coreG: Find core (primary) group
In ConvergenceClubs: Finding Convergence Clubs

coreG

R Documentation

Find core (primary) group

Description

Find the Core (primary) group according to step 2 of the clustering algorithm by Phillips and Sul (2007, 2009)

Usage

coreG(
  X,
  dataCols,
  time_trim,
  threshold = -1.65,
  HACmethod = c("FQSB", "AQSB"),
  type = c("max", "all")
)

Arguments

`X`	dataframe containing data (preferably filtered data in order to remove business cycles). Data must not contain any NA or NaN values, otherwise the clustering procedure will be stopped with an error.
`dataCols`	integer vector with the column indices of the data
`time_trim`	a numeric value between 0 and 1, representing the portion of time periods to trim when running log t regression model. Phillips and Sul (2007, 2009) suggest to discard the first third of the period.
`threshold`	numeric value indicating the threshold to be used to perform the one-tail t test; default is -1.65.
`HACmethod`	string indicating whether a Fixed Quadratic Spheric Bandwidth (`HACmethod="FQSB"`) or an Adaptive Quadratic Spheric Bandwidth (`HACmethod="AQSB"`) should be used for the truncation of the Quadratic Spectral kernel in estimating the log t regression model with heteroskedasticity and autocorrelation consistent standard errors. The default method is "FQSB".
`type`	one of "max" or "all"; "max" includes only the region with maximum t-value. The default option is "max"; "all" includes all units that pass the test t in the core formation (step 2).

Details

According to the second step of the Phillips and Sul clustering algorithm (2007, 2009), the log t regression should be run for the first k units 2 < k < N maximizing k under the condition that t-value > -1.65. In other words, the core group size k* is chosen as follows:

k* = argmax t(k)

subject to

min t(k) > -1.65

Such behavior is obtained with type="max"; if type="all", all units that satisfy t(k) > -1.65 are added to core group.

If the condition t(k) > -1.65 does not hold for k = 2 (the first two units), the algorithm drops the first unit and repeats the same procedure for the next pair of units. If t(k) > -1.65 does not hold for any couple of units, the whole panel diverges.