< Previous | Contents | Next >

Scottish Economic Statistics 2002

A4 Income Inequality in Scotland

Julie Goodall, Alan Fleming and Richard Murray, Scottish Executive

Introduction

The level of income inequality in the UK has increased considerably over the past two decades and has recently shown signs of being on the increase again. This has prompted much interest into how income inequality can increase when we have a strong economy and a government which is committed to reducing poverty. In addition, with increasing attention given to social justice and opportunities for all, there is the need for greater understanding of income inequality and how Scotland fares in relation to not just the UK but also the rest of the developed world.

This paper sets out some of the background to the debate on income inequality and points the way forward for future analysis. Although income inequality can exist between households and within households, the latter will be outwith the scope of this paper. Instead the aim of the paper is twofold: to examine the factors which influence income inequality, and to estimate the level of income inequality for Scotland. Furthermore, a detailed examination of the trend in inequality in the UK and what factors have had the greatest influence on this trend will be carried out.

Summary of main findings

• Since 1977 there has been a general increase in income inequality, with the steepest rise occurring in the mid-to-late 1980s. During the 1990s income inequality showed a slight downturn but has begun to rise again in recent years.
• The level of income inequality in Scotland is generally lower than it is in GB/UK but follows a similar trend over time.
• The level of income inequality in the UK is relatively high compared with other developed countries.

Does Income inequality matter?

Income inequality has again been thrust into the public limelight as figures published by the Office for National Statistics¹³show that there has been a slight rise in income inequality in the UK over the last two years. The most widely used indicator of income inequality, the Gini Coefficient¹⁴, increased from 33% in 1995-96 to 35% in 1999-00 for disposable income, and from 37% to 40% for post-tax income. Although the size of the increase is not dramatic in itself, it is the trend in income inequality which is of particular concern.

The level of income inequality in the UK remained relatively steady between 1977 and the mid-1980s, followed by a sharp rise which peaked around 1990. Inequality then fell slightly during the early 1990s but began to increase again after 1995-96 and is now almost back to the level seen in 1990.

This recent rise has occurred despite the current UK government introducing a number of schemes to address problems of low income¹⁵. However it is important to note that such measures will not have had a chance to impact fully upon the current income distribution figures as they were only recently introduced.

There are many reasons why one might be interested in changes in the personal income distribution. These include the following:

1 Understanding the distribution of income will help policy makers target policies at reducing income inequality and income poverty.

2 There is said to be a link between income and health/life expectancy. Studies have found that those on low income and those in poverty are, in general, more likely to have a lower life expectancy. Therefore, it is important to identify such factors as they will help target policies directed at reducing health problems within the country.

3 There is thought to be a strong link between inequality and economic growth. In the past, economists tended to assume a trade-off between equality and efficiency, arguing that any re-distribution would ultimately be at the expense of an economically productive factor. Therefore, income inequality was argued to lead to an enhancement of economic growth on the basis that the economy would be more efficient than when resources are being re-distributed. However, recent endogenous growth theory¹⁶ has re-considered this relationship as there is evidence that a poor income distribution might ultimately reduce economic growth. Growth would be affected through the channels of education, access to capital markets, as well as political and economic mechanisms.

4 Those in society who are socially excluded¹⁷, will tend to be at the bottom of the income distribution. Therefore, in order to create a more socially inclusive society, such income inequalities have to be identified and addressed.

5 Income inequality also gives an insight into the labour market and its operations. For example, skills shortages within a particular industry can act to push up wages, causing an increase in the overall dispersion of wages and will ultimately impact upon the degree of income inequality. Furthermore, if there are particular controls on the level of wages (e.g. a national minimum wage), then this will again affect the degree of overall income inequality.

6 Income inequality will have long-run implications and could influence the type of workforce we will have in the future. If income inequality prevents people from fulfilling their potential then it will have serious implications on future economic performance.

7 Fairness: what degree of divergence in earnings does society see as acceptable? Capitalist economies will always reward the highly skilled. However, many Governments seek, through taxation and expenditure, to try to re-distribute income so that these inequalities can be reduced, to the benefit of society as a whole.

Overlap between income inequality and income poverty

It is important to realise that although there is some overlap between income inequality and income poverty, the two are not the same thing.

