Use this page to post what you get/find in the way of data to pursue plans we made on 10/15/09
Employment specialization
both relative and absolute measures are used
absolute: Herfindahl--employment shares (by industry or sector) relative to "norm" of uniform (equal) shares; this is the "employment specialization index" featured in the La Plata, Colorado, information, whose link is on Resources page of this wiki
relative: compares employment shares not to some "normal" expectation of equal shares across sectors, but instead to the (weighted) average of the relevant comparison group (universe); this is a Krugman type index, as described in "Geographical Processes Behind Innovation" paper
if you run this STATA program, adjusting the path names etc., you will read in these 2 data files and produce a Herfindahl measure for La Plata county, Colorado; it should match the published figure (see Resources)
example of relative index calculations
data demands are greater, and I haven't yet been able to get a program running to do all, say, 100 or so counties of Illinois
instead, here is my synthetic example of the Krugman index we discussed yesterday in class
Here are some files for you to review on employment specialization indexes; the "raw data" were county-level employment stats, by industry & gender, from the year 2000 Census (Table P49)
The goal was to use these data to compute Herfindahl (absolute) and Krugman (relative) style indexes of employment specialization
I downloaded Census data, saved as *.csv file, cleaned it up a little by hand (dumping various headers; adding a gender column), and then re-saved it
Then I wrote a Stata program to read in that file and re-organize the data, to get "counties" as my unit of observation, with employment levels (by industry) as my variables
1. 1990 and 2000 ALL states and counties in the U.S., fips and coname/state
% of those 25 and older WITHOUT A HIGH SCHOOL DIPLOMA
% of those 25 and older WITH ONLY a high school diploma
% of those 25 and older with MORE THAN high school diploma
2. BEA PCPI 1998 to 2007 ALL states and counties in the U.S., fips and coname/state
3. INNOVATION INDEX – the individual items that are pulled for the index
4. 1998-2008 qcew data by 2-digit NAICS for all states and counties, fips and coname/state. This would be 16 items.
11/11/2009 - Geography Note from the IBRC
One Alaska county split into 2 counties effective June 20, 2007. One Virginia independent city merged with a county on July 1, 2001. See below copied from the popest documentation. Sorry the table didn’t come thru very well. The Vintage 2008 files have population estimates for 3,142 counties and county equivalents. The Federal Information Processing System (FIPS) codes on the Vintage.
2008 files differ slightly from those on earlier vintages: two new FIPS codes are included on the files (Skagway Municipality and Hoonah-Angoon Census Area) and one FIPS code is excluded (Skagway-Hoonah-Angoon Census Area). The table below summarizes differences in FIPS codes across the various bridged-race population files; for the most part, FIPS codes on the files are the same. Vintage 2002, Vintage 2005, Vintage 2006, and Vintage 2007 have estimates for 3,141 counties and county equivalents and the FIPS codes on these four vintages are consistent with each other. The Vintage 2003 and Vintage 2004 files have estimates for 3,140 counties and county equivalents and the FIPS codes on these two vintages are consistent with each other (unlike the other postcensal vintages, they do not have population estimates for Broomfield County, Colorado). The intercensal files and the April 1, 2000 files have estimates for 3,141 counties and county equivalents; these files do not have estimates for Broomfield, CO but unlike later files do have estimates for Clifton Forge, VA.
Counties and county equivalents that do not appear on all of the bridged-race population files: 1990-2008 Year
County of county equivalent
Number of counties
Skagway Municipality, AK
Hoonah-Angoon Census Area, AK
Skagway-Hoonah-Angoon Census Area, AK
Broomfield County, CO
Clifton Forge County, VA
Intercensal estimates
1990-1999
3,141
--*--
--*--
X
--*--
X
Census counts
April 1, 2000
3,141
--*--
--*--
X
--*--
X
Postcensal estimates
Vintage 2001
None, only national estimates
--*--
--*--
--*--
--*--
--*--
Vintage 2002
3,141
--*--
--*--
X
X
--*--
Vintage 2003
3,140
--*--
--*--
X
--*--
--*--
Vintage 2004
3,140
--*--
--*--
X
--*--
--*--
Vintage 2005
3,141
--*--
--*--
X
X
--*--
Vintage 2006
3,141
--*--
--*--
X
--*--
--*--
Vintage 2007
3,141
--*--
--*--
X
X
--*--
Vintage 2008
3,142
X
X
--*--
X
--*--
--*-- County or county equivalent is not on the file.
X County or county equivalent is on the file.
