Collection:
International macro reading list

22 papers
Written by Fable 5
Table of Contents

Twenty-two papers from the syllabus of Paulina Restrepo-Echavarría’s PhD course in International Macroeconomics (Washington University in St. Louis, Spring 2023), read and digested one by one. The syllabus organizes them into five topics, kept as the five blocks below. No lectures were recorded, so each note is written from the paper itself, plus a stand-in talk, lecture, or interview by the authors where one exists (nine of the twenty-two).

Read in order, the list turns out to be much more than five topics — it is a single argument, prosecuted over forty years. The frictionless neoclassical benchmark is built and immediately shown to fail in specific, catalogued ways; the field splits into two great repair programmes, one working on prices and one on promises; each programme develops its own instruments and its own anomalies; and the course ends where the instructor’s own research lives, with the tools of both programmes aimed at the two questions that started everything — why capital goes where it goes, and why poor countries’ business cycles look the way they do. The threads below are written after reading all twenty-two, and each block’s essay says how its papers advance the argument.

Introduction: intertemporal trade and small open economies

The course opens the way the field did: build the cleanest possible model, take it seriously, and write down exactly where it breaks. Backus, Kehoe and Kydland extend the Kydland–Prescott real business cycle engine to two countries with complete markets and discover the quantity anomaly: in the model consumption is nearly perfectly correlated across countries and output isn’t, because insurance decouples what you eat from what you produce; in the data the ordering is reversed, everywhere. Their honesty sets the template — every parameter chosen without regard to international implications, every failure labeled rather than patched, and the one patch they try (a small transport cost) fixing the quantities while leaving the correlation anomaly untouched, with the haunting side-finding that the gains from international asset trade in the model are worth only 0.3 percent of consumption.

Mendoza builds the one-country limit — the small open economy — and finds the same disease in miniature (frictionless investment is wildly too volatile; a tiny adjustment cost worth 0.1 percent of GDP cures it) while planting two seeds this list harvests later: the Feldstein–Horioka correlation reflects shock persistence rather than capital immobility, and interest-rate shocks are innocuous only for countries with small debt service — an exemption emerging markets will not enjoy. Schmitt-Grohé and Uribe then perform the field’s housekeeping: the small open economy model has no well-defined steady state, the five devices used to pin one down (Uzawa preferences, debt-elastic premia, portfolio costs, complete markets) are quantitatively interchangeable at business-cycle frequencies, and the researcher may pick whichever computes easiest. That certificate of innocuousness matters enormously later — because half the remaining syllabus consists of loading real economics onto exactly the objects this paper certified as technicalities. Obstfeld and Rogoff close the block with the grand unification of the benchmark’s failures: six famous puzzles — home bias in trade and portfolios, Feldstein–Horioka, the consumption correlations, PPP persistence, exchange-rate disconnect — and one suspect, plain goods-market trade costs, whose bite is a small number raised to a large elasticity. The quantity puzzles largely surrender; the two pricing puzzles do not, and their resistance is the bridge to the next block.

The real exchange rate and the terms of trade

The second block takes up the puzzles the trade-cost story couldn’t kill: real exchange rates are enormously volatile, persist for half-lives of three to five years, and track nominal rates almost one-for-one. Backus and Smith supply the block’s diagnostic instrument — under efficient risk sharing, relative consumptions and real exchange rates must move together; in OECD data the correlation is 0.045 against a theoretical value of one — a price-side restatement of the quantity anomaly so clean it became the litmus test every later model is graded on. The two Bretton Woods papers (Eichengreen’s retrospective, Bordo’s autopsy of 1965–73) look like history but function as identification: they establish that the 1973 collapse was caused by U.S. fiscal-monetary imbalance — the elephant in the room — and not by any change in real fundamentals, which is precisely what makes the regime switch a clean natural experiment. Eichengreen’s evidence that inflation persistence itself jumped at 1971 in four countries is the same lesson in another register: the nominal regime changes the statistical character of real and nominal series.

