## _wp06269

## Source details

**Canonical URL:** [_wp06269](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06269.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06269.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06269.pdf.json)

---

### I. INTRODUCTION
- Currency mismatch defined as borrower funding operations in one currency while earnings accrue in another currency.
- Context: corporate balance sheets in emerging markets, especially Latin America, are "heavily tilted towards foreign-currency-denominated debt and local-currency-denominated assets and/or earnings."
- Purpose: present tractable models for measuring default risk when assets and liabilities are denominated in different currencies.
- Methodology: models based on the structural approach to credit risk (Merton (1974)); calibrated using equity price and balance sheet data under assumptions commonly used in commercial structural implementations (e.g., Moody’s KMV).
- Contributions:
  - Firm-level quantification of how currency mismatches contribute to credit risk.
  - Mapping equity prices into default probabilities to obtain forward-looking measures of default risk when local equity markets are reasonably efficient.
  - Extension to use balance sheet proxies when equity prices are unavailable.
  - Models serve as building blocks for constructing system-wide vulnerability measures.
- Scope and calibration notes:
  - Paper does not directly model currency mismatches in cash flow (but techniques could be extended).
  - Calibration requires balance sheet data; off-balance-sheet derivative hedges can be incorporated by adjusting liabilities to reflect the unhedged share.

### II. WHY DO CURRENCY MISMATCHES MATTER?
- Financial-stability channel:
  - Currency mismatches create latent risk after sharp currency depreciation: sudden increases in debt service for corporate borrowers (especially nontradable sector) can lead to bankruptcies and nonperforming loans that undermine banking/financial systems.
  - Even without default, depreciations cause balance sheet effects that deteriorate borrowing capacity, reducing output and investment.
- Persistence and drivers:
  - Currency mismatches tend to be persistent; one exception: Argentina where "mandatory pesoization reduced currency mismatches from 80 percent to less than 7 percent in 2001" but with a slight increase in 2004.
  - High inflation episodes and low credibility of government policies increase prevalence and persistence of mismatches.
- Scope beyond emerging markets:
  - Corporations in developed countries also issue foreign-currency-denominated bonds; currency mismatches are not limited to emerging markets.
- Key statistic — Latin America: Foreign-Currency-Denominated Loans (In percent of total loans)
  - Argentina: 2001 80.0  2002 7.2  2003 7.1  2004 14.1
  - Bolivia: 2001 97.0  2002 97.3  2003 97.7  2004 97.7
  - Brazil: 2001 18.0  2002 19.4  2003 ...  2004 12.0
  - Chile: 2001 13.8  2002 13.0  2003 10.3  2004 10.3
  - Colombia: 2001 11.0  2002 11.6  2003 8.8  2004 6.1
  - Costa Rica: 2001 67.2  2002 53.0  2003 55.5  2004 53.3
  - Dominican Republic: 2001 27.6  2002 30.9  2003 37.0  2004 27.3
  - Ecuador: 2001 100.0  2002 100.0  2003 100.0  2004 100.0
  - El Salvador: 2001 100.0  2002 100.0  2003 100.0  2004 100.0
  - Guatemala: 2001 ...  2002 15.3  2003 16.7  2004 17.7
  - Honduras: 2001 22.2  2002 22.8  2003 26.4  2004 30.9
  - Mexico: 2001 20.5  2002 12.9  2003 12.3  2004 9.8
  - Nicaragua: 2001 83.6  2002 83.1  2003 84.3  2004 85.0
  - Paraguay: 2001 52.8  2002 52.8  2003 55.7  2004 51.7
  - Peru: 2001 80.5  2002 79.7  2003 77.9  2004 75.9
  - Uruguay: 2001 66.0  2002 81.0  2003 76.0  2004 70.0
  - Venezuela: 2001 0.7  2002 0.8  2003 0.7  2004 0.6
  - (Source in table: Rennhack and Nozaki (2006))

### III. THE STRUCTURAL APPROACH TO DEFAULT RISK
- Two modeling approaches:
  - Intensity-based (reduced form): default governed by an exogenous stochastic process, not linked to observable firm characteristics.
  - Structural approach (adopted): default when firm cannot service debt; default linked to capital structure and economic drivers; builds on Black and Scholes (1973) and Merton (1974).
- Option-pricing analogy:
  - Equity is a residual claim equivalent to a call option on firm assets with strike equal to debt owed under absolute priority.
  - Debt value = default-free bond − value of a put on assets.
- Default timing:
  - Default-at-maturity (Merton (1974)): firm issues a zero-coupon bond; default only at maturity — may underestimate default risk when currency mismatches induce early default.
  - First-passage (first stopping time) models: bankruptcy occurs when assets first fall below liabilities; map default to a first passage time problem and can produce probabilities that default occurs during a given period.
- Model choices for currency mismatches:
  - First-passage model with exchange rate as diffusion: closed-form solutions, easy calibration.
  - Default-at-maturity model with exchange rate following jump-diffusion: captures empirically observed jumps in exchange rates.
  - First-passage model with double exponential jump-diffusion (Kou, 2002): captures asymmetric return distributions and directional bias.