Income poverty relates to households with incomes below a certain threshold. The standard threshold used in Britain and in Europe is 60 per cent of national median income¹⁸. Income poverty measures concentrate on households at the bottom end of the distribution. Although incomes of all households are used to calculate the threshold, the measure is not unduly influenced by households at the very top of the income distribution, as the median is less susceptible to extreme values.

Income inequality is measured using a more complex technique, namely the Gini Coefficient. This calculates the overall degree of income inequality in a country, based on household income. Although both measures are relative measures in that they take into consideration other households' income, the Gini Coefficient looks at the entire income distribution rather than simply those below a certain threshold. Therefore the distribution of households in the middle and upper parts of the income distribution have a greater role to play in determining the level of income inequality than the level of income poverty.

The effects of policies directed at reducing income poverty are similar to those aimed at reducing the degree of income inequality, namely to raise the incomes of those people with low incomes. However as the degree of income inequality is influenced by those households on middle and higher incomes, such policies may have limited effect. For example, if there is an increase in income of the lowest income decile due to policies to reduce income poverty through encouraging more people to enter employment, then income inequality will fall, all other things being equal. However, if the increase in income in the lowest income decile is lower than the increase in income in the top decile, then income inequality will increase in spite of the policy to reduce income poverty.

The Gini Coefficient

The Gini Coefficient is a simple way of comparing different income percentiles using a single index. The index is derived from the Lorenz Curve which plots cumulative shares of the population against the cumulative shares of income. For income to be distributed evenly, the plot would trace a 45° line which represents the "line of perfect equality".

The Gini Coefficient is defined as the area between the Lorenz Curve and the 45° line, taken as a ratio of the whole triangle lying below the 45° line. In the extreme case of perfect equality, it will yield a value of zero, and in the case of perfect inequality, it will yield a value of 1. Therefore, an increase in the Gini Coefficient represents an increase in inequality.

One of the ways of expressing the Gini Coefficient algebraically is:

Here n is the number of income units, is mean income, i is the position in the distribution, and y_i are individual incomes ranked from poorest to richest.

There are alternative measures of income inequality to the Gini Coefficient. The Half Squared Coefficient of Variation and the 90/10 Ratio can also be used to measure income inequality. A full examination of these methods is outwith the scope of this article. However, it is important to note that different concepts of inequality typically suggest the use of different measures. In some circumstances, an alternative measure may be more appropriate.

Income Inequality in the UK

There are 4 separate measures of income against which the Gini Coefficient can be calculated. These are as follows:

• Original income - income before taxes and benefits. This includes income from earnings, occupational pensions and investments, but not income from benefits.
• Gross income - original income plus cash benefits.
• Disposable income - gross income minus direct taxes such as, for example, income tax.
• Post-tax income - disposable income minus indirect taxes (e.g. VAT).

It is important that when comparing Gini Coefficients between countries or over time, that a consistent definition of income is used across the board. The use of different types of income will influence the size of the Gini Coefficient and could lead to a misrepresentation of income inequality.

Original income takes no account of any Government intervention (i.e. taxes and benefits), whereas gross, disposable and post-tax income take account of varying degrees of Government intervention. Calculating inequality based on all four measures allows us to examine the effects of Government interventions on the income distribution.

Commentators tend to focus on the post-tax coefficient, as this shows the effect of the full weight of Government intervention. However, this paper mainly concentrates on disposable income as the data available for calculating Scottish estimates do not allow the measurement of post-tax income.

Chart A4.1 tracks the changes in income inequality for disposable income between 1977 and 1999/00. Between 1984 and 1990, the Gini Coefficient for disposable income rose from 28% to 36%. The level of income inequality fell steadily thereafter until reaching a low of 33% in 1995-6. Since then, it has begun to rise again and reached a level of 35% in 1999-00.

Chart A4.1: Income inequality in the UK, Disposable Income, 1977 to 1999/00

The OECD¹⁹ found that, across family types in the UK, relative incomes of persons living in households with children were lower, in particular those of single parents. However it is important to note that this covers the period from 1985 to 1995. This finding is mirrored in the latest income poverty statistics, which showed that, whilst 23% of people in Great Britain were living in income poverty in 1999-00, this proportion rose to 58% for those in single parent families.²⁰

Gini Coefficient for Scotland

The position of the UK is fairly clear, but what about Scotland? To calculate Gini Coefficients to measure income inequality in Scotland, an alternative data source is needed. The Family Expenditure Survey (FES) which is used for the calculation of UK estimates, has a small sample size for Scotland and there is too much uncertainty associated with Scottish estimates of Gini Coefficients derived from these data.

