Abstract
This article analyses the evolution of employment in the French regions, putting the accent on business service firms. A ‘shift and share’ type of analysis shows that a primary decentralisation in tertiary activities seems to emerge in the 1990s, essentially pertaining to ancillary producer services. An explanatory analysis backs up the general validity of the regional economic base theory: the basic activities, to which business services can quite legitimately be attached, certainly play a leading long-term role on global employment dynamics. Finally it shows that over the past 20 years, the regional offer of services to businesses has been the major discriminating variable between regions, greatly influencing the evolution of basic employment and thus confirming the vital driving force of this sector in regional dynamics.
Notes
The RESER is an European Research Network on Service Activities grouping university researchers and service professionals from many European countries. The 21 activities listed divide business services into two categories: (1) Consulting and advice services, subdivided into six detailed activities: Advertising, Architecture, Computer software engineering, Engineering and technical surveys, Legal advice and Management consulting. (2) Ancillary producer services, subdivided into 12 detailed activities: Advertising supports management, Car rentals, Cleaning of premises, Equipment rentals, Industrial laundry, Manpower hiring, Miscellaneous rentals, Photography, Secretarial piecework and data capture, Security and miscellaneous services, Sorting and routing and Surveyors/land surveyors. In the broad sense services to firms include these 18 business service activities but also five others more or less connected to this intermediary market: Transports and communication, Wholesale trade, Holding and consular organisations, Research and Foreign representatives.
The national component ( + 6.6 per cent between 1982 and 1999) is of no interest to us as such here, it only enables us to compare the two following components (share and shift effects which are variations compared to this national, average component).
The addition of their absolute values gives: 1.23 million jobs differentiating the regional dynamics due to the differences in activities composition (share effect) vs 1.32 million ‘displaced’ due to a shift effect.
The sum has to be divided by two because each job ‘displaced’ would be otherwise counted twice: once, positively, in the region it is displaced to, and once, negatively, in the region it is displaced from.
Due to the huge growth in some sectors, the percentage is established according to the average 1982-99 employment figures and not according to the 1982 employment figures.
The value of the constant term is always higher here (0.485 in 1982–1990 and 0.325 in 1990–1999) and the regression coefficient is lower than over a long period (0.541 and 0.680 respectively).
The basic sector groups together: agriculture, mining, manufacturing industry, transports, wholesaling, hotel trade, research, foreign representatives, holdings and consular organisations, merchant business services (consulting or others). Government service employment has been considered as being ‘induced’ unless it exceeds the average allowance per inhabitant (principally in the Île de France and PACA regions). The excess part in this case is added to the economic base.
Induced activities (or urban) group together: building, utilities, banks and insurance, real estate agencies, retail trading, services to households, teaching, health care, social and collective services. Government service employment is also counted as being ‘induced’ unless it is over-represented, and only for the excess part.
Regression with the two variables only leaves the urban employment development with a partial R2 of 0.073.
Brittany, Corsica and Languedoc-Roussillon.
The coefficient of representation compares the relative weight of jobs of one sector in one region to the same indicator calculated at the national level. A value of 100 per cent indicates a proportional representation, superior values for one region an over-representation in the sector and inferior values an under representation.