WORKING PAPERS
WORKING PAPERS
Submitted to the American Economic Journal: Macroeconomics
Existing theory prescribes FX intervention only against identified non-fundamental shocks, yet most emerging market central banks simply target volatility. This paper shows why that works. Heterogeneously informed FX dealers cannot decompose the exchange rate into fundamentals and unobservable capital flows; rational confusion cascades through higher-order beliefs, generating a 100-fold disconnect where capital flows drive 65% of exchange rate variance under HI versus less than 1% under CK. Intervention operates through the volatility–risk premium channel: absorbing capital flow variance lowers ασ², weakening the exchange rate's sensitivity to flows in a self-reinforcing loop. Targeting the Backus-Smith wedge dominates (−15% welfare loss); leaning against the wind is ineffective under HI because it dampens both components alike, preserving the signal-to-noise ratio. Monetary policy cannot substitute—the Taylor rule anchors more effectively under CK, widening the HI/CK ratio. The paper micro-founds the IPF's two-instrument architecture: monetary policy for inflation, FX intervention for the risk-sharing wedge.
This paper provides the first framework for optimally mixing spot and derivative (NDF) foreign exchange interventions. Banks intermediate capital flows in a segmented market and manage reserves in frictional interbank markets, so a single shock propagates through both a portfolio channel (risk-bearing) and a liquidity channel (FC reserve scarcity). Three key results: UIP deviations reflect both frictions, CIP deviations depend only on liquidity, and spot intervention relieves both channels while NDFs address only the portfolio margin. Evidence from Peru (2005–2023) confirms systematic variation in the BCRP's instrument mix across shock types. Embedded in a small open economy with FC working capital constraints and solved via a linear-quadratic approach, the optimal policy tilts toward NDFs for capital outflows (spot/NDF ≈ 0.55) and toward spot for liquidity shocks (ratio ≈ 0.03). Combining instruments strictly dominates either alone, with gains increasing in credit dollarisation—calibrated to Peru (μ_w = 0.40), the welfare improvement is concentrated in lower intervention costs and residual FC liquidity stress.
Emerging-market firms borrow in dollars because a constrained intermediary prices currency mismatch into a dollar-credit discount. Original sin emerges endogenously when intermediary frictions overwhelm precautionary and borrowing-limit forces. Central bank reserve sales generate two opposing channels via a split-variance structure: dealer variance σ²_d falls fast (shrinking the discount, cheapening pesos) while firm variance σ²_f falls slowly (weakening precaution). The volatility channel dominates where intermediation is shallow—precisely where original sin is worst. A heterogeneous-firm Bewley model calibrated to Peru reduces the dollar share by 11.5 pp, concentrated among middle-wealth firms; below κ ≈ 0.50 the risk-taking channel takes over, a self-limiting mechanism. Peruvian data support the full chain: BCRP intervention reduces FX volatility (GARCH-X), FX variance carries a sol sovereign premium loading on dealer friction rather than realized volatility, dollarization rises monotonically with firm size, and medium size firms de-dollarize fastest (−4.49 pp/year).
Central banks in dollarized economies accumulate dollar reserves at a negative carry and impose reserve requirements on dollar deposits. This paper argues these are two sides of a single insurance contract: the reserve requirement is the premium, the reserve stock is the fund. In a dynamic model with frictional interbank markets and rare disasters, the optimal reserve requirement is countercyclical—raised during capital inflows to build the buffer when credit conditions are easy, cut during crises to reduce the cost of dollar intermediation going forward. Reserve requirements are a double-edged instrument: they build the aggregate insurance buffer while aggravating the individual liquidity problem that makes insurance necessary, implying an interior optimum. The reserve requirement operates as a price instrument—reducing the cost of future dollar intermediation—while the central bank's own reserves are a quantity instrument, deployed instantly to banks already in deficit; the two are complements, not substitutes. Calibrated to Peru, the model generates a policy function spanning 6–31% with an ergodic mean of 27%, and the insurance scheme is self-financing over the cycle. Data from the BCRP (2010–2025) confirm the countercyclical pattern and a "slow in, fast out" asymmetry in policy changes.