That experiment is then cashed in twice. Chari, Kehoe and McGrattan stress-test the profession’s favorite story — monetary shocks plus sticky prices — and find it can produce the volatility (with year-long price fixity and risk aversion of five) and most of the persistence, but slams into the Backus–Smith wall: complete markets weld the exchange rate to relative consumption with correlation 1.00 against −0.35 in the data, and no goods-market friction, no bond-economy incompleteness, no habit formation can break the weld. Their closing sentence — the fix must come from “richer forms of asset market frictions” — is a purchase order. Itskhoki and Mukhin fill it two decades later: the Mussa facts (real exchange rate volatility jumps sixfold at 1973 while inflation, consumption and output change not at all) falsify sticky-price and flexible-price models alike, and the resolution moves monetary non-neutrality out of the goods market entirely, into a segmented financial market where the peg endogenously tames the risk premium. The Backus–Smith correlation flips sign with the regime, exactly as their model requires. Burstein and Gopinath’s handbook chapter is the block’s measurement backbone: pass-through into retail prices is tiny, border prices are sticky in their invoicing currency (mostly dollars) and pass exchange rates through only 28 percent over a good’s whole life — so goods-market frictions are real, measurable, and firm-chosen, yet precisely because prices barely respond, they explain how exchange-rate movements transmit while leaving why exchange rates move to the financial side. The block ends with the field’s center of gravity moved: from Dornbusch’s sticky prices to the currency risk premium.

Imperfections in international capital markets

The third block is the promises programme: if the first two blocks asked why prices misbehave, this one asks why anyone repays a sovereign loan at all. Eaton and Gersovitz found the modern answer in 1981 — reputation; default costs you future market access, so debt is sustained up to an endogenous credit ceiling — and thereby turned borrowing limits from assumptions into equilibrium objects. Bulow and Rogoff’s eight-page arbitrage then nearly demolished it: a defaulter who can still save abroad can replicate the insurance reputation provided, so reputation alone supports zero debt and enforcement must rest on direct sanctions. The field’s practical settlement — exclusion plus output costs of default — quietly concedes their point. Tomz and Wright move the fight to two centuries of data and embarrass both camps’ quantitative implementations: defaults do lean toward bad times, but only just (62 percent begin below trend; output at default averages a mere 1.6 percent below trend), far too loose a relationship for models in which default is insurance exercised on schedule.

The block’s second half builds with these bricks. Kehoe and Perri put limited enforcement inside the BKK world model and resolve the original quantity anomalies — consumption and output correlations nearly reconciled, cross-country investment and employment correlations finally positive — where exogenously incomplete markets (the single bond) had changed almost nothing: how markets are incomplete is everything, and the right incompleteness is derived from the commitment problem. Restrepo-Echavarría then aims the same machinery at a single historical wound — Latin America’s quarter-century of flat consumption after 1980 — reading binding participation constraints as renegotiations (capital keeps flowing through real defaults, so autarky-style default models start from the wrong fact), with high Volcker-era world rates slowing the approach to the constraint and prolonging the slide, and correcting the folklore along the way: the constraint binds in prolonged bad times, not only good ones. Hamann, Méndez-Vizcaíno, Mendoza and Restrepo-Echavarría bring the tradition to present-day commodity sovereigns: oil reserves are wealth that survives default, so the sovereign strategically hoards them before defaulting, and lenders — pricing the option — shave five points of GDP off sustainable debt. From Eaton–Gersovitz’s abstract ceiling to a barrel-denominated one, the through-line is unbroken: every credit limit in this block is a measured incentive, not an assumption.

Capital flows and capital controls

Block four zooms out from the contract to the direction of world capital. Lucas’s five pages set the terms: standard technology assumptions imply India’s marginal product of capital should be 58 times America’s, so something is drastically wrong — human capital (58 falls to 5), local human-capital externalities (5 falls to 1.04), political risk (undermined by the colonial-era evidence that enforcement was once perfect and capital still didn’t equalize), or monopoly. Gourinchas and Jeanne sharpen the question devastatingly: among developing countries, capital flows toward the slow growers — Korea got nothing, Madagascar got 6 percent of GDP a year — and no symmetric friction that merely throttles flows can reverse their direction; since the flow is saving minus investment and investment does track growth, the puzzle is really that fast growers out-save their own booms. The allocation puzzle is a saving puzzle wearing a capital-account costume.

