The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.
Generalized-mean price indexes are a large family of price indexes with nice properties, such as the mean-value and identity properties (e.g., Balk, 2008, Chapter 3). When used with value-share weights, these indexes satisfy the key homogeneity properties, commensurability, and are consistent in aggregation. This last feature makes generalized-mean indexes natural candidates for making national statistics, and this justifies the hierarchical structure used by national statistical agencies for calculating and disseminating collections of price indexes.
Almost all bilateral price indexes used in practice are either generalized-mean indexes (like the Laspeyres and Paasche index) or are nested generalized-mean indexes (like the Fisher index). The purpose of this vignette is to show some of the key functions for working with generalized-mean indexes. In what follows, everything is framed as a price index to avoid duplication; it is trivial to turn a price index into its analogous quantity index by simply switching prices and quantities.
A generalized-mean price index is a weighted generalized mean of
price relatives. Given a set of price relatives and weights, any
generalized-mean price index is easily calculated with the
generalized_mean() function. What distinguishes different
generalized-mean price indexes are the weights and the order of the
generalized mean. For example, the standard Laspeyres index uses
base-period value-share weights in a generalized mean of order 1
(arithmetic mean).
library(gpindex)
# Start with some data on prices and quantities for 6 products
# over 5 periods
price6
#> t1 t2 t3 t4 t5
#> 1 1 1.2 1.0 0.8 1.0
#> 2 1 3.0 1.0 0.5 1.0
#> 3 1 1.3 1.5 1.6 1.6
#> 4 1 0.7 0.5 0.3 0.1
#> 5 1 1.4 1.7 1.9 2.0
#> 6 1 0.8 0.6 0.4 0.2
quantity6
#> t1 t2 t3 t4 t5
#> 1 1.0 0.8 1.0 1.2 0.9
#> 2 1.0 0.9 1.1 1.2 1.2
#> 3 2.0 1.9 1.8 1.9 2.0
#> 4 1.0 1.3 3.0 6.0 12.0
#> 5 4.5 4.7 5.0 5.6 6.5
#> 6 0.5 0.6 0.8 1.3 2.5
# We'll only need prices and quantities for a few periods
p0 <- price6[[1]]
p1 <- price6[[2]]
p2 <- price6[[3]]
q0 <- price6[[1]]
q1 <- price6[[2]]
s0 <- p0 * q0
s1 <- p1 * q1
# Laspeyres index
arithmetic_mean(p1 / p0, s0)
#> [1] 1.4Changing the order of the generalized mean to \(1 - \sigma\), where \(\sigma\) is an elasticity of substitution, gives a Lloyd-Moulton index, whereas changing the weights to current-period value-shares gives a Palgrave index. This is the essence of the atomistic approach in chapter 2 of Selvanathan and Rao (1994).
# Lloyd-Moulton index (elasticity of substitution -1)
quadratic_mean <- generalized_mean(2)
quadratic_mean(p1 / p0, s0)
#> [1] 1.592692
# Palgrave index
arithmetic_mean(p1 / p0, s1)
#> [1] 2.268331Generalized-mean indexes can also be nested together to get indexes
like the Fisher, Drobisch, or AG mean index. The
nested_mean() function is a simple wrapper for
generalized_mean() for these cases.
# Fisher index
fisher_mean(p1 / p0, s0, s1)
#> [1] 1.592692
# Drobisch index
drobisch_mean <- nested_mean(1, c(1, 1))
drobisch_mean(p1 / p0, s0, s1)
#> [1] 1.834166
# Geometric AG mean index (elasticity of substitution 0.25)
ag_mean <- nested_mean(0, c(0, 1), c(0.25, 0.75))
ag_mean(p1 / p0, s0, s0)
#> [1] 1.358687On top of these basic mathematical tools are functions for making
standard price indexes when both prices and quantities are known.
Weights for a large variety of indexes can be calculated with
index_weights(), which can be plugged into the relevant
generalized mean to calculate most common price indexes, and many
uncommon ones. The price_index functions provide a simple
wrapper, with the quantity_index() function turning each of
these into its analogous quantity index.
Two important functions for decomposing generalized means are given
by transmute_weights() and factor_weights().
These functions augment the weights in a generalized mean, and can be
used to calculate percent-change contributions (with, e.g.,
contributions()) and price-update weights for
generalized-mean indexes.
quadratic_decomposition <- transmute_weights(2, 1)
arithmetic_mean(p1 / p0, quadratic_decomposition(p1 / p0, s0))
#> [1] 1.592692
quadratic_mean(p1 / p0, s0)
#> [1] 1.592692
quadratic_contributions <- contributions(2)
quadratic_contributions(p1 / p0, s0)
#> [1] 0.03110568 0.51154526 0.04832926 -0.03830484 0.06666667 -0.02665039
quadratic_update <- factor_weights(2)
quadratic_mean(p2 / p0, s0)
#> [1] 1.136515
quadratic_mean(p2 / p1, quadratic_update(p1 / p0, s0)) *
quadratic_mean(p1 / p0, s0)
#> [1] 1.136515Percent-change contributions can similarly be calculated for indexes that nest generalized means.
ag_decomposition <- nested_transmute(0, c(0, 1), 1, c(0.25, 0.75))
ag_mean(p1 / p0, s0, s0)
#> [1] 1.358687
arithmetic_mean(p1/ p0, ag_decomposition(p1 / p0, s0, s0))
#> [1] 1.358687
ag_contributions <- nested_contributions(0, c(0, 1), c(0.25, 0.75))
ag_contributions(p1 / p0, s0, s0)
#> [1] 0.03333243 0.29947183 0.04948725 -0.05377928 0.06536724 -0.03519241Balk, B. M. (2008). Price and Quantity Index Numbers. Cambridge University Press.
Selvanathan, E. A. and Rao, D. S. P. (1994). Index Numbers: A Stochastic Approach. MacMillan.
These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.