THIS CHAPTER LOOKS AT THE COMPLEX RELATIONSHIPS BETWEEN INEQUALITY and economic growth. Rather than accept the simple proposition that inequality is good (or bad) for growth, the chapter looks at how different types of inequality have different effects on growth and how some solutions to the inequality problems reduce growth while others increase it.
There have been many attempts to formalise a relationship between inequality and growth.
One theory is that income differences drive incentives. If that is so, the most highly incentivised should be those who are generating most growth, in which case one might expect high inequality to be combined with rapid growth. In 1975 Arthur Okun, an American economist, wrote Equality and Efficiency: The Big Tradeoff.1 He argued that societies cannot have both perfect equality and perfect efficiency, but must choose how much of one to sacrifice for the other. Many economists today continue to believe this. Okun wrote:
Contrast among American families in living standards and in material wealth reflects a system of rewards and penalties that is intended to encourage effort and channel it into socially productive activity. To the extent the system succeeds, it generates an efficient economy. But that pursuit of efficiency necessarily creates inequalities. Hence, society faces a trade-off between equality and efficiency. Tradeoffs are the central study of the economists. You can’t have your cake and eat it, too, is a good candidate for the fundamental theorem of economic analysis. We can’t have our cake of market efficiency and share it equally.
The theory has an intuitive plausibility but suffers from one major defect – the facts do not seem to fit it. Most cross-sectional examinations of the relationship between inequality and growth over a range of economies show either that there is no significant relationship or that the relationship is that the less the inequality, the faster the rate of growth.2
Indeed an IMF study concludes that a one-percentage-point increase in the income share of the top 20% will drag down growth by 0.08 percentage points over five years, while a rise in the income share of the bottom 20% actually boosts growth.
Given that the evidence seems to hint that inequality harms growth, research in recent years has concentrated on trying to explain the transmission mechanism. One theory is that inequality could impair growth if those with low incomes suffer poor health and low productivity as a result or if the poor struggle to finance investments in education.
An alternative theory is that inequality boosts pressure for bad economic policies (i.e. protectionism and limits on immigration or anti-capitalist policies) which were mentioned at the beginning of this book as likely to inhibit growth.
There is also some evidence to show that inequality can cause a savings glut. Well known economists such as Ben Bernanke and Larry Summers have argued that inequality contributes to the world’s ‘savings glut’ since the rich tend to spend a lower proportion of their incomes than the poor. And they argue that there are consequential results – higher savings cause interest rates to fall, boosting asset prices, encouraging borrowing and making it more difficult for central banks to manage the economy.
On the other hand, those advanced economies that have not had significant growth in inequality in the 1980-2008 period were those who also had the least GDP growth, such as Japan, Italy and France. Those with a larger rise in inequality, such as the UK and the US, showed stronger economic performance.
It seems probable that a range of mechanisms are at work, not all pointing in the same direction.
The Kuznets theory would imply that the early phases of economic development, when growth is often especially fast, are often phases where inequality is large and is increasing. At later phases when an economy is maturing and growth as a result is slowing, inequality would likely be diminishing.
It is important therefore to compare countries that are at the same phase of economic development. Yet even there, countries have widely differing social approaches. One typical contrast in advanced economies is between countries with a fairly homogeneous structure and ones which may be very open to migration but where the different groups are segregated. It is possible that the homogeneous structure, particularly if associated with a small underclass and a strong educational performance at the bottom end of the ability range, would generate both higher average productivity and faster GDP growth. It is possible that some of the Scandinavian countries have benefited from this in the past, though there are hints of evidence that increased migration is causing their social structures to become less homogeneous.
There is a fascinating analysis, so far available only in Dutch, that tries to unpick some of the social influences behind the different levels of equality and inequality in different but otherwise relatively comparable countries in Europe. Bert Bakker, a Dutch financial journalist, has written about how the economic history of different European economies has led to different social backgrounds that have resulted in different levels of equality.3
In a letter published in the Financial Times on 28 April 2018 Mr Bakker wrote:
The hypothesis I worked out in Onbegrepen Europa (loosely translatable as ‘Europe under the surface’) is that some regions of Europe have always been part of an empire (Roman, Frankish, Habsburg) governed along feudal, centralistic lines. Large landownership for a few and serfdom for the majority was the economics and societal model.
