Heterogeneous Global Booms and Busts

The paper asks why countries with seemingly similar fundamentals exhibit markedly different boom–bust cycles when hit by common global shocks. Motivated by the 2008 global financial crisis and the 2010 Eurozone crisis—where advanced economies displayed emerging‑market‑style vulnerabilities—the authors study how frictions in the global supply of capital produce heterogeneous macro‑financial outcomes. The key hypothesis is that investor skill in screening borrowers interacts with country opacity to endogenously segment countries into core (low‑exposure) and periphery (high‑exposure) groups, jointly determining the tightness of credit and the depth of recessions across countries.

The study relates to work on credit cycles driven by financial frictions (e.g., Kiyotaki–Moore, Lorenzoni, Mendoza, Gorton–Ordoñez), but differs by emphasizing information frictions in global capital supply rather than collateral constraints. It complements two‑country representative‑agent models of international risk sharing and flight‑to‑quality (Gourinchas–Rey; Maggiori) by modeling heterogeneous institutions and firm‑level interactions that generate cross‑sectional predictions for returns, debt ownership, and real effects. It connects to research on sudden stops in emerging markets (Aguiar–Gopinath; Rey–Martin; Eichengreen–Hausmann; Caballero–Krishnamurthy; Broner–Ventura) while showing similar mechanisms can arise among advanced economies. It also contributes to literatures on Eurozone crisis mechanisms (e.g., Reis; Gopinath et al.; Cuñat–Garicano; Fernandez‑Villaverde et al.; Schmitt‑Grohé–Uribe; Battistini et al.; Farhi–Tirole; Aguiar et al.; Martin–Philippon) by endogenizing loose pre‑crisis financing conditions in the periphery. Finally, it ties into safe‑asset determination (He–Krishnamurthy–Milbradt; Caballero–Farhi–Gourinchas; Farhi–Maggiori) by deriving safe assets from informational frictions, and builds on asymmetric information market structure (Kurlat) to allow heterogeneity on both demand and supply sides.

Three‑period model (t=0,1,2) with a single good. A continuum of firms (type τ∈{g,b} denoting pledgeability and opacity ω∈[0,1]) reside across a unit mass of countries indexed by their average opacity. Firms invest at t=0 (scale I), face a t=1 idiosyncratic liquidity shock with probability φ requiring per‑unit liquidity ξ to complete projects (β=1 in the full model), and produce at t=2 with output ρ_τ per completed unit (ρ_g≥ρ_b). “Good” (pledgeable) firms can pledge α per completed unit; “bad” cannot. Banks provide actuarially fair, state‑contingent saving for firms but cannot lend. International investors have unit wealth at t=1, heterogeneous skill s∈[0,1], and pay a per‑application testing cost (set to κ=0 in the main equilibrium). They can run tests of varying prudence: bold (identifies some bad firms but pools others with good; false positives) or cautious (identifies some good firms; false negatives). Investor evidence yields acceptance rules that are measurable with respect to the test outcomes. Credit markets at t=1 comprise many markets m, each posting an interest rate r(m); firms can demand across markets subject to an upper bound; investors select one market, the set of applications to consider, and an acceptance rule. A clearing algorithm (generalizing Kurlat’s LRF/NMR) determines rationing by investor selectivity. Many‑to‑many matching is allowed via rationing and allocation functions. Countries differ only by opacity distribution; mapping between names and opacity is uninformative to investors (baseline). Two environments are analyzed: (i) Section II endogenizes prudence in a simplified model (β=0, φ=0, α=1, three investor skill mass points) where a global productivity shock changes the fraction of good firms λ and thereby the optimal test (bold in high λ, cautious in low λ). (ii) The full model introduces an exogenous prudence shock across aggregate states θ∈{H,L}: investors are bold in H and cautious in L, with continuous w(s), β=1 and α=ξ. Equilibrium is solved by backward induction: at t=1, characterize r and rationing η(·) given firms’ scale; at t=0, firms choose I and saving subject to a budget constraint balancing investment and expected liquidity needs net of pledgeability. The paper derives equilibrium interest rate schedules r_H and r_L(ω), rationing for bad and good firms, and thresholds {ω_1,ω_2,ω_3} that partition countries into exposure groups. Comparative statics examine aggregate demand shocks (higher φ) and supply shocks (more low‑skill wealth), and implications for safe assets and portfolio composition.

