Just some numbers to bring reality into the general discussion:
Figure 1: Employment in government, in thousands, seasonally adjusted (blue), and excluding temporary Census workers (red). Total series is “USGOVT” from FREDII. NBER defined recession shsaded gray, assumes last recession ended 2009M06. Source: BLS, August employment situation release.
Update 8:50am Pacific, Thu 9/9:
Reader John Eckstein was kind enough to send me disaggregated government employment data, so that I wouldn’t have to do it myself. Here are some graphs thus generated. Note, these figures do not include uniformed military; they are based the establishment survey.
Figure 2: Federal (red), state (green), local ex.-educ. (orange) and local-education (purple), as a share of total nonfarm payroll employment (FRED II series PAYEMS). Source: BLS CES data via John Eckstein, and FRED II, and author’s calculations.
Figure 3: Federal (blue), state (red), local ex.-educ. (green) and local-education (purple) employment, seasonally adjusted, in thousands (FRED II series PAYEMS), all series seasonally adjusted. Source: BLS CES data via John Eckstein, and FRED II, and author’s calculations.
Update: 12:50pm Pacific, Thu 9/9:
For those afeared of the Federal government Leviathan, and unwilling to go the BEA website themselves, here is Federal nondefense consumption and gross investment as a share of GDP (see Table 3.9.5 for nominal series, then divide by nominal GDP, from 1.1.5).
Figure 4: Federal government nondefense consumption (blue), and nondefense gross investment (red), as a share of nominal GDP. Source: BEA, 2010Q2 2nd release, Table 3.9.5, and author’s calculations.
Update: 2:05pm Pacific, Friday, 9/10
Many readers have indicated their unhappiness that I did not plot their favored time series. Let me first say that one can search the Econbrowser site, and find many of the desired graphs, using the search facility. Second, many of your favorite series can be found in their most updated form (although not in the desired time period, since everybody seems to want a different sample plotted) at Calculated Risk, specifically this post, as well as this earlier post.