### IV. THE DIFFUSION MODEL — Assumptions and setup
- Applicability: suitable for relatively stable exchange rate regimes where sudden large moves are uncommon.
- Base assumptions (Merton-style) for tractability: no transaction costs/taxes, perfectly elastic markets, short-sales allowed, continuous trading, firm financed by equity and a zero-coupon bond.
- Additional assumptions:
  - A1: Firm issues one type of debt, a foreign-currency zero-coupon bond maturing at time T with face value D; face value in local currency L = D X, with X exchange rate (local currency per foreign currency).
  - A2: Existence of probability space (Ω, F, P) with filtration (F_t)_{0 ≤ t ≤ T}.
  - A3: Under P, dynamics:
    - dV_t = μ_V(t) V_t dt + σ_V(t) V_t dW_t^V
    - dL_t = μ_L(t) L_t dt + σ_L(t) L_t dW_t^L
    - W^V_t and W^L_t independent Wiener processes.
    - In credible fixed exchange rate regime, μ_L and σ_L = 0.
  - A4: μ_i(t) and σ_i(t), i = V, L are F-predictable real functions with boundedness ensuring existence and uniqueness of solutions.
- Lemma 1: Under A2 and A4, unique solutions exist for the SDEs for V and L (existence/uniqueness via Lipschitz and linear growth conditions).

### B. Main Results (Diffusion Model)
- Asset-debt ratio defined: Y_t = ln(V_t / L_t).
- Dynamics (Lemma 2):
  - dY_t = (μ_V(t) − μ_L(t) − 1/2(σ_V^2(t) − σ_L^2(t))) dt + (σ_V(t) − σ_L(t)) dW_t
  - Y_μ given by ()()22 1 ()()()()() YVLVL tttttμμμ σσ =−+ − (as stated in source).
  - Y_σ given by 222 YVLσσσ =+.
- Two factors causing declines in asset-debt ratio:
  - Exchange rate depreciates faster than firm asset growth: ( )( )    0 VL ttμμ−<.
  - Exchange rate more volatile than firm asset volatility: 22 ()()   0 VL ttσσ−<.
- Default event and time:
  - Default time τ = inf{s: Y_s ≤ 0, 0 ≤ s ≤ T}; default corresponds to V_t ≤ L_t.
- Closed-form conditional default probability (Lemma 3, equation (9)):
  - Conditional probability that default occurs before s given information at t expressed in terms of Φ (cumulative normal), Y_t, Y_μ, Y_σ and exact functional form preserved from source.
- Special case: credible pegged or fixed exchange rate by setting drift and diffusion coefficients in L dynamics to zero.

### C. Calibration Methodology (Diffusion Model)
- Equity as call option: E = max(A − D, 0).
- Parameter vector θ = {μ_V(), σ_V(), μ_L(), σ_L()}.
- Estimate L_θ from exchange rate log-returns:
  - Exchange rate log returns r_t = ln(X_t / X_{t−1}) normally distributed with mean /2 LL μσ =− and variance 2 L σ (notation preserved).
  - L_θ maximizes log-likelihood in equation (12).
- Estimate V_θ from market capitalization using exchange option approximation (Margrabe, 1978):
  - Market capitalization E_t = max(V_t − L_t, 0).
  - European exchange option approximation (Lemma 4, equation (14)):
    - E_t = exp(−r_f (T − t)) [ Φ(d1) V_t − Φ(d2) L_t exp(−(r − r_f)(T − t)) ], with definitions preserved and Y_σ = 222 YVLσσσ =+.
  - Implicit function g: E_t = g(V_t; σ_V, σ_L, μ_V, μ_L, r, r_f, T, t); invert to obtain V_t = g^{−1}(E_t; ...).
  - Log-asset returns r_{v,t} = ln(V_t / V_{t−1}) assumed normal with mean /2 V V μσ =− and variance 2 V σ; V_θ maximizes likelihood in equation (17).
- Estimation sequence:
  1. Estimate L_θ via (12).
  2. Estimate V_θ via (17).
  3. Compute default probability using equation (9).
- Remarks:
  - Equation (9) yields objective (real-world) default probabilities since (12) and (17) estimated under the physical measure.
  - MLE approach is consistent with assumed dynamics; contrasts with Moody’s KMV point-in-time approach.
  - Framework extends to firms with both local-currency and foreign-currency liabilities under Assumptions A1' and A3' by converting foreign debt and defining adjusted asset V and proceed analogously.

### V. THE JUMP-DIFFUSION MODEL
- Motivation:
  - Diffusion models understate short-horizon default risk and fail to capture fat tails; jump-diffusion accommodates infrequent large moves.
- Assumptions B1–B3:
  - B1: Single foreign-currency zero-coupon debt face value D.
  - B2: Probability space and filtration as usual.
  - B3: Asset value V follows jump-diffusion: dV/V = μ_V dt + σ_V dW + (J − 1) dN, with N Poisson(λ) and ln J ~ N(μ_J, σ_J^2).
- Solution (Lemma 5): V_t expressed as V_0 × exp( ... ) × Π_{j=1}^{N_t} J_j (exact expression preserved).
- Default definition:
  - τ = inf{s: V_s ≤ D, 0 ≤ s ≤ T}.
  - No closed-form first-passage default probability for general s; numerical methods required.
  - Closed-form default probability at maturity (Lemma 6, equation (21)) expressed as an infinite Poisson-weighted series involving Φ; exact series preserved.
- Calibration methodology:
  - Equity E_t = max(V_t − D, 0).
  - Equity approximated with European option; equity value expressed as infinite series (Lemma 7, equation (23)).
  - Implicit inversion V_t = g^{−1}(E_t; θ_J, r_f, T, t) with θ_J = {μ_V, σ_V, λ, μ_J, σ_J}.
  - Asset returns density conditional on number of jumps yields likelihood in equation (26); truncate infinite sum (practical upper bound suggested: 10 terms) or use approximations.
  - After estimation, compute default probability at maturity via Lemma 6.
  - Fixed exchange rate special case by setting μ_V = 0 and σ_V = 0.