However, in some other – often remote – regions, such as Scandinavia, the Low Countries and Switzerland, the conditions (lack of vast estates of arable land) prompted a large proportion of the people to make a living on their own, as small-scale farmers, tradespeople, traders, transporters, shipbuilders (Viking’s Frisians and later Hanseatic shippers).
The proportionately large middle class in these countries ruled themselves, knew material autonomy and thus laid the basis for the egalitarian, trust-based, individualistic attitudes we know now in the Nordic countries, the Netherlands and Switzerland. But also, for the same reasons, in the north of Italy and in Catalonia, Spain. This accounts for two very different European value systems.4
I suspect that a rigorous examination of different sets of circumstances would discover the following ‘stylised facts’ about the relationship between the level of inequality and GDP growth:
(1) Homogeneous societies with strong social networks and good education tend to have lower inequality and high GDP and fairly high rates of GDP growth.
(2) Highly capitalistic and free market societies with less strong social networks tend to have high levels of inequality and high rates of GDP growth.
(3) Economies with high levels of immigration and weak social networks tend to have high levels of inequality and high rates of GDP growth.
(4) Economies with high levels of inertia and historic inequality tend to combine high levels of inequality and low rates of GDP growth.
(5) Dictatorial and autocratic societies tend to have high levels of inequality. Their rates of growth can sometimes be high but over time tend to become lower and sometimes even negative.
(6) Communist societies can have low levels of inequality (though not all do) but typically have low rates of GDP growth and quite often have falling GDP (though measuring GDP and GDP growth in countries where there are no proper markets can be difficult).
(7) Emerging economies tend to have high rates of inequality and high rates of growth.
(8) Mature economies tend to have lower rates of inequality and lower rates of growth.
The key point about this is that even though on balance the data seem to show a correlation between less inequality and higher growth, it is not clear that reducing inequality, for example by redistributive taxation, will lead to higher economic growth. It is the underlying factors affecting the inequality, such as social structures, incentive systems, educational inequality and the quality of parenting that affect economic performance as well as the pervasiveness of social and economic incentives that are more important.
Inequality and economic development5
The relationship between aggregate output and the distribution of income is an important topic in macroeconomics.6 The World Bank Group has included among its key global objectives for development the eradication of extreme poverty and boosting the incomes of the bottom 40% of developing countries. The IMF has also involved itself in the discussion about the role of income distribution as a cause and consequence of economic growth.7
In a recent paper Brueckner and Lederman provided estimates of the within-country effect that income inequality has on aggregate output.8 Their empirical analysis started from the premise that the effect of changes in income inequality on GDP per capita may differ between rich and poor countries. This premise is grounded in economic theory.
In a seminal contribution, Galor and Zeira proposed a model with credit market imperfections and indivisibilities in investment to show that inequality affects GDP per capita in the short run as well as in the long term.9 Galor and Zeira’s model predicted that the effect of rising inequality on GDP per capita was negative in relatively rich countries but positive in poor countries. Brueckner and Lederman tested this prediction by introducing in the panel model an interaction term between income inequality and countries’ initial GDP per capita. Their empirical analysis showed that for the average country in the sample during 1970-2010, increases in income inequality reduced GDP per capita.
Specifically, they found that a one-percentage-point increase in the Gini coefficient reduced GDP per capita by around 1.1% over a five-year period and that the long-run (cumulative) effect was larger and amounted to about –4.5%.
The estimates from the interaction model thus suggest that in poor countries, increases in income inequality raise GDP per capita while the opposite is the case in high- and middle-income countries.
Redistribution and growth
Because the evidence seems to show that high inequality is correlated with lower growth, one might think that redistributing income and wealth from the rich to the poor would boost growth. But real life is messy and the studies seem to show that policies involving high rates of tax and redistribution in various ways if anything reduce growth.
