VaR is an estimate derived statistically from past data, so it is checked afterward against actual profit and loss1. With a 99% confidence level, a loss is expected to exceed VaR on 1 day in 100, or 2.5 times on average over 250 days. If the number of exceptions is well above this expectation, the model is judged to have a problem1.

Japan's capital adequacy notice calls it "バック・テスティング" (backtesting) and requires banks using the internal models approach for market risk to compare daily VaR with daily hypothetical and actual profit and loss over the most recent 250 business days2.

Notes

  1. Bank of Japan Financial System and Bank Examination Department, Center for Advanced Financial Technology, "Ⅱ．市場リスクの把握と管理" (June 2012), "４．バックテストによるＶａＲの検証" (Checking VaR by backtesting), page 39, and "バックテスト（２項検定）" (Backtesting (binomial test)), pages 42–43 (in Japanese). ↩ ↩2

  2. Financial Services Agency, "銀行法第十四条の二の規定に基づき、銀行がその保有する資産等に照らし自己資本の充実の状況が適当であるかどうかを判断するための基準" (FSA Notice No. 19 of 2006, consolidated version updated August 24, 2026), Article 271-3, paragraph 3, item 7 (page 436) and Article 275-2, paragraph 1 (page 460) (in Japanese). ↩