral for external debt; several studies suggest this fraction is small.
- Calibrated parameter values:
  - a = 0.2
  - θ = 0.001
  - κ = 0.03
- NFE steady-state outcomes with previously set γ, δ, β, and R:
  - α = 0.931
  - b = -4.825 (implies a debt-GDP ratio of about 0.62)
  - Foreign traders hold stationary equity position at price q = qf /(1+aθ), yielding Rq > R.
- Key analytic relation:
  - Ratio of Lagrange multipliers: η/λ = (Rq - R)/(Rq - Rκ).
  - With margin constraint binding, b = -κ α qk and α solves the displayed steady-state expression (equation (24) in source).

### Determinacy, portfolio allocation issues, and role of frictions
- Without credit constraints and with zero recurrent trading costs, stochastic RBC-SOE can lead to degenerate long-run distributions where domestic agents hold the smallest equity position (χ) and use bonds for consumption smoothing/precautionary saving because risk-averse domestic agents demand equity risk premium while risk-neutral foreign traders do not.
- Recurrent trading cost (θ>0) and binding margin constraint are critical to obtain unique, stationary portfolio allocations and to allow foreign traders to require an equity premium.
- With θ = 0 (and no price guarantees), a stationary equity position for foreign traders requires q = qf and return on equity equal to R, which implies η/λ = 0, preventing the borrowing constraint from binding.

### Stochastic simulation framework and grids
- Discrete state space:
  - Equity grid: 78 evenly-spaced nodes spanning [χ, 1] with χ = 0.84 (equity lower bound).
  - Bonds grid: 120 evenly-spaced nodes spanning [-5.2, 25.7].
- Bounds rationale:
  - Lower bound for bonds: -κ qmax k = -5.2, largest debt leverage holding largest equity at highest price.
  - Upper bound found by iterating until the grid captures support of ergodic distribution of bonds.
- Interpretation: resulting bond segment is relatively small; despite trading frictions domestic agents prefer riskless bonds for consumption smoothing and buffer-stock savings.
- Justification for χ = 0.84:
  - Mexico 1988–2000 average ratio stock market capitalization to GDP = 27.6 percent.
  - Calibration produced capital-output ratio ≈ 2.5, so publicly traded firms ≈ 11 percent of capital stock.
  - Much of physical capital lacks a liquid international market; calibration shows bond positions become positive and unrealistically large for α < 0.9, supporting high χ.

### Productivity shock process
- Modeled as a two-point, symmetric Markov process following the “simple persistence” rule.
- Targeted statistics from Mendoza (2006) for quarterly cyclical components of Mexico’s GDP:
  - Standard deviation = 2.64 percent
  - First-order autocorrelation = 0.683
- Resulting Markov process parameters:
  - σε = 1.79 percent
  - ρε = 0.683
  - Two points: εL = -εH = 0.0179 (source shows the two points as ±0.0179)
  - Long-run probability of each state = ½
  - Transition probabilities:
    - Staying in either state = ½(1-ρε)+ρε = 0.8415
    - Shifting across states = ½(1-ρε) = 0.1585

### Baseline experiments and key comparative setup
- Four simulations:
  1. NFE (nearly frictionless economy)
  2. BMR (binding margin requirements) with κ = 0.005
  3. NSCG (non-state-contingent guarantee): single guaranteed price set for all dates and states
  4. SCG (state-contingent guarantee): same guaranteed price but applied only in subset of state space (Sudden Stop region)
- Sudden Stop region definition: high-debt region where margin constraint binds (first 25 nodes of B grid and all 72 nodes of A grid in their discretization).

