Home / Wiki / E / Economic Capital Model

Economic Capital Model

Actuarial and Risk SciencesUpdated 17 Nov 2026
Share

An economic capital model is an internal model that a bank, insurer or other financial firm uses to estimate how much capital it needs to remain solvent after severe but plausible losses. It combines the firm’s main risks, such as credit, market, insurance and operational risk, into a single distribution of possible losses or changes in net asset value over a set horizon, usually one year, and reads off the capital needed at a chosen confidence level. The result, economic capital, reflects the firm’s own view of its risks rather than the rules set by a regulator.

Economic capital is related to, but distinct from, regulatory capital requirements. Banks use economic capital models in their ICAAP, and insurers use them in their ORSA. Under Solvency II in the European Union and the United Kingdom, an insurer may also seek approval to use a full or partial internal model to calculate its Solvency Capital Requirement instead of the standard formula, in which case the economic capital model and the regulatory model may largely coincide.

01

How economic capital is measured

Most models simulate many scenarios, often by Monte Carlo simulation, to build a distribution of losses LL. Economic capital is then commonly defined as the loss at the chosen percentile less the expected loss, which is already covered by pricing and provisions:

EC=VaRα(L)−E[L]\text{EC} = \text{VaR}_{\alpha}(L) - E[L]
  • VaRα(L)\text{VaR}_{\alpha}(L) is the value at risk of the loss distribution at confidence level α\alpha.
  • E[L]E[L] is the expected, or mean, loss.
  • α\alpha is the confidence level, such as 99.5% or 99.9%.

For example, if a simulation gives an expected loss of 50 million and a 99.5th percentile loss of 400 million, economic capital is 400−50=350400 - 50 = 350 million. Some firms instead use tail value at risk (TVaR), the average loss in the scenarios beyond the percentile, which captures the severity of the tail better and behaves more consistently when risks are combined. Insurers often measure the change in own funds rather than accounting losses, so that both assets and liabilities are valued on a market-consistent basis.

The confidence level usually links to a target credit standing. Solvency II calibrates the Solvency Capital Requirement to a 99.5% one-year value at risk, roughly a one-in-200-year event. Banks that aim for a strong credit rating often choose a higher level, such as 99.9% or above.

02

Risk types and diversification

  • Credit risk. Defaults and downgrades of borrowers, bonds and counterparties.
  • Market risk. Changes in interest rates, equity prices, spreads and exchange rates.
  • Insurance risk. Catastrophes, reserve deterioration, mortality, longevity and lapses.
  • Operational risk. Losses from failed processes, systems, fraud and legal events.

Because these risks are not perfectly correlated, the capital for the whole firm is less than the sum of the stand-alone figures. If market, credit and insurance risk need 300, 200 and 250 on their own, a total of 750, and the combined model gives 520, the diversification benefit is 230, about 31%. How the model captures dependence, through correlation matrices, copulas or common economic scenarios, is one of its most debated parts, because dependence tends to rise in a crisis. Firms then share total capital back to business units through capital requirement attribution.

03

Uses in banks and insurers

Economic capital supports decisions across the firm. It sets the capital charge in performance measures such as risk-adjusted return on capital (RAROC), so that business lines are judged on returns relative to the risk they add. It feeds product pricing, reinsurance and hedging choices, limits and the firm’s risk appetite. Supervisors expect firms to show that their models are actually used in such decisions; under Solvency II this is formalised as the use test for internal models.

04

Validation and limitations

Economic capital models are hard to validate. A one-in-200-year loss cannot be backtested over a few decades of data, so validation relies on sensitivity analysis, benchmarking against other models and the standard formula, stress testing, reverse stress tests, profit-and-loss attribution and independent review under model risk management. Weaknesses include reliance on historical data that may understate fat tails, unstable dependence assumptions, and the temptation to tune a model to produce lower capital. The 2007–2009 financial crisis showed that losses can exceed modelled extremes, which is one reason regulators use floors, standard formulas and stress tests alongside internal models.

Last updated 17 Nov 2026
Share this entry

Ref 1020000106