inequality in oecd countries: the facts and policies to curb it Rising income inequality risks leaving more people behind in an ever-changing world economy. by john p. martin and michael förster*

A condensed version of the Corden Lecture given at the University of Melbourne on 9 October 2013.

Why should we care about high and growing inequality?

returns by investors with rapidly-growing incomes might have contributed to asset-price bubbles.

The gap between rich and poor in OECD countries has reached its highest level in 30 years. Rising income inequality creates economic, social and political challenges and risks leaving more people behind in an ever-changing world economy. It can jeopardise social mobility: intergenerational earnings mobility is low in countries with high inequality and higher in countries where income is distributed more evenly. The resulting inequality of opportunities can negatively impact on economic performance and wellbeing. Inequality can also fuel protectionist sentiments. People will no longer support open trade and free markets if they feel that they are losing out while a small group of winners is getting richer. High inequality also raises political challenges because it breeds social resentment and generates political instability.

How unequal are OECD societies?

High and increasing inequality may also fuel economic instability. Aggregate demand can be reduced when resources are redistributed from poorer credit-constrained households to richer households with a higher propensity to save. The period of the Great Moderation prompted low interest-rate policies which helped trigger increases in household and sovereign debt beyond sustainable levels. In parallel, the search for high

There are large differences in income inequality levels across countries.1 Today in the OECD countries, the average income of the richest 10 per cent of the population is close to ten times that of the poorest 10 per cent. In Australia and New Zealand, the income gaps between the richest and poorest deciles are in line with the OECD average. In some European countries, the gap is much smaller, with the incomes of the richest decile being five times those of the poorest. In the US, the ratio raises to around 14:1, and in Chile and Mexico to a high of 27:1. In Japan, the rich have incomes more than 10 times those of the poorest. The Gini coefficient, a standard summary measure of inequality that ranges from 0 (when everybody has identical incomes) to 1 (when all income goes to only one person), stood at an average of 0.31 in OECD countries in 2010, ranging from a low of 0.25 in Iceland and Slovenia to a high of 0.50 in Chile. As Figure 1 shows, there are large differences in income inequality levels across OECD countries. How widespread is income poverty? For crosscountry comparisons, the standard practice is to treat ‘poverty’ as a relative concept in developed economies. The comparison of absolute incomes Insights Melbourne Business and Economics

05

between countries shows, for example, that the poorest 10 per cent in Japan have more money (in terms of purchasing power parities) than the average Mexican person. But what matters is the standard of living relative to other people in the country. Here, poverty is measured as half of the national median household income. This benchmark, of course, also varies over time. Around 11 per cent of the population across the 34 OECD countries fell below this poverty threshold (see the orange diamonds in Figure 1). Again, this differs hugely between countries: from 6 per cent in Denmark and the Czech Republic to 20 per cent and more in Israel and Mexico. Poor people, on this relative definition, make up 14.4 per cent of the population in Australia. Countries with wider distributions of income tend to have more widespread income poverty –

though measures of inequality and poverty do not necessarily go hand-in-hand. In New Zealand and the UK, for instance, inequality is higher than in Japan and Korea, but poverty is higher in the latter two countries.

Has the gap between rich and poor widened? Over a longer period of 25 years, the gap between rich and poor has widened in over three-quarters of OECD countries for which data series back to the 1980s are available. For the more recent period since the mid-1990s (Figure 2), it climbed by more than 3 percentage points in Canada, Denmark, Finland, Israel and Sweden. Only Mexico and Turkey recorded sizeable declines in their Gini coefficients. As Figure 2 shows,

Figure 1: Gini coefficients of disposable income and relative poverty rates, 2010 28%

0.55

Gini coefficient of inequality

Gini coefficient of income inequality (left-hand scale)

Relative poverty rate (right-hand scale)