### Main findings on globalization hazard and Sudden Stops (BMR vs NFE)
- Key result: financial frictions representing globalization hazard cause Sudden Stops when debt-to-market-value-of-equity ratio is high and equity market has liquidity (α > χ).
- Distributional shifts:
  - Bonds distribution shifts markedly to the right under BMR relative to NFE.
  - Equity distribution shifts further to the left under BMR due to equity fire sales triggered by margin calls.
- Precautionary saving effect: domestic agents build buffer stock of savings, lowering long-run probability of very large debt positions under BMR.
- Long-run probability of binding margin constraints in BMR: about 4 percent (Table 1 reports probability = 3.973%).
- Portfolio reallocation: long-run average bonds-output ratio:
  - NFE: 18 percent
  - BMR: 50 percent
- Financial frictions have negligible effects on standard business cycle moments (Table 1).

### Sudden Stop impact mechanics and conditional dynamics
- Two factors drive impact effects in Sudden Stop region:
  1. Leverage ratio = debt / market value of equity
  2. Equity market liquidity = α - χ
- Impact patterns:
  - High leverage + some asset-market liquidity → large drops in consumption and current account and fall in q (asset price).
  - High leverage + illiquid market (domestic agents near χ) → negligible asset price declines but still Sudden Stop in consumption/current account.
  - Low leverage + high liquidity → small consumption and current account drops but larger asset price declines (domestic agents sell equity to avoid consumption collapse).
- Example historic comparisons cited:
  - Mexico Q1 1995: current account reversal = 5.2 percent of GDP; drop in real equity prices ≈ 29 percent.
  - Korea Q1 1998: current account reversal twice as large; asset price drop ≈ 10 percent.
- Conditional forecasting (non-linear impulse responses) setup:
  - High leverage initial state: α = 0.938 and b = -4.68 (debt ratio 60 percent of GDP, leverage ratio 3 percent of GDP).
  - Low leverage initial state: same α but b = -3.38 (debt ratio 43 percent of GDP, leverage ratio 2 percent of GDP).
- Impact magnitudes (BMR vs NFE) in high leverage state, on impact from a one-standard-deviation negative productivity shock:
  - c drops by 1.5 percent more than in the NFE
  - q drops by 0.4 percent more than in the NFE
  - ca/y rises by about 1 percentage point of GDP
- In low leverage state, c and ca/y responses in BMR nearly identical to NFE though q still falls by about same amount as in high leverage scenario.
- Persistence: Sudden Stops are short-lived in this calibration; BMR responses converge to NFE in about 4 quarters.

### Price guarantees: non-state-contingent (NSCG) and state-contingent (SCG)
- Guarantee specification:
  - Non-state-contingent guarantee set ½ of a percentage point above the fundamentals price in the low productivity state (fundamentals price in low state = 2.185), hence guaranteed price = 2.196.
  - NSCG applies in all states; SCG applies only for (α,b) pairs inside the Sudden Stop region.
- Necessary condition for guarantees to be executed in some states (theoretical result shown in text): (1/(1+aθ)) qfL > qL a θ + ... (source presents the formal relation; guarantees set above fundamentals price in low productivity state).
- NSCG effects:
  - Reduces long-run probability of binding margin constraints to almost zero (Table 1: NSCG probability = 0.001%).
  - Long-run average of expected present value of excess prices (G) = 0.01 (about ½ of a percent above mean equity price in stochastic stationary state).
  - Asset-price variability, persistence and comovement with output decline slightly.
  - Mean equity price increases by 0.053 percent.
  - Consumption coefficient of variation falls by about 1/5 of a percentage point.
  - Equity price at date 0 in NSCG is about 2/3 of a percentage point higher than in BMR in both high and low leverage states.
  - In high leverage BMR vs NSCG: initial c and ca/y reversals smaller in NSCG; drop in equity prices turns into an increase of about ¼ of a percentage point.
  - Guarantees are executed in both high- and low-leverage states under NSCG.
  - Drawback: NSCG increases asset prices even in states far from Sudden Stop vulnerability, causing distortions in asset prices outside Sudden Stop region.
- SCG effects:
  - Produces long-run distribution of equity/bonds less skewed than NSCG.
  - Changes in long-run business cycle moments qualitatively similar to NSCG but smaller in magnitude.
  - Near-zero long-run probability of binding margin constraints (Table 1: SCG probability = 0.001%).
  - Date-0 consumption and current account responses in SCG are nearly identical to NSCG; recovery in asset prices slightly smaller.
  - SCG converges faster to NFE dynamics after date 0.
  - SCG yields higher asset prices mainly within Sudden Stop region, inducing smaller distortions outside Sudden Stop region relative to NSCG.