The seminal work suggesting that high levels of government (current) spending hinder economic growth was by Robert Barro in 1991.10
His study showed that ‘for 98 countries in the period 1960-1985, the growth rate of real per capita GDP is positively related to initial human capital (proxied by 1960 school-enrolment rates) and negatively related to the initial (1960) level of real per capita GDP’. Barro concluded:
Countries with higher human capital also have lower fertility rates and higher ratios of physical investment to GDP. Growth is inversely related to the share of government consumption in GDP, but insignificantly related to the share of public investment. Growth rates are positively related to measures of political stability and inversely related to a proxy for market distortions.
There is also an interesting concept called the Scully Curve.11 This is like the Laffer Curve but has a stronger scientific basis.
What it says is that there is a quadratic relationship between economic growth and public spending as a share of the economy. Economic growth varies positively with the share of public spending and negatively with the square of public spending. This gives a curve where initially growth rises with higher public spending but as the quadratic term starts to dominate, the relationship between growth and public spending first levels out and then shows lower growth for higher public spending.
When public spending is too low, the economy lacks the public infrastructural support to grow strongly. When it is too high, public spending crowds out the entrepreneurial sector and causes taxes to be so high that they destroy incentives. The latest estimates of the Scully Curve from the Fraser Institute in Canada show that the difference between public spending at 25% of GDP as in much of Asia and 45% as in the UK can be the difference between GDP growth at 1.5% per capita per annum and growth at 3%.12 If this really is the case, it points out that in this area there is likely to be a trade-off between inequality and growth since one of the purposes of public spending is in many circumstances to reduce public spending.
The analysis of public spending and growth has developed more dimensions since the initial work. In general, the latest research suggests that educational spending can boost growth and that infrastructural spending at least does not hinder growth, but that other areas of public spending remain generally negative for growth. Spending on welfare in particular can have a dramatic effect on inequality but appears to reduce growth both through the public spending mechanism and more directly through impacting on the labour supply.
And research confirms that the damage to growth from public spending rising as a share of GDP grows substantially as the share rises from 40%. The Institute of Economic Affairs in the UK has produced a useful table of studies on this and their conclusions (Table 5).
Table 5. Studies of public spending and growth13
| Author | Data coverage | Main explanatory variables | Comment |
| Barro (1991)14 | 98 countries in the period 1960-85 | Human capital, government consumption, political instability indicator, price distortion | 1 % point increase in tax-to-GDP ratio lowers output per worker by 0.12% |
| Koester and Kormendi(1989)15 | 63 countries for which at least five years of continuous rate, data exists for the 1970s | Marginal tax rates, average tax mean growth in labour force and population | 10% decrease in marginal tax rates would increase per capita income in an average industrial country by more than 7% |
| Hansson and Henrekson(1994)16 | Industry level data for 14 OECD countries | Government transfers, consumption, total outlays, education expenditure government investment | Government transfers, consumption and total outlays have a negative impact on growth whilst government investment is not significant |
| Cashin (1995)17 | 23 OECD countries over the 1971-88 period | Ratio of public investment to GDP, ratio of current taxation revenue to GDP, ratio of expenditure on transfers to GDP | 1% increase in tax to GDP ratio reduces output per worker by 2% |
| Engen and Skinner (1996)18 | US modelling Together with a sample of OECD countries | Marginal tax rates, human capital, investment | 2.5% rise in tax to GDP ratio reduces GDP growth by 0.2-0.3% |
| OECD (Leibfritz et al.) 199719 | OECD countries over the 1965-95 period | Tax to GDP ratio, physical and human capital formation, labour supply | 10 point increase in tax to GDP ratio reduces GDP growth by 0.5-1% |
| Alesina et al. (2002)20 | 18 OECD countries over the 1960-94 period | Primary spending, labour taxes, transfers, government wage consumption, indirect taxes (all as shares of GDP) | 1% increase in government spending relative to GDP lowers the investment to GDP ratio by 0.15% and a cumulative fall of 0.74% after 5 years |
| Bleaney et al. (2000)21 | 17 OECD countries over the 1970-94 period | Distortionary tax, productive expenditure, net lending labour force growth, investment ratio | 1% increase in distortionary tax revenue reduces GDP growth by 0.4% points |
| Folster and Henrekson (2000)22 | Sample of rich OECD/non OECD countries over 1970-95 period | Tax to GDP, government expenditure to GDP, investment to GDP, labour force growth, investment ratio | 10 point increase in tax to GDP ratio reduces GDP growth by 1% |
| Bassanini and Scarpetta (2001)23 | 21 OECD countries over the 1971-98 period | Indicators of government size and financing, physical capital, human capital, population growth | 1% point increase in tax to GDP ratio reduces per capita output levels by 0.3-0.6% |
In this section I draw heavily on Patrick Minford, ‘Tax and Growth’.24
An important set of work that became known as the ‘Barro growth regressions’ has been widely used to investigate policy impacts since Barro’s 1991 work).25 These take the form of a regression of either GDP growth or productivity growth on initial income (to control for convergence or ‘catch-up’), some factors to account for other influences, and a variable measuring the policy factor under investigation. Models of this type are often estimated in a panel of cross-country and time-series data with observations averaged over five-year periods to smooth out the impact of the business cycle. Barro used data for 98 countries between 1960 and 1985.
