TFP indices using prices
TFP indices can be constructed using prices to weight, and then
aggregate quantities of outputs and inputs into a value aggregate
for outputs Q(Y) and a value aggregate for inputs Q(X). Productivity
is the ratio of the output value aggregate Q(Y) to the input value
aggregate Q(X); i.e. TFP¼Q(Y)/Q(X). Productivity change is the ratio
of the productivity value aggregate in one time period to the
productivity value in a base time period, and is referred to as a TFP
index. The quantity indices of a TFP index typically use prices from
a base and/or a reference period to weight the significance of the
input and output quantities. There are several different types of
common quantity indices which differ by the period(s) from which
prices are chosen and how they are used to construct the weights
on quantity. An extensive review of indices in general can be found
in Balk [11] or Coelli et al. [4]. The two most basic indices are the
Laspeyers index and the Paasche index. These indices differ by the
period from which weights are constructed (base and reference,
respectively) and provide theoretical bounds for the “ideal” index
[12].
The Fisher index presents a middle ground and is the geometric
mean of the Laspeyeres and the Paasche index. Although, the
Fisher index is nonlinear it is approximately linear with weights
that are an average of the base and reference periods [13]. The
Fisher index has a number of desirable theoretical properties and
provides a second-order approximation to an arbitrary twice-differentiable
linearly-homogenous production function, a property
known as being superlative [14]. The Fisher index has been widely
used to measure productivity change in numerous industries. Recently,
the Australian government used the Fisher index to measure
productivity change for important Australian fisheries [15].
The Törnqvist index is another popular quantity index which
weights quantities by value shares from two time periods through
a geometric mean. The Törnqvist index is also both nonlinear and
superlative, and is the exact index implied by the popular translog
production function. Examples of some fisheries productivity
studies using the Törnqvist index include [3,16,17].
In spite of being one of the older indices, the Lowe index has enjoyed a recent revival and uses fixed prices to weight outputs
and inputs [2,11,18]. The critical feature that differentiates the
Lowe index from others is that the prices used as weights in the
index are fixed over time. Conveniently, the price weights may be
from the base or reference time period or from a time period
outside of either the base or reference period. This means that the
indices are transitive, which simplifies temporal comparisons of
productivity. The recent study of productivity change for U.S. catch
share fisheries used a Lowe index [8]. Furthermore, the Lowe index
permits a decomposition of the TFP index into technical, environmental,
and scale-mix components [2,19].