Click on the title to download the current (or preprint) version of a paper from SSRN.
Is Fraud Contagious? Co-Worker Influence on Misconduct by Financial Advisors, with William C. Gerken and Stephen G. Dimmock
The Journal of Finance, June 2018 Vol 73, Issue 3, Pages 1417-1450.
DOI:10.1111/jofi.12613 (links to published version)
We show that the propensity to commit financial misconduct is transmitted through social networks. We use a novel dataset consisting of financial advisors in the U.S., which includes employment histories and records of fraudulent behavior. We base our identification on changes in career networks caused by brokerage firm mergers; we include merger fixed effects in order to exploit the variation in changes to career networks across different branches of the same firm. Using this identification, we show that interacting with fraudulent co-workers increases the propensity to commit misconduct. Further, we show that this influence is stronger when the co-workers are demographically similar or when a fraudulent co-worker is in a supervisory position.
Presented at the Jim & Jack Conference (UK/UT conference in Spring 2014), FMA 2014, MFA 2015, SFS Cavalcade 2015, WFA 2015, European Finance Association 2015, the Federal Reserve Bank of New York's and JAE Conference on the Economics of Culture (2015, Invited), and Federal Reserve Bank of New York Conference on Culture, Performance and Financial Stability (2015, Invited) and AFA 2016.
Abstracted in Notre Dame's Center for the Study of Financial Regulation's Review of Financial Regulation Studies Winter 2015 Issue 14.
Various Versatile Variances: An Object-Oriented Implementation of Clustered Covariances in R, with Achim Zeileis and Susanne Köll
Journal of Statistical Software, 2020 Vol 95, Pages 1-36.
The current version is also available within the sandwich package
Clustered covariances or clustered standard errors are very widely used to account for correlated or clustered data, especially in economics, political sciences, or other social sciences. They are employed to adjust the inference following estimation of a standard least-squares regression or generalized linear model estimated by maximum likelihood. Although many publications just refer to "the" clustered standard errors, there is a surprisingly wide variety of clustered covariances, particularly due to different flavors of bias corrections. Furthermore, while the linear regression model is certainly the most important application case, the same strategies can be employed in more general models (e.g. for zero-inflated, censored, or limited responses).
In R, functions for covariances in clustered or panel models have been somewhat scattered or available only for certain modeling functions, notably the (generalized) linear regression model. In contrast, an object-oriented approach to "robust" covariance matrix estimation -- applicable beyond lm() and glm() -- is available in the sandwich package but has been limited to the case of cross-section or time series data. Now, this shortcoming has been corrected in sandwich (starting from version~2.4.0): Based on methods for two generic functions (estfun() and bread()), clustered and panel covariances are now provided in vcovCL(), vcovPL(), and vcovPC(). These are directly applicable to models from many packages, e.g., including MASS, pscl, countreg, betareg, among others. Some empirical illustrations are provided as well as an assessment of the methods' performance in a simulation study.
Financial advisors who have a fiduciary duty receive more than 2/3 of all formal customer complaints despite being slightly fewer in number than brokers. I show that the products and services their employers offer explains this difference; advisors at advisory firms and mutual fund dealers receive more complaints than brokers and advisors at insurance and private placement firms receive fewer. The higher incidence of complaints to advisors is not explained by selection at the beginning of a career, location or proxies for firm oversight and is similar in states that impose fiduciary duties uniformly on both groups. Because complaints appear to be related to product and service lines rather than the fiduciary or suitability standards, I find little evidence that imposing a fiduciary standard on brokers will address financial fraud.
Abstracted in Columbia Law School's Blue Sky Blog on October 13, 2015.
I examine whether the difficulty of estimating parameters, which is generally referred to as parameter uncertainty or estimation risk, is itself a priced risk. I primarily consider parameter uncertainty in the context of factor models such as the CAPM and Fama-French four factor model, using t-scores of estimated betas as a proxy for estimation risk. I show that this form of estimation risk is priced. I then construct a monthly risk factor from hedge portfolios and show that this factor is priced and has explanatory power. My results suggest that parameter uncertainty is an important component of market prices and can help to explain some of the alphas reported in previous studies.