Chile’s AFP interest rate swaps created large, correlated exposure under relative performance incentives that discouraged individual administrators from reducing risk, even as absolute exposure became increasingly dangerous.
As of early November 2026, Chile’s private pension administrators (Administradoras de Fondos de Pensiones, or AFPs) are facing substantial negative market valuations on their US interest-rate derivatives.
Foreign interest-rate swaps alone had recorded negative valuations of approximately $7.7 billion at the end of September. Including domestic swaps, the Superintendencia de Pensiones subsequently put the combined negative valuation at approximately $10.6 billion.
On April 9, the regulator introduced stricter rules governing derivative exposure and liquidity requirements, with full compliance scheduled for November 4. The intervention followed repeated warnings about growing concentrations of interest-rate risk.
According to reports, AFP executives had privately approached the regulator seeking restrictions on the very instruments they continued to use. Theirs was a problem of collective action; reducing exposure independently risked underperforming competitors that remained invested.
This was not simply a story of incorrect interest-rate forecasts. It was a failure of incentives and collective risk management.
How the funds built their positions
According to reports, Habitat was the first fund to enter the US interest-rate swap market, signing its first contract with Goldman Sachs for its Fund E in December 2013. This came shortly after the Federal Reserve’s taper tantrum triggered a sharp rise in US Treasury yields, although there is no clear evidence that the episode directly prompted the trade. Habitat’s initial positions remained relatively small, and the strategy expanded more significantly in the following years.
Capital began using the swaps in 2020, followed by Cuprum in December 2022, Provida and Modelo in the second half of 2023, and PlanVital and UNO in early 2024.
The strategy involved receiver swaps, which gained value when long-term US interest rates fell and lost value when they rose. These contracts allowed AFPs to increase their portfolios’ duration without purchasing equivalent quantities of long-term bonds. The positions initially proved profitable, particularly as US interest rates declined.
However, exposure expanded sharply during the second half of 2024. By December, the regulator estimated that a 40-basis-point increase in interest rates would reduce each Fund E portfolio value by approximately 2.1%, around four times the sensitivity recorded a year earlier.
The subsequent rise in US Treasury yields exposed the vulnerability. Citigroup analysis cited in reports indicated that by April 2026, the system’s sensitivity to foreign interest rates had reached approximately $131 million per basis-point movement, compared with approximately $10.6 million in early 2024.
Ignoring the warnings
The regulator had repeatedly raised concerns about derivative concentrations and their contribution to portfolio duration.
In January 2025, it issued Oficio No. 821, requiring administrators to strengthen their risk-management policies. The instructions were particularly revealing. They said that acceptable risk limits “cannot be limited to references to minimum returns or the relative position against the rest of the system”.
Superintendent Joaquín Cortez subsequently told legislators that the regulator had repeatedly flagged the problem and that he would have expected the administrators themselves to take action.
In an interview, León Fernández de Castro, executive president of the Association of AFPs, explicitly defended the use of derivatives and the long-term investment strategy.
He argued that swaps were not short-term bets, but tools to manage the duration and cash flows of pension portfolios. He said the shortage of long-dated assets in Chile, following the pandemic and pension withdrawals, had led AFPs to combine shorter-term foreign fixed-income instruments with derivatives to create longer-duration exposure.
The more troubling contradiction was that executives reportedly sought binding regulatory restrictions privately while continuing to maintain their positions. The apparent explanation was that no administrator wanted to reduce exposure unilaterally and sacrifice relative performance against competitors.
The herd effect and the illusion of near-zero relative risk
Under Chile’s pension system, each administrator has five investment fund categories, from Fund A, the riskiest, to Fund E, the most conservative.
Under Article 37 of Decree Law 3500, AFPs must meet minimum real-return requirements measured over rolling three-year periods. For Funds C, D and E, the minimum return is set relative to the industry average, with a permitted shortfall determined by a statutory formula. If an AFP falls below the required level, it may have to make up the difference from its own capital.
This creates a strong incentive for administrators to avoid underperforming their competitors.
The problem arose because the major AFPs increasingly held similar receiver swaps. If one administrator reduced its exposure while others maintained theirs, it risked underperforming if US interest rates subsequently fell and the swaps gained value.
But if everyone maintained or increased their positions, no individual administrator necessarily appeared particularly risky compared with its peers.
This created a dangerous disconnect between relative and absolute risk. As all the funds increased their exposure, their sensitivity to US interest rates grew substantially. Yet because they held similar positions, their returns tended to move together, keeping differences in performance relatively small.
This allowed the industry to accumulate enormous exposure to the same market risk while each administrator appeared relatively safe compared with its competitors. The perceived risk of holding additional swaps was therefore much smaller than the actual risk to the pension system as a whole.
The reported private requests for regulatory restrictions illustrate the resulting coordination problem. Administrators could recognise the collective danger but were reluctant to act independently for fear of underperforming against their competitors.
Conclusion: When relative risk conceals absolute risk
Chile’s AFP swap episode illustrates how competition under relative-performance evaluation can distort institutional risk management.
When managers are judged mainly against their peers rather than by absolute risk limits, they can become overexposed to the same risks while still appearing to perform well individually.
Chile’s approximately $10.6 billion negative valuation across domestic and foreign interest-rate swaps demonstrates the potential cost of that disconnect. A framework designed to prevent individual underperformance can, under certain conditions, encourage collective risk-taking.