The Family Resources Survey (FRS) and the derived Households Below Average Income (HBAI) dataset has a much larger Scottish sample and therefore produces more robust estimates at the Scottish level. The FRS, which currently covers Great Britain rather than the UK, began in the mid 1990s. Therefore, while data are available for each of the years 1994/95 to 1999/00, this does provide a relatively limited time series. It is not possible to calculate post-tax income estimates from the FRS as it does not collect data on expenditure. However, it does allow Gini Coefficients to be produced for original, gross and disposable income and gives broadly comparable results to those derived from the FES.

Table A4.1 shows how coefficients in Scotland compare with Great Britain. It can be seen that, for every income measure in every year (with only one exception), Scotland has a Gini Coefficient which is lower than the GB equivalent. Care should be taken when comparing Scottish figures against GB equivalents as there is a higher degree of uncertainty surrounding the Scottish estimates due to the sample size involved.

The differences between Scotland and GB for any single year are unlikely to be statistically significant; however the consistent picture of lower Scottish coefficients year-on-year suggests that Scotland does have a reduced degree of income inequality compared with Great Britain as a whole. To a large extent this difference may be due to high incomes in London and the South East increasing the level of inequality in Great Britain.

Chart A4.2 plots the coefficients for disposable income for Scotland and GB. The Scottish coefficients are consistently below those for GB, but the increasing trend is clear in both cases. The GB trend looks fairly smooth whilst the Scottish trend line is more volatile with small year-on-year fluctuations. This is likely to be a result of sampling variation, brought about by the smaller Scottish sample (this could partially explain the increase in 1999-00).

Chart A4.2: Gini Coefficients for Disposable Income, Scotland and Great Britain, 1994/95 to 1999/00

International Comparisons

The comparison of income inequality on an international scale is problematic. This includes having data for differing periods of time, the use of different income definitions, and the fact that there are inconsistencies between the Gini Coefficients calculated by the various organisations.

Despite these problems, it is important to compare our position with our international counterparts as this adds greater perspective when addressing the scale of the problem of income inequality. Caution however must be taken when interpreting the results as often the figures relate to differing time periods and can often become outdated very quickly.

Chart A4.3 illustrates the Gini Coefficients which were collated by OECD²¹ for a variety of countries using figures from three sources: Luxembourg Income Study (LIS), EUROSTAT, and OECD Questionnaire. An accurate comparison is limited by the fact that some figures are missing from particular studies for certain years²². Unfortunately, none of the international comparisons provide data for Scotland. However, given the similarity between the Scottish and UK figures, it is reasonable to use the UK as a proxy for Scotland when making these international comparisons.

Chart A4.3: International Comparison of Income Inequality

It is clear that the UK is amongst a group of countries, including Greece, Italy, Ireland and the US, with the highest level of income inequality with a Gini Coefficient at or around 35%. All 3 sources use data from 1995 for the UK, but dates are not consistent across other countries. Refer to the annex for details.

It is important to place income inequality in a world-wide perspective. The levels of income inequality shown in chart A4.3 are incomparable with the high degree of income inequality in the developing world. Indeed, the last time global income inequality was calculated (for 1993) the Gini coefficient was some 0.66²³.

What factors influence income inequality?

Employment: The dispersion of gross earnings has been identified as one of the main contributors to household income inequality. Consequently, the allocation of employment across and within households and how this allocation has changed over time will have an impact on income inequality. If new jobs are exclusively filled by persons from households in which other members are already employed, then employment growth will not be accompanied by a decrease in workless households. If, at the same time, worklessness is increasingly concentrated in low income households, then overall inequality and the share of low incomes clearly might increase despite economic and employment growth. This process of simultaneous increases in both worklessness and fully employed households has been described as the process of employment polarisation²⁴. The OECD found evidence of this process within 9 of the 11 European OECD member countries, including the UK.

Earnings Potential: There has been a move towards a "Transatlantic Consensus"²⁵ with regards to income inequality. The consensus has established the view that increased income inequality in the United States and high unemployment in Continental Europe are due to a shift in demand away from unskilled workers towards skilled workers. Furthermore, globalisation has also played its part in shifting labour demand in developed countries towards more skilled employment. The low relative wages on offer in developing countries has attracted many of the unskilled jobs previously held in developed countries. The goods and services produced by these unskilled jobs in developing countries are now imported to developed countries.

