2004 vs 2007 data

Since Virginia (and all counties and states data) is tied to the US - in that they are forced to sum to the US and some of the relationships come from US relationships I am using the US to describe what happened. Two factors came into play: 1) Reported output "Motion picture and video industries" dropped between 2004 and 2006 (2007 was projected based on employment and income). The BEA reports output based on benchmark sectoring. For 2004 the file GDPbyInd_GO_NAICs.xls was downloaded (~Nov 2006) and output was reported at $79,918 millions. For 20006 the file GDPbyInd_GO_NAICs1998-2006.xls (~July, 2008) reported $76,042 millions. Interestingly enough, the BEA revised the 2004 estimate down to $70,576 millions. 2) Our original estimate of 2004 employment was forced downward when we controlled to NIPA accounts (All US data is controlled to the current NIPA account. The upshot is that 2007 IMPLAN data matches REIS estimates (wage and salary plus proprietors) fairly closely while 2004 IMPLAN employment is low compared to REIS (~5%). Since output per worker is lower, indirect effects tend to be lower. This is a generalization as other factors could negate this. TheProprietor income dropped from $5,697 millions in 2004 to 778 millions in 2006 (2007 uses 2006 relationships) for the combined NAICS 512. This serves to reduce the induced effect as labor income drives the induced effect. I am quite comfortable with the 2007 IMPLAN data relationships for this sector. The simplest explanation for your audience would be "post 2004 analysis revisions to the 2004 data by the BEA and MIG data reduction techniques that exacerbated the apparent differences between 2004 and 2007". Hope this helps.