Non-parametric estimation of operational risk losses adjusted for under-reporting

Not all claims are reported when a database for financial operational risk is created. The probability of reporting increases with the size of the operational risk loss, and converges towards one for big losses. Losses in operational risk have different causes, and usually follow a wide variety of distributional shapes. Therefore, a method for modelling operational risk based on one or two parametric models is deemed to fail. In this paper, we introduce a semi-parametric method for modelling operational risk that is capable of taking under-reporting into account and being guided by prior knowledge of the distributional shape.     

Should the insurance industry be banking on risk escalation for solvency II?

Basel II introduced a three pillar approach which concentrated upon new capital ratios (Pillar I), new supervisory procedures (Pillar II) and demanded better overall disclosure to ensure effective market discipline and transparency. Importantly, it introduced operational risk as a standalone area of the bank which for the first time was required to be measured, managed and capital allocated to calculated operational risks. Concurrently, Solvency II regulation in the insurance industry was also re-imagining regulations within the insurance industry and also developing operational risk measures. Given that Basel II was first published in 2004 and Solvency II was set to go live in January 2014. This paper analyses the strategic challenges of Basel II in the UK banking sector and then uses the results to inform a survey of a major UK insurance provider. We report that the effectiveness of Basel II was based around: the reliance upon people for effective decision making; the importance of good training for empowerment of staff; the importance of Board level engagement; and an individual's own world view and perceptions influenced the adoption of an organizational risk culture. We then take the findings to inform a survey utilizing structural equation modelling to analyze risk reporting and escalation in a large UK insurance company. The results indicate that attitude and uncertainty significantly affect individual's intention to escalate operational risk and that if not recognized by insurance companies and regulators will hinder the effectiveness of Solvency II implementation.     

Mandatory Disclosure and Operational Risk: Evidence from Hedge Fund Registration

Mandatory disclosure is a regulatory tool intended to allow market participants to assess operational risk. We examine the value of disclosure through the controversial SEC requirement, since overturned, which required major hedge funds to register as investment advisors and file Form ADV disclosures. Leverage and ownership structures suggest that lenders and equity investors were already aware of operational risk. However, operational risk does not mediate flow‐performance relationships. Investors either lack this information or regard it as immaterial. These findings suggest that regulators should account for the endogenous production of information and the marginal benefit of disclosure to different investment clienteles.     

Quantitative models for operational risk: Extremes,dependence and aggregation

Due to the new regulatory guidelines known as Basel II for banking and Solvency 2 for insurance, the financial industry is looking for qualitative approaches to and quantitative models for operational risk. Whereas a full quantitative approach may never be achieved, in this paper we present some techniques from probability and statistics which no doubt will prove useful in any quantitative modelling environment. The techniques discussed are advanced peaks over threshold modelling, the construction of dependent loss processes and the establishment of bounds for risk measures under partial information, and can be applied to other areas of quantitative risk management.¹     

Scaling models for the severity and frequency of external operational loss data

According to Basel II criteria, the use of external data is indispensable to the implementation of an advanced method for calculating operational risk capital. This article investigates how the severity and frequencies of external losses are scaled for integration with internal data. We set up an initial model designed to explain the loss severity by taking into account potential selection bias in the external data. Estimation results show that many variables have significant power in explaining the loss amount. We use them to develop a normalization formula. We develop a zero-inflated count-data model to scale the loss frequency. We compute an operational VaR and we conduct out-of-sample backtesting.     

Modelling long-term investment returns via Bayesian infinite mixture time series models

This paper introduces the class of Bayesian infinite mixture time series models first proposed in Lau & So (2004) for modelling long-term investment returns. It is a flexible class of time series models and provides a flexible way to incorporate full information contained in all autoregressive components with various orders by utilizing the idea of Bayesian averaging or mixing. We adopt a Bayesian sampling scheme based on a weighted Chinese restaurant process for generating partitions of investment returns to estimate the Bayesian infinite mixture time series models. Instead of using the point estimates, as in the classical or non-Bayesian approach, the estimation in this paper is performed by the full Bayesian approach, utilizing the idea of Bayesian averaging to incorporate all information contained in the posterior distributions of the random parameters. This provides a natural way to incorporate model risk or uncertainty. The proposed models can also be used to perform clustering of investment returns and detect outliers of returns. We employ the monthly data from the Toronto Stock Exchange 300 (TSE 300) indices to illustrate the implementation of our models and compare the simulated results from the estimated models with the empirical characteristics of the TSE 300 data. We apply the Bayesian predictive distribution of the logarithmic returns obtained by the Bayesian averaging or mixing to evaluate the quantile-based and conditional tail expectation risk measures for segregated fund contracts via stochastic simulation. We compare the risk measures evaluated from our models with those from some well-known and important models in the literature, and highlight some features that can be obtained from our models.     

Taming Systemically Important Financial Institutions

We model a systemically important financial institution that is too big (or too interconnected) to fail. Without credible regulation and strong supervision, the shareholders of this institution might deliberately let its managers take excessive risk. We propose a solution to this problem, showing how insurance against systemic shocks can be provided without generating moral hazard. The solution involves levying a systemic tax needed to cover the costs of future crises and more importantly establishing a systemic risk authority endowed with special resolution powers, including the control of bankers’ compensation packages during crisis periods.     