### VI. THE DOUBLE EXPONENTIAL JUMP-DIFFUSION MODEL (KOU)
- Motivation:
  - Empirical returns exhibit leptokurtosis and asymmetry; double-exponential jumps capture asymmetric jump-size distribution.
- Assumptions C1–C3:
  - C1: Single foreign-currency zero-coupon debt D.
  - C2: Probability space and filtration.
  - C3: Asset dynamics with jumps where ln J_i has double-exponential density:
    - f_X(x) = p η_1 e^{−η_1 x} 1_{x≥0} + q η_2 e^{η_2 x} 1_{x<0}, with η_1 > 1, η_2 > 0, p + q = 1.
- Solution (Lemma 8): V_t analogous exponential-of-BM-times-product-of-jumps expression.
- Default and first-passage analysis:
  - τ = inf{s: V_s ≤ D, 0 ≤ s ≤ T}.
  - No closed-form first-passage distribution; Laplace transform of first passage time available (Lemma 9, equation (31)) expressed via positive roots α_{1,2} and β_{1,2} of G(α) = β (equations (32)–(33)); G(x) defined explicitly in source.
  - Numerical inversion of Laplace transform required to obtain default time distribution.
- Calibration methodology:
  - Equity approximated by European call with closed-form expression (Lemma 10, equation (34)) involving Υ (closed-form in Kou (2002)).
  - Implicit inversion V_t = g^{−1}(E_t; θ_K) with θ_K = {μ_V, σ_V, λ, p, η_1, η_2}.
  - Asset return density in closed form (equation (36)); estimate θ_K by maximizing likelihood (equation (37)).
  - Use Lemma 9 and numerical inversion to compute default probabilities after estimation.

### VII. SURVEILLANCE APPLICATIONS
- Application 1 — Measuring system-wide default risk in the corporate sector:
  - Steps:
    - Estimate individual firm default probabilities using diffusion, jump-diffusion, or double-exponential models.
    - Assess probability that a subset of firms default within a horizon, accounting for common shocks (e.g., exchange rate depreciation).
  - One-factor Gaussian copula (Vasicek/Li framework) used to capture default dependence:
    - Normalized asset value: x_i = a_i M + sqrt(1 − a_i^2) Z_i, with a_i ∈ [0,1].
    - Individual default threshold x_i determined by q_i(t) via x_i = Φ^{−1}(q_i(t)).
    - Conditional default probability given M: Prob(x_i ≤ x_i | M) = Φ((x_i − a_i M)/sqrt(1 − a_i^2)).
    - Distribution of number of defaults computed via recursive procedure and integrated over M.
  - Calibration: factor loadings via principal component analysis; empirical result cited that the first principal component explains around 70 to 80 percent of variance.
- Application 2 — Prudential leverage levels and regulatory capital in a dollarized economy:
  - Framework assumptions: single bank, independent corporate loan defaults, independence of PD, EAD, LGD, normal approximation to loss distribution, identical PD across loans for tractability.
  - Loss from loan i expressed as in equation (44) (exact piecewise expression preserved).
  - Value-at-Risk under normal approximation at confidence level α (equation (45)):
    - VAR_α = z_α × sqrt(Σ_i Σ_i (EAD_i LGD_i)^2 PD_i (1 − PD_i)) (structure preserved).
    - With constant PD = p (equation (46)): VAR_α simplifies accordingly.
  - Regulatory capital constraint (equation (47)) stated exactly in source; solving equality yields upper-bound max p for default probability consistent with minimum capital requirements and RWA.
  - max p determines an upper-bound for firm leverage Lev_max (prudential leverage ratio) so potential losses can be covered by aggregate banking capital.
  - Upper-bounds for default probabilities used in leverage calculations obtained via Lemma 3, Lemma 6, or inversion of Laplace transform in Lemma 9 depending on model.

### VIII. CONCLUSIONS
- Currency mismatches in the corporate sector increase likelihood and severity of financial crises through balance sheet effects.
- Paper proposes analytically tractable credit risk models for firms with currency mismatches: diffusion, jump-diffusion, and double-exponential jump-diffusion.
- Given equity price and balance-sheet data, models can be estimated by maximum likelihood to yield real-world default probabilities consistent with assumed asset dynamics.
- Models useful for:
  - Evaluating credit risk of individual firms with currency mismatches.
  - Building financial surveillance tools: systemic default distribution and assessing prudential leverage consistent with banking sector capital.
- Extensions:
  - Adaptations for banks by adjusting default barrier and incorporating off-balance-sheet exposures; prompt corrective action framework may require increasing liabilities multiplicatively.
  - Off-balance-sheet credit risk (loan commitments, letters of credit, derivatives) can be large relative to banking/trading books and should be considered.

*Source: _wp06269 - References.*

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

### _wp06269 - References..............................................................................................................

### I. INTRODUCTION
- Currency mismatch definition: borrower funds operations in one currency while earnings accrue in another currency.
- Problem context: In emerging market countries, and especially Latin America, corporate balance sheets are "heavily tilted towards foreign-currency-denominated debt and local-currency-denominated assets and/or earnings."
- Purpose: Present tractable models for measuring default risk when assets and liabilities are denominated in different currencies.
- Methodological grounding: Models are based on the structural approach to credit risk (Merton (1974)) and can be calibrated using equity price and balance sheet data under assumptions commonly used in commercial structural implementations (e.g., Moody’s KMV).
- Contributions:
  - Provides firm-level quantification of how currency mismatches contribute to credit risk, complementing empirical studies (Bleakley and Cowan (2005); Claessens, Djankov, and Xu (2000)).
  - Maps equity prices into default probabilities to obtain forward-looking measures of default risk, applicable when local equity markets are reasonably efficient.
  - Models can be extended when equity prices are unavailable by using balance sheet proxies of equity prices.
  - Models serve as building blocks for constructing system-wide vulnerability measures.
- Scope note: The paper does not address currency mismatches in cash flow directly, though techniques could be extended since cash flow can be represented as a sequence of zero-coupon bonds.
- Calibration note: Model calibration requires balance sheet data; off-balance-sheet derivative hedges can be incorporated by adjusting liabilities to reflect the unhedged share.

