## _wp07146

## Source details

**Canonical URL:** [_wp07146](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07146.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07146.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07146.pdf.json)

---

### I. Introduction — context and objectives
- Financial globalization in emerging economies began in the late 1980s (Edwards (2005); Chinn and Ito (2005)).  
- Waves of Sudden Stops began with the Mexican crisis of 1994–95; Table 1 lists 18 Sudden Stop episodes between 1994 and 2002.  
- Median increase in reserves in Sudden Stop countries was 7.7 percent of GDP.  
- Median increase in Asian Sudden Stop countries exceeded 13 percent of GDP.  
- Paper objective: quantitative assessment of the New Merchantilism via a dynamic stochastic general equilibrium framework for optimal precautionary demand for foreign assets in a small open economy with incomplete asset markets. Three driving factors examined:
  - (1) changes in business cycle volatility of output,
  - (2) financial globalization (reduction of distortions affecting international asset trading),
  - (3) self-insurance against Sudden Stops.

### Model framework and strategy (two-stage analysis)
- Stage 1: canonical one-sector endowment economy with non-insurable shocks to domestic income, access to frictionless credit market at world risk-free rate r (gross R≡1+r). Model calibrated to match variability and persistence of output in Sudden Stop economies; used to compute optimal short- and long-run dynamics of foreign assets in response to changes in output volatility and financial globalization.  
- Stage 2: two-sector production economy with liability dollarization and endogenous Sudden Stops. Tradable and nontradable sectors; nontradables produced with imported intermediate goods priced in world markets. Non-state-contingent debt is denominated in tradables (liability dollarization). Collateral constraint limits debt to a fraction κ of total income in tradable units; this combined with liability dollarization creates endogenous Sudden Stops via Fisherian debt-deflation. Precautionary demand accounts for how foreign asset holdings alter probability and magnitude of Sudden Stops (equilibrium outcomes).

### Preference specifications compared
- Bewley-Aiyagari-Hugget (BAH):
  - Constant, exogenous rate of time preference 0<ρ_BAH<1 (discount factor 1/(1+ρ_BAH)).  
  - Requires rate of time preference higher than interest rate to avoid infinite asset accumulation.  
  - Mean foreign assets increase as the gap between interest rate and time preference narrows; elasticity of mean foreign assets with respect to interest rate can be very large near equality.  
- Uzawa-Epstein (UE):
  - Endogenous rate of time preference v(c)=ln(1+ρ_UE c), with ρ_UE>0 measuring elasticity of rate of time preference wrt 1+c.  
  - Under condition ρ_UE γ ≤ 1 long-run rate of time preference converges to world interest rate; mean assets less sensitive to world real interest rate; elasticity of mean foreign assets approximately constant at high or low interest rates.

### Main quantitative findings — summary
- Financial globalization, even without Sudden Stops, can trigger significant increases in mean foreign asset holdings.  
- Risk of Sudden Stops can produce significant increases in foreign assets even when long-run variability of output is unaffected by Sudden Stops.  
- Changes in output volatility cannot explain the observed surge in foreign reserves:
  - Models predict large increases in foreign assets in response to higher variability of income, but data show no systematic increases in standard deviation of cyclical output for Sudden Stop economies during financial globalization.
  - Tradables GDP volatility rose in some Sudden Stop countries but not by magnitudes required by the model to explain observed reserve surges.
- Buildup of foreign assets in response to financial globalization or Sudden Stop risk is slow and gradual, characterized by current account surpluses and undervalued real exchange rates; does not require active central bank real exchange rate targeting.

### One-sector model — structure and baseline calibration
- Representative agent consumes composite good c; CRRA period utility with relative risk aversion coefficient γ. Time preference: UE v(c)=ln(1+ρ_UE c) or constant ρ_BAH in BAH. Income mean y subject to Markov shocks ε_t. Foreign assets b are one-period bonds at net real interest rate r (gross R≡1+r). Lump-sum exogenous absorption A introduced to match expenditure shares.  
- Numerical solution: Bellman equation solved by value function iteration on grid n=1000; Markov process for ε from Tauchen and Hussey (1991) quadrature.  
- Baseline calibration (matching Mexican statistics):
  - γ = 2
  - y = 1
  - steady-state net foreign assets-GDP ratio b = -0.44 (Mexico average 1985-2004)
  - consumption-GDP ratio c = 0.692
  - real interest rate R = 1.059 (5.9 percent, average including EMBI spread for Mexico)
  - resource constraint implies A = 0.282
- UE time preference elasticity: ln(R)/ln(1+ c) = 0.109 ⇒ subjective discount factor 0.944.  
- BAH calibration to match Mexico’s average net foreign assets (-44 percent) and consumption cyclical standard deviation (3.28 percent): φ = -0.51 and β = 0.94 ⇒ ρ_BAH = 0.064.  
- Income process: HP-filtered cyclical GDP in Mexican data (1965-2005): σ_y = 3.301% and ρ_y = 0.597. Underlying σ_e = 2.648 percent. THQA with 5 nodes yields σ_y = 3.285% with ρ = 0.550 and σ_e = 2.64 percent.

