Germany just paid 3.90% for thirty-year money

I skim them. The letters from my pension provider arrive once a year, and I file them without reading. You probably do the same.

This autumn, 2.8 million people in the Netherlands are getting one they will read.

It is a personal estimate of what their pension looks like under a new system. It comes from ABP, the fund for Dutch civil servants and teachers. ABP manages about €532 billion. It is the largest pension fund in Europe and one of the ten largest anywhere.

On 1 January 2027, ABP converts. €532 billion stops being one collective promise and becomes 2.8 million separate pots.

That sounds like a Dutch domestic story. Except for one thing.

Dutch funds have been the largest structural buyer of very long-dated euro interest-rate risk for decades. Under the old rules they had to be. A fixed promise stretching fifty years out has to be matched, so the funds bought thirty-year and fifty-year bonds and received fixed on thirty-year and fifty-year swaps.

The new rules remove the promise. No promise, nothing to match.

De Nederlandsche Bank, the Dutch central bank, put a number on it. It estimates the funds will cut their holdings of government bonds and interest-rate swaps maturing in twenty-five years or more by roughly €100 billion to €150 billion.

For scale, DNB notes there is about €900 billion of such bonds outstanding, and the net swap position is over €300 billion.

So the buyer is leaving. But look at what that money already costs while he goes.

On 16 September the German government sold more of the Bund that matures in August 2056. The average yield came in at 3.90%. A month earlier, the thirty-year sale cleared at 3.65%.

Germany is the safest borrower in the euro area. It now pays 3.90% for thirty-year money.

I have no idea whether the conversion drives the next move or just sits underneath it. The January 2026 wave went through without disruption. The IMF looked at it afterwards and found the funds hedged more than expected, not less.

That is the interesting part. The rehearsal went fine. The main act is 101 days away.

The Netherlands is converting its entire occupational pension system from fixed promises to individual pots. It is the largest change the country has made to its pensions, and the law sets a final deadline of 1 January 2028.

By July, 44 funds had regulatory approval and 42 had already gone across. Those included three of the five largest schemes in the country: PFZW, PMT and bpfBOUW.

Seven million of the thirteen million people in the system are already in the new arrangement, according to the Dutch pension federation.

Nothing broke. Transaction costs on the January hedging adjustments came in well below what the market had expected, and the IMF found afterwards that the funds held more interest-rate protection through the switch, not less.

So the market relaxed. The one still to come is ABP, the civil service fund, with about €532 billion and 2.8 million members. It is roughly a third of the whole system and it goes on 1 January 2027.

And here is where it spreads. The Dutch funds are not only a Dutch story, because what they hold is euro duration.

De Nederlandsche Bank expects the funds to reduce holdings of government bonds and swaps of twenty-five years and longer by roughly €100 billion to €150 billion. It measures that against about €900 billion of such bonds outstanding and a net swap position above €300 billion.

ING expects the swaps to go first. Thirty-year bonds yield more than thirty-year swaps, so a fund that wants to cut protection cuts the swap and keeps the bond.

The other side of the trade is supply. Germany's 2026 budget runs to €524 billion with close to €180 billion of new borrowing, and Berlin now sells fifteen-year, twenty-year and thirty-year paper at the same auctions.

More paper, fewer forced buyers. That combination shows up in the price of thirty-year money before it shows up anywhere else.

Meanwhile, traders have been positioning for this for two years, and the tell is how the euro curve reacts to administrative news.

ING notes that when a parliamentary approval tied to the reform went through in May 2025, the gap between ten-year and thirty-year euro rates widened by about five basis points that day. Updates on a single fund's readiness have moved it too.

Over a stretch of 2025, the swap curve steepened roughly fifteen basis points more than the Bund curve. ING reckons global spillovers explain only about a third of that.

The crowd is on the same side, waiting for the same morning.

Delays are the other half of it. Several funds have pushed their conversion back, sometimes announcing it a fortnight before the date. ING sees a fair chance that a material part of the 2027 wave slips into 2028.

So the position can be right and still lose. That is usually how these things go wrong.

Three more things worth keeping track of.

The private credit default rate depends on who counts it. Fitch's measure of US private credit defaults reached a record 6.0% in April. Proskauer's index, built on senior secured and unitranche loans, came in at 2.51% for the second quarter across 716 loans worth $195.6 billion. Same market, two answers, and the difference is definitions rather than data.

Borrowing to hold shares got more expensive while shares went up. Citadel Securities reported that one-month equity financing spreads reached as much as 138 basis points over SOFR in the first half of this year. It put that down to concentrated positioning, more demand for leverage, and less dealer balance sheet to supply it.

American large caps came apart last week. In the week to 18 September the Dow fell 1.65%, a third losing week in a row and its worst since March, while the Nasdaq 100 rose 0.95%. That is about 260 basis points between two indices that normally travel together. We'll see.

Autumn 2022. I kept hearing the same story from people in England. A mortgage offer withdrawn. A rate that was there on Monday and gone by Thursday.

They thought it was about the government. It was mostly about pension funds.

British company schemes had promised fixed payments decades ahead. To match those promises they used borrowed money and swap contracts. The industry called it liability driven investment.

The arrangement works one way. When long government bond yields rise, the hedge loses money, and the counterparty asks for cash that day.

On Tuesday 27 September, thirty-year gilt yields fell 20 basis points in the morning. By that evening they had risen 67 basis points from where they started the day.

Fund managers told the Bank of England that at those levels several funds would fall into negative net asset value and would begin winding up the following morning.

The funds were not speculating. They were hedging.

The Bank's own account of that week gives the scale. What it was hearing implied at least £50 billion of long-dated gilts for sale in a short space of time. Average trading volume in those maturities was about £12 billion a day.

Bank staff worked through the night designing something. The intervention came late on Wednesday morning, 28 September. Thirty-year yields fell more than 100 basis points on the announcement.

The Bank offered to buy up to £5 billion a day for thirteen days. It bought £19.3 billion in the end and stopped on 14 October. The Treasury underwrote the losses.

British pension funds owned about 28% of the gilt market at the time. That is the part I keep returning to. The biggest owner of an asset became a forced seller of it, and the cause was the hedge rather than the asset.

Nobody had written that down as a risk. It sat on the prudent side of the portfolio.

A hedge is still a position. Positions have to be funded. We'll see.