A Gentle Introduction to Effective Computing in Quantitative Research: What Every Research Assistant Should Know, by Harry J. Paarsch and Konstantin Golyaev
This book offers a practical guide to the computational methods at the heart of most modern quantitative research. It will be essential reading for research assistants needing hands-on experience; students entering PhD programs in business, economics, and other social or natural sciences; and those seeking quantitative jobs in industry. No background in computer science is assumed; a learner need only have a computer with access to the Internet. Using the example as its principal pedagogical device, the book offers tried-and-true prototypes that illustrate many important computational tasks required in quantitative research. The best way to use the book is to read it at the computer keyboard and learn by doing.
The book begins by introducing basic skills: how to use the operating system, how to organize data, and how to complete simple programming tasks. For its demonstrations, the book uses a UNIX-based operating system and a set of free software tools: the scripting language Python for programming tasks; the database management system SQLite; and the freely available R for statistical computing and graphics. The book goes on to describe particular tasks: analyzing data, implementing commonly used numerical and simulation methods, and creating extensions to Python to reduce cycle time. Finally, the book describes the use of LaTeX, a document markup language and preparation system.
Secure and Automated Enterprise Revenue Forecasting, by Jocelyn Barker, Amita Gajewar, Konstantin Golyaev, Gagan Bansal, and Matt Conners. Recepient of the Deployed Application Case Study paper at the Thirtieth Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-18).
Forecasting Sales of Consumer Devices Using Search Query Data, by Mayank Shrivastava, Konstantin Golyaev, Gagan Bansal, Matt Conners, Shahar Ronen, and Walter Sun. 2016 International Symposium of Forecasting, Santander, Spain.
Forecasting Multiple Time Series Using the baselineforecast R Package, by Konstantin Golyaev. 2016 Use R conference, Stanford, CA, USA.
Making Lemonade Out of Lemons: Model Stacking for Forecasting, by Konstantin Golyaev, Jocelyn Barker, and Gagan Bansal. 2017 International Symposium of Forecasting, Cairns, Australia.
The Economics of Mortgage Lending Regulations, bu Konstantin Golyaev.
In this dissertation, I consider various aspects of the U.S. residential mortgage lending market in 2005. In particular, I examine how existing regulations may have contributed to the mortgage default crisis that began in early 2007.
The first chapter of the dissertation is titled “The Impact of the Community
Reinvestment Act on the Home Mortgage Lending Industry”.
The Community Reinvestment Act (CRA) is a federal lending regulation that
creates incentives for depository institutions to lend in low- and
Only a subset of market areas is closely monitored by the regulators.
I exploit this CRA enforcement mechanism to identify its effect on the
banks’ loan approval decisions.
I employ a novel nonlinear Bayesian Instrumental Variables method to quantify
the above effect while admitting unobserved heterogeneity among mortgage
I find that, other things equal, loans in closely monitored areas have a
21.7 percent higher average chance of being approved.
This implies that more than
327,000 extra loans originated in 2005 in
California and suggests that banks’ responses to the CRA enforcement
mechanism sharply contradict the original CRA goals of providing credit
in all eligible neighborhoods.
Namely, CRA-induced incentives led banks to issue substantially more loans
to marginal borrowers in monitored areas.
The second chapter of the dissertation is titled “Race-to-the-Bottom In Home Mortgage Lending”. It explores the degree of strategic interactions among mortgage lenders and how these interactions differ depending on the regulatory agency. Conventional economic wisdom suggests that competition among mortgage lenders will result in overall welfare improvements. Recent theoretical research challenges this wisdom. Using the data concerning home mortgage loan applications, I test the “race-to-the-bottom” hypothesis that competition among lenders causes them to relax lending standards. I exploit the recently developed structural methods of estimating static games with incomplete information to identify how lenders form beliefs about the actions of their competitors. I find strong evidence supporting the “race-to-the-bottom” story among all types of mortgage lenders, with the exception of those regulated by the Federal Reserve System. Thus, my results provide a partial explanation for the subprime mortgage collapse of the early 2007.