The most extreme number in this quarter's emerging debt valuations is not any single metric. It is the distance between two of them. Hard currency spreads sit in the most expensive quintile the manager tracks, while local rates sit in the cheapest quartile. Each time in recent memory that emerging debt split this way, the gap closed through a dollar shock first, and the local holder had to survive that path before collecting the mean.
The manager's 3Q26 note resolves the split into a single instruction: blended portfolios tilted to the maximum local allocation, 70% local against 30% hard currency. On the valuation arithmetic the call is hard to argue with. The history of how such gaps resolve is where the reader should spend time, because the averages in the note describe destinations, not routes.
A quarter of tightening, and a gap that would not close
By late October 2026, the U.S. 10-year Treasury yielded 5.00% and the 2-year 4.69%. That left the curve positively sloped by 31 basis points as of October 23, a configuration that had not been typical for most of the preceding cycle. Against that backdrop, dollar-denominated sovereign and quasi-sovereign emerging debt had kept rallying. The manager measured the excess spread at 71 basis points, said it had "tightened significantly over the quarter," and placed it in the first, least attractive quintile. In the manager's own history, spreads in that quintile have preceded a mean two-year annualized credit return of minus 2.0% over the risk-free rate.
Local currency debt read the other way on every axis. The valuation gap between emerging local rates and U.S. rates stood at +1.4%, in the most attractive fourth quartile. Historically that bucket has been followed by a mean EM-over-U.S. return differential of +4.1%. The currency indicator, an expected spot return of +1.5%, landed near the top of the third quartile. Mean subsequent spot returns on the GBI-EM Global Diversified weights have been +5.1% from that quartile and +6.8% from the fourth. USD duration itself was rated neutral, slightly below fair value and marginally more attractive than a quarter earlier.
Taken alone, each reading makes sense. Put side by side, they describe a market that has paid up for emerging sovereign credit risk denominated in dollars while declining to pay for the same countries' risk in their own currencies. The note calls the relative valuations "unusually extreme." That split is the pattern worth tracing, because emerging debt has produced versions of it before, and the way those versions resolved matters more to a 70/30 tilt than the averages do.
The chart carries the mechanism the note leaves implicit. A max-local tilt, whatever else it is, is substantially a position against the dollar, since local bond returns to a dollar-based holder arrive through the currency line as much as the yield line. The valuation case and the dollar case are therefore the same case. This is what the earlier episodes test.
Mid-2007: rich everywhere, then the currency gave way
In the first half of 2007, dollar-denominated emerging sovereign spreads compressed to what were then historic tights. The asset class had spent four years being rerated from crisis-prone to investment-grade-adjacent, and the rerating had a real basis: reserve accumulation, current account surpluses in much of Asia and Latin America, and a wave of sovereign debt buybacks that shrank the supply of dollar paper. Local currency markets had become the new frontier, and a generation of dedicated local bond funds was raised on the argument that emerging currencies were structurally undervalued after the late-1990s devaluations.
The difference from 2026 is the one that matters most. In 2007 local debt was not cheap. Emerging currencies had appreciated for several years, real yields had compressed alongside nominal ones, and the carry trade into high-yielding local markets was among the most widely held positions in global macro. Both halves of emerging debt were expensive. The U.S. 10-year traded above 5% in June 2007, an echo of today's level, but the curve was flat to inverted rather than positively sloped.
When Lehman Brothers filed on September 15, 2008, the dollar became the only asset many holders wanted. Hard currency spreads blew out, but the drawdown in local currency debt, measured in dollars, was considerably deeper because the currency line and the rate line fell together. The Brazilian real lost roughly a third of its value against the dollar in the closing months of 2008. Holders of dollar bonds took a credit loss. Holders of local bonds took a credit loss, a rates loss, and a currency loss at the same time.
Then the recovery ran in the opposite order. Local markets, cheapened by the currency move, delivered some of the strongest returns in the asset class's history through 2009 and 2010. The investor who held local through the drawdown was eventually paid. The investor who had bought local at 2007 valuations needed most of two years to get back to even. Both outcomes sit inside a long-run average.
May 2013: the same yields, read two ways
The 2013 episode is the cleaner test of how a valuation signal behaves when the dollar regime turns underneath it. By early 2013 local currency debt had become the consensus emerging-market allocation. Developed-world yields were pinned by quantitative easing, the U.S. 10-year had fallen to roughly 1.6% in early May, and foreign ownership of local government bonds in markets such as Indonesia, South Africa, Turkey and Brazil had climbed steadily for several years.
On May 22, 2013, Ben Bernanke told Congress that the Federal Reserve could slow its asset purchases in coming meetings. Over the following months the 10-year moved toward 3%, and the adjustment in emerging local markets was violent precisely because the foreign holder base was so large. By late summer, Morgan Stanley's analysts had grouped Brazil, India, Indonesia, Turkey and South Africa as the "Fragile Five," economies whose current account deficits left their currencies exposed to the withdrawal of dollar liquidity.