### One-sector baseline results — key statistics and dynamics
- Baseline precautionary savings:
  - UE ≈ 2.5 percent of GDP
  - BAH ≈ 9.6 percent of GDP
- Foreign assets serial autocorrelation:
  - UE baseline 0.99
  - BAH baseline 0.959
- Transitional convergence times:
  - about 40 years (BAH) and more than 80 years (UE) from certain low-asset initial conditions.
- Sensitivity:
  - increases in ρ_y, σ_e, and γ raise precautionary savings
  - BAH always yields larger precautionary savings than UE
  - High-risk-aversion case γ = 5 yields precautionary savings of 10.4 percent (UE) and 23.8 percent (BAH) of GDP
- Volatility channel assessment:
  - increasing σ_e (σ_y from 1 to 8 percent) produces large increases in precautionary demand under both UE and BAH, but BAH yields higher levels.
  - Autocorrelation increases have weaker effects.
  - Empirical mismatch: median standard deviation of GDP in Sudden Stop countries fell from 3.5 percent (1965–1985) to 3.1 percent (1986–2005).
  - Model requires large volatility increases to explain reserve surge: UE needs >4 percentage points, BAH needs >1.5 percentage points.

### One-sector: financial globalization as a reduction in return distortions
- Financial globalization modeled as time-invariant distortionary tax τ on foreign asset returns; τ in [0, 27] percent considered. Tax revenue rebated lump-sum T_t = b_t r τ.  
- Approximation for UE using lump-sum transfer rbτ where b is long-run average accurate for τ ≤ 0.84.  
- Results:
  - cutting τ from 8 percent to 0.5 percent increases long-run average b/y sharply in both setups.
  - Examples reported:
    - BAH with natural debt limit: cut τ from 8% to 0.5% increases mean b/y from -10% to about -154% and precautionary savings from 81% to about 9.4 times GDP (extreme cases shown).
    - With φ = -1, same tax cut increases average b/y from -90% to about +20% and precautionary savings from 10% to 120% of GDP.
    - UE: cutting τ from 8% to 0.5% increases long-run average foreign assets from -156% to about -45% of GDP while precautionary savings approximately unchanged.
  - Interpretation: both BAH and UE predict large increases in foreign assets with financial globalization; UE suggests much of increase may be due to deterministic saving incentives (post-tax return effect) rather than pure precautionary motives. Whether financial globalization explains observed surge depends on initial τ and how much it fell.

### Two-sector production economy — structure and calibration
- Key structural features relative to one-sector:
  - consumption c aggregates tradables c_T and nontradables c_N with CES aggregator (elasticity of substitution 1/(1+μ), weight a).
  - nontradables produced using imported intermediate goods m with production technology parameter α (share of imported inputs).
  - liability dollarization: non-state-contingent debt denominated in tradables units.
  - collateral constraint: lenders limit credit to fraction κ of market value of total income in tradables, up to maximum Ω (12). Agents take p_N as given, introducing a credit-market externality.
- Collateral constraint binding introduces an endogenous Lagrange multiplier η_t in the Euler equation for tradables consumption, interpreted as an endogenous interest rate premium. Fisherian debt-deflation amplification mechanism described: tradables consumption falls → p_N falls → value of nontradables GDP falls → borrowing capacity tightens → further falls.
- Calibration (Mendoza (2002) for Mexico):
  - γ = 2
  - b = -0.44
  - R = 1.059
  - Steady-state p_N = 1, p_m = 1
  - μ = 0.316, a = 0.341
  - p_N y_N / y_T = 1.543 ⇒ y_T = 0.393 and (p_N y_N - p_m m) = 0.607
  - α = 0.2 ⇒ imported input ratio to GDP = 13 percent; gross output nontradables = 0.758; Z = 1.106; m = 0.152
  - A_T = -0.046; A_N = 0.328
  - UE time-preference elasticity ln(R)/ln(1( c_T, c_N )) = 0.1867 ⇒ discount factor 0.944
  - BAH: φ = -0.7, β = 0.94395 ⇒ ρ_BAH = 0.0594
- Shock identification: proxy nontradables TFP by nontradables GDP; VAR(1) estimated for tradables and nontradables cyclical components (1965-2005); THQA with 3 realizations for ε_T and z yields Markov process approximating sectoral moments; second-stage adjustment of VAR elements for matching moments.

### Two-sector baseline results — business cycle co-movements and precautionary savings
- Improvements over one-sector:
  - produces countercyclical current account and net exports (correlation of current account-output ratio with output in tradables: UE -0.14; BAH -0.48). Agents build debt during expansions.
  - produces large, procyclical, persistent fluctuations in real exchange rate (price of nontradables): coefficients of variation 6.7% (UE) and 8.1% (BAH); correlations with output in tradables 0.75 (UE) and 0.88 (BAH); first-order autocorrelation ≈ 0.5.
- Precautionary savings:
  - BAH two-sector ≈ 25.3 percent of long-run average GDP
  - UE two-sector ≈ 1.6 percent of long-run average GDP
  - UE precautionary savings lower in two-sector (2.5% in one-sector v. 1.6% in two-sector) due to implicit hedge from endogenous co-movement between tradables and nontradables incomes.

