Interest Rate Stress Testing for Fixed Income Portfolios: Parallel Shift, Twist, and Historical Scenarios

Why Stress Testing Goes Beyond VaR

Value at Risk is a statistical measure calibrated to recent market behavior. It works well when the next period looks like the past. It fails in three ways that stress testing addresses directly: (1) tail events are by definition underrepresented in historical data — a 99% VaR is a 1-in-100 event, and you cannot calibrate this reliably from 250 trading days of data; (2) VaR linearizes price sensitivity through duration, missing the convexity gains and losses that are material for large rate moves; (3) VaR does not capture the correlations between risk factors under stress — rate/spread correlations in normal markets behave very differently than in a 2008 or 2022 crisis.

Stress testing is mandatory under multiple regulatory frameworks: DFAST (Dodd-Frank Act Stress Testing) for US banks above $100B assets, ILAAP (Internal Liquidity Adequacy Assessment Process) under CRD IV/V for European banks, ICAAP (Internal Capital Adequacy Assessment Process), and EBA Guidelines on IRRBB (EBA/GL/2022/14). Beyond regulation, stress testing is the primary risk communication tool for fixed income portfolio managers presenting to investment committees and boards.

The Compliance and Risk templates and the Fixed Income Analysis framework both rely on scenario-based analysis as the operational foundation of risk management.

Standard Rate Shift Scenarios

The five canonical rate scenarios every fixed income risk manager must be able to run immediately:

Parallel Shifts

All yields move by the same amount across all maturities. Portfolio P&L = −ModD × P × Δy + ½ × Convexity × P × (Δy)². For large shifts, the convexity term becomes significant: a +300bp shift on a portfolio with $200M AUM, ModD 6.8, convexity 55 produces: Duration loss = 6.8 × $200M × 0.03 = $40.8M; Convexity gain = ½ × 55 × $200M × 0.03² = $495,000. Net loss = $40.3M — convexity provides less than 1.2% offset at 300bps.

• "Parallel shift stress scenarios — full KRD analysis: Portfolio KRD profile (in $/bp): 2Y = $18,500, 3Y = $9,200, 5Y = $42,800, 7Y = $21,600, 10Y = $35,400, 20Y = $8,100, 30Y = $5,200. Total rate DV01 = $140,800/bp. Run all five parallel shift scenarios: +100bp, +200bp, +300bp, −100bp, −200bp. For each scenario: (1) First-order P&L = total DV01 × Δy in bps; (2) add convexity correction using portfolio convexity = 48.5: ΔP_convexity = ½ × 48.5 × $250M × (Δy)²; (3) compute total P&L as percentage of $250M portfolio; (4) present results in a table: Scenario | Δy | DV01 P&L | Convexity P&L | Net P&L | % of AUM. At what parallel shift does the portfolio lose more than 10% of AUM?"

Bear Steepener

Short-end rises sharply, long-end rises less. Classic pre-tightening scenario. 2Y +150bps, 5Y +100bps, 10Y +75bps, 30Y +50bps. Hurts portfolios with short-duration overweight; benefits receive-fixed short-maturity swap positions.

• "Bear steepener scenario analysis: Using the same KRD profile (2Y = $18,500, 3Y = $9,200, 5Y = $42,800, 7Y = $21,600, 10Y = $35,400, 20Y = $8,100, 30Y = $5,200). Bear steepener scenario: 2Y +150bps, 3Y +125bps, 5Y +100bps, 7Y +85bps, 10Y +75bps, 20Y +60bps, 30Y +50bps. Calculate: (1) P&L at each tenor bucket: KRD_i × Δy_i × 100 (since KRD is in $/bp), (2) total portfolio P&L, (3) decompose P&L: how much comes from the front-end (2Y-3Y bucket) vs. intermediate (5Y-7Y) vs. long-end (10Y+)?, (4) compare to the equivalent parallel shift (+87.5bps = average of the bear steepener moves weighted by KRD) — is the bear steepener worse or better than the equivalent parallel shift for this portfolio?, (5) what positioning would reduce bear steepener losses without materially increasing bull flattener exposure?"