Leach26 and OECD document the empirical literature on the effects of taxation on growth and output levels. Table 5 sets out a selection of the major studies, noting their data set, the explanatory variables used and the main effects of tax on growth that are found. All control for various factors other than taxation (usually different variables across different studies). Some of these studies use tax as the explanatory variable and others use government spending.
In theory, the latter is preferable because government spending measures the total claim on the economic resources of government. Barro did, in fact, consider that some government spending, in theory at least, might have a positive impact on growth.
He believed that there may be a positive effect of education spending on human capital formation and therefore on growth and takes that into account. As such, he examined the impact of real government consumption net of spending on both education and defence as a percentage of real GDP over the period 1970-85 on both real economic growth (averaged over the period 1960-85) and on private investment. He found a negative correlation between net government spending (so defined) and growth.
Koester and Kormendi examined the effect of measures of the marginal and average tax rates, as well as population and labour-force growth on economic growth. In a cross-country analysis for the 1970s, they found a significant negative effect of marginal tax rates on the level of real GDP per capita, but not on the rate of growth when this was controlled for the initial level of income. They suggested that, holding average tax rates constant, a 10 percentage point decrease in marginal tax rates would increase per capita income in an average industrial country by more than 7% (and in an average developing country by more than 15%). Thus, a revenue-neutral tax reform which reduced tax progressivity would raise incomes.
Alesina et al. focussed on the extent of government spending of various sorts on the investment-to-GDP ratio (and hence by implication on growth). They concluded that, via the effect of raising private sector labour costs, a 1 percentage point increase in government spending relative to GDP resulted in a decrease in the investment-to-GDP ratio of 0.15 percentage points and a cumulative fall of 0.74 percentage points after five years. In general these studies, with their varying methodologies, find that there are measurable negative effects of higher tax rates on growth.
The order of magnitude of this effect was estimated to be around 0.5-1.0% for a 10% rise in the ratio of taxation (or government spending) to GDP. The OECD’s own conclusion from its survey was as follows:
A number of studies, influenced by the new growth theories, have taken a top-down approach to assess the impact of taxes on per capita income and growth at the macro level. Several of them purport to demonstrate a significant negative relationship between the level of the tax/GDP ratio (or the government expenditure ratio) and the growth rate of GDP per capita, implying that high tax rates reduce economic growth … our estimates [using a top-down cross-country regression] suggest that the increase in the average (weighted) tax rate of about 10 percentage points over the past 35 years, may have reduced OECD annual growth rates by about 0.5 percentage points.
The OECD study suggested that a 10 percentage point cut in the tax-to-GDP ratio could increase economic growth by 0.5-1.0 percentage points. Thus they also argued that ‘up to one third of the growth deceleration in the OECD [over the 1965-95 period] would be explained by higher taxes. In some European countries, tax burdens increased much more dramatically than the OECD average, which would imply correspondingly larger effects on their growth rates’.
Taken together, all these studies seem to show that subsidies and government current expenditure other than on education have the worst negative effects on growth on the spending side. Overall, the tax (or government spending) and growth studies indicate a strong association between the two variables.
