Explore the platform

Portfolio, Investment & Restructuring

Value at Risk and Expected Shortfall for Shipping Portfolios

Reviewed 2026-11-04

Summary

Value at Risk (VaR) estimates a loss quantile at a specified horizon and confidence level. Expected Shortfall (ES), also called Conditional VaR in some market-risk literature, measures the average loss in the tail beyond the VaR threshold. In shipping they can be applied to freight exposures, vessel or equity portfolios, and - with appropriate modelling - collateral or credit portfolios.

Why this matters in ship finance

Shipping portfolios combine highly cyclical freight markets, volatile vessel values, concentrated asset classes and potentially illiquid positions. VaR provides a compact threshold measure, while ES adds information about the severity of outcomes beyond that threshold. The portfolio boundary is decisive: freight-price risk, vessel-value risk, default losses and liquidity costs should not be mixed without an explicit valuation convention.

The concept

VaR answers a threshold question: at a chosen confidence level and horizon, what loss is not expected to be exceeded under the model? ES answers a tail-severity question: conditional on being in the worst tail, how large is the average loss? Both require an explicit loss definition, horizon, valuation basis and treatment of dependence.

For vessel-loan portfolios, these measures should normally sit beside - not replace - expected credit loss, probability of default, loss-given-default, collateral and covenant analysis. The Basel Expected Shortfall framework was developed for trading-book market risk; its parameters should not be imported mechanically into ship-lending portfolios.

Quantitative formulation

VaR_alpha(L) = inf { l : P(L <= l) >= alpha }
ES_alpha(L) = (1 / (1 - alpha)) * integral from alpha to 1 of VaR_u(L) du
For continuous loss distributions: ES_alpha can be interpreted as the mean loss in the tail beyond VaR_alpha

Worked example

Suppose a hypothetical shipping portfolio has one-year VaR at 95% of EUR 18 million and one-year ES at 95% of EUR 28 million. The VaR identifies the 95th-percentile loss threshold under the model; the ES summarises the average severity of the worst 5% outcomes. These numbers are illustrative only and say nothing about whether the model is well calibrated.

Data and calibration

Portfolio positions and exposures; freight and vessel-return series; debt balances; default and recovery assumptions where credit loss is included; hedge positions; dependence structure; revaluation frequency; liquidity horizon and transaction-cost assumptions. Tail estimates should be stress-tested because extreme observations are sparse.

Practical issues and limitations

  • VaR does not describe the severity of losses beyond the selected quantile.
  • ES is more informative about the tail but can be statistically unstable when data are limited.
  • Results are highly sensitive to horizon, valuation method, distributional assumptions and correlation structure.
  • A Basel trading-book ES calibration should not be presented as a regulatory capital requirement for a vessel-loan portfolio.

How ShipFinance.ai uses this concept

ShipFinance.ai can use VaR and ES as portfolio-level overlays for market, collateral or investment exposures, with clearly separated risk boundaries. For a bank portfolio, the platform can show how tail loss changes with segment concentration, vessel-value shocks, freight stress and hedge assumptions, while preserving conventional credit metrics at the loan level.

Key takeaways

  • VaR is a loss threshold; Expected Shortfall measures average tail severity.
  • The portfolio boundary and loss definition must be explicit before the metric has meaning.
  • Tail-risk measures complement - but do not replace - credit underwriting, covenant and liquidity analysis.

Primary methodological and shipping sources