Bull Flattener

Long-end falls more than short-end. Flight-to-safety scenario. 2Y −25bps, 5Y −50bps, 10Y −75bps, 30Y −100bps. Typically benefits long-duration positions and hurts floating-rate notes and short-duration portfolios relative to their benchmarks.

• "Bull flattener — compare three portfolios: Portfolio A (barbell): 40% in 2Y bonds, 60% in 30Y bonds. Portfolio B (bullet): 100% in 10Y bonds. Portfolio C (laddered): equal weights across 2Y, 5Y, 10Y, 20Y, 30Y. All three portfolios have $100M AUM and approximately equal modified duration of 7.5 years (confirm this). Bull flattener scenario: 2Y −25bps, 5Y −50bps, 10Y −75bps, 20Y −88bps, 30Y −100bps. For each portfolio: (1) approximate KRD distribution implied by the portfolio structure, (2) P&L under the bull flattener, (3) P&L as % of AUM, (4) P&L relative to the benchmark (Portfolio B, bullet at 10Y). Which portfolio structure best captures the bull flattener? This is the classic reason why a bullet 10Y portfolio outperforms a barbell in a flight-to-quality rally."

Twist and Butterfly Scenarios

• "Yield curve twist and butterfly stress scenarios: Same KRD profile as above. Run two additional scenarios: (A) Twist: 2Y +100bps, 5Y +50bps, 10Y 0bps, 20Y −25bps, 30Y −50bps. (B) Butterfly (mid-curve dip): 2Y 0bps, 5Y −75bps, 7Y −50bps, 10Y 0bps, 30Y 0bps. For each scenario: (1) compute P&L by tenor bucket and total, (2) identify which scenario produces a larger P&L impact on this portfolio, (3) what type of fixed income investor is most exposed to the butterfly scenario? (Answer: investors with concentrated exposure in the 5Y-7Y belly of the curve, e.g., many liability-driven investment portfolios matched to pension liabilities in the 5-10Y range), (4) in the twist scenario, a portfolio that is long the 2Y and long the 30Y (barbell) is particularly exposed — explain the mechanism using the KRD attribution."

Historical Stress Scenarios

Historical scenarios use actual rate moves from named market events applied to the current portfolio structure.

• "1994 rate shock scenario: In 1994, the Federal Reserve raised the Fed Funds rate from 3.0% to 5.5% in 12 months (February 1994 to February 1995). The 2-year Treasury yield rose approximately +250bps, 5-year +220bps, 10-year +185bps, 30-year +160bps over the same period. Monthly moves were highly concentrated — February and March 1994 alone saw the 10Y move +80bps. Apply this scenario to a $500M fixed income portfolio: KRDs are 2Y $85,000/bp, 5Y $140,000/bp, 10Y $95,000/bp, 30Y $32,000/bp. Calculate: (1) total P&L loss over the 12-month horizon, (2) worst single-month loss (use February 1994: assume 2Y +55bps, 5Y +50bps, 10Y +40bps, 30Y +30bps), (3) compare the 1994 rate shock to the 2022 rate shock for the same portfolio (2022: 2Y +369bps, 5Y +290bps, 10Y +237bps, 30Y +175bps), (4) what does this comparison tell you about the relative severity of the two tightening cycles?"
• "2013 Taper Tantrum scenario: In May–August 2013, following Ben Bernanke's May 22 taper announcement, the 10Y Treasury yield rose from 1.63% to 2.99% (+136bps in ~3 months). The move was concentrated in intermediate-to-long tenors: 2Y +30bps, 5Y +100bps, 10Y +136bps, 30Y +120bps. This was a sell-off driven by duration extension risk in MBS and real-money rebalancing. Apply to a $300M IG corporate bond portfolio with spread duration partially offsetting: KRDs (Treasury): 2Y $12,000, 5Y $38,000, 10Y $52,000, 30Y $18,000. Spread DV01: $62,000/bp (IG OAS also widened +60bps during Taper Tantrum on flight-to-quality reversal). Calculate: (1) rate VaR loss from KRD × Δy attribution, (2) spread loss from OAS widening +60bps, (3) total loss, (4) what portion of the loss was rate-driven vs. spread-driven?, (5) how would a 25% portfolio hedge using 10Y Treasury futures (short) have affected the outcome?"
• "2022 rate shock — the fastest tightening since 1981: In 2022, the Fed raised rates 425bps (March 2022 to December 2022). Yield moves: 2Y +369bps, 5Y +290bps, 7Y +260bps, 10Y +237bps, 20Y +205bps, 30Y +175bps. Simultaneously, IG OAS widened approximately +80bps and HY OAS widened approximately +290bps. Apply to a multi-sector bond portfolio: Government bonds $400M (KRD 10Y $120,000/bp), IG corporates $200M (KRD 10Y $60,000/bp, spread DV01 $52,000/bp), HY bonds $100M (KRD 5Y $22,000/bp, spread DV01 $45,000/bp). Calculate: (1) government bond loss (rate only), (2) IG corporate loss (rate + spread), (3) HY loss (rate + spread), (4) total portfolio P&L, (5) total loss as % of $700M AUM, (6) what was the approximate 2022 full-year return of the Bloomberg US Aggregate Bond Index? (Answer: approximately −13%, the worst calendar year on record for US investment-grade bonds.) How does your model result compare?"

