Weak auctions on both shores on the same day: the US paid 5.033% on 5-year notes, Germany could not fill the Bobl; term premium or rate expectations?

23 September: a same-day demand test on both shores

On 23 September the US Treasury sold 70 billion dollars of 5-year notes at 5.033%. With the when-issued yield at 5.002% before the auction, this produced a 3.1 basis-point tail; the average for the last six auctions was 0.6 basis points. The bid-to-cover ratio fell to 2.21 (one-year average 2.33), and indirect bidding, which represents foreign central banks, to 54.31% (average 65.2%, previous auction 61.51%). Primary dealers took 15.77% of the auction, roughly 11 billion dollars. The previous auction on 26 August had cleared at 4.393%; the difference is 64 basis points. Sources give the auction size as two different figures, 70 and 80 billion dollars.

On the other side of the Atlantic the picture is similar. According to Bundesbank data, the 22 September Bobl auction drew 4,530 million euros in bids against a 5,000-million-euro target, 3,735 million euros were allotted and the average yield was 3.28%; the Federal Finance Agency retained 25.3% of the increase for its own account. On 23 September the 2047 bond was allotted at 3.78% with a cover ratio of 2.1 and the 2056 bond at 3.80% with 1.6; 1.65 billion of the 2 billion euros of supply was allotted.

Neither finding should be overstated. In German auctions the agency's retention of part of the issue is a routine tool, and a cover ratio of 1.2 does not on its own amount to a buyers' strike. What is meaningful is that two major issuers ran into weak demand at the same maturity, 5 years, within 48 hours.

From the front end to the far end? What the curve says

Until 22 September tightening was working through price at the front end: the 2-year auction cleared at 4.787%, 58.3 basis points above August's 4.204%, but its bid-to-cover of 2.63 was ordinary. On 23 September, however, yields rose across the whole curve: according to Wolf Street data the 2-year rose 13 basis points to 4.90%, the 5-year 18 basis points to 5.03%, the 10-year 13 basis points to 5.10% (the highest since June 2007) and the 30-year 9 basis points to 5.39% (the highest since July 2004).

This distribution narrows the answer. The largest move was at the 5-year maturity, where the auction took place; the 2-year and 10-year shifted in parallel by the same 13 basis points, and the 2s10s spread stayed at 20 basis points. Were term premium moving out to the far end, the 30-year would be expected to lead; instead it rose only half as much as the 5-year. The Treasury's same-day announcement of a nominal 6-billion-dollar buyback of 20–30-year bonds may also be partly supporting the far end. The evidence shows tightening spreading from the front end into the belly; it does not yet show it reaching the far end.

Germany is a partial counter-example to this reading. There weakness also appeared at the far end: the 2056 maturity's cover of 1.6 fell short of the 2047's 2.1, and the German 30-year yield rose from 3.81% on 22 September to 3.85% on 23 September. On the level of the 10-year Bund, sources conflict: one gives above 3.50%, the other 3.46%.

The competing reading: policy expectations, not term premium

The trigger for the yield jump was a data release. S&P Global's flash composite PMI rose from 56.0 in August to 58.4 in September, its highest since July 2021; services came in at 58.7, manufacturing at 57.0, and input costs rose at the steepest pace in four years. The same day, the probability of a hike at the 27–28 October meeting on CME FedWatch rose from about 55% to 70%; a month earlier it had been about 9%. Fed Governor Barr said further hikes will likely be needed, and according to Reuters 16 of 18 Fed officials project at least one more hike by year-end.

On this reading, weak demand stems from investors' reluctance to commit to the 5-year maturity before a new hike is fully priced; the rise in yields is not extra compensation for duration risk but an upward shift in the expected policy path. Markets are pricing roughly 93 basis points of Fed tightening by the end of 2027. The same mechanism is at work in Europe: the composite PMI rose to 53.1, and the probability of an ECB hike in October rose from about 45% to 48%.

The available data are consistent with both readings and are not enough to declare a regime change: there is a single day and three auctions in two countries; the result of the 24 September US 7-year auction and the 23 September CDS closes were not available at the time of writing. The discriminating test is whether the 30-year breaks above 5.39% while the 2-year yield holds steady.

Divergence: 58.4 against 53.1, and the dollar

The growth gap between the two shores was priced through the currency. When the US composite PMI came in at 58.4 and the euro area's at 53.1, the euro fell 0.3% to about 1.1412 dollars. The dollar index rose above 101 for the first time since late July, and the dollar/peso rose 1.25% to 17.50.

This divergence can reach emerging-market borrowing through three channels. The first is the benchmark rate: dollar-denominated sovereign debt is priced as the US yield plus a premium, and the 5-year auction yield has risen 64 basis points in a month. The second is the dollar: an index above 101 increases the dollar burden of borrowers earning in local currency. The third is the risk premium; this channel was not active as of 22 September. That day 5-year CDS fell from 84.77 to 84.04 in Mexico, from 121.12 to 119.75 in Brazil and from 125.73 to 122.88 in South Africa.

Transmission to Türkiye: which channel is open

Türkiye's 5-year CDS fell to 239.61 basis points on 22 September; it had been 242.43 on 21 September, leaving 10.39 basis points to the 250 threshold. The 23 September close was not available at the time of writing. As a rough indicator, the sum of the 5-year US yield (5.03%) and CDS (2.40 points) comes to about 7.43%; this is not an actual eurobond yield, only a calculation to separate the shares of the base rate and the premium. In this calculation the weight of the past month's increase lies in the base rate.

On the lira side, the 10-year yield was flat at 32.52% on 22 and 23 September. The pressure is building in the currency and reserves: the dollar/lira hit a new high of 48.84 on 23 September, and according to market calculations the CBRT's gross reserves fell to 174.5 billion dollars in the week of 18 September, shrinking by 13.9 billion dollars in 4 weeks; official data will be released on 24 September. On 23 September the OECD cut its 2026 growth forecast for Türkiye from 3.1% to 2.7% and raised its inflation forecast to 31.5%. Brent futures stood at 103.31 dollars on 23 September according to investing.com; Reuters reported about 101.09 dollars the same day.

Conclusion: Türkiye's external financing cost is rising today from the base rate and the dollar, not the risk premium. The breaking point would be CDS ceasing to move together with peers such as Mexico and Brazil and rising above 250 basis points on its own.