### Two-sector: effects of volatility, financial globalization, and Sudden Stops
- Volatility and financial globalization effects qualitatively mirror one-sector:
  - increases in volatility of tradables or nontradables raise long-run average foreign assets.
  - data do not show systematic volatility increases across Sudden Stop countries.
- Financial globalization (τ reductions) results:
  - both one- and two-sector models show large increases in foreign assets as τ falls.
  - UE exhibits near-linear increases; BAH exhibits sharply increasing rate as τ → 0.
  - Precautionary savings rise sharply in BAH but are nearly invariant to τ in UE.
  - Two-sector models yield smaller precautionary savings than one-sector due to implicit hedge: equilibrium correlation between tradables income and nontradables income (in tradables units) is about -0.4.
- Sudden Stop risk (collateral constraint κ = 0.5):
  - Long-run average assets:
    - BAH rises from -44.7% (without Sudden Stops) to -24.3% (with Sudden Stops).
    - UE rises from -42.4% to -37.8% when Sudden Stops possible.
  - Precautionary-savings effect:
    - UE: increase of -37.8 - (-42.4) = 4.6 percent of long-run GDP due to Sudden Stop risk.
    - BAH: long-run average assets rise 20.4 percentage points, but the debt limit itself rises by 20 percentage points; overall change in precautionary savings ≈ 0.5 percentage point relative to corresponding credit limits. Transitional build-up required if Sudden Stop risk is newly introduced would be 20.4 percentage points of GDP.
  - Sudden Stop dynamics (conditional on initial binding debt ratio b = -48.7% of long-run GDP; long-run probability of reaching this debt ratio 0.9% (BAH) and 1.1% (UE)):
    - BAH (UE) feature larger (smaller) Sudden Stop magnitudes:
      - current account reversal 3.5 (2.5) percentage points of GDP
      - fall in nontradables output about 2.5 (1.5) percent
      - real depreciation 9 (6) percent
      - decline in aggregate consumption 5 (3) percent
      - drop in output 7 (5) percent
    - Convergence of foreign assets slow:
      - BAH gap of about -4 percentage points of GDP remains after 50 years
      - UE gap closes in ~5 years then slowly returns to zero (moderate overshooting)
    - Long-run Sudden Stop region:
      - for b ≥ -0.482 constraint does not bind
      - for -0.578 < b < -0.482 constraint binds and Sudden Stops occur
      - cumulative probability of all long-run Sudden Stop states 3.88% (BAH) and 7.85% (UE)
    - Transitional recurrence probabilities (starting from b = -48.7%):
      - probability of hitting that ratio again after 2 years: 21% (UE) and 40% (BAH)
      - after 15 years: 3.4% (UE) and 4.7% (BAH)
    - Building the war-chest to minimize Sudden Stop probability is long and gradual: more than 15 years to appreciably lower recurrence probability; more than 50 years to attain minimal long-run probability.

### Conclusions — synthesis and policy-relevant implications
- Three mechanisms drive precautionary demand for foreign assets: (i) increases in cyclical output volatility; (ii) international financial integration; (iii) Sudden Stop risk. All can produce large changes in foreign assets in the model.  
- Empirical evidence: output volatility has not systematically increased in Sudden Stop countries → volatility channel unlikely to explain reserve surge. By contrast, financial globalization and Sudden Stop risk coincide with reserve build-up and can generate quantitatively comparable increases.  
- Dynamics: buildup of assets, current account surpluses, and undervalued real exchange rates are protracted and gradual; slow decline in Sudden Stop probability (over decades). BAH setup implies nonlinear response to financial globalization: small reductions in distortions near full integration can produce very large asset adjustments.  
- Two-sector implicit hedge: endogenous co-movement yields negative correlation between tradables income and nontradables income in tradable units (about -0.4), reducing precautionary savings relative to one-sector outcomes; however, financial globalization still a potent driver of reserve accumulation.  
- Caveats and normative considerations:
  - analysis abstracts from efficiency gains (capital accumulation distortions may induce investment booms after opening)
  - sovereign default considerations may alter implications (Alfaro and Kanczuk (2006))
  - welfare implications mixed: Lucas-style results suggest small average welfare cost of consumption volatility; Durdu and Mendoza (2006) show conditional welfare costs in Sudden Stop states are high; Caballero and Panageas (2006) find contingent financing arrangements can be welfare-improving under high growth aspirations
  - distributional concerns: underdeveloped domestic financial systems imply large, regressive costs of precautionary saving and financial globalization (Mendoza et al. (2007))
- Overall judgment: New Merchantilism is only partially correct. Minimizing Sudden Stop risk can lead to reserve accumulation; but the surge is not explained by rising business cycle volatility. Financial globalization itself is a strong driver of reserve increases; self-insurance behavior is part of but not the sole mechanism — the removal of barriers to global asset trading interacts with underdeveloped financial markets to produce large reserve buildups.

### Key model parameter lists (baseline entries as listed)
- One-sector parameters:
  - ρBAH: 0.064
  - ρUE: 0.109
  - γ: 2.000
  - π (ad-hoc debt limit φ): -0.510
  - R: 1.059
  - y: 1.000
  - c: 0.692
  - b: -0.440
  - σe: 0.026
  - θy: 0.597
  - A: 0.282
- Two-sector parameters:
  - ρBAH: 0.059
  - ρUE: 0.187
  - γ: 2.000
  - μ: 0.316
  - a: 0.341
  - π (ad-hoc debt limit φ): -0.700
  - α: 0.200
  - R: 1.059
  - b: -0.440
  - pN: 1.000
  - pm: 1.000
  - yT + pN yN: 1.000
  - cT/yT: 0.665
  - cN/yN: 0.710
  - pN yN / yT: 1.543
  - AT: 0.106
  - AN: 0.176

*Source: _wp07146 - References (content unit provided).*

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

### _wp07146 - References..............................................................................................................

