You have Stochastic Frontier Analysis (SFA) as a quite stablished alternative to DEA.

I remember one of my master thesis juri members many years ago asking me why didn't I use it to compare results with the DEA approach.

I think this technique is parametric but there are some recent works developing a non-parametric version of it. Check in wikipedia and see for example the following paper:

http://www.nowpublishers.com/article/Details/ECO-023

Good luck.

See, e.g., Free Disposal Hull by Deprins, Simar and Tulkens (1984) or SroNED (which is semi-parametic) by Kuosmanen and Kortelainen (2011).

DEA itself is a broad term quite rich in terms of details of its application. That is why you can find, e.g., CRS-DEA, VRS-DEA or non-decreasing returns to scale-DEA dependently on your assumption about returns to scale. Also, since DEA produces biased estimates of efficiency scores, there are also DEA methods that include bootstrapping to account for that.

Other methods that come to my mind at the moment are purely parametric (e.g., deterministic frontier analysis based on corrected OLS or stochastic frontier analysis using maximum likelihood or Bayesian inference).

The Alternatives to DEA are Stochastic Frontier Analysis. Even you may combine the two approaches under SDEA i.e., Stochatic Data Envelopment Analysis Framework. StoNED is also a solution. All the best

Basically all frontier methods can be splitted into parametric and non-parametric methods. The latter accurately calculate the scores of efficiency on the basis of empirical (in form of piecewise envelop) efficiency frontier built on observed objects of analysis. On the contrary, parametric methods stochastically estimate the efficiency scores and the user has to a priori set the functional form of transformation resource (inputs) into results (outputs). SFA is probably the best solution, although the choice would depend on the particular case you have to analyze.

There is a stochastic frontier function(based upon truncated production function or cost function. See Greene-Econometrics)  that you can use in lieu of DEA. Since the estimation of a stochastic frontier is a statistical technique,  it requires good number of observations. Or if you have observations at two points of time(sufficiently distant points to justify technical change), you can compute  tornqvist index.

Just a quick note regarding the number of observations. Being a nonparametric method DEA requires far more observations than SFA. That is the price you pay for not specifying a functional form. See Fried Lovell Shmidt (2008).

THE MALMQUIST INDEX OF TOTAL FACTOR PRODUCTIVITY. This index contains two components: Total factor productivity, technical efficiency. See Page 71 in article,

Productivity Growth, Technical Progress, and Efficiency Change in Industrialized Countries, By Fare, GROSSKOPF, NORRIS, AND  ZHANG*

Tornqvist index used for productivity analysis and assumes that the productivity change equals technical progress i.e., there is no inefficiency in production process. Erm Erm wants to calculate efficiency not productivity.

There are two basic approaches to measuring efficiency and they are based either on parametric or nonparametric methods. A good example of parametric approach is the Stochastic Frontier Analysis (SFA), a statistical method to fit the frontier and which finds its root in econometric methods. Examples of the non-parametric methods are the Free Disposal Hull (FDH) technique and the Data Envelope Analysis (DEA) approach. Do search the literature on these approaches, along with their strengths and weaknesses.

What I meant to say was, THE MALMQUIST INDEX OF TOTAL FACTOR PRODUCTIVITY. This index contains two components:Total factor productivity, technical efficiency.

See Page 71 in article,

Productivity Growth, Technical Progress, and Efficiency Change in Industrialized Countries, By Fare, GROSSKOPF, NORRIS, AND  ZHANG*

Please note my corrections in my above replies

You can find a simple paper on this subject, a comparison between DEA and SFA in an old paper I wrote some years ago (french language): "la qualité de la justice...", in Economie Appliquée, Tome LIV, 2001, N°3, pp. 181-210.

DEA and SFA are applied to French Judicial Courts. Bibliography included.

Dear Erm,

Check the book: A Practitioner's Guide to Stochastic Frontier Analysis Using Stata. (2015). Wide selection of models covering both panel/cross-sectional data; radial/distance approach; productivity/efficiency issues  and a lot more is available/explained. 

Best, Gordana

Apology for last answer.

Its a parametric technique for estimation of efficiency known as SFA

See this paper for instance

Thank you for all answers.

find my research paper on banking efficiency using DEA in the banking sector - " Revisiting the Efficiency of Indian Banking Sector: An Analysis of Comparative Models Through Data Envelopment Analysis ".