Case studies
Video text transcript
The aims of this research were to disentangle the contribution of people and place to overall levels of wage inequality. By place, I mean the impact of differences between labour markets and the area specific effects on individual wages. By people, I mean the differences between individuals in terms of their characteristics, the things that make them productive.
To do this analysis, what we needed was data on individuals, data that would allow us to link these individuals to their workplaces, data that had information on wages, occupation, other characteristics of the individuals. And we used the Annual Survey of Hours and Earnings and New Earnings Survey, which is an excellent data source for this; it allows us to track people over time for a period of up to ten years.
Using this data what we did was essentially try to decompose the inequality in earnings across the individuals, the variance, into a sort of share that's attributable to differences across places and a share that's attributable to differences across individuals within those places.
When we do this analysis, if you just kind of look at the maximum share that could be attributed to difference across labour markets, that's around six to seven percent of the overall variance in earnings across individuals -- so quite a small amount, even at the outer end.
Once you control for basic occupational skills and other characteristics, that comes down a lot, and if you look very carefully at individuals who move from area to area over time -- look at how much the differences in earnings that they experience as they move from place to place -- you can compare that with the variance across individuals. Then the share attributable to places is very very small, less than one percent.
Wage disparities in Britain: People or place?
How much do the labour markets where people live and work contribute to wage inequality?
The study
This study aimed to provide evidence on the relative contribution of 'people' and 'place' effects to wage inequality in Britain. 'People' effects are the effects of an individual's characteristics on wages that are not dependent on where the individual lives and works. 'Place' effects are the influence of a geographical area on an individual's wage that does not depend on the characteristics of that individual.
The study used variance decomposition methods to find the relative contribution of people and place to individual wage inequality, estimating the share of overall wage inequality attributable to differences between labour markets and observing how this estimate is reduced with control for observable and unobservable characteristics of individual workers.
The data
The study used a combination of data from both the New Earnings Survey (NES) and the Annual Survey of Hours and Earnings (ASHE) for the period 1998-2008. The researchers accessed the data via the Virtual Microdata Laboratory at the Office for National Statistics.
Other surveys of earnings exist, but the ASHE (and NES) data are considered more accurate. This is because the data are collected from employers about their staff - rather than from employees themselves - so accurate wages and hours of work data are collected.
The dataset also provides researchers with a large sample of observations. ASHE is roughly a one-percent sample of the working population, so there are data on nearly 160,000 individuals.
Furthermore, the same individual workers are generally surveyed each year, making it possible to track the history of an individual's job and wages throughout time, allowing researchers to undertake panel data analysis.
Also, since the data are collected from employers, the unique business reference numbers are also available to researchers, allowing them to link the data to business data (such as the Annual Business Inquiry). For example, researchers could analyse the effects of productivity on wages.
Both the NES and ASHE contain a number of variables for individual workers, including:
- age
- occupation
- hourly earnings
- bonuses
- hours of work
- pensions
The results
The findings indicate that while geography affects individual wage inequality, this contribution is very small relative to the contribution of individual-specific factors. Observed geographical differences in wages are mostly due to sorting of workers with different skills into different labour markets.
The researchers concluded that trying to address area effects directly will not have a large impact on wage disparities in the UK. Addressing total wage disparities instead of area effects and outcomes should be the primary policy objective.
The authors
Stephen Gibbons is a senior lecturer on geography and environment at the London School of Economics.
Henry Overman is a professor and Director of the Spatial Economics Research Centre at the London School of Economics.
Panu Pelkonen is an economics lecturer at the University of Sussex.