### Quantitative long-run business cycle moments (selected figures from Table 1)
- Panel I. NFE Economy:
  - GDP: mean = 7.833, std. dev. = 2.644, std. dev. (in %) = 1.000, correlation w/ GDP = 1.000, first-order auto-correlation = 0.683
  - Consumption: mean = 5.366, std. dev. = 2.185, std. dev. (in %) = 0.826, correlation w/ GDP = 0.853, first-order auto-correlation = 0.770
  - current account-GDP ratio: mean = 0.000, std. dev. = 1.347, std. dev. relative to GDP = 0.509, correlation w/ GDP = 0.979, first-order auto-correlation = 0.660
  - equity price: mean = 2.187, std. dev. = 0.121, std. dev. relative to GDP = 0.046, correlation w/ GDP = 0.961, first-order auto-correlation = 0.606
  - foreign debt-GDP ratio: mean = 0.177, std. dev. = 56.694, std. dev. relative to GDP = 21.442, correlation w/ GDP = -0.076, first-order auto-correlation = 0.997
  - debt-equity ratio: mean = 0.010, std. dev. = 2.888, std. dev. relative to GDP = 1.092, correlation w/ GDP = 0.000, first-order auto-correlation = 0.001
- Panel II. BMR Economy (probability of binding margin constraints = 3.973%):
  - foreign debt-GDP ratio: mean = 0.499, std. dev. = 37.964, std. dev. relative to GDP = 14.358
  - debt-equity ratio: mean = 0.026, std. dev. = 2.007, std. dev. relative to GDP = 0.759
- Panel III. NSCG Economy (probability of binding margin constraints = 0.001%):
  - equity price: mean = 2.198, std. dev. = 0.099
  - foreign debt-GDP ratio: mean = 1.336, std. dev. = 41.053
  - debt-equity ratio: mean = 0.071, std. dev. = 2.186
- Panel IV. SCG Economy (probability of binding margin constraints = 0.001%):
  - equity price: mean = 2.188, std. dev. = 0.121
  - foreign debt-GDP ratio: mean = 1.114, std. dev. = 34.141
  - debt-equity ratio: mean = 0.059, std. dev. = 1.810

### Normative implications and welfare analysis
- Welfare metric: compensating variation in date-0 consumption W(α,b,ε) that equates expected lifetime utility in BMR/NSCG/SCG with NFE for each (α,b,ε).
- Mean welfare costs (ergodic distribution) of deviating from NFE are small:
  - Average welfare losses in BMR, NSCG and SCG relative to NFE < 0.07 percent in c0.
- Welfare in Sudden Stop (high-leverage) state:
  - Short-run welfare costs (percent cut in c0 to equate u(0) with NFE):
    - BMR: 1.4 percent
    - NSCG: 0.9 percent
    - SCG: 1.0 percent
  - Long-run welfare effects (percent cut in c0 equating exp(-v(0))E0[V(1)] with NFE) are negative (i.e., long-run welfare gains) because consumption after date 0 increases temporarily.
    - High-leverage long-run welfare gains:
      - BMR: about 1 percent
      - NSCG: 1.9 percent
      - SCG: 1.3 percent
  - Total welfare (short-run + long-run):
    - NSCG: total welfare gain of 1 percent (domestic agents obtain total welfare gains)
    - SCG: (implied positive total welfare but smaller than NSCG)
- Interpretation:
  - Short-run costs reflect immediate consumption drops at date 0 in Sudden Stop states; price guarantees reduce these short-run costs relative to BMR.
  - Long-run gains reflect consumption tilting and, in price-guarantee economies, higher asset prices that finance larger short-run consumption increases (and larger early external deficits financed by equity sales).