### I. Introduction — context and objectives
- Financial globalization in emerging economies began in the late 1980s (Edwards (2005); Chinn and Ito (2005)).  
- Waves of Sudden Stops began with the Mexican crisis of 1994–95; Table 1 lists 18 Sudden Stop episodes between 1994 and 2002.  
- Median increase in reserves in Sudden Stop countries was 7.7 percent of GDP (cross-country median of differences between each country’s average reserves-to-GDP ratio from the year of the country’s Sudden Stop to 2004 and the average from 1985 to the year of the Sudden Stop).  
- Median increase in Asian Sudden Stop countries exceeded 13 percent of GDP.  
- Paper objective: quantitative assessment of the New Merchantilism via a dynamic stochastic general equilibrium framework for optimal precautionary demand for foreign assets in a small open economy with incomplete asset markets. Three driving factors examined: (1) changes in business cycle volatility of output, (2) financial globalization (reduction of distortions affecting international asset trading), (3) self-insurance against Sudden Stops.

### Model framework and strategy (two-stage analysis)
- Stage 1: canonical one-sector endowment economy with non-insurable shocks to domestic income, access to frictionless credit market at world risk-free rate r (gross R≡1+r). Model calibrated to match variability and persistence of output in Sudden Stop economies; used to compute optimal short- and long-run dynamics of foreign assets in response to changes in output volatility and financial globalization.  
- Stage 2: two-sector production economy with liability dollarization and endogenous Sudden Stops. Tradable and nontradable sectors; nontradables produced with imported intermediate goods priced in world markets. Non-state-contingent debt is denominated in tradables (liability dollarization). Collateral constraint limits debt to a fraction κ of total income in tradable units; this combined with liability dollarization creates endogenous Sudden Stops via Fisherian debt-deflation. Precautionary demand accounts for how foreign asset holdings alter probability and magnitude of Sudden Stops (equilibrium outcomes).

### Two preference specifications compared
- Bewley-Aiyagari-Hugget (BAH): constant, exogenous rate of time preference 0<ρ_BAH<1 (discount factor 1/(1+ρ_BAH)). Requires rate of time preference higher than interest rate to avoid infinite asset accumulation; mean foreign assets increase as the gap between interest rate and time preference narrows; elasticity of mean foreign assets with respect to interest rate can be very large near equality.  
- Uzawa-Epstein (UE): endogenous rate of time preference v(c)=ln(1+ρ_UE c), with ρ_UE>0 measuring elasticity of rate of time preference wrt 1+c. Under conditions (ρ_UE γ ≤ 1) long-run rate of time preference converges to world interest rate; mean assets less sensitive to world real interest rate; elasticity of mean foreign assets approximately constant at high or low interest rates.

### Main quantitative findings (summary)
- Financial globalization, even without Sudden Stops, can trigger significant increases in mean foreign asset holdings.  
- Risk of Sudden Stops can produce significant increases in foreign assets even when long-run variability of output is unaffected by Sudden Stops.  
- Changes in output volatility cannot explain the observed surge in foreign reserves: models predict large increases in foreign assets in response to higher variability of income, but data show no systematic increases in standard deviation of cyclical output for Sudden Stop economies during financial globalization (see Table 2, Figures 2–3). Tradables GDP volatility rose in some Sudden Stop countries (Table 3) but not by magnitudes required by the model to explain observed reserve surges.  
- Buildup of foreign assets in response to financial globalization or Sudden Stop risk is slow and gradual, characterized by current account surpluses and undervalued real exchange rates; does not require active central bank real exchange rate targeting.

### I. One-sector endowment economy — structure and calibration
- Representative agent consumes composite good c; CRRA period utility with relative risk aversion coefficient γ. Time preference: UE v(c)=ln(1+ρ_UE c) or constant ρ_BAH in BAH. Income mean y subject to Markov shocks ε_t. Foreign assets b are one-period bonds at net real interest rate r (gross R≡1+r). Lump-sum exogenous absorption A introduced to match expenditure shares.  
- Aiyagari Natural Debt Limit and ad-hoc debt limit φ imposed: min{(1)}/R etc. (document preserves natural debt limit expressions).  
- Bellman equation solved by value function iteration on grid n=1000; Markov process for ε from Tauchen and Hussey (1991) quadrature.  
- Baseline calibration (matching Mexican statistics): γ=2; mean income y=1; steady-state net foreign assets-GDP ratio b=-.44 (Mexico average 1985-2004); consumption-GDP ratio c=0.692; real interest rate R=1.059 (5.9 percent, average including EMBI spread for Mexico); resource constraint implies A= y + b(R-1) - c = 0.282.  
- UE time preference elasticity: ln(R)/ln(1+ c) = 0.109 ⇒ subjective discount factor 0.944.  
- BAH calibration to match Mexico’s average net foreign assets (-44 percent) and consumption cyclical standard deviation (3.28 percent): φ = -0.51 and β = 0.94 ⇒ ρ_BAH = 0.064.  
- Income process: HP-filtered cyclical GDP in Mexican data (1965-2005): σ_y = 3.301% and ρ_y = 0.597. Underlying σ_e = 2.648 percent. THQA with 5 nodes yields σ_y = 3.285% with ρ = 0.550 and σ_e = 2.64 percent.