IRRBB: Economic Value of Equity Scenarios

The IRRBB (Interest Rate Risk in the Banking Book) framework (Basel BCBS 368, EBA/GL/2022/14) requires banks to report EVE (Economic Value of Equity) sensitivity under the six standardized shocks. EVE = PV(Assets) − PV(Liabilities), and ΔEVE measures the change in this equity-equivalent value under a rate scenario. Banks must hold capital if ΔEVE under any standardized shock exceeds 15% of Tier 1 capital (the EVE outlier test).

• "IRRBB EVE calculation for a bank's banking book: Assets: $800M fixed-rate mortgage loans (average life 7.5 years, effective duration 6.2Y, weighted average rate 5.20%); $200M floating rate commercial loans (duration 0.25Y, SOFR-linked). Liabilities: $600M fixed-rate retail deposits (average repricing 2.2 years, duration 2.0Y, cost of funds 1.80%); $300M wholesale funding CDs (average maturity 1.5 years, duration 1.4Y, cost 4.90%); $100M equity (duration 0). Run IRRBB Parallel Shock Up (+200bps): (1) compute ΔPVA (change in PV of assets): sum of -duration_i × PV_i × 0.02 for each asset class, (2) compute ΔPVL (change in PV of liabilities): sum of -duration_j × PV_j × 0.02 for each liability class, (3) ΔEVE = ΔPVA − ΔPVL, (4) Tier 1 capital = $95M. Is ΔEVE/(Tier 1) within the 15% EVE outlier threshold? (5) Run the Short Rate Shock Up scenario: 3M +250bps, 1Y +200bps, 5Y +100bps, 10Y +50bps, 30Y +25bps and recompute ΔEVE."

Reverse Stress Testing: Finding the Breaking Point

Reverse stress testing asks: "What scenario causes a predefined adverse outcome?" For a bond portfolio, the natural adverse outcome is a loss of 10% of AUM, breach of a VaR limit, or a negative EVE exceeding the outlier threshold. Regulators (PRA SS31/15, EBA GL/2018/04) require documented reverse stress tests.

• "Reverse stress test — find the rate scenario that causes a 10% loss: Portfolio: $250M AUM, total rate DV01 $140,800/bp. A 10% loss = $25M. For a parallel shift, the required Δy = $25M / $140,800 = 177.6bps. For a bear steepener where the 10Y move is β times the 2Y move (β = 0.5): assume the portfolio is more exposed to the 2Y (high KRD there relative to duration-weighted average). Solve for the 2Y shock that produces a $25M loss: $18,500 × Δy_2Y + $42,800 × (0.67 × Δy_2Y) + $35,400 × (0.50 × Δy_2Y) + $5,200 × (0.33 × Δy_2Y) = $25,000,000. Solve for Δy_2Y. (1) What is the required 2Y yield shock under this bear steepener to breach the 10% loss threshold?, (2) Compare to the historical precedent — did the 2013 Taper Tantrum or 2022 rate shock produce a larger 2Y move than this threshold?, (3) What specific market event would trigger this scenario? Document as a named reverse stress scenario."