*Source: _wp0673 - 0.10 for the margin constraint to bind for at least some of the multiple steady-state pairs of (α,b).*

### 0.3 percent in the SCG.

### _wp0673 - 0.3 percent in the SCG.

### Key quantitative findings (Table 2 — long-run averages and Sudden Stop states)
- Long-run averages (Domestic agents / Foreign traders):
  - Domestic agents: Welfare cost (long-run average, compensating variation) — NFE: 0.017; BMR: 0.062; NSCG: 0.057.
  - Foreign traders: Present value of traders' returns (long-run average) — NFE: 17.920; BMR: 20.519; NSCG: 27.350; SCG: 25.529.
    - Percent change w.r.t. NFE (long-run average) — BMR: 14.499; NSCG: 52.621; SCG: 42.460.
  - Foreign traders: Returns (long-run average) — NFE: 0.280; BMR: 0.321; NSCG: 0.427; SCG: 0.399.
    - Dividend earnings (a) — NFE: 0.280; BMR: 0.320; NSCG: 0.427; SCG: 0.399.
    - Trading costs (b) — NFE: 1.1E-05; BMR: 2.3E-05; NSCG: 8.8E-05; SCG: 1.6E-04.

- High leverage Sudden Stop state (date-0 and present values):
  - Domestic agents: Welfare cost (compensating variation) — NFE: 0.367; BMR: -1.036; NSCG: -0.317.
    - Short-run cost (date-0 compensating variation) — NFE: 1.376; BMR: 0.862; NSCG: 0.978.
    - Long-run cost (difference) — NFE: -1.010; BMR: -1.899; NSCG: -1.295.
    - Date-1 rate of time preference — NFE: 1.58; BMR: 1.56; NSCG: 1.57; SCG: 1.57.
  - Foreign traders:
    - Present value of traders' returns (conditional high leverage) — NFE: 10.931; BMR: 10.940; NSCG: 10.858; SCG: 10.901.
      - Percent change w.r.t. NFE — BMR: 0.082; NSCG: -0.670; SCG: -0.276.
    - Returns at date 0 — NFE: -0.192; BMR: -3.770; NSCG: -3.796; SCG: -3.791.
      - Dividend earnings (a) — all four economies: 0.166.
      - Value of trades (b) — NFE: 0.359; BMR: 3.927; NSCG: 3.953; SCG: 3.948.
      - Trading costs (c) — NFE: 1.6E-04; BMR: 9.8E-03; NSCG: 9.8E-03; SCG: 9.8E-03.

- Low leverage Sudden Stop state:
  - Domestic agents: Welfare cost (compensating variation) — NFE: 0.215; BMR: -1.194; NSCG: -0.393.
    - Short-run cost — NFE: 0.055; BMR: -0.356; NSCG: -0.293.
    - Long-run cost — NFE: 0.160; BMR: -0.838; NSCG: -0.100.
    - Date-1 rate of time preference — 1.59 in all economies.
  - Foreign traders:
    - Present value of traders' returns — NFE: 10.931; BMR: 10.937; NSCG: 10.854; SCG: 10.901.
      - Percent change w.r.t. NFE — BMR: 0.055; NSCG: -0.707; SCG: -0.272.
    - Returns at date 0 — NFE: -0.192; BMR: -3.056; NSCG: -3.077; SCG: -3.074.
      - Dividend earnings (a) — all four economies: 0.166.
      - Value of trades (b) — NFE: 0.359; BMR: 3.215; NSCG: 3.237; SCG: 3.234.
      - Trading costs (c) — all economies: 1.6E-04 for NFE; BMR/NSCG/SCG: 6.7E-03.

