Stress testing valuations using scenarios

Prev Next

A scenario is a named set of market moves, such as equities down 20%, sterling down 10% and US rates up 100bp, that LUSID applies dynamically when you value a portfolio. See how to create a scenario.

LUSID revalues every holding in full under the shifted market inputs, without touching your stored market data, and you can line several scenarios up next to your base valuation to compare them:

What is a scenario?

A scenario is a stored definition, identified by a scope and a code, that contains one or more shifts. Each shift changes one kind of input to a valuation:

  • Interest rate curves, inflation curves and credit spread curves

  • FX spot rates

  • Instrument prices and other quotes

  • Volatility surfaces

  • The parameters of certain pricing models

  • The market data rules your recipe uses, for example to value at bid rather than mid

When you apply a scenario to a valuation, LUSID locates market data in the normal way using a recipe, applies the scenario’s shifts to that market data in memory, and then revalues every holding. Your stored quotes, curves and surfaces never change.

Note the following:

  • You can mix shifts of different types in the same scenario, up to 1,000 shifts in total. See how LUSID combines shifts.

  • Each scenario has a scenario type of Historical, Regulatory or Hypothetical. The type helps you organise your scenarios, and the LUSID web app groups them by type. It has no effect on valuation, and the list of types is fixed.

  • Scenarios are versioned. Each time you save a scenario, LUSID records a new version. A valuation uses the version that was current at the valuation’s as-at time, so you can reproduce past results exactly.

Why use scenarios?

  • Stress test across asset classes at once. A single scenario can move equities, currencies, rates and volatility together, the way markets move in a crisis.

  • Revalue in full. LUSID reprices every holding under the shifted market rather than estimating the change from sensitivities, so large moves and options are handled properly.

  • Target exactly what you mean. Shift every equity, a single named instrument, or any set of instruments you can describe with a filter on asset class, currency, identifiers or instrument properties.

  • Compare side by side. In the LUSID web app, each scenario you add appears as its own colour-coded column group next to the base valuation.

  • See analytics under stress. As well as PV, you can see analytic metrics such as modified duration under a scenario.

  • Reverse stress test. You can work backwards from a target loss to find out how severe a scenario might need to be. More information.

Scenarios and risk metrics

Scenarios work well together with LUSID’s risk metrics, but are designed to answer different questions:

Scenarios

Risk metrics

Question answered

What is my portfolio worth if this particular set of moves happens?

How sensitive is my portfolio to a small move in one risk factor?

Size of move

Any size you choose

Small, standard bumps, for example 1bp

Number of factors

As many as you like, all at once

One at a time

Result

PV and other metrics, revalued under the shifted market

A sensitivity per unit of the risk factor

What can move

See the shift types

The curves, FX rates, volatility, credit and inflation data each pricing model depends on

Note: You can apply scenarios to risk metrics to stress test sensitivities in certain circumstances. More information.

Because LUSID revalues the whole portfolio under all the shifts together, a scenario captures the way moves interact. For example, consider a sample multi-asset portfolio worth 2,349,065 USD, holding US and UK shares, a US Treasury bond, an equity option and some sterling cash:

Scenario

Portfolio value (USD)

Change (USD)

Base (actual)

2,349,065

Equities −20%, banks −30%

1,933,809

−415,256

GBP/USD −10%

2,297,715

−51,350

USD rates +100bp

2,288,713

−60,353

Equity volatility +5 points

2,387,390

+38,324

All four together

1,853,879

−495,187

Applied separately, the four scenarios would add up to a loss of 488,635 USD, but applying all four together loses 495,187 USD. The difference comes from interactions, chiefly the equity option reacting to its underlying price, volatility and interest rates moving at the same time. Adding up separate sensitivities would miss this. Work through a tutorial on this.