### I. One-sector baseline results — key statistics and dynamics
- Baseline precautionary savings: UE ≈ 2.5 percent of GDP; BAH ≈ 9.6 percent of GDP (nearly 5 times larger).  
- Foreign assets serial autocorrelation: UE baseline 0.99; BAH baseline 0.959 — implies slow, gradual adjustment of foreign assets. Transitional convergence: about 40 years (BAH) and more than 80 years (UE) from certain low-asset initial conditions (Figures 4–5).  
- Sensitivity: increases in ρ_y, σ_e, and γ raise precautionary savings; BAH always yields larger precautionary savings than UE. High-risk-aversion case (γ=5) yields precautionary savings of 10.4 percent (UE) and 23.8 percent (BAH) of GDP.  
- Consumption can be more volatile than income under some calibrations; self-insurance under incomplete markets can help explain this (Kose, Prasad, Terrones (2003) puzzle).

### I. One-sector: effects of volatility and financial globalization
- Variability effects (Figure 6a): increasing σ_e (σ_y from 1 to 8 percent) produces large increases in precautionary demand under both UE and BAH, but BAH yields higher levels. Autocorrelation increases have weaker effects (Figure 6b).  
- Data mismatch: median standard deviation of GDP in Sudden Stop countries fell from 3.5 percent (1965–1985) to 3.1 percent (1986–2005); many countries saw volatility fall rather than rise. Model requires large volatility increases to explain reserve surge: UE needs >4 percentage points, BAH needs >1.5 percentage points (Figure 6b). Conclusion: higher long-run business cycle volatility is unlikely to explain surge in reserves.

- Financial globalization modeled as time-invariant distortionary tax τ on foreign asset returns; τ in [0, 27] percent considered. Tax revenue rebated as lump sum T_t = b_t r τ. BAH: equilibrium with τ equivalent to lower discount factor exp( -v_c(c) )(1-τ_R) adjustment. UE: approximation using budget constraint with lump-sum transfer rbτ where b is long-run average. Approximation accurate for τ ≤ 0.84.  
- Results (Figures 7a–7b): cutting τ from 8 percent to 0.5 percent increases long-run average b/y sharply in both setups. Examples: BAH with natural debt limit: cut τ from 8% to 0.5% increases mean b/y from -10% to about -154% and precautionary savings from 81% to about 9.4 times GDP (extreme cases shown). With φ = -1, same tax cut increases average b/y from -90% to about +20% and precautionary savings from 10% to 120% of GDP. UE: cutting τ from 8% to 0.5% increases long-run average foreign assets from -156% to about -45% of GDP while precautionary savings approximately unchanged.  
- Interpretation: both BAH and UE predict large increases in foreign assets with financial globalization; UE suggests much of increase may be due to deterministic saving incentives (post-tax return effect) rather than pure precautionary motives. Whether financial globalization explains observed surge depends on initial τ and how much it fell.

### II. Two-sector production economy — structure and calibration
- Key differences relative to one-sector: (1) consumption c aggregates tradables c_T and nontradables c_N with CES aggregator (elasticity of substitution 1/(1+μ), weight a); equation (5) preserved. (2) Nontradables produced by representative firm using imported intermediate goods m with production technology in (6); α is share of imported inputs. Profit maximization condition (7). (3) New budget/resource constraints in tradables and nontradables sectors; liability dollarization (foreign assets denominated in tradables units). (4) Collateral constraint: lenders limit credit to fraction κ of market value of total income in tradables, up to maximum Ω (12). Agents take p_N as given — introduces credit-market externality.  
- Collateral constraint binding introduces endogenous Lagrange multiplier η_t in Euler equation for tradables consumption (13), interpreted as an endogenous interest rate premium when binding. Debt-deflation amplification: tradables consumption falls → p_N falls → value of nontradables GDP falls → borrowing capacity tightens → further falls (Fisherian amplification).  
- Calibration follows Mendoza (2002) for Mexico: γ=2; b=-0.44; R=1.059. Steady-state p_N=1, p_m=1, y_T + p_N y_N - p_m m =1. μ=0.316, a calibrated to match p_N=1 given c_T=0.261, c_N=0.431. Steady-state shares: nontradables-to-tradables GDP = 1.543 ⇒ y_T = 0.393 and (p_N y_N - p_m m) = 0.607. α=0.2 ⇒ imported input ratio to GDP = 13 percent; gross output nontradables = 0.758; Z=1.106; m = 0.152. A_T = -0.046; A_N = 0.328. UE time-preference elasticity ln(R)/ln(1( c_T, c_N )) = 0.1867 ⇒ discount factor 0.944. BAH: φ = -0.7, β = 0.94395 (ρ_BAH = 0.0594) to match Mexico’s mean net foreign assets and aggregate consumption variability.  
- Shock identification: proxy nontradables TFP by nontradables GDP; VAR(1) estimated for tradables and nontradables cyclical components (1965-2005) with matrix estimates reported; THQA with 3 realizations for ε_T and z yields Markov process approximating sectoral moments; second-stage adjustment of VAR elements involving nontradables TFP to better match moments (resulting unconditional moments approximations reported).

