Olin Business School Study Finds Insider Information Helps Analysts at Big Banks

    ST. LOUIS, MO, April 01, 2011 /24-7PressRelease/ -- A new study from the Olin Business School at Washington University in St. Louis finds it's difficult to maintain "Chinese walls" - self-regulated firewalls - intended to keep client information from leaking between loan and investment departments when they are housed in the same financial institution.

The repeal of the depression era Glass-Steagall Act 11 years ago allowed financial institutions to engage in commercial and investment activities under the same roof.

The study "Do Bank-affiliated Analysts Benefit from Lending Relationships?" by Xiumin Martin, PhD, assistant professor of accounting at Olin, and Ting Chen of CUNY Baruch College, is forthcoming in the Journal of Accounting Research.

When the walls that had previously separated lenders from equity analysts came down, according to the new study, information about borrowers (that was private and restricted under Glass-Steagall), began to flow from commercial to equity research divisions within financial conglomerates.

Martin and Chen's research finds analysts' forecasts improve when their banking institution provides loan services to the companies they are following.

Evidence that divisions within large financial conglomerates share information and that bank-affiliated analysts benefit from the information spillover is based on analysis of sample bank loans and analyst forecasts from 1994-2007.

While prior studies have shown that commercial banks have superior information about borrowers, this paper investigates whether banks' information advantage benefits their affiliated security analysts by helping them make more accurate earnings forecasts.

"Analysts are supposed to be on the public domain side and they are not supposed to get what we call private information, or proprietary information," Martin says. "All their forecasts should be based on all the publicly available information."

Martin and her research partner focused on 16 conglomerate banks, including JP Morgan, Bank of America, Credit Suisse and Deutsche Bank, for the period of 1994-2007.They made a pre- and post-loan comparison in the accuracy of conglomerate forecasts relative to benchmark forecasts issued by the same analysts for firms that did not borrow from the affiliated bank.

Martin found that the accuracy of bank-affiliated analyst forecasts does increase following a loan inception at the same institution.

The information advantage of analysts, the study finds, is concentrated among borrowers with high credit risk and negative or bad news related to the company.

The information advantage exists only when conglomerates serve as lead financing arrangers, not merely as participating lenders.

"We think our paper probably will have some policy implications," Martin says. "The trend towards large conglomerate banks is probably going to persist. Maybe there are some benefits in housing these two divisions within one organization, but our paper provides one piece of evidence that this sharing of information probably is not fair to normal, individual retail investors."

Xiumin Martin, PhD is an assistant professor of accounting at Olin Business School. Watch her discuss her research in this video. Learn more about Olin's Masters in Accounting program at http://www.olin.wustl.edu.

About Olin Business School:
Olin Business School at Washington University in St. Louis is an institution of leaders: distinguished business faculty... exhilarated, brilliant students... and successful, energized alumni. Our 12 business degree and non-degree programs emphasize rigorously analytical, critical-thinking skills; applied learning; global competence; and communication and collaboration skills -- advancing today's business world and tomorrow's global leaders. Learn more about Olin Business School on the Web at:

Website: http://www.olin.wustl.edu/Pages/default.aspx
Facebook: http://www.facebook.com/OlinBusinessSchool
YouTube: http://www.youtube.com/user/OlinBusinessSchool

# # #