0.50

24%

0.45

20%

0.40

16%

0.35

12%

0.30

8%

0.25

4%

0.20

0%

d ia ay rk lic d lic m ria n rg ry ny s d ce d d a d ly ia a lia n ce in m al el es ey o ile 9 lan ven rw ma ub lan ub lgiu st ede ou ga a land rlan an lan lan ore lan Ita ston nad tra apa ree Spa gdo rtug Isra tat urk exic Ch D-2 S T M C n o Ice Slo No en Rep Fin Rep Be Au Sw mb Hun ermher itze Fr Po Ire K Zea E Ca us J G e Ki P D h G et w A ed OE ux ew ak c nit N S ed L v t e N i U o Cz Sl Un

Note: 2009 for Hungary, New Zealand, Japan and Turkey; 2011 for Chile. Data refer to household equivalised disposable income. Poverty rates refer to the percentage of persons living in households with less than half the median income, in each country. Information on data for Israel: http://dx.doi.org/10.1787/888932315602. Source: OECD Income Distribution Database (www.oecd.org/social/income-distribution-database.htm), accessed in December 2013.

06

Inequality in OECD countries: the facts and policies to curb it

Figure 2: Gini coefficients of disposable income, mid-1990s to 2010 0.55

Increasing inequality mid 1990s

0.55

Little change in inequality

2010s

0.50

0.50

0.50

0.45

0.45

0.45

0.40

0.40

0.40

0.35

0.35

0.35

0.30

0.30

0.30

0.25

0.25

0.25

0.20

0.20

0.20

Decreasing inequality

0.55

k d d rg den el pan ralia ada nce any ar an ite s ra t ou a n Ja inl nm rm mb Swe Fr us Un tate Is F e e Ca A D G S xe Lu

s h y in ita land zecblic rwa C u No er p h a t e e Re Gr N

r tB

ile o ey d aly ary Ch exic urk alan It ng M T Ze Hu w Ne

Source: OECD Income Distribution Database (www.oecd.org/social/income-distribution-database.htm), accessed in December 2013.

income inequality (Gini coefficients) increased in a majority of OECD countries during this period. Income inequality followed different patterns across OECD countries over time. It first started to increase in the 1980s in some English-speaking countries, notably the UK and the US, but also in Israel. The trends in the 2000s showed a widening gap between rich and poor not only in some of the already high-inequality countries like Israel and the US, but also – for the first time – in traditionally low-inequality countries, such as Germany and the Nordic countries, where inequality grew more than anywhere else in the 2000s. At the same time, Chile, Mexico, Greece, Turkey and Hungary reduced income inequality considerably – often from very high levels. There are thus signs of a possible convergence of inequality levels towards a higher average level across OECD countries. In most countries, increasing inequality was due to rich households faring much better than both low- and middle-income families. There has been a marked increase in the share of top incomes, especially the top 1 per cent of earners. This has often been attributable to higher shares of labour, not capital, income – partly due to the development of stock options which are reported as part of wages and salaries. Figure 3 shows how the share of very high incomes increased in many

countries from 1980 to 2010. The rise at the top was most marked in the US where the share of the richest 1 per cent in all pre-tax income reached 18 per cent. However, it was also large in a number of other English-speaking countries (Australia, Canada, Ireland and the UK). Elsewhere, increases tended to be greater in the Scandinavian and Mediterranean countries than in Continental Europe. Even within the group of top-income earners, incomes became more concentrated. In the US, for instance, the share of the top 0.1 per cent in total pre-tax income quadrupled in the 30 years to 2010 from 2 per cent to over 8 per cent of total pre-tax incomes. The top 0.1 per cent accounted for some 4 to 5 per cent of total pre-tax incomes in Canada, the UK and Switzerland, and close to 3 per cent in Australia, New Zealand, and France.