### II. Two-sector baseline results — business cycle co-movements and precautionary savings
- Two-sector model improvements over one-sector: (1) produces countercyclical current account and net exports (correlation of current account-output ratio with output in tradables: UE -0.14; BAH -0.48). Agents build debt during expansions. (2) Produces large, procyclical, persistent fluctuations in real exchange rate (price of nontradables): coefficients of variation 6.7% (UE) and 8.1% (BAH); correlations with output in tradables 0.75 (UE) and 0.88 (BAH); first-order autocorrelation ≈ 0.5.  
- Precautionary savings: BAH two-sector ≈ 25.3 percent of long-run average GDP; UE two-sector ≈ 1.6 percent. Across one- vs two-sector: UE precautionary savings lower in two-sector (2.5% in one-sector v. 1.6% in two-sector) due to implicit hedge from endogenous co-movement between tradables and nontradables incomes.

### II. Two-sector: effects of volatility, financial globalization, and Sudden Stops
- Effects of volatility and financial globalization qualitatively mirror one-sector: increases in volatility of tradables or nontradables raise long-run average foreign assets; data do not show systematic volatility increases across Sudden Stop countries (Table 3).  
- Financial globalization effects (Figure 8): both one- and two-sector models show large increases in foreign assets as τ falls; UE exhibits near-linear increases; BAH exhibits sharply increasing rate as τ → 0. Precautionary savings rise sharply in BAH but are nearly invariant to τ in UE. Two-sector models yield smaller precautionary savings than one-sector due to implicit hedge: equilibrium correlation between tradables income and nontradables income (in tradables units) is negative (~ -0.4), reducing need for precautionary savings.

- Sudden Stop risk quantified with collateral constraint κ=0.5 (debt limit 50% of GDP in tradable units). Long-run average assets: BAH rises from -44.7% (without Sudden Stops) to -24.3% (with Sudden Stops). UE rises from -42.4% to -37.8% when Sudden Stops possible.  
  - UE: precautionary savings increase by -37.8 - (-42.4) = 4.6 percent of long-run GDP due to Sudden Stop risk.  
  - BAH: long-run average assets rise 20.4 percentage points, but the debt limit itself rises by 20 percentage points; overall change in precautionary savings ≈ 0.5 percentage point relative to corresponding credit limits. The transitional build-up needed if Sudden Stop risk is newly introduced would be 20.4 percentage points of GDP.  
- Sudden Stop dynamics (Figure 9): conditional on initial debt ratio at which credit constraint binds (initial b = -48.7% of long-run GDP; long-run probability of reaching this debt ratio 0.9% (BAH) and 1.1% (UE)):  
  - BAH (UE) feature larger (smaller) Sudden Stop magnitudes: current account reversal 3.5 (2.5) percentage points of GDP; fall in nontradables output about 2.5 (1.5) percent; real depreciation 9 (6) percent; decline in aggregate consumption 5 (3) percent; drop in output 7 (5) percent.  
  - Convergence of foreign assets slow: BAH gap of about -4 percentage points of GDP remains after 50 years; UE gap closes in ~5 years then slowly returns to zero (moderate overshooting).  
  - Long-run Sudden Stop region: for b ≥ -0.482 constraint does not bind; for -0.578 < b < -0.482 constraint binds and Sudden Stops occur; cumulative probability of all long-run Sudden Stop states 3.88% (BAH) and 7.85% (UE). Some Sudden Stops are so large they are ruled out by precautionary behavior in the long run.  
  - Transitional CDFs (Figure 11): starting from b = -48.7% initial position, probability of hitting that ratio again after 2 years: 21% (UE) and 40% (BAH); after 15 years: 3.4% (UE) and 4.7% (BAH). Building the war-chest to minimize Sudden Stop probability is long and gradual (more than 15 years to appreciably lower recurrence probability; >50 years to attain minimal long-run probability).

### IV. Conclusions — synthesis and policy-relevant implications
- Three mechanisms drive precautionary demand for foreign assets: (i) increases in cyclical output volatility; (ii) international financial integration; (iii) Sudden Stop risk. All can produce large changes in foreign assets in the model.  
- Empirical evidence: output volatility has not systematically increased in Sudden Stop countries → volatility channel unlikely to explain reserve surge. By contrast, financial globalization and Sudden Stop risk coincide with reserve build-up and can generate quantitatively comparable increases.  
- Dynamics: buildup of assets, current account surpluses, and undervalued real exchange rates are protracted and gradual; slow decline in Sudden Stop probability (over decades). BAH setup implies nonlinear response to financial globalization: small reductions in distortions near full integration can produce very large asset adjustments.  
- Two-sector implicit hedge: endogenous co-movement yields negative correlation between tradables income and nontradables income in tradable units (~ -0.4), reducing precautionary savings relative to one-sector outcomes; however, financial globalization still a potent driver of reserve accumulation.  
- Caveats and normative considerations: analysis abstracts from efficiency gains (capital accumulation distortions may induce investment booms after opening) and sovereign default (Alfaro and Kanczuk (2006) suggest defaultable-debt portfolio choice may imply zero reserves). Welfare implications mixed: Lucas-style results suggest small average welfare cost of consumption volatility; Durdu and Mendoza (2006) show conditional welfare costs in Sudden Stop states are high and state-contingent credit facilities can be complex; Caballero and Panageas (2006) find arrangements providing financing during Sudden Stops can be welfare-improving if growth aspirations are high. Distributional concerns: underdeveloped domestic financial systems imply large, regressive costs of precautionary saving and financial globalization (Mendoza et al. (2007)). Policy priority: promote domestic financial development; identify policies that mitigate adverse distributional implications without full financial development remains an open question.  
- Overall judgment: New Merchantilism is only partially correct. Minimizing Sudden Stop risk can lead to reserve accumulation; but the surge is not explained by rising business cycle volatility. Financial globalization itself is a strong driver of reserve increases; self-insurance behavior is part of but not the sole mechanism — the removal of barriers to global asset trading interacts with underdeveloped financial markets to produce large reserve buildups.

