For most of the post-war period, emerging market sovereigns could not borrow in their own currencies. Foreign investors would not hold the paper, so governments issued in dollars instead. That created a mismatch: revenues in local currency, obligations in dollars. When the currency fell, the debt burden rose at exactly the moment the capacity to service it fell. Eichengreen and Hausmann called this original sin, and it sat behind most of the crises of the 1980s and 1990s.
The problem has largely been solved as the share of emerging market government debt denominated in foreign currency has fallen substantially since those crises. Major emerging markets now hold more than two-thirds of total government debt in local currency and have avoided significant net foreign currency issuance since 2010. In some major emerging markets the local currency share now exceeds 90%. Across emerging and developing economies the foreign currency share sits at roughly 20% excluding China and India, against 6% in OECD countries, where much of that exposure is hedged.
This is a real achievement, and it has worked with sovereign balance sheets genuinely less exposed to currency depreciation than they were thirty years ago.But the currency risk did not vanish and below we will explore what happened.
Where the risk went first: foreign investors
Local currency bond markets did not deepen purely through domestic demand. A substantial share of the new issuance was absorbed by global portfolio investors, who bought local currency sovereign paper and took the currency exposure onto their own balance sheets.
Carstens and Shin called the result original sin redux. The mismatch migrated from borrower to lender, and in doing so changed the transmission mechanism rather than removing it.
The new mechanism works like this: when the dollar strengthens or global risk appetite deteriorates, foreign holders of local currency debt face losses in dollar terms before anything happens to the underlying credit. They reduce exposure and their selling pushes local currency yields up, which tightens domestic financial conditions, which weakens the currency further. The feedback loop runs through the investor base rather than through the sovereign balance sheet.
The magnitude is measurable, with caveats. Working paper research across eleven emerging economies finds that an average 8% currency depreciation against the dollar is associated with capital outflows that reduce foreign holdings of local currency bonds by 3.7 percentage points and push local currency bond yields up by 154 basis points.
The same research reports significant heterogeneity between the eleven, with the degree of insulation depending on each country’s financial development and debt level. A mean of 154 basis points across a sample that varies that widely is not a figure any individual sovereign should expect to face. It implies the effect is concentrated, and identifying where it concentrates is more useful than the headline number.
That is a meaningful tightening, delivered by investor behaviour rather than by any domestic policy decision or deterioration in fiscal fundamentals. The sovereign did nothing wrong and still faces higher borrowing costs.
The BIS puts it directly in its 2026 Annual Economic Report: currency and duration risk are transferred to foreign investors, whose outflows push up local currency yields when the dollar appreciates or risk appetite falls.
Where the risk went next: domestic banks
When foreign investors retreat, the issuance still has to be bought. In economies with shallow institutional investor bases and thin hedging markets, the buyer of last resort is the domestic banking system.
The cleanest measure is banks’ sovereign holdings as a share of total bank assets, which the IMF tracks because it controls for the size of the banking system. On that measure the exposure has risen steadily since 2014.
The GDP series points the same way. Domestic banks’ sovereign holdings across emerging and developing economies rose from around 9% of GDP in 2010 to roughly 14% by 2020, though that series is harder to read on its own, since a rising ratio can reflect growing public debt rather than growing bank concentration.
The sharpest shift is in low-income countries, where domestic banks went from holding 30% of domestic debt to 50%. There the measure is unambiguous, because it is a share of the debt itself rather than of GDP.
The IMF identifies three drivers behind banks’ growing sovereign exposure since 2014: liquidity management requirements, attractive yields in a high interest rate environment, and, in its own phrasing, moral suasion from authorities in some cases.
Where a government leans on domestic banks to absorb its issuance, the resulting demand is not a market signal. It is a form of financial repression that makes the debt appear more sustainable than it is, because the price is not being set by investors who could decline.
The channel that actually matters
The consequence is not primarily a sovereign risk problem but rather a private credit problem, though the mechanism is more conditional than it first appears.
