How important are regional income disparities in Scotland?

In this blog post I want to discuss some ongoing analysis into income inequality in Scotland. It might seem intuitive to assume that location plays an important role on people’s incomes: however, some of our ongoing work is - perhaps surprisingly - suggesting other factors matter more.

Policy and public debates about income inequality are generally focused on regional disparities. It is true that average aggregate incomes can vary greatly across regions. I can demonstrate this using the maps below, which show the average annual gross pay in 2010 for the 32 Local Authorities in Scotland. There are noticeable differences both for workplace and residence measures of earnings. For example, average workplace gross pay in Aberdeen City is about 1.62 times higher than average gross pay in Angus and the Scottish Borders, while average residence gross pay in East Renfrewshire is about 1.57 times higher than average gross pay in the Western Isles.

However, these maps don’t tell us anything about income disparities within regions.  To better understand this, I and some colleagues within the Social, Economic and Geographical Sciences (SEGS) group are developing research funded by the Scottish Government Strategic Research Programme 2011-2016 on the sources of income disparities in Scotland. We have been using individual microdata from the ‘Scottish Household Survey’ (SHS) for the period 1999-2008 and regression techniques to investigate the key factors underlying income disparities across individuals. The first part of our work has explored the contribution of inter- and intra-regional differences to overall income inequality, and in this post I’d like to highlight some of the key findings arising from this.

Our work is indicating that much of the overall income inequality in Scotland can be attributed to within-region disparities rather than between-region disparities.  This is illustrated by the graphs below. These show changes in income inequality over time, from 1999 to 2008.  I’ve used the solid line to show the overall change – fairly flat as you can see.  The nearby dotted line shows the inequality found within regions, whilst the dashed line (very near the bottom) shows the inequality attributed to differences between regions. We find a similar pattern whether we define regions by using the Scottish Local Authority areas (left hand graph), or by using the Scottish Government’s rural-urban classification, which defines areas as urban, small towns, or rural and accessible or remote (right hand side graph). So, in both cases, it seems the large majority of income inequality is located within regions.

Figure 2. Overall, within, and between region income inequality in Scotland
*Large urban, other urban, accessible small towns, remote small towns, accessible rural, and remote rural.

These graphs suggest that differences between people within the same region (“people effects”) may be more important than differences between people in different regions (”place effects”). However, in order to  estimate the importance of “place” effects, we need to compare the incomes of “equivalent” people living in different areas.   

We have explored all these “people effects” and “place effects” using regression models.  The “people effects” that we consider include gender, age, education, employment status (employee vs. self-employed) and regime (full-time vs. part-time), occupation, and industry. “Place effects” include the economic size of the area, local economic structure (i.e. industry mix), qualifications of the local workforce, cost of local housing, and type of rural-urban area.

Our models indicate that much of the differences in individual incomes are driven by people rather than place.  For example, accounting for “place effects” reduces the variability in income levels only by 2.3%, whereas “people effects” are greater:  accounting for differences in individuals’ occupational group reduces the variability in income levels by 21.5%, while accounting for differences in individuals’ education and industry sector reduces the variability in income by 10.0% and 9.1% respectively.

These findings might have some important implications for those seeking to tackle income inequality. Our analysis suggests that focusing solely on regional differences in income levels might not be beneficial for policy making aimed at reducing income disparities. Instead, average regional incomes result from the interaction of both people and place effects and hence need to be considered simultaneously by policy-makers and others working on this issue.

One area of particular importance relates to people’s skills and job-related qualifications. However, highly skilled workers are generally more mobile than low skilled workers, and will consider area-specific effects when deciding where to live and work. So, this adds an additional complication to the design of regional policies. Spatial planning (including housing planning, transport planning, and consumption amenities) could play an important role in attracting skilled workers. Our ideas about how to consider combinations of people and place policies, however, shall be the topic of a future post!

Author: Patricia Melo

Disclaimer: The views expressed in this blog post are the views of the author(s), and not an official position of the institute or funder.