### Mechanisms and interpretation
- Decomposition of foreign traders’ net returns:
  - Three components: dividend earnings, net value of trades, and trading costs.
  - Lump-sum taxes paid by foreign traders cancel with the value of executed guarantees (IFO budget constraint), so price guarantees affect traders’ first-order condition via the moral hazard effect without direct income effects.
- Long-run vs state-conditional effects:
  - Long-run average E[D] and Eπ̂ are higher in BMR, NSCG, SCG relative to NFE: BMR +14.5 percent; NSCG +52.6 percent; SCG +42.5 percent (Table 2 narrative).
  - The long-run increase in foreign traders’ firm value arises mainly from increased average equity holdings and higher dividend earnings, not from changes in value of trades (unconditional mean of α_t equals α_t+1) or trading costs (negligible).
  - Conditional on Sudden Stop states, foreign traders’ payoffs fall in economies with price guarantees: present value of profits is nearly unchanged in BMR but falls by about 0.7 and 0.3 percent in NSCG and SCG respectively (high- and low-leverage conditional results).
  - Date-0 decline in net returns during Sudden Stops is driven almost entirely by increases in the value of trades (fire sales): value of trades rises sharply from 0.36 in NFE to around 3.9 in BMR/NSCG/SCG in high-leverage states.
  - Price guarantees reduce downside risk and raise long-run average returns for foreign traders, but increase moral hazard that raises equilibrium prices and foreign equity demand.

### Distributional and accounting implications
- Resource constraint and long-run averages:
  - Long-run average consumption is nearly identical across NFE, BMR, NSCG, SCG.
  - Increased long-run average foreign traders’ dividend earnings are offset by domestic agents’ greater bond holdings (higher interest income), so GNP is unaffected while foreign traders capture a larger share of domestic GDP.
  - SOE assumption and independence of GDP from financial frictions are crucial for these results.
- Short-run redistribution at Sudden Stops:
  - GNP and GDP unchanged across simulations, but current account reallocates income between foreign traders and domestic households via equity market transactions.
  - In high-leverage Sudden Stop states, redistribution from foreign traders to domestic agents via the equity market is often more than offset by loss of credit market access (larger current account reversal), reducing domestic share of world output.
  - Effects are stronger in economies with price guarantees.