Despite this consensus, Atkinson finds, using the continuum of earnings capacity, there has been a tilt in the UK's earnings distribution: although those at the lower end of the distribution moved closer to the median between 1989 and 1999, increased wage dispersion is due more to what has happened in the upper part of the distribution.

Understanding wage dispersion is the key to understanding the routes of income inequality. One of the most common explanations for increasing income inequality is that declining trade union membership and diminution of union power may have reduced the union wage premium in the lower part of the distribution. Alternatively, increased dispersion may be due to a decline in centralised bargaining. However centralised wage setting is clearly not enough in itself to limit the growth in earnings inequality²⁶.

Types of Income: The measure of household income can be divided into six different sources of income: earnings from employment, self-employment income, occupational pensions, income from savings and investments (including personal pensions), state benefits and a catch-all "other" category which includes miscellaneous income sources such as private health benefits, student loans and maintenance payments. Income from the labour market - both employment and self-employment - is responsible for a great bulk of income inequality in Britain: some 86 per cent in 1997-98.

Breaking income down into these components, three particular features become apparent when examining UK income inequality²⁷:

• Overall inequality has increased substantially;
• All sources of income now make a larger absolute contribution to overall inequality than at the start of the period;
• Sources of income inequality are more diverse than previously: in 1979 the great majority of income inequality reflected inequality of earnings, but by the late 1990s, the combined effect of other income sources has grown to be almost as important.

Of particular interest is the share of inequality that is due to self-employment income. This has more than doubled, while the contribution of investment income has increased almost fourfold. The Gini Coefficient for employment income was 35.7%, compared with 71.0% for income from self-employment. Despite concerns over the quality of self-employment income data in general, this difference is nevertheless striking. The reason for such a high degree of inequality is that there is a high amount of risk associated with entrepreneurial ventures which increases the volatility of earnings from these activities.

Transfer Payments²⁸: One of the effects of fiscal policy²⁹ is to alter the distribution of income within the economy. The OECD found that, in most countries, between one third and 40% of those transfers went to the lower income groups in the working-age population (bottom three deciles), and between 20% and 25% to the higher income groups (top three deciles). This progressive pattern, whereby a higher proportion of benefits go to those in the lower end of the income distribution than to the higher end, was strongest in, amongst other countries, the UK where 50% to 60% went to the lower income groups, and only 10% to 20% to the higher income groups.

Taxation: Generally speaking there are two main types of taxation: direct taxation and indirect taxation. The UK has a progressive tax system with regards to direct taxes as the revenue collected rises more than proportionally to income. Income tax is made progressive by having exemptions for very small incomes, low rates for the first slice of the taxable income, and higher rates for the largest incomes. The progressive nature of direct taxation acts as a form of redistribution from those on higher earnings to those on lower incomes, redistributed in the form of the various benefits offered by the state.

Indirect taxes can be progressive if there are exemptions or low tax rates for goods heavily consumed by the poor, and higher rates on luxury items more heavily consumed by the rich. However generally indirect taxes tend to be regressive in nature whereby the ratio of tax paid to income falls as income rises. Regressive taxes counteract the other measures governments use to redistribute income from the rich to the poor. Therefore a policy of increasing the level of indirect taxation instead of direct taxation, will not help reduce the income inequalities within an economy.

Conclusions

Greater understanding is needed of the determinants of income inequality before policy can be targeted at reducing such inequalities. Few governments have a stated policy to reduce the level of income inequality within their society as an essential part of the fight against social exclusion. At the moment, policy is primarily directed at reducing the level of income poverty within society. Such policies are not guaranteed to lower the degree of income inequality as changes in the middle and upper sections of the income distribution will act to counter such policies.

Recent figures show that the level of income inequality in Scotland has risen slightly. However, it is important to place this recent increase into perspective - the recent increase (since 1995-96) is small in relation to the overall level for Great Britain. Certainly there is no strong upward trend. Instead, what is of importance is that the UK has a relatively high degree of income inequality which has not appeared to fall significantly since the dramatic rise of the 1980s.

Figures indicate that levels of inequality have been consistently smaller in Scotland than in Great Britain, suggesting more equality in the structure of the Scottish income distribution. However, the fact that Scotland follows the increasing trend experienced by Great Britain gives some cause for concern.

< Previous | Contents | Next >