Download
1 / 34
340 likes | 483 Views
Operational Risk and Reputation in the Financial Industry. Roland Gillet (Sorbonne, Solvay) Georges Hübner (ULg, UM and LSF) Séverine Plunus (HEC-ULg) . AGENDA. Basel II : Operational risk and reputational risk Litterature Review Sample : construction and descriptive statistics Methodology
E N D
Operational Risk and Reputation in the Financial Industry Roland Gillet (Sorbonne, Solvay) Georges Hübner (ULg, UM and LSF) Séverine Plunus (HEC-ULg)
AGENDA • Basel II : Operational risk and reputational risk • Litterature Review • Sample : construction and descriptive statistics • Methodology • Results • Evidence from other data • Conclusion
Operational risk • Basel II: “The risk of losses resulting from inadequate or failed internal processes, people and systems or from external events. This definition includes legal risk, but excludesstrategic risk andreputational risk.” (BIS, 2005, p140, n°644)
Reputational risk • Basel II: “the risk of significant negative public opinion that results in a critical loss of funding or customers” (BIS, 1998, p7). • More generally, reputational losses are all events that, once known by the market, negatively affect the cash-flows of the company, either due to losses in the client base, lack of confidence of external parties materializing in increased discount rate, funding rates, or decreased cash-flows. But do the markets really dissociate reputational risk from operational risk?
Literature Review (1) • Murphy et al. (2004) • Their contribution builds on a previous line of results showing significant negative price impacts of firms accused of fraudulent activities (Skantz et al., 1990; Karpoff and Lott, 1993; Reichert et al., 1996). • Results: • significant declines in reported earnings, increased stock return volatility, and declines in analyst’s estimates. • larger firms experience smaller negative impacts since losses behave as fixed costs. • A strong brand name mitigates the impacts and is interpreted as a protection against reputational damage.
Literature Review (2) • Only two papers examine the reputation impact on market returns of operational events affecting financial institutions. • Cummings, Lewis and Wei (2004) • Results: • Banks experience smaller negative impact than insurance companies. • Both types of companies however experience significant negative price reactions • market value drops exceeding the amount of the operational losses • de Fontnouvelle and Perry (2005) • Results • the announcement date only has a significant, negative impact on the price • negative price impacts are larger when the operational loss is due to internal fraud
Our study • Stock market reaction after the announcement of operational losses in listed financial companies. • 154 financial companies listed on major Stock Exchanges • Three events per firm: • First press release, • Explicit recognition by the company, and, • Settlement date. • Reputational risk: difference between the market value loss and the announced loss amount of the firm.
Sample construction • OpVantage First, provided by the Fitch Group. • criteria to filter this data collection: • company group incorporated either in United States or in Europe; • companies of the financial industry; • operational losses higher than 10 millions US dollars; • loss settled no sooner than January 1994. • companies publicly listed • “September 11th” events removed. • final sample: • 103 largest losses having occurred in American companies • 51 largest losses in European companies.
Methodology • Abnormal return for firm i: ARit = Rit – αi – βi Rmtfor t= -20 to 20 • Abnormal return due to reputational effect: • Average and cumulative average abnormal return: and ARi0(Rep) =Ri0 – αi – βi Rm0+ loss / Market Cap
Sub-sample analysis according to the knowledge of the losses – First press release
Sub-sample analysis according to the knowledgeof the losses – Recognition by the company
Sub-sample analysis according to the knowledge of the losses – Settlement
Sub-samples according to the event typeRecognition by the company
Evidence from other data • Volumes • average volume for each companies on a 250 days basis, • daily variation of the volume • -20 + 20 days around the three announcement date. • Garch • Student test : • H0: the average conditional volatility during the 10 days following the announcement is the same than the average conditional volatility of the estimation period. • H1: conditional volatilities are different.
Evidence from other data (2) • Cusum of squares • The CUSUM of squares test (Brown, Durbin, and Evans, 1975) aims at assessing the constancy of the parameters of a model and is based on the test statistic: where w is the recursive residual defined as: The expected value of S under the hypothesis of parameter constancy is :
Conclusion • CAR < 0 around the first press release and the recognition by the company date. • As far as reputational risk is concerned, it is significantly negative before the first press release, and significantly positive after. • CAR significantly positive around the settlement date • The investors overreact when they do not know about the loss size. automatic correction of the stock returns 10 days after the recognition by the company date. • if the loss is due to frauds: market reaction significantly worse and negative effect on the reputation of the company.
Conclusion Volumes variations: significant peak in trades whenever the company recognizes the loss event, which corresponded to changes in market alphas and betas confirmed by a cusum of squares test. anticipation before the first disclosure happens, correction of the settlement returns for the initial effect. The timing of the resolution of uncertainty also matters to a very large extent, especially when one has to assess at what moment the market perceives a shift in the risk profile of the financial institution that has suffered from a large operational loss.