Quantification of Operational Risk: A Scenario-Based Approach

In this article, I identify challenges to the loss distribution approach in modeling operational risk. I propose a scenario-based methodology for operational risk assessment, which recognizes that each risk can occur under a number of wide-ranging scenarios and that association between risks may behave differently for different scenarios. The model that is developed internally in the company provides a practical quantitative assessment of risk exposure that reflects a deep understanding of the company and its environment, making the risk calculation more responsive to the actual state, ensuring that the company is attending to its key operational risks. In this model qualitative and quantitative approaches are combined to build a loss distribution for individual and aggregate operational risk exposure. The model helps to portray the company's internal control systems and aspects of business environment. These features can help the company increase its operational efficiency, reduce loss from undesirable incidents, and maintain the integrity of internal control.     

Quantitative measures of operational risk: an application to funds management

Basel II defines operational risk as the risk of direct or indirect loss resulting from inadequate or failed internal processes, people or systems or from external events. In the past decade, there have appeared a number of quantitative approaches to measuring this risk, approaches that abstract from market risk and reputational risk. The challenge is to develop operational risk measures in an asset management context where there is only limited information available about the incidence and severity of operational loss events. We survey different approaches to this problem and argue that managing this risk through operational due diligence is a source of alpha in this funds management context.     

Entrepreneurs and new ideas

We study how early‐stage new ideas are turned into successful businesses. Even promising ideas can be unprofitable if they fail on one dimension, such as technical feasibility, correspondence to market demand, legality, or patentability. To screen good ideas, the entrepreneur needs to hire experts who evaluate the idea along their dimensions of expertise. Sharing the idea, however, creates the risk that the expert would steal it. Yet, the idea‐thief cannot contact any other expert, lest he should in turn steal the idea. Thus, stealing leads to incomplete screening and is unattractive if the information of the other expert is critical and highly complementary. In such cases, the entrepreneur can form a partnership with the experts, thus granting them the advantage of accessing each other's information. Yet, very valuable ideas cannot be shared because it is too tempting to steal them.     

A Markov-switching generalized additive model for compound Poisson processes,with applications to operational loss models

This paper is concerned with modelling the behaviour of random sums over time. Such models are particularly useful to describe the dynamics of operational losses, and to correctly estimate tail-related risk indicators. However, time-varying dependence structures make it a difficult task. To tackle these issues, we formulate a new Markov-switching generalized additive compound process combining Poisson and generalized Pareto distributions. This flexible model takes into account two important features: on the one hand, we allow all parameters of the compound loss distribution to depend on economic covariates in a flexible way. On the other hand, we allow this dependence to vary over time, via a hidden state process. A simulation study indicates that, even in the case of a short time series, this model is easily and well estimated with a standard maximum likelihood procedure. Relying on this approach, we analyse a novel data-set of 819 losses resulting from frauds at the Italian bank UniCredit. We show that our model improves the estimation of the total loss distribution over time, compared to standard alternatives. In particular, this model provides estimations of the 99.9% quantile that are never exceeded by the historical total losses, a feature particularly desirable for banking regulators.     

Accurate Short-Term Yield Curve Forecasting using Functional Gradient Descent

We propose a multivariate nonparametric technique for generatingreliable short-term historical yield curve scenarios and confidenceintervals. The approach is based on a Functional Gradient Descent(FGD) estimation of the conditional mean vector and covariancematrix of a multivariate interest rate series. It is computationallyfeasible in large dimensions and it can account for nonlinearitiesin the dependence of interest rates at all available maturities.Based on FGD we apply filtered historical simulation to computereliable out-of-sample yield curve scenarios and confidenceintervals. We back-test our methodology on daily USD bond datafor forecasting horizons from 1 to 10 days. Based on severalstatistical performance measures we find significant evidenceof a higher predictive power of our method when compared toscenarios generating techniques based on (i) factor analysis,(ii) a multivariate CCC-GARCH model, or (iii) an exponentialsmoothing covariances estimator as in the RiskMetrics^TM approach.     

Internal loss data collection implementation: evidence from a large UK financial institution

The paper conducts a critical analysis of internal loss data collection implementation in a UK financial institution. We use elite semi-structured interviews, with a sample of 15 operational risk consultants from a leading international financial institution. Using content analysis, the data covers a wide range of business areas, with particular attention drawn towards the development of internal loss collection and operational risk management. The results suggest that the development of operational risk management as a function stems from external compliance (Basel II) and the internal pressure to add value to the business portfolio. This need for compliance was augmented as a driver of internal loss data collection; however, participants also recognised that the function of loss data collection is a tool of solid internal risk management and enhances managerial decision-making. The research also highlights the problems in cleansing data in order to ensure that all information implemented in the capital allocation model is valid and reliable.     

