The causes and consequences of operational risk: some empirical tests

Jiang, X 2016, The causes and consequences of operational risk: some empirical tests, Doctor of Philosophy (PhD), Economics, Finance and Marketing, RMIT University.

Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Jiang.pdf Thesis application/pdf 2.41MB
Title The causes and consequences of operational risk: some empirical tests
Author(s) Jiang, X
Year 2016
Abstract The thesis provides empirical evidence on the causes and consequences of operational risk. First, by including operational losses endured by firms across all sectors worldwide, we investigate the determinants that potentially explain cross-country differences in operational risk. These determinants are based on country-level information. They can be broadly classified into three categories measuring three unique dimensions of a country: macroeconomic, regulatory, and social. To circumvent model-specification issues and variable-selection bias, we carry out the empirical work according to extreme bounds analysis (EBA), which is an econometric modelling approach suggested by Leamer (1983, 1985) and further extended by Granger and Uhlig (1990), as well as Sala-i-Martin (1997).

The empirical results show that operational-loss severity, on average, rises as a country’s GDP level and the cost of living increase. In addition, a country as a whole is more likely to experience catastrophic losses with a poorer regulatory and governance standard, particularly against the background of the rigorous process by which a country’s government is selected, monitored, and replaced; and also on the capacity of that government to formulate and implement sound policies effectively. Furthermore, the overall development of a country’s citizens—including their life expectancy, education, and income levels—also plays a role when comparing operational-loss severity from one country to another.

Second, to address the consequences of operational risk, we use an event-study approach to examine the economic impact of operational-loss announcements on firms’ stock market value and the potential reputational damage that follows. We distinguish operational-loss settlement news from its initial press release to detect potential discrepancies in market reactions to the two announcement types, and we examine the effect of gradual information release. We account for the nominal amount of operational losses to separate the reputational effect of the loss announcement from its direct monetary impact, hence refining the measures of reputational risk. We scope the empirical estimation at a firm-level for 331 operational-loss events settled by commercial banks headquartered in the United States, the United Kingdom, and Canada during the period 1995 to 2008.

The findings reveal that the stock market reacts negatively to the initial press release of operational-loss events, as well as to its settlement news across all three countries analysed. This negative reaction is more abrupt surrounding the event dates, highlighting the strong initial reaction to loss news, although it fades quickly after the announcements are made to the public. This suggests that the market selloff may be short-lived. Reputational risk is consistently evident in the global and in all of the sub-regional samples, indicating that the market tends to overreact to operational-loss announcement. In addition, the market appears to be more sensitive to announcements of (i) losses resulting from internal fraud; (ii) losses of a bigger magnitude with an undisclosed loss figure; (iii) losses that result in restitutions, and (iv) losses that are consequences of regulators’ investigation.
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Economics, Finance and Marketing
Subjects Financial Institutions (incl. Banking)
Investment and Risk Management
Keyword(s) Operational risk
Operational loss announcement
Extreme bounds analysis
Reputational risk
Event study approach
Version Filter Type
Access Statistics: 1714 Abstract Views, 365 File Downloads  -  Detailed Statistics
Created: Wed, 26 Oct 2016, 14:48:15 EST by Keely Chapman
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us