### II. WHY DO CURRENCY MISMATCHES MATTER?
- Financial-stability channel:
  - Currency mismatches create latent risk following sharp currency depreciation; corporate borrowers (especially nontradable sector) face sudden increases in debt service, leading to bankruptcies and accumulation of nonperforming loans that can undermine banking/financial systems.
  - Even without default, depreciations cause balance sheet effects that deteriorate borrowing capacity, reducing output and investment (cited: Aghion, Bacchetta, and Banerjee (2001); Krugman (1999); Chang, Céspedes, and Velasco (2001)).
- Persistence: Currency mismatches tend to be persistent (Table 1), with the notable exception of Argentina where "mandatory pesoization reduced currency mismatches from 80 percent to less than 7 percent in 2001" but with a slight increase in 2004.
- Drivers:
  - One main driver is uncertainty associated with high inflation periods; Table 1 suggests currency mismatches are more prevalent where high inflation occurred.
  - Despite falling inflation, mismatches persist due to low credibility of government policies and expectations that stabilizing policies may be discontinued (e.g., election of populist governments weakening fiscal discipline).
- Scope beyond emerging markets: Corporations in developed countries also issue foreign-currency-denominated bonds (eurobond, samurai bond markets; investors' appetite for emerging-market currency-denominated bonds), so currency mismatches are not limited to emerging markets.

### Key statistic: Table 1 — Latin America: Foreign-Currency-Denominated Loans (In percent of total loans)
- Argentina: 2001 80.0  2002 7.2  2003 7.1  2004 14.1
- Bolivia: 2001 97.0  2002 97.3  2003 97.7  2004 97.7
- Brazil: 2001 18.0  2002 19.4  2003 ...  2004 12.0
- Chile: 2001 13.8  2002 13.0  2003 10.3  2004 10.3
- Colombia: 2001 11.0  2002 11.6  2003 8.8  2004 6.1
- Costa Rica: 2001 67.2  2002 53.0  2003 55.5  2004 53.3
- Dominican Republic: 2001 27.6  2002 30.9  2003 37.0  2004 27.3
- Ecuador: 2001 100.0  2002 100.0  2003 100.0  2004 100.0
- El Salvador: 2001 100.0  2002 100.0  2003 100.0  2004 100.0
- Guatemala: 2001 ...  2002 15.3  2003 16.7  2004 17.7
- Honduras: 2001 22.2  2002 22.8  2003 26.4  2004 30.9
- Mexico: 2001 20.5  2002 12.9  2003 12.3  2004 9.8
- Nicaragua: 2001 83.6  2002 83.1  2003 84.3  2004 85.0
- Paraguay: 2001 52.8  2002 52.8  2003 55.7  2004 51.7
- Peru: 2001 80.5  2002 79.7  2003 77.9  2004 75.9
- Uruguay: 2001 66.0  2002 81.0  2003 76.0  2004 70.0
- Venezuela: 2001 0.7  2002 0.8  2003 0.7  2004 0.6
- Source stated in table: Rennhack and Nozaki (2006).

### III. THE STRUCTURAL APPROACH TO DEFAULT RISK
- Two main modeling approaches:
  - Intensity-based (reduced form): default is governed by an exogenous stochastic process; default event not linked to observable firm characteristics.
  - Structural approach (adopted here): default/bankruptcy occurs when firm cannot service debt, linked to capital structure and economic drivers (business cycle, etc.); starts from Black and Scholes (1973) and Merton (1974).
- Option-pricing analogy:
  - Under absolute priority, equity is a residual claim and equivalent to a call option on firm assets with strike equal to debt owed.
  - Debt value = default-free bond − (value of a put option on assets).
- Default timing distinctions:
  - Default-at-maturity models (Merton (1974) benchmark): firm issues a zero-coupon bond; default only at maturity; risk of underestimating default when currency mismatches exist because firms may default before maturity.
  - First-passage (first stopping time) models: bankruptcy is the first time asset value falls below liabilities; these map default to a first passage time problem and can provide probabilities that default occurs during a given period.
- Relevance to currency mismatches:
  - First-passage models may better capture earlier defaults triggered by sharp depreciations (e.g., East Asia 1997, Argentina 2001).
  - First-passage models are more complex and harder to estimate than default-at-maturity models.
- Models introduced in the paper:
  - First-passage model with exchange rate as a diffusion process (yields closed-form solutions and is easy to calibrate).
  - Default-at-maturity model with exchange rate following a jump-diffusion process (empirically supported for exchange rates by Jorion (1988), Dumas et al. (1995), Bates (1996)).
  - First-passage model based on double exponential jump-diffusion (Kou, 2002) to capture asymmetric return distributions and directional bias in exchange rate moves.