### Sensitivity analysis highlights (Table 3)
- Parameter changes analyzed: larger and more persistent productivity shocks (σ_ε =0.024 and ρ_ε =0.8), higher guaranteed price (1 percent above q_f(ε_L)), higher recurrent trading costs (θ =0.01), higher per-trade costs (a =2). Column (I) is baseline.
- BMR panel (robustness):
  - Probability of binding margin constraint changes with shocks: rises to 4.2 percent when ρ_ε increases; falls to 2.3 percent when σ_ε increases.
  - Long-run average value of foreign traders’ firms falls to 12.2 and 4.4 percent in Columns (II) and (III) respectively (relative changes reported in Table 3).
  - Higher θ or a produce larger Sudden Stops: with θ =0.01 probability of margin calls reaches 16.740 percent; with a =2 it reaches 19.444 percent.
  - BMR with a =2 can account for large consumption and current account reversals and large drops in asset prices; also increases long-run variability and domestic welfare costs.
- NSCG and SCG panels (guarantees and trading costs):
  - Price guarantees virtually eliminate long-run probability of binding margin constraints in all panels.
  - Long-run probability of executing guarantees:
    - NSCG: ranges 29–34 percent (falls to 15 percent when σ_ε higher).
    - SCG: below 1 percent in baseline scenarios; rises to 6.063 and 3.245 percent with θ =0.01 and a =2 respectively.
  - Raising guaranteed price to 1 percent above q_f(ε_L) (Column IV) further dampens Sudden Stops:
    - Fall in consumption c in high-leverage Sudden Stop: NSCG -0.5 percent; SCG -0.6 percent (vs baseline NSCG -0.8 percent; SCG -0.9 percent).
    - Reversal in ca/y: NSCG 0.3 percentage points; SCG 0.4 percentage points (vs baseline 0.6 and 0.7).
    - Equity price increases in Sudden Stop: NSCG 0.7 percent; SCG 0.6 percent.
    - Domestic welfare gain in high-leverage Sudden Stop increases to 2.4 percent (NSCG Column IV) and 0.9 percent (SCG Column IV).
    - Decline in foreign traders’ payoff in Sudden Stop grows to 1.4 percent (NSCG) and 0.9 percent (SCG) with higher guarantees; long-run average of traders’ payoff still increases.
  - High per-trade costs (a =2, Column VI):
    - Baseline guarantee 0.5 percent above q_f(ε_L) insufficient to prevent marked reversals; domestic agents suffer substantial welfare loss in high-leverage Sudden Stop: NSCG 1.506 percent; SCG 2.596 percent.
    - Long-run average welfare costs increase about 5 (NSCG) and 2 (SCG) times relative to a=0.2 baseline.
    - In SCG with a =2, foreign traders’ firm value in high-leverage state can exceed NFE counterpart.
- Key sensitivity conclusions:
  - Price guarantees always reduce long-run probability of binding margin constraints.
  - State-contingent guarantees (SCG) greatly reduce frequency of IFO trades relative to non-state-contingent guarantees (NSCG).
  - The per-trade cost parameter a (and hence the implied price elasticity of world demand) is crucial: higher a magnifies moral hazard and can reverse long-run welfare gains from guarantees.
  - Higher guaranteed prices and more state-contingent design mitigate Sudden Stops but increase moral hazard distortions; optimal guarantee design should consider trading costs and demand elasticity.

### Policy implications and recommendations (derived from model results)
- Ex-ante price guarantees can be effective to contain Sudden Stops when:
  - Guaranteed prices are set moderately above fundamentals (baseline: 0.5 to 1 percent above q_f(ε_L) in low productivity state).
  - World demand elasticity for the SOE’s equity is sufficiently high (per-trade costs implying elasticity around 5 in calibrated cases).
- State-contingent guarantees (applying only at high debt levels) are preferable to non-state-contingent guarantees when feasible:
  - Equally effective at reducing margin-call probability.
  - Result in much lower long-run probability of IFO intervention (reducing persistent asset overvaluation and frequency of trades).
- Design tradeoffs:
  - Increasing guaranteed prices weakens Sudden Stops but exacerbates international moral hazard, raising long-run average welfare costs if trading frictions are large.
  - When per-trade costs a are large (low demand elasticity), higher guarantees can improve Sudden Stop outcomes but may yield long-run welfare losses; raising guarantees further or adjusting state-contingency can offset this.
- Practical challenge for IFOs:
  - Need a "useful" quantitative asset-price model (able to explain Sudden Stops) to set sustainable external-debt levels and credible guaranteed prices and to evaluate moral hazard vs globalization hazard tradeoffs.

### Overall conclusions
- Ex-ante price guarantees can substantially reduce Sudden Stop real effects and nearly eliminate long-run probability of margin calls under reasonable calibration (guarantees ½ to 1 percent above low-productivity fundamentals price).
- Guarantees create international moral hazard that raises foreign demand and long-run average value of foreign traders’ firms, benefiting foreign traders on average but potentially hurting them conditional on Sudden Stops.
- Welfare implications depend critically on the price elasticity of world demand for the SOE’s equity (linked to per-trade cost parameter a):
  - High elasticity (example elasticity of 5): guarantees yield domestic welfare gains at Sudden Stops and negligible long-run welfare change; foreign traders are much better off on average.
  - Low elasticity (example elasticity of 1/2): guarantees may reduce Sudden Stop severity but international moral hazard can dominate, producing long-run welfare losses for the domestic economy unless guarantees are optimally adjusted.
- Recommendation: design guarantees that are state-contingent and sized conditional on credible asset-price models; adjust guarantee levels upward when trading frictions (per-trade costs) are higher, but account for increased moral hazard.