- Prudence regime and market integration: When investors are bold (boom, θ=H), there is a single global interest rate r_H; all good firms obtain funding up to capacity; some bad firms obtain credit (especially in opaque countries) due to false positives. When investors are cautious (bust, θ=L), credit markets fragment and the borrowing rate r_L(ω) is weakly increasing in opacity; bad firms are screened out; good firms in opaque countries face higher rates and rationing. - Endogenous core vs periphery: Thresholds {ω_1,ω_2,ω_3} partition countries. Core (ω≤ω_1) face low, often zero, rates in busts and have low exposure; periphery (ω≥ω_3) face the maximal rate and rationing in busts with high exposure. This partition arises from the interaction of investor skill distribution w(s), country opacity, and firm credit demand. - Real outcomes: In booms, output and credit rise broadly with similar rates across countries and larger credit/output growth in the periphery. In busts, output collapses disproportionately in the periphery due to expensive/insufficient liquidity and abandonment of production; core experiences modest declines. Firms in periphery “gamble” by choosing larger scales and saving less against liquidity risk, amplifying busts. - Prices and spreads: Booms exhibit compressed cross‑country spreads; busts display fragmented markets with higher yields in periphery countries for comparable firms. - Portfolio rebalancing and ownership concentration: Low‑skill investors overweight opaque countries in booms and rebalance toward core in busts; high‑skill investors rotate toward periphery in busts to capture high returns. Debt ownership is more concentrated in busts, especially in periphery. - Non‑performing debt and misallocation: Most non‑performing debt is originated in booms, disproportionately in periphery and funded by low‑skill investors. Productivity dispersion among funded firms is higher in periphery during booms, consistent with misallocation evidence. - Returns: Realized average returns on bonds issued in booms are higher in core versus periphery (fewer defaults); in busts, bonds issued by periphery offer higher realized returns (higher yields on good borrowers), especially for skilled investors. - Safe assets: Only sufficiently transparent countries issue assets whose yields fall in bad times; safe‑asset status is endogenous to opacity and the prudence regime. - Comparative statics: (i) Higher global credit demand (higher φ, e.g., pandemic) shifts more countries out of core (smaller ω_1), suppresses booms and deepens busts overall, with the sharpest busts in the most opaque countries. (ii) Greater low‑skill capital supply (global saving glut) lowers r_H, amplifies booms (especially in periphery via bad lending), increases non‑performing debt, and indirectly shrinks the core (both ω_1 and ω_3 move left), exacerbating periphery busts despite more abundant capital in booms.

The findings substantiate the central mechanism: informational frictions in global capital supply, mediated by investor skill and country opacity, translate common shocks into heterogeneous national cycles. Endogenous investor prudence maps aggregate conditions into screening intensity: boldness integrates markets with widespread credit at homogeneous prices, while cautiousness fragments markets with steeply increasing prices by opacity, reallocating capital toward countries where screening is most effective. This mechanism jointly explains (i) similar pre‑crisis pricing across core and periphery, (ii) sudden spread divergence and flight‑to‑quality in crises, and (iii) deeper recessions in opaque countries. The model aligns with stylized facts from the Eurozone crisis and broader sudden‑stop episodes, and generates new cross‑sectional predictions on portfolio shifts, ownership concentration, misallocation, and returns. It also provides a coherent framework linking safe‑asset supply to transparency and illustrating how aggregate demand/supply shocks in funding reconfigure the global core–periphery anatomy.

The paper develops a tractable equilibrium framework with two‑sided heterogeneity (firms by opacity/pledgeability and investors by skill) to explain heterogeneous global booms and busts. Core and periphery emerge endogenously from informational frictions in the international supply of capital. Booms feature integrated markets and compressed spreads; busts entail market fragmentation, capital reallocation to the core, and severe real contractions in the periphery. The model yields testable implications for spreads, flows, ownership concentration, return patterns, non‑performing debt, misallocation, and safe‑asset issuance. It also clarifies how global shocks to credit demand (e.g., pandemics) and to capital supply (global saving glut) reshape the set of core/periphery countries. Future research could bring the framework to quantitative discipline with calibration and empirical tests, incorporate dynamic feedback of learning and reputation, allow partially informative priors about country opacity, model government/sovereign interactions explicitly, and study policy tools (e.g., macroprudential or liquidity backstops) that target prudence shifts and cross‑border spillovers.

- Countries are modeled as sets of firms with a single opacity parameter and identical fundamentals across countries; real‑world heterogeneity in technology, institutions, and policy is abstracted from to isolate the supply‑side mechanism. - Investors’ priors about country opacity are uninformative in the baseline; while extended in an appendix, richer information structures may alter quantitative results. - Agents are risk‑neutral, and dynamic aspects are compressed into three periods; prudence is exogenous in the main equilibrium (endogenous only in the simplified section), and test costs are often set to zero for tractability. - Governments and sovereign decisions are not modeled; implications for sovereign debt are inferred from corporate spreads. - Results are theoretical; several predictions (e.g., debt ownership concentration dynamics, return decomposition across skills) await systematic empirical tests.