### IV. THE DIFFUSION MODEL — Assumptions and setup
- Applicability: Appropriate for relatively stable exchange rate regimes where sudden large moves are uncommon ("normal" regime; day-to-day changes relatively "small").
- Base Merton (1974) assumptions used for tractability:
  a) No transaction costs nor taxes; assets indivisible.
  b) Markets perfectly elastic.
  c) Short-sales allowed; lending and borrowing rates identical.
  d) Continuous trading; firm’s assets are tradeable.
  e) Firm financed by equity and a zero-coupon bond.
- Additional model assumptions:
  - Assumption A1: Firm issues one type of debt, a foreign-currency zero-coupon bond maturing at time T with face value D. Face value in local currency is L = D X, where X is exchange rate (local currency per foreign currency).
  - Assumption A2: There exists a probability space (Ω, F, P) with filtration (F_t)_{0 ≤ t ≤ T} capturing all economy information.
  - Assumption A3: Under P, dynamics of firm value in local currency V and face value of debt L follow:
    - dV_t = μ_V(t) V_t dt + σ_V(t) V_t dW_t^V
    - dL_t = μ_L(t) L_t dt + σ_L(t) L_t dW_t^L
    where W^V_t and W^L_t are independent Wiener processes w.r.t. the filtration F.
    - Note: In a credible fixed exchange rate regime, μ_L and σ_L in equation for L are equal to zero.
  - Assumption A4: μ_i(t) and σ_i(t), i = V, L are F-predictable real functions satisfying:
    - i) For some finite real C: |μ_i(t)| + |σ_i(t)| ≤ C, (expressed in source as ()(),        , ii tCiVLμσ+≤= ).
    - This ensures existence and uniqueness of solutions to stochastic differential equations for V and L.
- Lemma 1: Under Assumptions A2 and A4, unique solutions exist for the stochastic differential equations (3) and (4) for V and L (proof outline: Lipschitz and linear growth conditions satisfied; reference Karatzas and Shreve (1991)).

*Source: _wp06269 - References..............................................................................................................*

### chapter 5).

### chapter 5)

### B. Main Results (Diffusion Model)
- Asset-debt ratio Yt defined as ln(Vt/Lt).
- Dynamics of asset-debt ratio (Lemma 2):
  - dYt = (μV(t) − μL(t) − 1/2(σV^2(t) − σL^2(t))) dt + (σV(t) − σL(t)) dWt (stated in equation (6) in the source).
  - Yμ (equation (7)): ()()22 1 ()()()()() YVLVL tttttμμμ σσ =−+ −.
- Two factors contributing to declines in the asset-debt ratio:
  - Exchange rate depreciates faster than firm asset growth: ( )( )    0 VL ttμμ−<.
  - Exchange rate more volatile than firm asset volatility: 22 ()()   0 VL ttσσ−<.
- Default event and default time (Definitions):
  - Default time τ defined as τ = inf{s: stY ≤ 0, 0 ≤ s ≤ T} (equation (8)).
  - Default corresponds to Vt ≤ Lt.
- Closed-form expression for default probability (Lemma 3, equation (9)):
  - Conditional probability that default occurs before s given information at t:
    - {}()2()()|exp2 tYtY tYt YY YstYst PsFY stst μμ τμσ σσ − ⎛⎞⎛⎞ −− −−+ − =Φ+− Φ ⎜⎟⎜⎟ ⎜⎟⎜⎟ −− ⎝⎠⎝⎠,
    - where Φ is the cumulative normal distribution, Yμ given by equation (5), and Yσ given by equation (10): 222 YVLσσσ =+.
- Special case: credible pegged or fixed exchange rate regime obtained by setting drift and diffusion coefficients in equation (4) to zero.

### C. Calibration Methodology (Diffusion Model)
- Equity as call option on assets: E = max(A − D, 0) (equation (11)).
- Parameter vector to estimate: θ = {μV(), σV(), μL(), σL()} (notation preserved from source).
- Estimate Lθ from exchange rate log-returns:
  - Exchange rate process: dX = μL(t) X dt + σL(t) X dW (implied).
  - Log returns rt = ln(Xt/Xt−1) are normally distributed with mean /2 LL μσ =− and variance 2 L σ.
  - Lθ maximizes log-likelihood in equation (12) (exact expression preserved).
- Estimate Vθ from market capitalization series using European option to exchange approximation (Margrabe, 1978):
  - Market capitalization E t = max(Vt − Lt, 0) (equation (13)).
  - European exchange option approximation (Lemma 4, equation (14)):
    - Et = exp(−r f (T − t)) [ Φ(d1) Vt − Φ(d2) Lt exp(−(r − r f )(T − t)) ] with d1, d2 defined as in source and Yσ = 222 YVLσσσ =+.
  - Implicit function g (equation (15)): Et = g(Vt; σV, σL, μV, μL, r, r f , T, t).
  - Solve for Vt as function of σV (equation (16)): Vt = g−1(Et; ...).
  - Log-asset returns rv,t = ln(Vt/Vt−1) assumed normal with mean /2 VV μσ =− and variance 2 V σ; Vθ maximizes likelihood in equation (17).
- Estimation sequence:
  1. Estimate Lθ via (12).
  2. Estimate Vθ via (17).
  3. Compute default probability using equation (9).
- Remarks:
  - Equation (9) yields objective (real-world) default probabilities since (12) and (17) estimated under physical measure.
  - Maximum likelihood estimation is consistent with assumed dynamics; contrasts with Moody’s KMV point-in-time approach which implies time-varying asset volatility inconsistent with Merton structural model assumption of constant asset volatility.
  - Framework extends to firms with both local-currency and foreign-currency liabilities under Assumptions A1' and A3' (described in source): local debt C and foreign debt D converted to L = D X; define Z and V = Z − C and proceed with analogous dynamics.