The links between high inequality and social mobility The trends highlighted so far are based on crosssectional income-distribution data over time. While these are very informative, they do not tell us anything about the dynamics of inequality over time, that is, the extent of equality of opportunities across generations. We now turn briefly to this important intergenerational dimension of inequality. Insights Melbourne Business and Economics

07

Figure 3: Shares of top 1 per cent incomes in total pre-tax income, 1980 – 2010 (or closest) 24

% of total pre-tax income

21

2010

18

1980

15 12 9 6 3 0

a y in d d d al an an ad ite s rita ug Jap lan lan m an er Un tate at B Ire Port er C z t i G S re Sw G

ly

Ita

a

ali

str

Au

ain

Sp

ce

an

Fr

ay

rw

No

d

lan

Fin

w

Ne

nd

ala

Ze

en

ed

Sw

k

ar

nm

De

ds

lan

er

h et

N

Source: OECD calculations derived from the World Top Incomes Database.

When measures of the intergenerational earnings elasticity are plotted against measures of income inequality, this generates what Alan Krueger has called the ‘Great Gatsby Curve’.3 Figure 4 plots such a relationship for a sample of OECD countries for which we have comparable estimates of the two variables, and shows how high inequality can hinder social mobility (the inequality measure is the Gini coefficient). This shows that the ‘Gatsby Effect’ is alive and well: intergenerational earnings mobility is lower in high-inequality countries. At one extreme, we have countries such as the US, UK and Italy with relatively low intergenerational earnings mobility and high inequality while the Nordics are clustered at the other end, with relatively high intergenerational earnings mobility and low inequality. Both Australia and New Zealand are in mid-way positions, as is Canada; this may partly reflect the greater shares 08

Inequality in OECD countries: the facts and policies to curb it

of immigrants in their populations as well as their greater selectivity in choosing immigrants who tend to exhibit greater earnings mobility than their native-born peers. While there is some disagreement in the literature about the interpretation of the negative correlation in Figure 4 and the mechanisms underpinning intergenerational mobility, there is no doubt that Figure 4: Income inequality and intergenerational earnings mobility, mid-2000s 40 38

United States

36 34

Gini coefficient

At OECD we have been studying the phenomenon of social mobility and how it is associated with income inequality across countries by focussing on the concept of intergenerational earnings mobility: in the literature this is proxied by the degree to which sons’ earnings are correlated with those of their fathers.2 The key empirical measure here is the intergenerational earnings elasticity: the higher the value of this elasticity, the lower the degree of earnings mobility between generations.

Italy New Zealand

Great Britain

32

Spain

30 28

France

Canada Japan Australia South Germany Korea Norway

Switzerland 26

Finland

24

Sweden

22

Denmark

20 45 50 55 60 65 70 75 80 85 90

Intergenerational earnings mobility Source: OECD (2008) and d’Addio (2007).

it focuses policy debate on a very important topic: the degree to which high and rising inequality may hinder or help achieve equality of opportunity across countries.

Possible culprits for the growing divide Globalisation is often blamed for growing inequality. Over the past decades, OECD countries underwent significant structural changes, driven by their closer integration into the global economy and to rapid technological progress. Trade integration doubled in many OECD countries and outward stocks of foreign direct investment (FDI) increased steeply – from an average of less than 5 per cent of GDP in 1980 to nearly 50 per cent in the late 2000s. The increased productivity and opportunities for trade and FDI have contributed to raising the growth potential but these changes often brought highly skilled workers greater rewards than low-skilled ones and thus affected the way earnings from work were distributed. Further, technological progress shifted production technologies in both industries and services in favour of skilled labour and away from unskilled and middle-skilled workers. These structural changes got underway in the early 1980s and accelerated from the mid-1990s. But changes in policy choices, regulations and institutions have also often been cited as a culprit for increasing inequality. These changes can shape how globalisation and technological changes affect the product markets, wage-setting mechanisms and workers’ bargaining power. Most OECD countries carried out regulatory reforms to strengthen competition in the markets for goods and services, and to make labour markets more adaptable. All countries significantly relaxed anticompetitive product-market regulations and many also loosened employment protection legislation for workers with temporary contracts. Minimum wages also declined relatively to median wages in a number of countries. Wage-setting mechanisms also changed: the share of union members among workers fell across most countries, although