*Source: _wp0673 - 0.3 percent in the SCG. https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0673.pdf*

### REFERENCES

### REFERENCES

### Cited works
- Arellano, C., and E.G. Mendoza, 2003, “Credit Frictions and ‘Sudden Stops’ in Small Open Economies: An Equilibrium Business Cycle Framework for Emerging Markets Crises,” in Dynamic Macroeconomic Analysis,” ed. by S. Altug, J. Chadha and C. Nolan (Cambridge, United Kingdom: Cambridge University Press).
- Aiyagari, R.S., and M. Gertler, 1999, “Overreaction of Asset Prices in General Equilibrium,” Review of Economic Dynamics, Vol. 2, pp. 3–35.
- Caballero, R.J., and A. Krishnamurthy, 2001, “International and Domestic Collateral Constraints in a Model of Emerging Market Crises,” Journal of Monetary Economics, Vol. 48, pp. 513–48.
- Calvo, G.A., 1998, “Capital Flows and Capital-Market Crises: The Simple Economics of Sudden Stops,” Journal of Applied Economics, Vol. 1, pp. 35–54.
- Calvo, G.A., 2002, “Globalization Hazard and Delayed Reform in Emerging Markets,” Economia, Vol. 2 (Spring), pp. 1–29.
- Calvo, G.A., and E.G. Mendoza, 2000a, “Capital-Market Crises and Economics Collapse in Emerging Markets: An Informational-Frictions Approach,” American Economic Review: Papers & Proceedings, Vol. 90 (May), pp. 59–64.
- Calvo, G.A., and E.G. Mendoza, 2000b, “Rational Contagion and the Globalization of Securities Markets,” Journal of International Economics, Vol. 51, pp. 79–113.
- Fisher, I., 1933, “The Debt-Deflation Theory of Great Depressions,” Econometrica, Vol. 1, pp. 337–57.
- Greenwood, J., Z. Hercowitz, and G.W. Huffman, 1988, “Investment, Capacity Utilization and the Real Business Cycle,” American Economic Review, Vol. 78 (June), pp. 402–17.
- Izquierdo, A., 2000, “Credit Constraints, and the Asymmetric Behavior of Asset Prices and Output under External Shocks” (unpublished; Washington: World Bank).
- Kiyotaki, N., and J. Moore, 1997, “Credit Cycles,” Journal of Political Economy, Vol. 105, pp. 211–48.
- Krugman, P., 2000, “Fire-Sale FDI,” in Capital Flows and the Emerging Economies, ed. by Sebastian Edwards (Chicago: University of Chicago Press).
- Lerrick, A., and A.H. Meltzer, 2003, “Blueprint for an International Lender of Last Resort,” Journal of Monetary Economics, Vol. 50, pp. 289–303.
- Ljungqvist, L., 2000 “Government Guarantees on Assets and Volatility” (unpublished; Stockholm: School of Economics).
- Mendoza, E.G., 1991, “Capital Controls and the Dynamic Gains from Trade in a Business Cycle Model of a Small Open Economy,” Staff Papers, International Monetary Fund, Vol. 38 (September), pp. 480–505.
- Mendoza, E.G., 2006 “ ‘Sudden Stops’ in an Equilibrium Business Cycle Model with Credit Constraints: A Fisherian Deflation of Tobin’s Q,” (unpublished; College Park: Department of Economics, University of Maryland).
- Mendoza, E.G., and K.A. Smith, 2006, “Quantitative Implications of a Debt-Deflation Theory of Sudden Stops and Asset Prices,” Journal of International Economics, forthcoming.

*Source: _wp0673 - REFERENCES*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0673.pdf_