### V. The Jump-Diffusion Model
- Motivation:
  - Diffusion-only models underestimate short-horizon default risk and fail to capture fat tails in exchange rate returns.
  - Jump-diffusion accommodates infrequent large moves; relevant for emerging markets and fixed/pegged regimes under attack.
- Assumptions B1–B3 (Merton-style jump-diffusion):
  - Single foreign-currency zero-coupon debt with face value D (B1).
  - Probability space and filtration as usual (B2).
  - Asset value V follows jump-diffusion (equation (18)): dV/V = μV dt + σV dW + (J − 1) dN, where N is Poisson(λ) and ln J ~ N(μJ, σJ^2) (B3).
- Solution for asset value (Lemma 5, equation (19)):
  - Vt = V0 exp(Σ terms) × Π_{j=1}^{Nt} Jj (exact expression preserved as in source).
- Default definitions in jump-diffusion:
  - Default time τ = inf{s: Vs ≤ D, 0 ≤ s ≤ T} (equation (20)).
  - No closed-form first-passage default probability for general s; numerical methods required.
  - Closed-form default probability at maturity (Lemma 6, equation (21)):
    - (|) t PT τ = Σ_{j=0}^{∞} (e^{−λ(T−t)} (λ(T−t))^j / j!) × Φ( ... ), exact series expression preserved from source.
- Calibration methodology (jump-diffusion):
  - Equity in foreign currency E t = max(V t − D, 0) (equation (22)).
  - Approximate equity with European option; equity value representation (Lemma 7, equation (23)):
    - Et = Σ_{n=0}^{∞} e^{−λ(T−t)} (λ(T−t))^n / n! × BS(V, σ√(T−t), r f , D) × adjustment (exact series expression preserved).
  - Implicit inversion (equation (24)): Vt = g−1(Et; θJ, r f , T, t).
  - Asset returns density conditional on number of jumps leads to likelihood in equation (26) for θJ = {μV, σV, λ, μJ, σJ}.
  - Truncate infinite sum (practical upper bound suggested: 10 terms) or use closed-form likelihood approximations.
  - Once parameters estimated, compute default probability at maturity via Lemma 6.
  - Fixed exchange rate special case by setting μV = 0 and σV = 0.

### VI. The Double Exponential Jump-Diffusion Model (Kou)
- Motivation:
  - Empirical returns show leptokurtosis and asymmetry; double-exponential jumps allow asymmetric jump-size distribution.
- Assumptions C1–C3 (Kou model):
  - Single foreign-currency zero-coupon debt D (C1).
  - Probability space and filtration (C2).
  - Asset dynamics (equation (27)): dV/V = μV dt + σV dW + Σ_{i=1}^{N} (Ji − 1) with ln Ji distributed double-exponential (density in equation (28)):
    - fX(x) = p η1 e^{−η1 x} 1_{x≥0} + q η2 e^{η2 x} 1_{x<0}, with η1 > 1, η2 > 0, p + q = 1.
- Solution (Lemma 8, equation (29)): analogous exponential-of-BM-times-product-of-jumps expression as in source.
- Default definitions:
  - τ = inf{s: Vs ≤ D, 0 ≤ s ≤ T} (equation (30)).
  - No closed-form first-passage time distribution; Laplace transform of first passage time available (Lemma 9, equation (31)):
    - Laplace transform expressed in terms of positive roots α1,2 and β1,2 of equation G(α) = β (equations (32)–(33)); G(x) defined explicitly in source.
  - Invert Laplace transform numerically to obtain default time distribution.
- Calibration methodology (Kou model):
  - Equity approximated by European call with closed-form expression (Lemma 10, equation (34)), involving Υ (closed-form in Kou (2002), referenced).
  - Implicit inversion (equation (35)): Vt = g−1(Et; θK) with θK = {μV, σV, λ, p, η1, η2}.
  - Asset return density given in closed form (equation (36)); maximize likelihood (equation (37)) to estimate θK.
  - Once parameters are estimated, compute Laplace transforms via Lemma 9 and invert to obtain default probabilities.

### VII. Surveillance Applications
- Two applications illustrated:
  1. Measuring system-wide default risk in the corporate sector.
     - Steps:
       - Estimate individual firm default probabilities using any of the three models.
       - Assess probability that a subset of firms default within horizon, accounting for common shocks (e.g., exchange rate depreciation).
     - One-factor Gaussian copula (Vasicek/Li framework):
       - Normalized asset value xi = ai M + sqrt(1 − ai^2) Zi (equation (38)), with ai ∈ [0,1].
       - Individual default threshold xi determined by qi(t) via xi = Φ−1(qi(t)).
       - Conditional default probability given M: Prob(xi ≤ xi | M) = Φ((xi − ai M)/sqrt(1 − ai^2)) (equation (39)).
       - Distribution of number of defaults computed by recursive procedure (equations (40)–(42)) and integrated over M (equation (43)).
     - Calibration of factor loadings via principal component analysis; empirical results cited: first principal component explains around 70 to 80 percent of variance (result from Chan-Lau and Gravelle (2005) as stated).
  2. Prudential leverage levels and regulatory capital in a dollarized economy.
     - Framework assumptions:
       - Single bank, independent corporate loan defaults, independence of PD, EAD, LGD, normal approximation to loss distribution, identical PD across loans for tractability.
     - Loss from loan i (equation (44)):
       - xi = EADi × LGDi with probability PD i; xi = −PROB? (source shows piecewise; exact expression preserved as in equation (44) format).
     - Value-at-Risk under normal approximation at confidence level α (equation (45)):
       - VARα = zα × sqrt(Σi Σi (EADi LGDi)^2 PD i (1 − PD i)), exact structure preserved.
       - With constant PD = p (equation (46)): VARα simplifies accordingly.
     - Regulatory capital constraint (equation (47)):
       - VARα ≤ MCR × RWA − PROV, with equation (47) exact:
         - 22 11 (1) LL NN iii i ii VAR p EAD LGD z p p EAD LGD αα == ×+ − × ≤ × + ∑∑ = MCR RWA PROV (format preserved).
     - Solving equality yields upper-bound max p for default probability consistent with minimum capital requirements and RWA.
     - max p determines upper-bound for firm leverage Levmax (prudential leverage ratio) so potential losses expected to be covered by aggregate banking capital.
     - Upper-bounds obtained via Lemma 3, Lemma 6, or inversion of Laplace transform in Lemma 9 depending on model.