the coverage of collective bargaining generally remained rather stable over time. A number of countries cut unemployment benefit replacement rates and, in an attempt to promote employment among low-skilled workers, some also reduced taxes on labour for low-income workers. Apart from economic globalisation and regulatory change, other societal changes may also have had a direct impact on increasing inequality. In particular, changing family structures made household incomes more diverse and reduced economies of scale. Populations are ageing and there are more single-headed households with and without children today than ever before: in the mid-2000s, they accounted for 20 per cent of all working-age households, on average, in OECD countries. In couple households, employment rates of the wives of top earners increased the most. And in all countries, marriage behaviour has changed. People are now much more likely to choose partners in the same earnings bracket: so rather than marrying nurses, doctors are now increasingly marrying other doctors.

What OECD evidence tells us about the main culprits The OECD’s 2011 report Divided We Stand: Why Inequality Keeps Rising took a fresh look at the drivers and remedies for increasing inequality and reveals a number of surprising findings. It finds that neither rising trade integration nor financial openness has a significant impact on either wage inequality or employment trends within the OECD countries. The wage-inequality effect of trade appears neutral even when only the effects of increased import penetration from emerging economies, such as China and India, are considered – a finding that runs counter to the expectation that trade flows from such countries should drive down wages of workers in manufacturing and/or services in OECD countries.5 At the same time, technological progress – for example, in information and communications – has exhibited a bias in favour of high-skill workers and this has been reflected in widening gaps in earnings between high-skilled, middle-skilled and low-skilled workers. Insights Melbourne Business and Economics

09

On the other hand, regulatory reforms and institutional changes increased employment opportunities but, at the same time, contributed to wider wage disparities as more low-paid people were brought into employment. Thus, the increase in part-time employment in atypical labour contracts and a decline in the coverage of collectivebargaining arrangements in many countries also contributed to disparities in earnings. However, the rise in the supply of skilled workers helped offset the increase in wage inequality resulting from technological progress, regulatory reforms and institutional changes. The upskilling of the labour force also had a significant positive impact on employment growth. The growth in average educational attainment among the working-age population thus appears to have been the single most important contributor not only to reducing wage dispersion among workers but also to increased employment rates. On the basis of these results, the evolution of wage inequality across OECD countries over the past few decades can be best viewed mainly as the difference between the demand for and supply of skills, or as neatly summarised by the Dutch economist Jan Tinbergen almost 40 years ago, the outcome of a ‘race between education and technology’. This explanation, however, does not satisfactorily cover the rapid rise in top-income shares. For the latter, other factors need to be taken into account too – such as the growth of the financial sector, the spread of a ‘winner-takes all’ culture, cuts in marginal tax rates on high incomes, and lobbying of the political elites by high earners in order to preserve their rents.

The importance of tax/benefit systems Public cash transfers, as well as income taxes and social security contributions, played a major role in all OECD countries in reducing market-income inequality. Together, they were estimated to reduce inequality among the working-age population by an average of about one-quarter across OECD countries. This redistributive effect was larger in the Nordic countries, Belgium and Germany, but well below average in Chile, Iceland, Korea, Switzerland 10

Inequality in OECD countries: the facts and policies to curb it

and the US (Figure 5). In Australia and New Zealand, the redistributive impact of the tax/social transfer system is just below the OECD average. As the Figure shows, market incomes are distributed much more unequally than net incomes. In most countries, tax-benefit policies traditionally offset some of the large increases in marketincome inequality, although they became less effective at doing so from the mid-1990s. Until the mid-1990s, tax-benefit systems in many OECD countries offset more than half of the rise in market-income inequality. However, while market-income inequality continued to rise after the mid-1990s, the offsetting effect of taxes and benefits on household income inequality declined. Only a few countries bucked this trend. In Japan, for instance, the extent of redistribution continued to increase slightly though its level is still lower than in most other OECD countries. The main reason for the decline in redistribution lies on the benefits side: the real value of many social benefits fell, eligibility rules were tightened to contain spending on social protection, and transfers to the poorest failed to keep pace with earnings growth in many countries. In addition, spending on out-of-work benefits shifted towards ‘inactive’ benefits, which resulted in reduced activity rates and thus exacerbated the trend towards higher market-income inequality. At the same time – and despite the substantial gains of high-income earners in some countries – income taxes played a relatively minor role in moderating trends towards higher inequality. The reason is that trends towards lower income taxes, on the one hand, and more progressive taxation, on the other, had opposite effects on redistribution and partly cancelled each other out.

