Industrial Production and Capacity Utilization - G.17
Documentation for the Revision History of Industrial Production
This file provides data showing the seasonally adjusted indexes for total industrial production, manufacturing, mining, and utilities, as they were originally published and subsequently revised during the normal six-month estimation period. For each industry group, the file includes a time series showing the level of the index as it was initially published as well as six time series showing its level when published one, two, three, four, five, and six months after its initial release. The estimate published six months after the initial release is usually the same as that published five months after the initial release, except when an annual or major revision was issued. Each time series, by itself, has discrete breaks at the time of annual revisions. Some come when IP has been estimated using a new base year.
In addition to the revision history for the index levels, the revision history for the reliability estimates found in table 15 of the G.17 release are included in this file.
II. Data format and file information:
The data file name is revh_sa.txt.
Data begin in January 1972 for all of the index levels except the
fifth and sixth revisions, which begin in 2007.
The ending date for the initial estimate
is the year and month of the most recent Industrial Production Index
(designated in the data file naming convention, YYMM). Data are
written in ASCII format with four decimal places, twelve months of
data per line for complete years.
|1-25||quoted series code|
|26-29||year in YYYY format|
|30-149||12 months or 4 quarters of data (for a complete year), 10 positions per value including decimal and 4 decimal places. The first value on a line is January and the data for the current year continue as far as has been published. Months with no values are reported as "na". If the initial estimate of IP has been published for May, the first revision of IP is available for April; the second revision is available for March; the third revision is available for February; the fourth revision is available for January; the fifth revision is available for December; and the sixth revision is available for November.|
The series codes are composed of three fields, a prefix, an industry code, and a suffix, separated by a period. For this revision history, only total IP, manufacturing, mining, and utilities, are included. Only estimates for the seasonally adjusted data are included.
The prefixes used in the variable names (i.e., init., rev1., rev2., rev3., rev4., rev5., and rev6.) denote the timing of the publication of the data.
init. Initial index level, published about 15 days after the end of the month. rev1. first revision, published about 45 days after the end of the month. rev2. second revision, published about 75 days after the end of the month. rev3. third revision, published about 105 days after the end of the month. rev4. fourth revision, published about 135 days after the end of the month. rev5. fifth revision, published about 165 days after the end of the month. rev6. sixth revision, published about 195 days after the end of the month. (Usually identical to fifth revision, except when there was an annual or major revision.)
The industry codes are:
B50001: Total industrial production B00004: Manufacturing G21: Mining G2211A2: Electric and gas utilities
The third fragment of the name is one of the following:
S: Index level S85: 85th percentile for index level S15: 15th percentile for index level PCT: Monthly percentage change PCT85: 85th percentile for monthly percentage change PCT15: 15th percentile for monthly percentage change QPCT: Quarterly percentage change QPCT85: 85th percentile for quarterly percentage change QPCT15: 15th percentile for quarterly percentage change
For example, the initial February 2008 value was published in the middle of March 2008. At the same time, the first revision of the January 2008 value was being published.
III. Calculation of growth rates:
Computing growth rates similar to those actually published requires some care. For example, one should divide the initial estimate of February 2008 (published in mid-March 2008) by the first revision of January 2008 IP (also published in mid-March 2008) to get the initial estimate of growth in February 2008. Use the following formulas to calculate the published growth rates for a series:
Initial growth rate: 100 * ((init[t] / rev1[t-1]) - 1) First revision of growth rate: 100 * ((rev1[t] / rev2[t-1]) - 1) Second revision of growth rate: 100 * ((rev2[t] / rev3[t-1]) - 1) Third revision of growth rate: 100 * ((rev3[t] / rev4[t-1]) - 1) Fourth revision of growth rate: 100 * ((rev4[t] / rev5[t-1]) - 1) Fifth revision of growth rate: 100 * ((rev5[t] / rev6[t-1]) - 1)
Estimates of the index levels prior to 1992 are only available to 1 decimal place so percent changes calculated from them may sometimes round differently than the growth rates actually published, which are based on unrounded data. The four decimal accuracy provided in the file after 1992 will usually yield a growth rate that rounds to the one published.