As a rule of thumb, the conclusion of the studies is that it would appear that a 10 percentage point fall in the share of national income taken in tax would lead to around a 1 percentage point increase in the growth rate – results of this order of magnitude occur over and over again. This does not mean that the US or the UK can automatically expect to increase their growth rate by 1 percentage point if the government reduced the proportion of national income it spends from (say) 45% to (say) 35%. First, there is the issue of causality. For example, does high government spending lead to low growth or the other way round? Second, the results are averages using data taken from a wide variety of situations. Third, the effects are likely to be non-linear. The impact of a 1 percentage point change in government spending relative to national income cannot simply be multiplied by ten to find the impact for a 10 percentage point change. Intuitively, it is likely that the gains in terms of extra growth at the margin will reduce as the tax burden falls and rise as the tax burden increases.
The analysis of the Scully curve suggests that the growth-maximising average spending ratio is 17.5-22.5% of national income and the welfare-maximising point is 27.5-32.5% of national income. Broadly, the model suggests that a cut in government spending to the welfare-maximising point from current levels might imply a rise in the economic growth rate by around 1 percentage point; and a cut to the growth-maximising point would imply a rise in the economic growth rate by a further 0.8 percentage points. These are clearly orders of magnitude and the effects of cutting government spending are unlikely to be a simple linear function of the extent of the cut. However, there is a lot of evidence from different sources that points towards figures of this order.
Profits and investment
The relationship between profits and investment is a complicated one. And extremists on all sides make seemingly plausible assertions which do not stand up to the test of empirical economic analysis.
One extreme view is that there is no relationship between profits and investment. The opposite view is that all profits get ploughed back into investment. Neither is correct, though those that argue with the view that profits lead to investment often forget that redistributed profits are often reinvested and that high income earners have a high investment ratio.
The most highly cited paper on profits and investment argues that there is a very strong relationship between profits and investment.27 The argument is not a cashflow argument (although this probably provides an element of the correlation) but more that profits act as a signal to investors and hence encourages them. This seems intuitively plausible, especially if one has some experience of investment decisions in the financial sector or at board level.
Redistributing to the poor
Obviously if a sum of money is taken from a rich person and given to a poor person it initially reduces both inequality and poverty. But whether redistribution to the poor helps reduce poverty in the long term is more controversial.
In the United States, members of both the Republican and Democratic Party (as well as third parties such as the Libertarians) have favoured reducing or eliminating welfare. The landmark piece of legislation which reduced welfare was the Personal Responsibility and Work Opportunity Act passed under the Clinton administration though on the basis of strong pressure from the Republicans who controlled both Houses of Congress at that time.
Conservative and libertarian groups such as the Heritage Foundation28 and the Cato Institute29 assert that welfare creates dependence and a disincentive to work, and reduces the opportunity of individuals to manage their own lives. This dependence is called a ‘culture of poverty’ which is said to undermine people and prevent them from finding meaningful work. Many of these groups also point to the large budget used to maintain these programmes and assert that it is wasteful.
In the book Losing Ground, Charles Murray argues that welfare not only increases poverty but also increases other problems such as single-parent households and crime.30
All this is not to say that any kind of redistribution is bound to fail. What it says instead is that when redistributing two things are critical: first, any redistribution has to be in a form that minimises the disincentives to work; and second, any redistribution needs to ensure that it has as little impact as possible on the tax burden, particularly the tax burden on the rich who are both mobile and typically have enough income and wealth to make choices about how many hours they work.
Conclusions
All this shows that the relationship between inequality and growth is a mixed one. The trick is not to make crude assumptions based on correlations. For example, it seems to be generally the case that higher inequality is associated with less growth. So one might be tempted to recommend raising taxes to redistribute from rich to poor to boost growth. But then one discovers that higher taxes are (generally, although it does depend on the starting point) also associated with lower growth. What to do?
The answer is to redistribute in the right way. First, concentrate on universal benefits. These do not create the negative work incentives that other forms of redistribution can. Second (because affordability is critical here) get the cost of living down, especially for those elements which affect the cost of living of the poorest groups. Third, redistribute while keeping taxes down to the extent possible. There probably has to be a trade-off here. What this also means is that government has to be rigorous about avoiding spending on non-mission-critical items and on eliminating waste. It is possible that one escape from the impact of high compulsory taxes on growth is to look at voluntary taxes, and I devote a section of Chapter 17 to this.