What was the impact of the recent Great Recession? 6 The significant increase of inequality was occurring before the Great Recession when many countries were undergoing a period of fairly steady economic expansion. What will happen now that the ILO estimates that 200 million people are out of work worldwide? The jobs crisis is affecting the

most vulnerable groups particularly hard, amid growing long-term unemployment and mounting youth unemployment, and this has put additional pressures on the distribution of incomes. And let us not forget that many governments are embarking on a path of fiscal consolidation to rein back publicsector deficits and rising public-debt/GDP ratios. Unfortunately, comparable income distribution data are only available with a lag and the latest data refer to 2010. This means that we are only able to chart the initial effects of the Great Recession on income inequality. These data suggest that the initial short-term impact of the crisis on inequality may have been smaller than commonly suggested in most OECD countries, with some notable exceptions like Ireland and Spain (Figure 6).

This apparently small effect was due to a number of causes. First, to stimulus packages and additional public support through the tax and benefit system in 2008 and 2009, which cushioned falls in household income levels at the bottom end. Second, households adopted coping strategies – for example, young people returned to live with their parents, or second earners increased working hours. Third, at the top end of the distribution, income shares fell due to declines in stock prices and interest rates, and a big drop in capital gains. This fall in top incomes was temporary, however, and has not reversed the preceding increase in topincome shares. Further, previous recessions have increased inequality in the mid-term because of an increasing employment divide between rich and poor. Finally, there is a risk of increasing inequality

Figure 5: Inequality (Gini coefficient) of market income and disposable (net) income in the OECD area, late 2000s 0.55

Inequality of market income

Inequality of disposable income

0.50

Gini coefficient of inequality

0.45

0.40

0.35

0.30

0.25

0.20

l e 9 l y ) e y y c a s n s a c n a k a n a re and and bli de ni ar wa bli ni nd pa and and ai tri ium ad alia an ary nc and urg te om ga Ital srae hil -2 C D I Ko zerl Icel epu we love enm Nor epu Esto erla Ja eal Finl Sp Aus elg an str erm ung Fra Pol bo Sta ingd ortu m C t Z B S u i d P S EC h R R D G (H A xe nite d K O w et h u Sw e ak c N e L v t e N U ni o z l C S U Note: Data refer to working-age persons (18-65 years old). Source: OECD.

Insights Melbourne Business and Economics

11

and, particularly, poverty if fiscal consolidation and austerity policies are not well balanced today.

What can policies do to reduce toohigh inequality? The most promising way of reducing too-high inequality is by boosting employment and career prospects. Hence, the focus for policy-makers should be on fostering more and better jobs, enabling people to escape poverty and offering real career prospects. Within current budgets, policies to address growing inequality could be made more efficient, for example, by making more use of inwork benefits which encourage people to take up paid work and give additional income support to low-income households. Such benefits are in place in about half of all OECD countries. Another important policy challenge is to improve access to,

and the quality of, education and training which will enable workers to take up better-paid jobs and thus reduce inequality. Investing in human capital is key. This must include the vital early childhood period and be sustained through compulsory education. This will ensure equality of opportunity for children from disadvantaged backgrounds. Once the transition from school to work has been accomplished successfully, there must be sufficient incentives for workers and employers to invest in skills throughout the working life. Reforming tax and benefit policies is the third key to promote a better distribution of income – taxes and benefits are the most direct instruments to redistribute income. As top earners now have a greater capacity to pay taxes than before, governments may consider re-examining their

Figure 6: Trends in income inequality (Gini coefficients) in selected countries, 1975 – 2010 0.40

USA

0.38 0.36

UK

0.34

Spain 0.32

OECD

0.30

Germany France

0.28

Sweden 0.26

Norway

0.24 0.22 0.20 0.18

1975 1980 1985 1990 1995 2000 2005 2010 Note: Break in series for Spain in 2000. Source: OECD.

