I want to use DEA techniques to benchmark national industrial sectors using harmonized input-output tables, assuming that output is homogeneous across countries (as long as the output is traded, it is not such an heroic assumption).
Efficiency here would look at how sectors of activities use inputs (as given by the IO tables) in producing output compared to their trade partners.
When looking at the literature, I was surprised to find only one paper directly related to this topic (Guido Cella & Giovanni Pica: Inefficiency Spillovers in Five OECD Countries, 2001).
This rings a few bells, as either I did not look where I should, or the DEA approach is not suitable (indeed, most DEA analysis are done at firm level and not on aggregated data).
So, this is my question: Has DEA been used for my purpose. If yes, where could I find the papers (if possible publicly available)? If not, why? Thanks in advance. Hubert
Dear Hubert, As far as I understand, DEA's use is not limited to firm level. Although there may not be papers on your area but if you can be pretty sure that outputs and inputs are measured similarly across all industries you can use the same. DEA is nothing but ratio of weighted outputs to inputs where the weights are not decided in advance and varies with each DMU. There fore its applicability is not limited and i am sure u could be the first to apply the same in your area and guide others. All the best
Dear Hubert, I'am affraid that I am not entitled to duly answer your question because I am not specialzed in this matter. For this reson, please don't take as nothing else than an informal opinion of mine the one that Mr. Goel's comment looks as very reasonnable to me. Best regards,
Dear Hubert, I endorse what Mr. Goel has said. You can use DEA also on aggregated data. For example Fare (1994) has used DEA (with the aim of calculating the Malmquist index) considering as DMUs the OECD countries. The important thing is the homogeneity of units under evaluation.
Differently from the other answers I have got, I would be very prudenti in using DEA in this context. The problem lies in aggregation. You may not be able to find harmonized country I-0 tables at the apropriate NACE level to be able to avoid a major composition effect. Suppose you have a disaggregation such that you identify 'furniture' as an industry. This is already fairly detailed in terms of data requirement, but is still inadequate for what you want to do do because the technology of producing e.g. upholstered furniture is much more labor intensive than producing kitchen cabinets. Thus you risk to misunderstand a country that specializes in upholstered furniture, perhaps the best one, as less efficient that a country that specializes in kitchen cabinets.The existence of trade says nothing about this error, because it will be trade of different things. This is why DEA is good (to a certain extent) only when you are comparing similar things, and with microdata you may (or may not) be able to make such assumption.
Dear Hubert, I agree with Florio.Using DEA in a global perspective may result in aggregation bias.
Thanks to all of you. Aggregation bias is a big issue when using input-ouput or supply-use tables in trade statistics: firms that actively participate in trade are usually very different than firms working only for the domestic market. Especially in developing countries. So I need to think a bit more about clustering the observations into more homogeneous groups to limit this bias. Cheers. HE
Data envelopment analysis (DEA) is a technique for measuring the relative efficiency of a set of decision making units (DMUs) which apply multiple inputs to produce multiple outputs in a period of time. Since the pioneering work of Charnes et al. (1978), numerous studies discussing the methodology and application of DEA have been published; see, for example Cook and Seiford (2009). Not only non-profit organizations, but also for-profit companies find this technique effective in identifying inefficient DMUs, as well as the factors which cause inefficiency.
References
Charnes, A., Cooper, W.W., Rhodes, E., 1978. Measuring the efficiency of decision making units. European Journal of Operational Research 2, 429–444.
Cook, W.D., Seiford, L.M., 2009. Data envelopment analysis (DEA) – thirty years on. European Journal of Operational Research 192, 1–17.
Dear Hubert,
In efficiency and productivity analysis, aggregation bias is closely associated to technology heterogeneity surrounding the examined DMUs. If one can handle effectively technology heterogeneity issues the worries about aggregation bias become less worthy. In this line, the most prominent approaches are (1) the metafrontier framework and (2) the second stage analysis of the estimated efficiency scores. I think that there are some papers which undertake efficiency analysis at a country level (macroeconomic performance) by Fare, Grosskopf and Margaritis. The quality of the data are of major importance. Of course there are several papers at a sectoral and industry level. For the metafrontier approach you may look at my papers in EJOR (2010), JBF (2010), Energy Economics (2012) and Economic Modelling (2012). I hope the above are helpful.
Dear Hubert,
Fully agree with the previous comments. My suggestion is the metafrontier approach for analysing heterogeneous DMU (due to for example country specific characteristics), so you wil be able to estimate both the country specific and the overal efficiency.
Battese, G.E., D.S.P. Rao and C.J. O'Donnell "A Metafrontier Production Function for Estimation of Technical Efficiencies and Technology Gaps for Firms Operating Under Different Technologies." Journal of Productivity Analysis, Vol. 21, (2004) pp. 91-103.
I agree with Maximo Florio and also with Manevska. There are two problems: one is the aggregation and the other is the heterogeneity. I also recommend a metafrontier or alternatively a SFA approach .
- Badges
- Science topic
- Similar topics
- Academic Journals