Assessing the relative efficiency of agricultural production units in the Blackland Prairie, Texas

DEA-R: ratio-based comparative efficiency model, its mathematical relation to DEA and its use in applications

“…To illustrate the use of DEA-R model in applications, we will now take an example problem borrowed from Haag et al (1992). The example will be used merely to illustrate some features of the DEA-R model with no intention to claim that what is being done is an appropriate analysis for this particular data set.…”

Global and local scale characteristics in convex and nonconvex nonparametric technologies: A first empirical exploration

“…Third, the detailed sources of conflict in classification vary a lot among the different samples. While for the Haag, Jaska, and Semple (1992) study the CRS-IRS conflict dominates for about 20% of observations, for Cesaroni (2011), Färe, Grosskopf, andLogan (1983) and Fan, Li, and Weersink (1996) the IRS-DRS case is dominant: for a small about 7% for the first two cases to a substantial about 14% of observations for the third sample.…”

“…Second, apart from the first study with the smallest sample size, all other samples yield some minimal to moderate conflict in classification between convex and nonconvex technologies. This conflict varies from a modest about 7% for the Färe, Grosskopf, and Logan (1983) sample to a quite substantial about 40 % for the Haag, Jaska, and Semple (1992) case, all three cases confounded. Third, the detailed sources of conflict in classification vary a lot among the different samples.…”

“…9 Note that this Li-test cannot be computed for the nonconvex TE component of Haag, Jaska, and Semple (1992), since all observations are technically efficient and hence the kernel density cannot be estimated.…”

“…Table 1 summarizes some key features of each data set: sample size, number of inputs and outputs, and the sector. There is one small unbalanced panel (Färe, Grosskopf, and Logan (1983)) and four cross sections (Cesaroni (2011), Fan, Li, andWeersink (1996), Färe, Grosskopf, Logan, and Lovell (1985), and Haag, Jaska, and Semple (1992)). Note that the time dimension in the panel is ignored: this amounts to assuming there is no technical change over the five time periods.…”