{smcl}
{* 23oct2004}{...}
{hline}
help for {hi:bking8}
{hline}
{title:Apply Baxter-King band pass filter to time series}
{p 8 17}{cmd:bking8}
{it:varlist}
[{cmd:if} {it:exp}]
[{cmd:in} {it:range}]
{cmd:,} {cmdab:plo(}{it:#}{cmd:)}
{cmdab:phi(}{it:#}{cmd:)}
{cmdab:stub(}{it:abbrev}{cmd:)}
[
{cmdab:k(}{it:#}{cmd:)}
]
{p 4 4}{cmd:bking8} is for use in Stata 8.2. Users of more recent versions of Stata should use {cmd:bking}, which makes use of Mata.
{p 4 4}You must {cmd:tsset} your data before using {cmd:bking8}; see help {cmd:tsset}. If a
panel calendar is in effect, the filter can be applied if a single panel is specified using {cmd:if} or {cmd:in} qualifiers.
{p 4 4}{cmd:varlist} may contain time-series operators; see help {cmd:varlist}.
{p 4 4}{cmd:bking8} filters one or more time series using the Baxter-King symmetric band-pass filter described in Baxter and King (1995).
If a panel calendar is in effect, the filter can be applied if a single panel is specified using if or in qualifiers.
The {cmd:plo()} and {cmd:phi()} arguments specify the minimum period of oscillation and maximum period of oscillation of the desired component of the time series, where 2 < plo < phi < infinity.
The "Burns-Mitchell" settings for this filter, applied to quarterly data, are 6 and 32 quarters, with k=12.
{p 4 4} You should {cmd:set matsize 800} before calling this routine. Due to the matrix size limits, a timeseries longer than 800 periods cannot be handled by this routine in Intercooled Stata.
{title:Options}
{p 4 8}{cmd:plo(}{it:#}{cmd:)} and {cmd:phi(}{it:#}{cmd:)} specify the minimum and maximum period of oscillation to be retained in the time series, and must be given. For quarterly data, common values are 6 and 32, which preserve the components of the data with period between 1.5 and 8.0 years, with k=12 (which loses three years' data from each end of the filtered series). For monthly data,
common values are 18 and 96, which preserves the component of the data
with period between 1.5 and 8.0 years. For annual data, common values are
2 and 8, with k=3.
{p 4 8}{cmd:stub(}{it:abbrev}{cmd:)}, which must be provided, specifies the "stub" from which new variable names will be created. Variables created by {cmd:stub} must be new variables. If the {it:varlist} contains time-series
operators, the dots in their names are replaced by underscores so that the resulting new variables' names are legal.
{p 4 8}{cmd:k(}{it:#}{cmd:)} specifies the lead-lag length of the filter. This many observations will be lost at each end of the filtered series. If not specified, {cmd:k} is set to 12 (Baxter and King's recommended value for quarterly data).
{title:Examples}
{p 4 8}{stata "webuse lutkepohl,clear" :. webuse lutkepohl,clear}{p_end}
{p 4 8}{stata "bking8 investment, plo(6) phi(32) stub(F)" :. bking8 investment, plo(6) phi(32) stub(F)}{p_end}
{p 4 8}{stata "bking8 investment income consumption, plo(6) phi(32) k(8) stub(filt)" :. bking8 investment income consumption, plo(6) phi(32) k(8) stub(filt)}{p_end}
{p 4 8}{stata "bking8 D.investment, plo(4) phi(12) stub(fl)" :. bking8 D.investment, plo(4) phi(12) stub(fl)}{p_end}
{title:Author}
{p 4 4}Christopher F. Baum, Boston College, USA{break}
baum@bc.edu
{title:References}
{p}Marianne Baxter and Robert G. King, Measuring Business Cycles: Approximate
Band-Pass Filters for Economic Time Series. Review of Economics and Statistics,
1999, 81(4), 575-593. {p_end}
{p}Lawrence J. Christiano and Terry J. Fitzgerald, The Band Pass Filter, International
Economic Review, 2003, 44(2), 435-465.{p_end}
{title:Acknowledgements}
{p}This code is based on Eduard Pelz' MATLAB implementation of the Baxter-King
filter, available from http://www.clev.frb.org/research/workpaper/1999/bpassm.txt
and has been validated against that code for selected series. {p_end}
{title:Also see}
{p 4 13}On-line: {help hprescott} (if installed), {help tsset}