Why We Don't Have to Choose... - related material

November 28, 2000

Why We Don't Have to Choose between Social Justice and Economic Growth: The myth of the equity/efficiency trade-off

Section II
The Equity/Efficiency Trade-off in Practice:
Empirical evidence for North American and European OECD countries in the 1990s

(A) An Overview of GDP Growth and Income Inequality

Table 1 presents data for per capita GDP growth in the 1990s and for household after-tax income inequality (adjusted for household size) in 1995 for Canada, the U.S., Australia and nine European countries. The income inequality (Gini) data are from the Luxembourg Income Study database, and they dictate the size of the total sample of OECD countries. (The Gini co-efficient is the most widely used statistical measure of inequality.) Chart 1 presents the data in such a way as to indicate the relationship between the two variables.

Source: Growth data are from the report of the OECD Statistical Working Party, Economic Growth in the OECD Area: Are the Disparities Growing? (OECD, November 1999). American and Canadian GDP data have since been revised, but this is still the most reliable source of comparative data, since GDP data for other countries have not been revised on the same basis. It should be acknowledged that the U.S. has turned in a very good growth performance in the second half of the 1990s.The Gini data are from Lars Osberg and Andrew Sharpe, International Comparisons of Trends in Economic Well-Being (Centre for the Study of Living Standards, January 2000).

Chart 1: Taxes and Growth in OECD Countries

The key point that emerges from this comparison is that relatively low income inequality – Gini less than .300 – is associated with either low or high growth. Two low inequality countries – the Netherlands and Denmark – did very well in terms of per capita growth in the 1990s (2.1% and 2.3%, respectively) and indeed, growth in these countries over the period was more rapid than in the U.S. or the UK, the two countries with the highest levels of after-tax income inequality. There was no statistically significant relationship between growth and income inequality in the 1990s.

Nor was there a consistent or significant relationship between per capita GDP growth in the 1990s and the change in inequality between 1990 and 1997, as measured by the change in the Gini (data provided below in Table 9). It is notable that there was, in fact, a negligible increase in after-tax income inequality in Denmark and the Netherlands in the 1990s, while the increase in inequality by this measure was much greater in the U.S. than it was elsewhere. Thus, relatively high growth was associated with markedly divergent trends in terms of the change in the income distribution.

The statistically insignificant relationship between growth and inequality, and between growth and changes in inequality, is not surprising, given that GDP growth can be associated with very different patterns of employment performance, and given the wide variations in initial conditions among the countries, not least of which is the

nature of the tax/transfer system, the fiscal position of governments, wage structures, and so on. Some of these differences are explored below.

The conclusion to be drawn is that it has clearly been possible in the 1990s for relatively equal societies to achieve relatively strong rates of economic growth, and to do so without increasing inequality.

(B) Taxes and Growth

The tax share of GDP is a key indicator of the combined size of a country’s redistributive tax/transfer system and the extent of access to (equality enhancing) non-market services.

Table 2 provides data on per capita GDP growth and productivity growth as well as data on the total tax burden for an almost complete sample of OECD countries in the 1990s. There was no statistically significant relationship between the tax "burden" and either economic growth or productivity growth. Per capita GDP growth has been higher in the U.S. than in most major European countries in the 1990s, but a number of high-tax countries (such as Denmark and the Netherlands) have grown as fast as, or faster than, the U.S., despite – or even because of – a higher tax burden. Many countries with high tax burdens have had strong productivity performance.

Table 2
Taxes and Economic Performance

	Growth of GDP per capita 1990 to 1998	Taxes as % GDP 1994	Growth of GDP per Hour Worked 1990 to 1998
Australia	2.2	28.7	2.3
Austria	1.5	43.3	1.8
Belgium	1.5	45.9	2.4
Canada	1.0	35.1	1.2
Denmark	2.3	49.9	2.1
Finland	1.0	46.7	3.0
France	1.0	43.7	1.5
Germany	0.9	38.4	2.9
Greece	1.2	31.7	1.3
Iceland	1.1	30.9	na
Ireland	5.5	35.7	4.2
Italy	1.1	41.4	1.9
Japan	1.0	27.8	1.9
Netherlands	2.1	44.7	1.3
New Zealand	0.6	36.7	0.2
Norway	3.2	41.3	2.7
Portugal	2.2	32.6	3.2
Spain	1.9	35.0	1.8
Sweden	0.6	49.0	1.9
Switzerland	-0.3	33.0	0.8
United Kingdom	1.7	34.5	2.2
United States	1.7	28.4	1.2

na=not available
Note: Korea, Mexico, Czech Republic, Hungary and Poland are excluded.
Source: OECD Statistical Working Party, Economic Growth in the OECD Area: Are the Disparities Growing? November 1999 (DTSI/EAS/IND/SWP(99)3.)

However, there is a strong and significant statistical relationship between the size of the tax "burden" and relatively low income polarization, as indicated by the ratio of the top to the bottom deciles of the income distribution (see data in Table 3). The Spearman’s Correlation Coefficient between taxes and the decile ratio is –0.818, significant at the 0.01 level.

The decile ratio (the ratio of the after-tax incomes of the top 10% of households compared to the bottom 10%) clearly indicates continuing and very large differences in the income distribution among advanced industrialized countries in North America and Europe. The ratio varies from a minimum of more than 6 in the U.S., to 4 in Canada, to 3 or less in the Benelux and Scandinavian countries.

Table 3
Taxes and Inequality

	Taxes**	Decile Ratio***
United States	28.4	6.44
Japan	27.8	4.17
Germany	38.4	3.84
France	43.7	4.11
Italy	41.4	3.14
UK	34.5	4.56
Canada	35.1	3.93
Belgium	45.9	2.79
Denmark	49.9	2.86
Netherlands	44.7	3.05
Norway	41.3	2.85
Sweden	49.0	2.78

** Taxes as a per cent of GDP in 1994 (OECD).
***The Decile Ratio is the ratio between the top of the bottom decile and the bottom of the top decile, of the household after-tax income distribution and thus slightly understates the gap between the top and the bottom of the income distribution. Luxembourg Income Survey data for the mid-1990s. Data presented by Timothy Smeeding to the 1999 Canadian Economics Association Meetings.

To summarize, the relationship between taxes as a share of GDP and growth and productivity is statistically insignificant, while the relationship to inequality is strong and significant.⁷ Seen from this perspective, relatively high taxes represent a social choice in favour of public as opposed to private consumption, and in favour of redistribution through the tax/transfer process, and it not true to say that this choice comes at the cost of reduced national income.

Table of Contents | Previous Section | Next Section

Why We Don't Have to Choose... - Related Material

Canadian Council on Social Development, 190 O'Connor Street, Suite 100, Ottawa, Ontario, K2P 2R3
Tel: (613) 236-8977, Fax: (613) 236-2750, Web: www.ccsd.ca, Email: council@ccsd.ca