### VIII. Conclusions
- Currency mismatches in the corporate sector increase likelihood and severity of financial crises through balance sheet effects.
- Paper proposes analytically tractable credit risk models for firms with currency mismatches (diffusion, jump-diffusion, double-exponential jump-diffusion).
- Given equity price and balance-sheet data, models can be estimated by maximum likelihood to yield real-world default probabilities consistent with assumed asset dynamics.
- Models useful for:
  - Evaluating credit risk of individual firms with currency mismatches.
  - Building financial surveillance tools: systemic default distribution and assessing prudential leverage consistent with banking sector capital.
- Extensions:
  - Models can be adapted for banks by adjusting default barrier and incorporating off-balance-sheet exposures; prompt corrective action framework may require increasing liabilities multiplicatively (Chan-Lau and Sy, 2006).
  - Off-balance-sheet credit risk (loan commitments, letters of credit, derivatives) can be large relative to banking/trading books and should be considered.

*Source: _wp06269 - chapter 5).*

### REFERENCES

### _wp06269 - REFERENCES

### Theoretical foundations and stochastic calculus
- Harrison, J. Michael, 1985, Brownian Motion and Stochastic Flow Systems (New York: John Wiley).
- Karatzas, Ioannis, and Steve Shreve, 1991, Brownian Motion and Stochastic Calculus, 2nd edition (New York: Springer).
- Protter, Philip, 1992, “Stochastic Integration and Differential Equations” (New York: Springer Verlag).
- Bichteler, Klaus, 2002, Stochastic Integration with Jumps, Encyclopedia of Mathematics and Its Applications (Cambridge, U.K.: Cambridge University Press).
- Lo, Andrew, 1988, “Maximum Likelihood Estimation of Generalized Ito Processes with Discretely Sampled Data,” Econometric Theory, Vol. 4, pp. 231–47.
- Atiya, Amir F., and Steve A.K. Metwally, 2002, “Efficient Estimation of First Passage Time Density Function for Jump Diffusion Processes” (unpublished; New York: Columbia University).
- Kou, Steven G., and Hui Wang, 2003, “First Passage Times of a Jump Diffusion Process,” Advances in Applied Probability, Vol. 35, pp. 504–31.
- Gukhal, Chandrasekhar R., 2001, “Analytical Valuation of American Options on Jump-Diffusion Processes,” Mathematical Finance, Vol. 11, pp. 97–116.
- Yu, Jialin, 2005, “Closed-Form Likelihood Estimation of Jump-Diffusions with an Application to the Realignment Risk of the Chinese Yuan” (unpublished; New York: Columbia University).

### Jump-diffusion, jumps, and option pricing
- Bates, David S., 1996, “Jumps and Stochastic Volatility: Exchange Rate Processes Implicit in Deutsche Mark Options,” Review of Financial Studies, Vol. 9, pp. 69–107.
- Jorion, Philippe, 1988, “On Jump Processes in the Foreign Exchange and Stock Markets,” Review of Financial Studies, Vol. 1, pp. 427–55.
- Kou, Steven G., 2002, “A Jump-Diffusion Model for Option Pricing,” Management Science, Vol. 48, pp. 1086–101.
- Merton, Robert C., 1976, “Option Pricing When Underlying Stock Returns are Discontinuous,” Journal of Financial Economics, Vol. 3, pp. 125–44.
- Duffie, Darrell, and Kenneth J. Singleton, 2003, Credit Risk: Pricing, Measurement, and Management (Princeton, NJ: Princeton University Press).
- Zhou, Chunsheng, 2001, “The Term Structure of Credit Spreads with Jump Risk,” Journal of Banking and Finance, Vol. 25, 2015–40.
- Bates, David S., 1996, “Jumps and Stochastic Volatility: Exchange Rate Processes Implicit in Deutsche Mark Options,” Review of Financial Studies, Vol. 9, pp. 69–107.

