OCC faces VAR vetting questions over JP Morgan loss
The post-mortem into the $2 billion loss at JP Morgan’s chief investment office (CIO) is expanding to draw in the bank’s regulators – and questions about whether and when they should have been told of changes to the value-at-risk (VAR) model used by the CIO.
The unit adopted a new VAR model at the start of this year, only to switch back to the previous iteration after losses spiralled – the old model showed the unit’s $129 million average VAR during the first quarter was almost twice as high as the $67 million the bank had publicly reported.
Because the CIO is classified as the branch of a national bank, it falls under the purview of the Office of the Comptroller of the Currency (OCC) – including its London arm, where the trades took place. The OCC is only required by statute to validate models used for regulatory capital calculations, and the regulator says in an emailed statement that it “has systematic procedures in place for those approvals, which include a combination of targeted exams, periodic meetings with the bank, and ad hoc meetings and evaluation. OCC economists provide an opinion on the technical aspects of the models and examiners evaluate other aspects”.
But a former senior regulator at the OCC says it is impossible for the agency to examine in detail every model used by a large bank such as JP Morgan. “The OCC doesn’t validate every specific model, it validates the framework by which the institutions construct and validate their own models,” says Jeff Brown, a managing director at Promontory Financial Group, and former senior deputy comptroller for international and economic affairs at the OCC. “It’s a fundamental issue. For a multinational bank, we’re not talking about one risk model, we’re talking about many VAR models across the business lines and many different versions of these over time. It requires a framework and a process. When you are reviewing that framework, it is not possible to dig deeply into each model.”
Both the OCC and JP Morgan declined to comment on whether the model used to estimate the CIO’s VAR for the purposes of the bank’s earnings reports was also used to calculate regulatory capital. The bank’s annual report contains a lengthy discussion of its VAR modelling, saying at one point that the results “are reported to senior management, and they are used in regulatory capital calculations”. The report also states that the VAR calculation is based on risk factors that are selected to fit the risk profile of the portfolio in question, raising the possibility that the VAR understatement was the result of a key risk being omitted.