The financial crisis of 2008 led to devastating consequences such as bankruptcies and recession in the US economy. Many big banks were at the forefront owing to their risk exposures and open positions. Prior research documents that bank financial statements did not provide adequate lead indicators on the looming crisis in reducing information asymmetry. However, there is no prior research focused on the sufficiency of risk disclosures around this time period. This paper seeks to address this gap using Bank Value at Risk (VAR), a single number publicly disclosed in the annual reports of banks. Bank VAR attempts to quantify the worst possible loss the bank expects to have on its trading portfolios under normal market conditions. Using hand-collected data from the annual reports of the top twelve US banks, this study documents that the change in VAR was steady and positive until the point of the crisis and then decreased in the years thereafter. A repeated-measures analysis of variance model is used to study whether two indicators of VAR (year-to-year change in VAR and log-transformed ratio of VAR to the total trading revenue) differ from pre-crisis to the post-crisis levels. Both VAR indicators reveal an increasing trend pre-crisis and are significantly higher pre-crisis compared to post-crisis. This opens the possibility that the trend of VAR might have information content as a potential leading indicator of the crisis. The finding sheds light on efficacy of risk analysis in bank trading portfolios and could have implications for governance.
Improving the value at risk forecasts: Theory and evidence from the financial crisis