Ohanian, Restrepo-Echavarría and Wright deliver the autopsy with the wedge-accounting method inherited from the business-cycle accounting tradition: three regions, 1950–2007, wedges on labor, capital, and international assets, model matching the world data exactly. The verdict indicts a market no one had charged: labor. Asia’s labor markets were heavily distorted in 1950 — hours low, returns to capital therefore unremarkable — so the world rationally declined to send capital; as the distortions unwound, Asia grew fast and financed itself. Labor wedges account for 60–70 percent of the flows; international capital-market distortions matter mainly through slow-moving asset positions and, crucially, in Latin America’s 1980s — the same crisis blocks three’s papers modeled from the inside. The three papers form a tidy syllogism: frictions must be enormous (Lucas), they cannot be symmetric throttles (Gourinchas–Jeanne), and when finally measured, the big one sits in the labor market (ORW) — which also means fifty years of policy built on shipping capital to poor countries was, on most readings, pushing on a string that domestic reform could actually pull.

Emerging markets

The final block is where every friction in the course becomes first-order at once. The emerging-market business cycle is a genuinely different object: twice the output volatility of developed small open economies, consumption more volatile than output, strongly countercyclical trade balances, and interest rates that are countercyclical and lead the cycle. Neumeyer and Perri explain it through prices of credit: decompose the interest rate into a world rate plus country risk, let firms borrow working capital so the rate walks into labor demand, and let domestic fundamentals drive the spread — eliminating country-risk fluctuations would cut Argentine output volatility by 27 percent, while stabilizing world rates buys less than 3. This is the debt-elastic premium from block one’s closing-device catalogue, resurrected with three orders of magnitude more economics, and it is the reduced form of block three’s default premia. Aguiar and Gopinath answer with shocks instead of frictions: emerging markets’ defining feature is a volatile trend — regime switches, nationalizations, reforms — and a frictionless model with mostly-permanent shocks produces the whole syndrome, Tequila crisis included, because a country whose good news is chronically about the trend rationally borrows against every boom. Permanent shocks account for 82 percent of Mexican output variance against 50 for Canada: the cycle is the trend.

Restrepo-Echavarría’s closing paper interrogates the data both stories calibrate to: consumption in the national accounts is a residual, the informal sector — a third of GDP in developing countries — largely escapes it, and cyclical substitution between formal and informal activity makes measured consumption more volatile than output even when true consumption is smooth. The correlation between informal-sector size and excess consumption volatility is 0.75, and the tell is Scandinavia, Spain and Portugal — developed economies with developing-sized informal sectors and developing-style consumption volatility that no country-risk or trend-shock story can reach. It is the perfect final reading for a PhD course: after twenty-one papers of models judged against moments, the twenty-second asks how the moments were made.

The threads, pulled together

Three threads run the length of the list. The first is the career of the risk-sharing condition. The complete-markets benchmark makes two linked promises — consumptions comove (BKK), and relative consumption tracks the real exchange rate (Backus–Smith) — and the whole syllabus can be read as the field’s forty-year interrogation of their failure: measured in block one, restated as prices in block two, tested against sticky prices (CKM) and regimes (Mussa), and finally derived as an equilibrium wedge — from enforcement constraints in Kehoe–Perri and Restrepo-Echavarría, from segmented financial markets in Itskhoki–Mukhin. What began as an anomaly ends as a measured object with two competing microfoundations, promises and portfolios.

The second thread is the promotion of closing devices into economics. Schmitt-Grohé and Uribe certified the small open economy’s stationarity devices as interchangeable technicalities; the rest of the course is a parade of those same objects returning with content — the debt-elastic premium as Neumeyer–Perri’s country risk, the ad hoc debt limit as Eaton–Gersovitz’s incentive-compatible ceiling, Kehoe–Perri’s enforcement constraint, and the resource-backed default option of the Hamann et al. paper. The certificate of innocuousness is what makes the promotions credible: because the plumbing was shown not to matter, whatever matters in the later papers must be the economics.

The third is the instructor’s own line, which the syllabus quietly builds toward: limited commitment read as renegotiation for Latin America’s lost decade (16), wedge accounting for the postwar allocation of world capital (19), strategic resources inside the default tradition (14), and the measurement critique of the whole enterprise (22). Its common conviction — visible also in the choice to close the course with it — is that the interesting frictions are measurable, that Latin America and Asia are the decisive testing ground, and that before theorizing about a moment one should ask who measured it and how. A student who reads the list in order arrives, by the last page, at the frontier the course was designed to hand over: the benchmark’s failures are catalogued, the two repair programmes are in hand, and the open question is the one every block sharpened — how much of what we call international macroeconomics is frictions in markets, and how much is frictions in what we know about them.