The fiscal theory of the price level consolidates the central bank and treasury, so all government liabilities are backed by fiscal surpluses. This paper argues consolidation is inappropriate when central bank assets are foreign reserves, importing asset partitioning from corporate law: reserve-backed liabilities are insulated from fiscal shocks just as secured bonds are insulated from a firm's earnings deterioration. A treasury exposure ratio α ∈ [0, 1] governs transmission—at α = 1 (the Fed) inflation absorbs fiscal shocks; at α = 0 (the BCRP) shocks reduce bond prices without moving the price level. The relationship is nonlinear: the market value of money is bounded below by its book value, and only when this floor binds does α matter, explaining why the effect is stronger in EMs. Calibrated to Peru (α ≈ 0.03) and the Fed (α ≈ 0.95), a 2pp fiscal shock generates 8% inflation in the Fed economy but under 1% in Peru. Panel evidence (23 countries, 2001–2020) finds a significant fiscal-inflation-α interaction in EMs (β̂ = 1.75, p = 0.034) but not in advanced economies.
State-guaranteed lending programs prevent firm bankruptcy, not just smooth consumption. In a two-sector DSGE model with endogenous exit—firms die when net worth falls below a threshold—and an informal sector comprising 72% of employment, the guarantee breaks a doom loop: defaults erode bank capital, widen spreads, raise borrowing costs, and trigger further defaults. Calibrated to Peru's 30% GDP contraction in 2020Q2, Reactiva Perú saved approximately 5 pp of GDP, compressed credit spreads by 103 bp, and reduced formal death rates from 7.7% to 6.5%, with low-productivity firms (Q1) saved disproportionately (8.7 pp vs. 2.0 for Q5). The extensive margin (exit prevention) accounts for two-thirds of the total welfare gain; doom-loop prevention alone contributes one-quarter. Informal workers benefit through wage spillovers, demand linkages, and formalization—the formal employment share rises from 25% to 26%—even though they cannot access the program. Optimal coverage is 30–40% of firms, slightly above Reactiva's actual 22%. The guarantee and conventional rate cuts are complements, and the guarantee's value roughly doubles at the zero lower bound.
Peru capped microenterprise lending rates at 83% in 2021, but the BCRP's own data show that 90% of marginal borrowers' interest rates reflect real intermediation costs, not markups. In a heterogeneous firm model with dual lender cost structures and judicial efficiency, a rate cap partitions formal borrowers into margin-compressed and rationed zones. At 83.4%, the cap compressed margins for 18% of firms (−16 pp) but rationed another 18% into gota a gota lending at 104 pp more; welfare fell 6.6%. An optimal cap at ~102%—just above the bank cost frontier—compresses all margins without rationing, yielding +2.1%. The BCRP's formula-based cap self-corrected from 83.4% (−6.6%) to 109.8% by 2024 (+0.8%, zero rationing), validating the adaptive approach. Extensions with endogenous wages and firm dynamics confirm the baseline. The 542,900 affected borrowers reflect the cost of starting too low; rate caps treat a symptom while the disease is weak institutions.
Existing models of credit spreads treat the loan demand elasticity as a structural parameter, but in the data pass-through from policy rates to lending rates is state-dependent. This paper replaces Calvo pricing in the banking sector with customer search: firms endogenously shop for cheaper credit, and banks set lending rates knowing that aggressive pricing triggers customer attrition. The equilibrium spread equals the retention rate divided by the marginal sensitivity of attrition to pricing—an endogenous, market-determined elasticity that collapses to the standard CES markup when search costs are infinite. Financial shocks erode the customer base directly, generating persistent spread dynamics: the half-life of the credit spread rises from 24 to 29 quarters, a 21% amplification. A spread-augmented Taylor rule (φ_μ ≈ −0.04) closes 68% of the gap to Ramsey; the optimized rule (φ_π = 3.75, φ_μ = −0.24) closes 98%. Search makes the optimal spread response smaller than in constant-elasticity models because spreads partially self-correct endogenously. Using Peruvian banking data (2003–2024), anticipated policy rate changes pass through at 45 cents on the dollar while surprises pass through at only 17 cents—a 2.6:1 ratio that the search model predicts but constant-elasticity models cannot generate.
