Cobb-Douglas or Translog Production Function in Efficiency Analysis?

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00027251%3A_____%2F20%3AN0000019" target="_blank" >RIV/00027251:_____/20:N0000019 - isvavai.cz</a>

Result on the web

<a href="https://mme2020.mendelu.cz/wcd/w-rek-mme/mme2020_conference_proceedings_final_final.pdf" target="_blank" >https://mme2020.mendelu.cz/wcd/w-rek-mme/mme2020_conference_proceedings_final_final.pdf</a>

DOI - Digital Object Identifier

—

Result language

angličtina

Original language name

Cobb-Douglas or Translog Production Function in Efficiency Analysis?

Original language description

The aim of this paper is to assess two types of specification of production function when estimating the technical efficiency with Stochastic Frontier Analysis – Cobb-Douglas (CD) and translog (TL) production functions and their influence on the results. CD production function is easy to be estimated. It is a special case in a more general class of production functions with constant elasticity of substitution, where is the mostly used a TL function. We compared both specifications on the unbalanced panel data of agricultural holdings in the Czech Republic. There were 517 farms and 1708 observations for years 2013 to 2016. Accounting data were taken from Albertina database purchased from Bisnode, s. r. o., company and data about acreage were taken from Land Parcel Identification System. Production was approximated by sales, production factors were: consumption of material and energy, fixed assets, number of employees and acreage of agricultural land. Technical efficiencies were calculated by JLMS (Jondrow, Lovell, Materov, Schmidt, 1982 [11]) method and their differences tested by Wilcoxon signed-rank test and correlated by Spearman’s rank coefficient. The technical efficiency in CD specification was slightly higher (85.69%) than in TL specification (85.12%), but there were found statistically significant differences. Spearman’s ???? = 0.9597 was close to 1 pointing on high positive dependence. Using likelihood ratio test, the CD specification was rejected in favor of more general specification of TL function. We can conclude that TL function is more appropriate to be used as a production function in SFA, but the effect on the technical efficiency is mild and almost negligible (despite that the test revealed statistically significant differences, the dependence was very high.)

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

50202 - Applied Economics, Econometrics

Project

—

Continuities

I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Publication year

2020

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

38th International Conference on Mathematical Methods in Economics (MME 2020)

ISBN

978-80-7509-734-7

ISSN

—

e-ISSN

—

Number of pages

7

Pages from-to

557-563

Publisher name

Mendel University Publishing Center, Brno

Place of publication

Brno

Event location

Brno

Event date

Sep 9, 2020

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

—