Limit Breach and Model Risk in a SARON OIS Receiver Book
The risk committee of a Swiss cantonal bank reviews a CHF 250 million seven-year receiver SARON OIS swap. The bank receives fixed 0.85% and pays compounded SARON; the trade was put on to hedge a fixed-rate asset but the asset was sold last month, leaving the swap unhedged. The bank's risk limit for the OIS book is a DV01 of CHF 150,000 per basis point; the current reported DV01 is CHF 175,000. A 50 bp parallel downward shift in SARON OIS would produce a gain, but a 25 bp bear steepening (short rates +25 bp, long rates +10 bp) would produce a loss. The treasury proposes hedging with three-month SARON futures, while the risk manager prefers paying fixed in a 10-year CHF Confederation bond future. Liquidity in SARON OIS beyond five years is thin, and the bank's model uses a normal distribution calibrated to the last two years, a period that includes negative SARON prints. The committee asks for a recommendation before the next limit breach report.
1. Assess the limit breach and the suitability of the two proposed hedge instruments.
Read the worked answer
The book is CHF 25,000 DV01 above the limit, so the breach must be escalated and reduced promptly; it is not a borderline rounding issue. A seven-year receiver swap has most of its DV01 in the seven-year area. Three-month SARON futures hedge the very short end and will leave substantial curve and duration risk; they are not a close match for a seven-year exposure. A 10-year CHF Confederation bond future has long duration but introduces credit spread, bond basis, and curve risk; it may overhedge the seven-year point and does not directly hedge SARON OIS. The better approach is to reduce the swap notional or pay fixed in a matching or laddered SARON OIS, using five-year and ten-year tenors if seven-year liquidity is thin. The committee should require a hedge-effectiveness test before execution and should not accept a hedge that leaves the DV01 above limit on a look-through basis.
2. Explain why the normal-distribution model is inadequate and propose stress scenarios and controls.
Read the worked answer
A normal distribution calibrated to the last two years may understate tail risk and regime shifts. The period includes negative SARON prints, but a two-year window may still be too short to capture a sharp hiking cycle, liquidity stress, or a change in the SNB's reaction function. Normal models also assume stable correlations, which can break down in a steepening or flattening move. The committee should add stress tests rather than rely on a single distribution: parallel shifts of plus and minus 100 bp, bear steepening and bull flattening scenarios, a return to negative SARON, and a liquidity scenario in which the seven-year OIS market cannot be hedged without a wide spread. Controls should include independent model validation, daily limit monitoring with escalation, a documented hedge-effectiveness policy, and a contingency plan to reduce notional if hedging instruments are illiquid. The model output should be treated as one input, not the sole risk measure.