Inflation redistributes wealth from creditors to debtors through the Fisher channel, but only on the local-currency portion of portfolios. This paper solves a Ramsey problem in a continuous-time HANK model with four extensions: heterogeneous dollarization across wealth, an exchange-rate balance sheet channel, endogenous denomination choice via a precautionary motive, and an intermediary dollar-credit discount. Endogenous dollarization creates a feedback loop—inflation raises FX variance, which changes the denomination mix the planner redistributes through—making inflation partially self-defeating. The intermediary wedge widens with volatility, amplifying original sin. A critical threshold emerges at 50% portfolio dollarization: below it, the Fisher channel dominates and inflation is redistributive; above it, the balance sheet channel dominates and inflation is regressive. Country applications to Peru (30%), Chile (48%), and Brazil (2%) show that these forces reduce optimal inflation by 30–90% relative to the closed-economy benchmark, with the effect concentrated in dollarized economies while leaving the low-dollarization case essentially unchanged.
Peru has enacted eight extraordinary pension fund withdrawals since 2020, draining roughly US$30 billion (57% of pre-pandemic AFP assets); Chile authorized three rounds totaling US$49 billion. In a closed-economy OLG model with mandatory DC savings, borrowing constraints, and idiosyncratic shocks calibrated to Peru's AFP system, temporarily reducing the contribution retention rate improves welfare during crises but harms it otherwise (CEV = −0.58% absent a shock). An extension with three income types and Markov employment shocks (30% crisis job-loss probability) reveals a hump-shaped dose–response: welfare peaks at a 5 pp cut—a halving of the rate, not full suspension—yielding +0.20% CEV. Full suspension is suboptimal because the pension annuity loss outweighs the additional liquidity gain. The policy is progressive: low earners gain +0.26% CEV while high earners lose −1.04%, explaining its persistent political appeal. A fire-sale sensitivity exercise shows that a permanent 25 bp reduction in the fund's capitalisation rate suffices to eliminate the welfare gain, making the optimal dose an upper bound and reinforcing the case for flow-based retention-rate cuts over the lump-sum balance withdrawals actually used. Results are robust across parameterisations and confirmed under stochastic value function iteration.
We examine how unrealized gains and losses on U.S. banks’ Available-for-Sale (AFS) securities are linked to deviations from Covered Interest Parity (CIP) in FX swap markets. Using monthly data for G-10 currencies versus the U.S. dollar from 2001–2022, we construct an aggregate measure of large U.S. banks’ AFS unrealized gains/losses and test its relationship with cross-currency bases. Our mechanism is balance-sheet capacity: under post-crisis Basel III, AFS mark-to-market changes (through AOCI for many large banks) affect regulatory capital and, in turn, banks’ willingness to engage in balance-sheet-intensive arbitrage that enforces CIP. To address endogeneity, we isolate the regulation-driven component of AFS valuation changes in a two-step framework and estimate panel specifications with currency fixed effects. We find a strong post-crisis association in which weaker AFS valuations predict wider CIP deviations, with effects more pronounced when CIP is measured using OIS rates rather than LIBOR. The relationship is also stronger around quarter-ends, consistent with binding reporting-date constraints. Overall, the results suggest that interest-rate-driven swings in banks’ AFS portfolios transmit regulatory balance-sheet pressures into global dollar funding conditions and FX swap pricing.
Fiscal councils are widely seen as key institutions for promoting fiscal discipline, yet measuring their effectiveness remains difficult. Existing indices emphasize de jure features—mandates, independence, resources—while largely overlooking de facto behavior. This paper proposes a complementary, behavior-based approach by quantifying how fiscal councils communicate their assessments and warnings. Using a large language model, we analyze all reports and communiqués issued by the Peruvian Fiscal Council (FC) between 2016 and 2025 to construct an index of “fiscal tone,” capturing the severity of concern expressed in each document. The results reveal a marked shift toward more critical positions in recent years, consistent with a gradual weakening of fiscal rules and oversight. Beyond documenting Peru’s fiscal deterioration, the study shows how LLM-based textual analysis can provide a replicable tool for evaluating the behavior, vigilance, and credibility of fiscal councils, complementing traditional institutional indicators.
