Peter,

You are correct in noting the distinction between "enterprise" and "establishment". The overwhelming majority of "enterprises" (businesses) have but a single "establishment" (and most establishments are synonymous with a small enterprise). However, the Economic Census asks respondents to report their headquarters and whether they are a subsidiary. All of this is based on a 2009 master list. [I've not looked at the actual wording of the survey, but it's available on the www.stat.go.jp site, at least in Japanese -- the data are also reported on the English-language site at www.e-stat.go.jp].

For many enterprises, that's fine, but subsidiaries and closely held affiliates may not report their parent correctly -- though clearly the survey is designed to elicit that. And do you aggregate individual franchises as a larger enterprise? While the franchisor (say McDonald's, the largest franchise in Japan) controls many details, and may set guidelines for employment conditions, stores are individually owned (how many stores the average franchisor owns is likely reported on their investor links on the company site). That's a conceptual issue, and almost surely McDonald's establishments are aggregated by franchisee and not by franchisor.

So the "Enterprise" numbers likely understate the number of people who work in large enterprises. It is also based on an earlier survey that tried to get a master list to which surveys could be sent, and to educate them in advance on reporting whether they were a subsidiary etc.

So in timing and the nuances of questions answers would differ from ones focusing on employment. But that's the nature of data collection. I put misc on the Economic Census below. The survey does try to count non-regular employment. It excludes family farms and in-house businesses (家事サービス業 which I assume includes "naishoku" -- I knew women who sewed neckties or made kimono fittings -- and not just cleaning) and foreign government offices.

mike smitka