*Source: _wp07146 - References (content unit provided).*

### 1.  One-sector model

### _wp07146 - 1.  One-sector model

### Model parameters — One-sector (baseline entries as listed)
- ρBAH Rate of time preference in the BAH setup: 0.064
- ρUE Rate of time preference elasticity in the UE setup: 0.109
- γ Coefficient of relative risk aversion: 2.000
- π Ad-hoc debt limit: -0.510
- R Gross world interest rate: 1.059
- y Mean output: 1.000
- c Consumption-output ratio: 0.692
- b Net foreign assets-output ratio: -0.440
- σe Standard deviation of output innovations: 0.026
- θy Autocorrelation of output: 0.597
- A Lump-sum absorption: 0.282

### Statistical moments of the stochastic stationary state — One-sector (Table 6)
- Precautionary savings (measured as defined in the text) shown across scenarios with values: 0.02, 0.10, 0.04, 0.12, 0.05, 0.22, 0.01, 0.05, 0.10, 0.24
- Means
  - Output: 1.00 (repeated entries: 1.00)
  - Consumption: 0.69, 0.69, 0.69, 0.70, 0.70, 0.70, 0.69, 0.69, 0.70, 0.70
  - Foreign assets: -0.42, -0.42, -0.41, -0.39, -0.39, -0.30, -0.43, -0.46, -0.34, -0.28
  - Trade balance (percent of output): 0.02 repeated across scenarios
  - Discount factor: 0.94 repeated across scenarios
- Coefficients of variation (in percent)
  - Output: 3.28, 3.28, 3.63, 3.63, 4.97, 4.97, 2.49, 2.49, 3.28, 3.28
  - Consumption: 3.13, 3.26, 3.92, 3.92, 4.72, 4.66, 2.38, 2.59, 4.11, 3.11
  - Foreign assets: 24.41, 10.11, 29.73, 13.39, 36.97, 20.28, 18.52, 6.33, 40.92, 20.10
  - Current account (percent of output): 2.68, 2.02, 2.77, 2.08, 4.08, 3.42, 2.03, 1.40, 2.81, 2.48
  - Trade balance (percent of output): 3.04, 2.11, 3.27, 2.23, 4.62, 3.66, 2.30, 1.44, 3.72, 2.78
  - Discount factor (variation): 0.14, 0.00, 0.18, 0.00, 0.21, 0.00, 0.11, 0.00, 0.18, 0.00
- Normalized coefficients of variation (relative to output)
  - Consumption: 0.95, 0.99, 1.08, 1.08, 0.95, 0.94, 0.96, 1.04, 1.25, 0.95
  - Foreign assets: 7.43, 3.08, 8.19, 3.69, 7.43, 4.08, 7.44, 2.55, 12.46, 6.12
  - Current account (relative): 0.82, 0.62, 0.76, 0.57, 0.82, 0.69, 0.82, 0.56, 0.86, 0.75
  - Trade balance (relative): 0.92, 0.64, 0.90, 0.61, 0.93, 0.74, 0.93, 0.58, 1.13, 0.85
  - Discount factor (relative): 0.04, 0.00, 0.05, 0.00, 0.04, 0.00, 0.04, 0.00, 0.06, 0.00
- Output correlations
  - Consumption: 0.42, 0.75, 0.48, 0.78, 0.42, 0.67, 0.42, 0.81, 0.26, 0.54
  - Foreign assets: 0.32, 0.56, 0.34, 0.53, 0.32, 0.44, 0.32, 0.62, 0.19, 0.33
  - Current account (percent of output): 0.97, 0.85, 0.97, 0.83, 0.97, 0.89, 0.97, 0.81, 0.99, 0.93
  - Trade balance (percent of output): 0.76, 0.70, 0.68, 0.63, 0.76, 0.73, 0.76, 0.67, 0.66, 0.74
  - Discount factor correlation: -0.42, 0.00, -0.48, 0.00, -0.42, 0.00, -0.42, 0.00, -0.26, 0.00
- Autocorrelations
  - Output: 0.59 repeated across scenarios (0.59)
  - Consumption: 0.97, 0.84, 0.97, 0.88, 0.97, 0.88, 0.97, 0.81, 0.99, 0.93
  - Foreign assets: 0.99, 0.96, 0.99, 0.98, 0.99, 0.98, 0.99, 0.94, 1.00, 0.99
  - Current account (percent of output): 0.57, 0.51, 0.67, 0.62, 0.57, 0.54, 0.57, 0.49, 0.59, 0.56
  - Trade balance (percent of output): 0.67, 0.55, 0.76, 0.67, 0.67, 0.59, 0.67, 0.52, 0.76, 0.64
  - Discount factor autocorrelation: 0.98, 0.00, 0.98, 0.00, 0.98, 0.00, 0.97, 0.00, 0.99, 0.00
- Additional scenario labels appearing in the table header: Risk Aver. 5.0; Baseline; Auto Corr 0.7; Std Dev. 5%; Std Dev. 2.5%

