poobly (version 0.1.2)

hsiao: Hsiao test

Description

Hsiao poolability test, as described by Hsiao (1986;2022).

Usage

hsiao(formula, data, index = NULL, ...)

Value

hsiao and list object with hypotheses with their corresponding F-statistics, degrees of freedom, and p-values.

Arguments

formula

formula object for plm::plm function.

data

plm::pdata.frame or data.frame object for input. Note that for the data.frame an additional index should be given.

index

An index input for plm::pdata.frame in case that data.frame is given instead of plm::pdata.frame.

...

Rest arguments passed to plm::plm. Note that 'model' and 'effect' can not be used in the current hsiao function and in such a case an error will be triggered.

Details

Hsiao (1986;2022) poolability/homogeneity test consists of three consecutive tests. One for testing if the slope and constant coefficients are same across the panel. If it is not the case, then a second test is conducted, with the heterogeneity of both slope and constant coefficients as alternative hypothesis. If this second hypothesis is not rejected, then the final third hypothesis is tested, where the null of the same slope and constant against the alternative of same slopes but different constants is tested.

References

Hsiao, C. (1986) Analysis of Panel Data. 1st edn. Cambridge: Cambridge University Press (Econometric Society Monographs).

Hsiao, C. (2022) Analysis of Panel Data. 4th edn. Cambridge: Cambridge University Press (Econometric Society Monographs), pp. 43-49.

Examples

library(plm)
data("Gasoline", package = "plm")

x <- hsiao(lgaspcar ~ lincomep + lrpmg + lcarpcap, Gasoline)
print(x)
##
##                     Hsiao Homogeneity Test
##
## Hypothesis| Null |                 Alternative
## ----------+------+---------------------------------------------
##     H1    |Pooled|                    H2
##     H2    |  H3  |      Heterogeneous intercepts & slopes
##     H3    |Pooled|Heterogeneous intercepts & homogeneous slopes
## ===============================================================
##
## formula: lgaspcar ~ lincomep + lrpmg + lcarpcap
##
##     Hypothesis  F-statistic     df1         df2       p-value
##   1     H1       129.3166       68          270       < 0.001
##   2     H2        27.3352       51          270       < 0.001
##   3     H3        83.9608       17          321       < 0.001