The straightforward version runs like this. When domestic banks hold a large and rising share of government debt, sovereign stress transmits directly into bank balance sheets. Bank capital is finite, so every unit of sovereign exposure displaces lending capacity that would otherwise have gone to firms.
Banks holding government bonds is not automatically crowding out. Sovereign paper is liquid, carries a zero risk weight under Basel, and can be pledged as collateral. A bank holding treasuries can borrow against them and lend the proceeds. Where liquidity rather than capital is the binding constraint, a larger sovereign book may expand lending capacity rather than consume it.
The crowding out binds under narrower conditions: where capital is the constraint, and where the sovereign exposure is held to maturity rather than circulated through repo markets.
This is what the research actually finds. Kim’s result is not that sovereign holdings crowd out credit generally. It is that banks in countries with low financial development relative to their debt levels disrupt private credit significantly more when foreign capital exits the local currency bond market.
In those economies, the banking system is simultaneously the only available buyer of sovereign debt and the only available lender to firms, so the two uses of capital compete directly.
If crowding out is real, private credit growth should fall as bank sovereign holdings rise, controlling for credit demand. If it is not, banks should expand both sides of the balance sheet together. That test is straightforward to run on any individual country and would settle the question far more cleanly than cross-country averages.
The old crisis was acute and visible. The new one, where it exists, is chronic and diffuse, and it shows up in growth data rather than in default statistics.
What this does not mean
Sovereign default risk from currency mismatch is genuinely lower. A government that owes money in the currency it prints has options that a government owing dollars does not, and the research confirms that local currency issuance has provided real short-term resilience to global shocks by reducing exposure to both original sin and its redux.
The improvement is also unevenly distributed, and concentrated where domestic investor bases are deep enough to absorb the issuance.
Hofmann and Patel put the underlying point more bluntly: the root cause is the shallowness of domestic financial markets, which creates dependence on external funding regardless of what currency the debt is written in.
Argentina is the useful control case, because it has not made the transition. Still substantially dependent on foreign currency issuance, it carries the original version of the risk rather than the relocated one.
That has a consequence for how its spreads should be read. Argentina’s EMBI+ spread, analysed separately in this research series, is a measure of sovereign default risk, because dollar debt is what a sovereign actually defaults on. A local-currency-dominant sovereign’s yield measures something structurally different: foreign investor flow, domestic absorption capacity, and the inflation the government might choose rather than the default it might be forced into.
Comparing spreads across the two groups therefore compares different things. A compressed spread on a dollar-denominated credit is a statement about default probability. A compressed yield on a local currency credit is partly a statement about who happens to be holding it. The convergence of the two numbers does not imply convergence of the underlying risk.
What to track
Four observable measures would settle whether the relocation has reduced aggregate vulnerability or merely moved it.
- Foreign holdings as a share of local currency debt. If this falls while issuance continues, domestic absorption is rising and the bank channel is tightening.
- Banks’ sovereign exposure as a share of total assets. The IMF tracks this specifically because it is the cleanest single measure of the nexus. Rising exposure in a country with a thin institutional investor base is the warning signal.
- The spread between sovereign yield moves and private credit growth during dollar strength episodes. If sovereign yields rise 150 basis points during a dollar rally and private credit growth falls simultaneously, the transmission is confirmed.
- Whether moral suasion becomes regulation. Informal pressure on banks to hold government paper is one thing. Capital requirements or liquidity rules that mandate it are another, and they would make the repression explicit and the pricing permanently uninformative.
Disclaimer
This piece is published by Howden Research for informational and educational purposes only. It is not investment advice, a personal recommendation, or an investment recommendation within the meaning of UK market abuse rules, and it is not an offer or solicitation to buy or sell any security. Howden Research is not authorised or regulated by the Financial Conduct Authority. Views expressed are those of the author at the date of publication and are subject to change. Contributors may hold positions in the securities or instruments discussed. The value of investments can fall as well as rise. Anyone considering an investment decision should seek advice from an appropriately authorised professional.