### Option theory, corporate securities, and credit risk models
- Black, Fischer, and Myron S. Scholes, 1973, “The Pricing of Options and Corporate Liabilities,” Journal of Political Economy, Vol. 81, pp. 637–54.
- Black, Fischer, and Myron S. Scholes, 1973, “The Pricing of Options and Corporate Liabilities,” Journal of Political Economy, Vol. 81, pp. 637–59.
- Black, Fischer, and John C. Cox, 1976, “Valuing Corporate Securities: Some Effects of Bond Indenture Provisions,” Journal of Finance, Vol. 31, pp. 351–67.
- Geske, Robert, 1977, “The Valuation of Corporate Liabilities as Compound Options,” Journal of Financial and Quantitative Analysis, Vol. 12, pp. 169–76.
- Merton, Robert C., 1974, “On the Pricing of Corporate Debt: the Risk Structure of Interest Rates,” Journal of Finance, Vol. 29, pp. 449–70.
- Longstaff, Francis, and Eduardo S. Schwartz, 1995, “A Simple Approach for Valuing Risky Fixed and Floating Rate Debt,” Journal of Finance, Vol. 50, pp. 789–819.
- Kim, In-Joo, Krishna Ramaswamy, and Suresh Sundaresan, 1993, “The Valuation of Corporate Fixed Income Securities,” Financial Management, Vol. 22, 117–31.
- Saá-Requejo, Jesús, and Pedro Santa-Clara, 1999, “Bond Pricing with Default Risk” (unpublished; Los Angeles: University of California).
- Nielsen, Lars, Jesús Saá-Requejo, and Pedro Santa-Clara, 1993, “Default Risk and Interest Rate Risk: The Term Structure of Credit Spreads” (unpublished; Los Angeles: University of California).
- Crosbie, Peter, and Jeffrey R. Bohn, 2003, “Modeling Default Risk,” (unpublished; San Francisco: Moody’s KMV Company).
- Vacisek, Oldrich, 1987, “Probability of Loss on Loan Portfolios” (unpublished; San Francisco: Moodys KMV).
- Duan, Jin-Chuan, Genevieve Gauthier, and Jean-Guy Simonato, 2004, “On the Equivalence of the KMV and Maximum Likelihood Methods for Structural Credit Risk Models” (unpublished; Toronto: University of Toronto).
- Li, David X., 2000, “On Default Correlation: A Copula Function Approach,” Journal of Fixed Income (March), pp. 115–18.
- Rubinstein, Moshe, 1991, “One for Another,” Risk, Vol. 4, pp. 30–32.
- Longstaff, Francis, and Eduardo S. Schwartz, 1995, “A Simple Approach for Valuing Risky Fixed and Floating Rate Debt,” Journal of Finance, Vol. 50, pp. 789–819.

### Currency, dollarization, balance-sheet effects, and sudden stops
- Calvo, Guillermo A., Alejandro Izquierdo, and Luis-Fernando Mejía, 2003, “On the Empirics of Sudden Stops: The Relevance of Balance SheetsEffects,” NBER Working Paper No. 10520 (Cambridge, MA: National Bureau of Economic Research).
- Krugman, Paul, 1999, “Balance Sheet Effects, the Transfer Problem and Financial Crises,” in International Finance and Financial Crises, Essays in Honour of Robert Flood, ed. by Peter Isard, Assaf Razin, and Andrew Rose (Dordrecht: Kluwer).
- Chang, Roberto, Luis F. Céspedes, and Andrés Velasco, 2001, “Dollarization of Liabilities, Financial Fragility, and Exchange Rate Policy,” in Currency Unions, ed. by Alberto Alesina and Robert Barro (Stanford: Hoover Institution Press).
- De Nicolo, Gianni, Patrick Honohan, and Alain Ize, 2005, “Dollarization of Bank Deposits: Causes and Consequences,” Journal of Banking and Finance, Vol. 29, pp. 1697–727.
- Cowan, Kevin N., 2004, “Financial Dollarization,” in Economic and Social Progress in Latin America— 2005 Report (Washington: InterAmerican Development Bank).
- Bleakley, Hoyt, and Kevin Cowan, 2005, “Corporate Dollar Debt and Depreciations: Much Ado About Nothing?,” Working Paper 532 (Washington: Inter-American Development Bank).

### Banking, corporate vulnerability, and market indicators
- Chan-Lau, Jorge A., and Toni Gravelle, 2005, “The END: A New Indicator of Financial and NonFinancial Corporate Sector Vulnerability,” IMF Working Paper No. 05/231 (Washington: International Monetary Fund).
- Chan-Lau, Jorge A., Arnoud Jobert, and Janet Q. Kong, 2004, “An Option-Based Approach to Bank Vulnerabilities in Emerging Markets,” IMF Working Paper No. 04/33 (Washington: International Monetary Fund).
- Chan-Lau, Jorge A., and Amadou N.R. Sy, 2006, “Distance-to-Default in Banking: A Bridge Too Far?” IMF Working Paper No. 06/215 (Washington: International Monetary Fund).
- Chan-Lau, Jorge A., and André O. Santos, 2006a, “Currency Mismatches and Corporate Default Risk: An Application to Emerging Markets” (unpublished, Washington; International Monetary Fund).
- Chan-Lau, Jorge A., and André O. Santos, 2006b, “Off-Balance Sheet Exposure and Default Risk: Conceptual Issues and Modeling Approaches” (unpublished, Washington: International Monetary Fund).
- Gropp, Reint, Jukka Vesala, and Giuseppe Vulpes, 2006, “Equity and Bond Markets Signals as Leading Indicators of Bank Fragility,” Journal of Money, Credit and Banking, Vol. 38, pp. 399–428.
- Gibson, Michael S., 2004, “Understanding the Risk of Synthetic CDOs,” Board of Governors of the Federal Reserve System, Washington.

### Exchange options, currency option valuation, and related methods
- Margrabe, William, 1978, “The Value of an Option to Exchange One Asset for Another,” Journal of Finance, Vol. 33, pp. 177–86.
- Bjerksund, Petter, and Gunnar Stensland, 1993, “American Exchange Options and Put-Call Transformation: A Note,” Journal of Business, Finance, and Accounting, Vol. 20, pp. 761–64.
- Dumas, Bernard, L. Peter Jennergren, and Bertil Naslund, 1995, “Siegel’s Paradox and the Pricing of Currency Options,” Journal of International Money and Finance, Vol. 14, pp. 213–23.
- Rubinstein, Moshe, 1991, “One for Another,” Risk, Vol. 4, pp. 30–32.

*Source: _wp06269 - REFERENCES*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06269.pdf_