Specific details:
Broomfield County, Colorado (FIPS code=08014) was created effective November 15, 2001 from parts of four Colorado counties: Adams, Boulder, Jefferson, and Weld. There are estimates for this county on some, but not all, of the bridged-race files. Note that data for Broomfield County do not appear on NCHS birth or mortality files until data year 2003.
Clifton Forge County, Virginia (FIPS code = 51560). On July 1, 2001, Clifton Forge city, Virginia, formerly an independent city, merged with Alleghany county (FIPS code=51005). There are no estimates for this county on the bridged-race postcensal population files. Note that data for Clifton Forge city appear on NCHS birth and mortality files prior to data year 2003; beginning with the 2003 data year, no data for Clifton Forge city appear on the birth and death files.
Skagway Municipality, AK (FIPS code = 02-230). Skagway Municipality was created from part of the former Skagway-Hoonah-Angoon Census Area (FIPS code = 02-232) effective June 20, 2007; boundaries are identical to the Skagway census subarea; population: 862. The remainder of the former Skagway-Hoonah-Angoon Census Area was established as the new Hoonah-Angoon Census Area (02-105). This county equivalent appears only on the Vintage 2008 files. Note that no data for Skagway Municipality appear on NCHS birth and mortality files.
Hoonah-Angoon Census Area, AK (FIPS code = 02-105). The Hoonah-Angoon Census Area was created from the remainder of the former Skagway-Hoonah-Angoon Census Area (FIPS code = 02-232) when Skagway Municipality (FIPS code = 02-230) was created effective June 20, 2007; population: 2,574. This county equivalent appears only on the Vintage 2008 files. Note that no data for Hoonah-Angoon Census Area appear on NCHS birth and mortality files.
Skagway-Hoonah-Angoon Census Area, AK (FIPS code = 02-232). Effective June 20, 2007, Skagway-Hoonah-Angoon Census area was split to create Skagway Municipality (FIPS = 02-230) and Hoonah-Angoon Census Area (FIPS code = 02-105), This county equivalent appears on the bridged-race files prior to Vintage 2008 (except for Vintage 2001 for which no county-level data are available). Note that data for Skagway-Hoonah-Angoon Census Area appear on NCHS birth and mortality files for 1994 – present.
11/9/2009
Items 1, 2, & 4 from the request below are available at https://iuv.indiana.edu/slashtmp/html/download.php?FILE=falewis/25019B8.tmp.
The download is a zip file with several Excel and CSV files. Password - november
1. 1990 and 2000 ALL states and counties in the U.S., fips and coname/state
% of those 25 and older WITHOUT A HIGH SCHOOL DIPLOMA
% of those 25 and older WITH ONLY a high school diploma
% of those 25 and older with MORE THAN high school diploma
2. BEA PCPI 1998 to 2007 ALL states and counties in the U.S., fips and coname/state
3. INNOVATION INDEX – the individual items that are pulled for the index
4. 1998-2008 qcew data by 2-digit NAICS for all states and counties, fips and coname/state. This would be 16 items.
11/11/2009 - Geography Note from the IBRC
One Alaska county split into 2 counties effective June 20, 2007. One Virginia independent city merged with a county on July 1, 2001. See below copied from the popest documentation. Sorry the table didn’t come thru very well. The Vintage 2008 files have population estimates for 3,142 counties and county equivalents. The Federal Information Processing System (FIPS) codes on the Vintage.2008 files differ slightly from those on earlier vintages: two new FIPS codes are included on the files (Skagway Municipality and Hoonah-Angoon Census Area) and one FIPS code is excluded (Skagway-Hoonah-Angoon Census Area). The table below summarizes differences in FIPS codes across the various bridged-race population files; for the most part, FIPS codes on the files are the same. Vintage 2002, Vintage 2005, Vintage 2006, and Vintage 2007 have estimates for 3,141 counties and county equivalents and the FIPS codes on these four vintages are consistent with each other. The Vintage 2003 and Vintage 2004 files have estimates for 3,140 counties and county equivalents and the FIPS codes on these two vintages are consistent with each other (unlike the other postcensal vintages, they do not have population estimates for Broomfield County, Colorado). The intercensal files and the April 1, 2000 files have estimates for 3,141 counties and county equivalents; these files do not have estimates for Broomfield, CO but unlike later files do have estimates for Clifton Forge, VA.
X County or county equivalent is on the file.
Specific details:
12/2/2009
Index data - https://iuv.indiana.edu/slashtmp/html/download.php?FILE=falewis/130D87E.tmp
Secret word: indexdata