Two kinds of reader faced the same data that spring. The first looked at local yields relative to Treasuries and saw a spread of several hundred basis points that more than compensated for currency risk, a carry that a gradual taper would barely dent. That reasoning was sound on a two-year mean. It assumed, without saying so, that the dollar would stay inside the range of the preceding years. The second reader looked at who owned the bonds and how they had come to own them. Foreign real money had bought local paper as a substitute for zero-yielding developed debt. If developed yields rose, the substitution would reverse, and it would reverse through the currency first because currencies trade faster than bonds. The second reader was right about the next six months. Over a longer horizon, the first reader's yield cushion did much of what it promised for the stronger credits and very little for the weakest.
The structural point is that a valuation gap measured against U.S. rates is only as stable as the U.S. rate it is measured against. In 2013 the reference point itself moved by well over a percentage point, and the "attractive" spread was attractive relative to a level that no longer existed.
2018: when the average hid the dispersion
The third rhyme is quieter and more instructive for a portfolio that is tilted rather than concentrated. Through 2017 dollar-denominated emerging spreads had compressed again, local carry had performed strongly as the dollar weakened, and the asset class entered 2018 with low realized volatility and broad participation.
What unwound it was not a single global shock but a sequence of idiosyncratic ones that the market then generalized. Argentina turned to the International Monetary Fund in the late spring of 2018 after a run on the peso. Turkey followed in August, when the lira, which had started the year near 3.8 per dollar, briefly traded through 7. Neither country was large enough on its own to define the benchmark, but both forced holders of local debt across the index to reduce exposure at the same time, and currencies with no Argentine or Turkish problem sold off alongside them.
Hard currency holders in that episode suffered spread widening that was meaningful but survivable. Local holders saw the currency line once again do most of the damage. The lesson of 2018 is not that the local tilt was wrong. It is that a mean subsequent return computed across an index tends to be built from a handful of very bad currency outcomes and many ordinary ones, and the tilt's realized return depends on whether the bad ones arrive early or late in the holding period.
What the pattern says about a 70/30 tilt now
Across 2007-08, 2013 and 2018, a recurring sequence appears. Dollar spreads compress until they price very little default risk. Local markets look comparatively generous. Then some change in dollar conditions, whether a funding panic, a shift in Fed guidance, or a contagion from one weak credit, hits the local holder through the currency before it hits the hard currency holder through spreads. Afterward the cheapened local market often delivers the strong returns the valuation model anticipated. Not always: some currencies never recovered their pre-shock levels, and weak credits concentrated the losses.
The way 2026 does not fit the pattern is the reason the manager's tilt deserves more credit than a simple "late-cycle EM" reflex would give it. In 2007 and early 2013 both halves of emerging debt were expensive, and local cheapness was relative at best. In this quarter's readings only one half is expensive. Local rates sit in the cheapest quartile against a U.S. 10-year already at 5.00%, which means the valuation gap is measured from a high reference rate rather than from the 1.6% base that made 2013 so fragile. A repeat of the 2013 mechanism would need U.S. yields to rise substantially from a level that is already restrictive. That is a heavier lift than it was then, and the note's own reading that USD duration is slightly cheap points the same way.
The positive 2s10s slope of 31 basis points cuts in two directions, and the read should not pretend otherwise. A steepening curve with the long end near 5% can reflect term premium being rebuilt rather than growth being expected, and term-premium shocks have historically tightened global dollar liquidity in ways that land on emerging currencies first. That is the specific channel through which this quarter's setup could rhyme with 2013 despite the better starting valuations. It is an inference from the curve shape, not something the manager's note claims.
There is also a limitation in the evidence the note offers. The return figures are mean outcomes by quintile and quartile, and the note does not report their dispersion, their worst cases, or how many independent episodes sit inside each bucket. A +4.1% mean differential built from a small number of cycles can be dominated by one or two recoveries such as 2009. The 71 basis point excess spread is a model construct whose fair-value anchor the excerpt does not define, so it should be read as a ranking within the manager's own history rather than as an absolute statement about default compensation.
The hard currency side of the call is where conviction is easiest. A first-quintile spread that has kept tightening offers a poor asymmetry: limited room to compress further and a historical mean that is negative after the risk-free rate. Cutting that leg to 30% of a blended book removes the part of the portfolio with the least to gain. The local leg is where the episode history earns its keep. The tilt is right on the two-year arithmetic, and the earlier episodes say how it tends to be tested: through the currency, early, and across the index at once.
The sequence that recurs
Emerging debt does not usually reprice its two halves together. The dollar half tends to look safest exactly when it pays least, and the local half often looks cheapest just before a dollar shock makes it cheaper still. Valuation signals that point to local at such moments have frequently been right about where returns end up and silent about the drawdown required to get there. The pattern worth carrying forward is the order of events, not the numbers: spreads compress, the currency breaks, the carry repays the patient. A 70/30 tilt is a wager on the last step, and its holder should size it for the second.
