News

New appointments for Kuosmanen and Kortelainen

July 2009: Timo Kuosmanen has been appointed as Professor of Quantitative Methods at Helsinki School of Economics. He starts in the new position in August 2009. Mika Kortelainen has been appointed to a lectureship at the University of Manchester, the School of Social Sciences. For more detailed news, click HERE.

Journal of Productivity Analysis invites revised version of the StoNED paper

June 2009: Timo Kuosmanen and Mika Kortelainen have revised the StoNED paper to address the comments from an anonymous Associate Editor and two reviewers of the Journal of Productivity Analysis. Novel features of the revised manuscript include a log-transformed multiplicative variant of the StoNED model, an extension to panel data, and a new set of Monte Carlo simulations. As result of these revisions, the title of the paper has been changed to "Stochastic Nonparametric Envelopment of Data: Frontier Estimation Subject to Shape Constraints (the revised version can be downloaded by clicking the title). The revised manuscript was submitted to review in June 2009.

Operations Research publishes the foundational article of Kuosmanen and Johnson

March 2009: The article titled "Data Envelopment Analysis as Nonparametric Least Squares Regression", co-authored by Timo Kuosmanen and Andrew Johnson, has been accepted for publication in the Operations Research. Kuosmanen and Johnson's article proves that DEA can be formulated as convex nonparametric least squares (CNLS) regression subject to shape constraints on frontier and sign constraints on residuals. This reinterpretation reveals the classic parametric programming model by Aigner and Chu (1968) as a constrained special case of DEA. Applying these insights, Kuosmanen and Johnson develop a nonparametric variant of the corrected ordinary least squares (COLS) method, named as Corrected Concave Nonparametric Least Squares (C2NLS). The new C2NLS estimator is shown to be consistent and asymptotically unbiased.

The least squares formulation of DEA establised in Kuosmanen and Johnson's article forms the foundation of the StoNED method. Specifically, by relaxing the sign-constraint on residuals, the DEA model transforms to the CNLS regression, which is used for estimating the shape of the frontier in the StoNED method.

Operations Research is the flagship journal of the Institute of Operations Research and Management Sciences (INFORMS), and one of the most prestigious journals in its field.

The Scandinavian Journal of Economics publishes article applying CNLS to testing inefficiency hypothesis

11 July 2008: The article titled ”Neoclassical versus Frontier Production Models? Testing for the Presence of Inefficiencies in the Regression Residuals”, co-authored by Timo Kuosmanen and Mogens Fosgerau, has been accepted for publication in the The Scandinavian Journal of Economics. This article develops nonparametric tests that enable one to test for the presence of inefficiency in the regression residuals at the sample level. CNLS regression (step 1 of the cross-sectional StoNED method) is considered as one possible estimation technique in the paper. Thus, the tests developed in this article can be readily used as specification tests for the StoNED model.

The Scandinavian Journal of Economics is one of the oldest and most distinguished economics journals in the world.

The Econometrics Journal publishes article on CNLS

2 July 2008: Timo Kuosmanen’s article ”Representation Theorem for Convex Nonparametric Least Squares” has been published in the 2nd (July 2008) issue of volume 11 of the The Econometrics Journal. This article points out the connection between DEA and the nonparametric least squares regression subject to shape constraints, referred to as CNLS. Importantly, CNLS regression forms the first step of the two-step estimation routine of the StoNED approach. Kuosmanen’s article shows that the infinite dimensional CNLS problem can be converted to an equivalent quadratic programming (QP) problem, which is solvable by standard QP solver software.

The Econometrics Journal is a relatively new journal, established in 1998 by the Royal Economic Society with the aim of creating a top international field journal for the publication of econometric research. The Royal Economic Society is one of the oldest and most prestigious economic associations in the world.

1 July 2007: StoNED has won three prizes at the 10th European Workshop on Efficiency and Productivity Analysis (EWEPA X), organized at the IESEG School of Management in Lille, France, June 27-30, 2007. Timo Kuosmanen and Mika Kortelainen were awarded the prize for the most original paper of the conference for their study "Stochastic Nonparametric Envelopment of Data: Cross-sectional frontier estimation subject to shape constraints". The jury (p. Agrell, K. Kerstens, and P. Vanden Eeckaut) selected the prize winning paper from more than 200 papers accepted for a presentation in the conference. The prize - Limdep 9 software - was sponsored by the Econometric Software Inc. The paper by Kuosmanen and Kortelainen was also awarded the 1st prize of the poster presentation in Session III, based on anonymous votes by conference participants. The prize-winning poster can be viewed in the pdf format HERE. The poster prize - the fortcoming book by Fried, Lovell, and Schmidt - was sponsored by the Oxford University Press. In addition, Mika Kortelainen was awarded with the 3rd prize of the Young Researcher (YR) session II for his presentation titled "Semiparametric Estimation of Stochastic Frontiers under Regularity Conditions: A Single-Index Approach". The prizes of YR sessions were awarded based on votes by the conference participants. The YR prize - Frontier Analyst software - was sponsored by Banxia Software. Many thanks for the sponsors for the valuable prizes, and especially to all of the conference participants who voted our papers!

StoNED homepage opens

4 June 2007: Stochastic nonparametric envelopment of data (StoNED) is a new frontier estimation framework that combines the virtues of nonparametric DEA frontier with the stochastic noise term of SFA. The purpose of the StoNED homepage is to provide the latest news and developments about the StoNED approach. From this homepage, you can download working papers, preprints, posters, presentations, and computation codes related to the StoNED methodology free of charge. Save the StoNED homepage in your favorites and stay tuned for the latest developments!