*Source: Authors' calculations.*

### Model parameters — Two-sector (baseline entries as listed)
- ρBAH Rate of time preference in the BAH setup: 0.059
- ρUE Rate of time preference elasticity in the UE setup: 0.187
- γ Coefficient of relative risk aversion: 2.000
- μ Elasticity of substitution: 0.316
- a CES weight of tradable consumption: 0.341
- π Ad-hoc debt limit: -0.700
- α Share of imported inputs: 0.200
- R Gross world interest rate: 1.059
- b Net foreign assets-output ratio: -0.440
- pN Relative price of nontradables: 1.000
- pm Price of imported input: 1.000
- yT + pN yN Output in units of tradables: 1.000
- cT/yT Tradable consumption-output ratio: 0.665
- cN/yN Nontradable consumption-output ratio: 0.710
- pN yN /yT Nontradable-tradable output ratio: 1.543
- AT Lump-sum absorption of tradables: 0.106
- AN Lump-sum absorption of nontradables: 0.176

### Statistical moments of the stochastic stationary state — Two-sector (Table 7)
- Scenario categories: Econ w/ perfect credit markets; Econ w/ binding credit constraints; Baseline; UEBAH; Precautionary savings values shown (examples): 0.02, 0.06, 0.25, 0.26
- Means
  - Consumption of tradables: 0.262, 0.264, 0.261, 0.269
  - Consumption of nontradables: 0.431, 0.432, 0.430, 0.435
  - Consumption (aggregate): 0.360, 0.362, 0.359, 0.366
  - Price of nontradables: 1.005, 1.011, 1.003, 1.027
  - Net foreign assets: -0.424, -0.378, -0.447, -0.243
  - Current Account-GDP ratio: 0.000, 0.000, 0.000, 0.000
  - Tradables GDP: 0.393, 0.393, 0.393, 0.393
  - GDP in units of tradables: 1.002, 1.006, 1.000, 1.019
  - Nontradables GDP: 0.607, 0.608, 0.606, 0.610
  - Imported input: 0.152, 0.153, 0.152, 0.156
- Coefficients of variation (in percent)
  - Consumption of tradables: 1.643, 1.523, 3.139, 2.993
  - Consumption of nontradables: 5.369, 5.379, 4.926, 4.896
  - Consumption (aggregate): 3.626, 3.634, 3.169, 3.109
  - Price of nontradables: 6.622, 6.550, 8.099, 8.053
  - Net foreign assets: 19.824, 16.436, 42.960, 76.574
  - Current Account-GDP ratio: 1.453, 1.416, 1.950, 1.939
  - Tradables GDP: 3.345 repeated across columns
  - GDP in units of tradables: 2.213, 2.184, 3.292, 3.271
  - Nontradables GDP: 3.050, 3.059, 2.797, 2.793
  - Imported input: 3.805, 3.688, 5.846, 5.754
- Correlation with GDP
  - Consumption of tradables: 0.374, 0.341, 0.804, 0.803
  - Consumption of nontradables: -0.579, -0.579, -0.562, -0.591
  - Consumption (aggregate): -0.471, -0.482, -0.247, -0.293
  - Price of nontradables: 0.749, 0.740, 0.867, 0.873
  - Net foreign assets: 0.266, 0.151, 0.437, 0.408
  - Current Account-GDP ratio: -0.141, -0.142, -0.474, -0.497
  - Nontradables GDP: -0.579, -0.579, -0.562, -0.591
  - Imported input: 0.833, 0.826, 0.929, 0.931
- First-order autocorrelation
  - Consumption of tradables: 0.961, 0.920, 0.908, 0.897
  - Consumption of nontradables: 0.483, 0.484, 0.505, 0.502
  - Consumption (aggregate): 0.528, 0.525, 0.663, 0.650
  - Price of nontradables: 0.460, 0.451, 0.489, 0.485
  - Net foreign assets: 0.985, 0.974, 0.995, 0.994
  - Current Account-GDP ratio: 0.489, 0.477, 0.456, 0.455
  - Tradables GDP: 0.587 repeated across columns
  - GDP in units of tradables: 0.674, 0.653, 0.673, 0.660
  - Nontradables GDP: 0.483, 0.484, 0.505, 0.502
  - Imported input: 0.514, 0.484, 0.577, 0.565

*Source: Authors' calculations.*

*Italic source attribution: _wp07146 - 1.  One-sector model (PDF chapter/section) — Authors' calculations._*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07146.pdf_