Where to Start

Interest rate stress testing for bond portfolios is foundational to regulatory capital adequacy (ILAAP, ICAAP) and investment risk governance. The Compliance & Risk templates include a full Interest Rate Stress Testing module covering all six IRRBB scenarios, EVE and NII impact calculation, and historical scenario replay. For the underlying duration and KRD analytics, see Bond Duration and Convexity with AI. For VaR methodology that complements stress testing, see Fixed Income VaR. For yield curve analysis, see Yield Curve Analysis with AI. The Quant Finance category includes MCP tools for automating scenario generation and portfolio revaluation.

Frequently Asked Questions

What is the difference between DFAST and ILAAP/ICAAP stress testing for fixed income?

DFAST (Dodd-Frank Act Stress Testing) is a US regulatory requirement for banks with $100B+ in assets, requiring annual stress tests under three macroeconomic scenarios (baseline, adverse, severely adverse) defined by the Federal Reserve. DFAST focuses on net income and regulatory capital adequacy under each scenario, across a 9-quarter projection horizon. ILAAP (Internal Liquidity Adequacy Assessment Process) and ICAAP (Internal Capital Adequacy Assessment Process) are European frameworks under CRD IV/V where banks self-design their stress scenarios and test them against their own capital adequacy and liquidity positions. For fixed income portfolios, DFAST typically focuses on NII impact (how rate scenarios affect the banking book's spread income) and capital adequacy. ILAAP also requires liquidity stress scenarios — what if the bank needs to liquidate its bond portfolio at stressed prices? This is where fixed income liquidity risk (bid-ask widening, market depth reduction) intersects with interest rate stress testing.

How is net interest income (NII) stress testing different from EVE stress testing?

EVE (Economic Value of Equity) stress testing measures the change in the present value of all future cash flows of assets minus liabilities under a rate scenario — it is a balance sheet measure that captures the long-run economic impact. NII (Net Interest Income) stress testing measures the change in interest income and expense over a short horizon (typically 12 months) as assets and liabilities reprice. EVE is more relevant for long-duration fixed income portfolios and measures structural interest rate risk. NII is more relevant for banks with significant variable-rate lending and deposit gathering — a rising rate environment may increase NII (if assets reprice faster than liabilities) even while reducing EVE (because fixed-rate asset values fall). The IRRBB framework requires reporting of both: EVE captures the balance-sheet economic perspective; NII captures the near-term earnings impact that affects P&L and dividends.

What is the supervisory outlier test under IRRBB?

The EBA IRRBB guidelines (EBA/GL/2022/14) specify that a bank is an "EVE outlier" if the maximum loss of EVE under any of the six standardized shocks exceeds 15% of Tier 1 capital. Being an outlier triggers a supervisory discussion and may result in a Pillar 2 capital add-on. As of 2024-2025, many European banks with long-duration asset books (particularly those holding significant government bonds or fixed-rate mortgages with low repricing) have been near or exceeding the outlier threshold due to the 2022 rate shock's impact on long-duration assets. The parallel shock up scenario is typically the most punishing for banks with typical asset/liability structures (long fixed assets, shorter-term liabilities). Banks use interest rate derivatives (payer swaps, interest rate caps) to manage their EVE sensitivity.

How should a fixed income portfolio manager present stress test results to an investment committee?

Investment committee presentations should lead with the dollar loss under each scenario, expressed as both a nominal dollar amount and a percentage of AUM. Include a scenario attribution table showing how much of each scenario's loss comes from each rate bucket (level/slope/curvature contribution). Compare stress results to historical precedents with dates — "this bear steepener scenario produces a $28M loss, comparable to what the portfolio would have lost during the 2013 Taper Tantrum ($24M)." Always present the reverse stress result: "A loss of 10% of AUM requires a 178bp parallel rate increase — which has been exceeded once in the past 30 years (2022)." End with the risk/mitigation: what hedges are in place, what is their cost, and how do they change the stress P&L. Claude AI can automate this presentation generation once you supply the KRD profile and scenario parameters.