12

Inequality in OECD countries: the facts and policies to curb it

tax systems to ensure that wealthier individuals contribute their fair share of the tax burden. This aim can be achieved in several different ways – not only via raising marginal tax rates on the rich but also improving tax compliance, eliminating tax deductions, and reassessing the role of taxes on all forms of property and wealth, including the transfer of assets.

1 Throughout this article we focus on income inequality. Arguably, wealth inequality is more important. However, comparable cross-country wealth data are much harder to collect than data on income. Where data exist they show that wealth is much more unequally distributed than income. See OECD (2008, Chapter 10) for a discussion.

Redistribution is not only about cash. Governments spend as much on public social services, such as education, health and care services, as they do on all cash benefits taken together – Australia spends as much on such services as a proportion of GDP as the OECD average while New Zealand spends more than the average. While the prime objective of such services is not redistribution, they reduce income inequality by a fifth. Public services such as high-quality education furthermore constitute a longer-term social investment to foster upward mobility and create greater equality of opportunities in the long run.

3 For a comprehensive review of the burgeoning literature on the topic, see Corak (2013).

The new OECD work shows that there is nothing inevitable about growing inequalities. Globalisation and technological changes offer opportunities but also raise challenges that can be tackled with effective and well targeted policies. Any policy strategy to reduce the growing divide between rich and poor should rest on three main pillars: inclusive employment promotion; a more intensive human capital investment; and welldesigned tax/transfer redistribution policies. John P. Martin is a Former Director, and Michael Förster a Senior Economist, in the Directorate for Employment, Labour and Social Affairs in the Organisation for Economic Cooperation and Development (OECD).   * This article is a revised and updated version of ‘Balancing economic efficiency and social equity’ which was published in the November/December 2012 issue of Japan SPOTLIGHT. The authors are extremely grateful to the editors of Japan SPOTLIGHT for giving permission to reproduce this material. The views expressed are our own and cannot be held to represent those of the OECD or its member governments.

2 Daughters’ earnings are excluded because of the great changes in women’s labour force participation rates across generations linked to rising educational attainments and different choices about work and unpaid activities made by women across different generations.

4 For more details on these estimates, see OECD (2008, Chapter 8). 5 This relatively benign conclusion about the impact of globalisation on earnings inequality needs to be nuanced, however, as it may vary depending on the institutions and policies of the countries in question. Evidence from OECD (2011) suggests that growing import competition from low-income developing countries such as China and India tended to widen wage dispersion though only in countries with less-strict employment protection legislation. Autor et al. (2012) argue that the rapid growth in China-US trade accounted for about 15 to 20 per cent of jobs lost in the US though they find no significant effect on wage dispersion. 6 This part draws heavily on OECD (2013).

References Autor, DH, Dorn, D and Hanson, GH 2013, ‘The China Syndrome: Local Labour Market Effects of Import Competition in the United States’, IZA Discussion Paper No. 7150, January. Corak, M 2013, ‘Income Inequality, Equality of Opportunity, and Intergenerational Mobility’, IZA Discussion Paper, No. 7520, July. D’Addio, AC 2007, ‘Intergenerational Transmission of Disadvantage: Mobility or Immobility across Generations’, OECD Social, Employment and Migration Working Papers No. 52, OECD Publishing. OECD (2008), Growing Unequal?, OECD Publishing. OECD 2011, Divided We Stand: Why Inequality Keeps Rising, OECD Publishing. OECD (2013), Crisis Squeezes Income and Puts Pressure on Inequality and Poverty, See www.oecd.org/social/inequality.htm

Insights Melbourne Business and Economics

13