For each k, consider the following definitions: (i) πuk ≡ πk xfk represents the solution
to (8), i.e., observation k’s solution to the profit maximization problem with no revenue,
expenditure and employment constraints; (ii) πrk ≡ πrk xfk ,Rck represents the solution to
(8) and the additional revenue constraint (6); (iii) πek ≡ πek xfk ,Rck,Ek represents the
solution (8) along with the revenue and expenditure constraints (6) and (5); and (iv) πnk =
πnk xfk ,Rck ,Ek ,Nk represents (8) along with the revenue, expenditure, and employment
constraints (6) — (7). Finally, denote the actual observed profits for observation k as πak.
The overall efficiency measure is denoted TEk , and defined as
TEk
πak
πuk
By definition, πak is always less than or equal to πuk . Therefore TEk will be less than or
equal to one, with TEk =1only if the enterprise k is overall efficient. We note, without
proof, that the measures πuk , πrk , πek , and πnk are nested relationships, and hence, allow
us to decompose TEk in into four components: actual efficiency AEk , revenue efficiency
REk , financial efficiency FEk , and employment efficiency EEk . These measures are
defined as follows:
AEk = πak ,eek = πnk ,FEk = πek and REk = πrk ,
πnk πek πrk πuk
and the overall efficiency can be expressed as the product of these four sources of efficiencies
TEk = AEk * EEk * FEk * REk.
The actual efficiency could be further decomposed into technical efficiency and allocative
efficiency.
3.1 Data and Empirical Results
The project requires data on inputs, outputs, variable input prices, output prices, tax pay-
ment, expenditure and revenue for TVEs and private enterprises. Since we treat TVEs and
13