The financial crisis and hedge fund returns

The financial crisis has focused the lens of politicians and regulators on hedge funds as a source of systemic and operational risk in asset markets. We examine the extent to which available data can provide useful information regarding the impact of hedge funds on the financial system. Using data from January 1994 through September 2008, we find dramatic changes in the exposures of hedge funds to risk factors, accompanied by a significant and widespread increase in correlation between hedge fund and factor returns. Lastly, the discontinuity at zero in the cross-sectional distribution of hedge fund returns persists throughout the sample.     

A Theoretical Framework for Incorporating Scenarios into Operational Risk Modeling

In this paper, I introduce a theoretically justified framework that incorporates scenario analysis into operational risk modeling. The basis for the framework is the idea that only worst-case scenarios contain valuable information about the tail behavior of operational losses. In addition, worst-case scenarios introduce a natural order among scenarios that makes possible a comparison of the ordered scenario losses with the corresponding quantiles of the severity distribution that research derives from historical losses. Worst-case scenarios contain information that enters the quantification process in the form of lower bound constraints on the specific quantiles of the severity distribution. The framework gives rise to several alternative approaches to incorporating scenarios.     

Bayesian estimation of truncated data with applications to operational risk measurement

Data insufficiency and reporting threshold are two main issues in operational risk modelling. When these conditions are present, maximum likelihood estimation (MLE) may produce very poor parameter estimates. In this study, we first investigate four methods to estimate the parameters of truncated distributions for small samples—MLE, expectation-maximization algorithm, penalized likelihood estimators, and Bayesian methods. Without any proper prior information, Jeffreys’ prior for truncated distributions is used. Based on a simulation study for the log-normal distribution, we find that the Bayesian method gives much more credible and reliable estimates than the MLE method. Finally, an application to the operational loss severity estimation using real data is conducted using the truncated log-normal and log-gamma distributions. With the Bayesian method, the loss distribution parameters and value-at-risk measure for every cell with loss data can be estimated separately for internal and external data. Moreover, confidence intervals for the Bayesian estimates are obtained via a bootstrap method.     

What do a bank’s legal expenses reveal about its internal controls and operational risk?

Excessive (substantially above peer) litigation against a bank is indicative of operational risk because it often suggests failure to maintain a strong system of internal control. We examine the relation between bank performance and weak internal control using legal expense as a proxy. We find that legal expense is a strong determinant of loan losses and stock returns. Bank regulators should require reporting of legal expense on call reports to help identify institutions with weaknesses in internal control. Current reporting creates unnecessary information asymmetries because investors are not well informed about operational risk, leading to mispricing of bank securities.     

Sharia supervisory boards,governance structures and operational risk disclosures: Evidence from Islamic banks in MENA countries

This paper examines the impact of Sharia supervisory board (SSB) and governance structures on the extent of operational risk disclosures (ORDs), using a sample of 63 Islamic banks from 10 (i.e., Bahrain, Egypt, Jordan, Kuwait, Lebanon, Oman, Qatar, Saudi Arabia, Syria, and the UAE) countries in the Middle East and North Africa (MENA) region for the fiscal years 2006 to 2013. Drawing on Sharia compliance, Islamic banking and corporate governance literature, our findings are as follows. We find that SSB, block ownership, board independence, and country-level governance quality are statistically significant and positively associated with ORDs. Our results are robust when controlling for several bank- and country-level variables. Our study has implications for policy-makers and regulators in the MENA region with respect to the development and implementation of SSB and governance mechanisms that can improve operational risk disclosures. Finally, the findings highlight the need to enhance current understanding of SSB structures and governance mechanisms that can best help Islamic banks towards engaging in effective compliance with recent governance and accounting reforms.     

Does it help firms to secretly pay for stock promoters?

We examine deals between listed firms and promoters who have been secretly hired to increase their stock prices. This behavior by the secret promoter is illegal (and leads to prosecution) but the actions of the hiring firm are legal. We use data from these prosecutions to analyze the behavior and motivations of the hiring firms. We find that secret promotion leads to an initial increase in the price and trading volume of the firms on the date that the secret promotion started. Subsequently, however, we find that this increase in price is reversed when regulators (e.g. SEC or NASD) take action against these promoters for not disclosing their relationships with the hiring firms. We find that the main motives behind these relationships are to maximize the private benefits of the firm’s managers and owners through pumping the share prices and subsequently dumping their shareholdings.     

Harmful events and misconducts in financial organizations: Human biases and root causes

We conduct a large scope field investigation of 19 major incidents in 19 large European insurance and banking institutions, based on 116 post-event interviews with managers and top executives over a two-year period. We demonstrate the power of the Root Cause Analysis (RCA) method for detecting human biases documented by the behavioral finance, the organizational behavior and occupational psychology literatures. These biases constitute key operational risk factors these organizations. We find that organizational biases (such as a breach of psychological contract) take center stage as root causes of incidents in these organizations. We also find that banks are more exposed to emotional biases (fear and greed) and insurance companies more subjected to cognitive conservatism as root cause biases. This research has direct implications regarding how banks and insurance companies may cope with regulations that put a greater emphasis on measuring and controlling operational risk and specifically misconduct risk.