This paper studies how geographic disadvantage and labor market informality interact to shape rural–urban migration in developing economies. We develop a spatial equilibrium model that integrates firm dynamics, household migration and participation decisions, and informality margins, building on Ulyssea (2018) and Lagakos et al. (2018). We estimate the model using a two-step Simulated Minimum Distance (SMD) approach with Peruvian microdata. The model matches sixteen empirical moments - including informality by skill level, firm size distributions, and informal hiring within formal firms—and captures both extensive and intensive margins of informality across regions. Our findings show that rural migration rates (6.2%) exceed those in urban areas (5.4%), driven by higher formal entry costs and lower informal barriers in rural markets. Informal firms in rural areas face lower operating costs but higher volatility and exit rates. These mechanisms sustain regional gaps in formality and productivity. The results underscore the limits of spatially neutral formalization policies. Targeted interventions that reduce rural formalization costs, improve infrastructure, and support mobility are needed to break the cycle of informality and geographic inequality.
This paper extends the Real Business Cycle (RBC) model with incomplete financial markets by incorporating capital outflow shocks and financial intermediaries, as introduced by Itskhoki and Mukhin (2021). This framework introduces endogenous deviations from Uncovered Interest Parity (UIP), driven by the risk premium required by risk-averse intermediaries when absorbing the economy’s foreign currency position. We demonstrate how this setup enhances many stylized facts observed in small open economies, particularly the countercyclicality of the current account. Additionally, we argue that this framework significantly improves the explanatory power of small-open economy RBC models, as it not only aligns with the key stylized facts of business cycles but also addresses the exchange rate puzzles documented in the literature. While the classical RBC model with productivity shocks captures key stylized facts, it falls short in explaining the complexities of exchange rate dynamics and puzzles. By integrating financial frictions, this paper bridges those gaps and advances the empirical alignment of small-open economy RBC models.
This paper develops a general equilibrium framework that integrates heterogeneous firms with both idiosyncratic productivity and subjective beliefs about public goods provision—specifically, confidence in institutional quality—alongside endogenous informality. We examine the impact of tax policy on the formalization process and highlight the crucial role of firms' trust in public institutions. Our findings reveal that when firms perceive the government as credible, an increase in both tax rates and tax revenues enhances public goods provision, fostering greater formalization. However, in environments with weak institutional trust, formalization policies may yield suboptimal economic outcomes—potentially even worsening conditions compared to scenarios with higher trust levels. This underscores how institutional confidence influences the productivity of formal firms and facilitates their transition into the formal sector. In the long run, effective tax policy can improve overall welfare, but its success is contingent on government credibility. Our research contributes to the literature on informality by providing novel insights for policymakers seeking to enhance formalization and economic welfare, particularly in settings where skepticism about government commitment and institutional capacity prevails.
Nominal GDP targeting (NGDP) rules have gained attention as a potential alternative to traditional models of monetary policy. In this paper, we extend the analysis of the welfare implications of NGDP rules within a New Keynesian model with nominal price and wage rigidities. Using a welfare function derived from the utility of consumers, we compare the NGDP target with a domestic inflation target, a CPI inflation target, and a Taylor rule in a small open economy scenario. Our simulations reveal that NGDP rules confer advantages on a central bank when the economy faces supply shocks, while their performance against demand shocks is comparable to that of a CPI target rule. These findings suggest that NGDP targeting could be a useful policy framework for central banks seeking to enhance their ability to stabilize the economy.
We propose a dynamic general equilibrium model to study the optimal reaction to terms of trade shocks when international financial markets are imperfect and the composition of capital flows affects the exchange rate determination. These elements allow us to showcase the interactions between commodity prices and international financial market inefficiencies. Positive commodity price shocks will generate a real over-appreciation of the currency and an inefficiently large shift of factors between the tradable and non-tradable sectors. We study the welfare implications of foreign exchange intervention through optimal simple rules and find support for leaning-against-the-wind foreign exchange intervention. Our setup, allows us to rationalize the reserve accumulation episodes commonly observed during periods of high commodity prices in resource-rich economies.