Increasing income inequality in Sweden is not a (big) problem.
Despite the facts that developments of real wages exceed labour productivity developments and that the labour share in GDP is increasing, many claim that income inequality is a big problem in Sweden. That is wrong.
One of the most debated issues in the economic-political debate following the Financial Crisis has been the increased inequality in many countries. In Sweden, both income and wealth inequality are said to have increased. It is difficult to say anything about wealth inequality since data on that is scarce. Wealth is not taxed in Sweden since 2007 which means that data on wealth inequality must be estimated.[1]
Increased income inequality is often discussed in terms of increasing shares of capital income at the expense of labour income. In the previous post, I noticed that real labour income, productivity and the labour share all increased in Sweden since 2000.
The two figures below show these developments again. To begin with, between 2000 and 2017, Swedish real wages per hour worked outgrew labour productivity, cf. Figure 1.[2]
These developments show that one of the often-cited causes of increased income inequality, “de-coupling” of real wages and productivity, does not seem to have occurred on the aggregate level for Sweden. A de-coupling of the two entities imply a decreasing labour share of income and a redistribution of income from labour to capital. See for example for a discussion and further references. But as Figure 2 below shows, in Sweden, the labour share in GDP has increased.
No, changes in income inequality are not due to wage developments…
No, it’s not a one-to-one correspondence but the wage share has increased between 2000 and 2017 which also the real wage relative to labour productivity has.[3] But it may be that wage dispersion has increased if technological development would have caused a “polarization” of the labour force meaning that the low and top ends of the wage spectrum increase at the expense of the medium part. That could increase income inequality. Developments of this kind has had large effects in the United States as demonstrated by David Autor here. But the skills and education premia are relatively low in Sweden yielding a low wage dispersion which does not seem to be a major cause of increased income inequality, c.f. Figure 3.
And here is another study confirming that conclusion. However, income inequality in Sweden has increased. This is obvious if one looks at the development of incomes shares in equalised disposable incomes, c.f. Figure 4.
This post definitely needs a picture of a furry animal.
…but to increased capital income and decreased transfers…
As types of income differ across the population, a look at the different types of incomes before taxes can provide some explanation to the pattern above. Adding wages and entrepreneurial income shows that this share is relative constant while the capital income share increases and the transfer share decreases. Since capital income holds a larger share of disposable income for people and households in the upper part of the income distribution and transfers are more important for people and households in the lower part, the developments in Figure 5 below should lead to increased inequality.
What about the wage share developments from the two sources of statistics? The wage share above cannot be directly compared to the one in Figure 1 since that also included social security contributions. Adding pensions would not change the picture since that share is constant during the time period. Also, wages in National Accounts are based on what employees, some 4.5 million people, earn and related to GDP. Wages (“förvärvsinkomst”) in Income Statistics are earned by some 8 million people. I am too old to go deeper into this, but it seems that an increased wage share in GDP is compatible with a constant (and perhaps even slightly decreasing) wage share in gross income.
So back to income inequality in general. If neither a decreased wage share nor an increased wage dispersion have caused the income inequality to increase, what has or have? The figure above suggests that increasing capital income together with a lower unemployment and sickness benefits have something to do with it. As suggested capital incomes have increased their share in income in the top income decile, c.f. Figure 6.
…demographic changes have contributed but to a smaller extent. But hey, increased income inequality has occurred at rising real incomes for everyone…
These factors are also pointed out in this study. To those the authors add a third, namely demographic changes and changes in the composition of households. They find that changes in the composition of households and immigration also affect inequality though for Sweden, these factors had a smaller effect than those mentioned above.[4]
After rambling so much about income inequality it might be interesting to see that in the case of Sweden, all income deciles have experienced increased real incomes since at least 2000, c.f. Table 1.
So, income inequality is not a big problem
Now the question arises, is the increased income inequality a big problem and if it is, what should be done about it? If inequality is caused by widespread corruption causing rent-seeking and diverting resources away from production, then there is a problem. However, according to Transparency International’s Corruption Perception Index, Sweden is among the least corrupt countries in the World, so there may not be any large gains from addressing corruption.
Another problematic source of inequality is weak competition among firms. One recent study by OECD, shows that higher profits emanating from increased market power leads to higher inequality of wealth and income. Eurostat data on shares in value added, employment and number of firms for different industries do not indicate any tendencies towards increasing concentration with the exception for the energy industries and Wholesale.[5] And a higher concentration among firms would probably also show up in a larger wage dispersion.
That may not be the best ways to assess this problem. The Swedish Competition Authority performs an annual review of the situation in Sweden. In this report, the authority finds that some markets, especially construction and financial markets are dominated by the large firms. Also, digital platform markets is pointed out as suffering from weak competition. Addressing those weaknesses may also decrease inequality. (Yes, that report is in Swedish but as D.S, M.J. and E.M. would tell you, Swedish is the new lingua franca.)
Inequality can be good.
So, my conclusion is that income inequality is not a big problem in Sweden. Even if it may be tempting to change the income distribution through taxes and transfers, that also impacts on incentive to work and save. Lower unemployment and sickness benefits may in the short run, and surely during a slump, have negative effects but could be beneficial for growth in the long run. And if the increasing capital income is the result of returns from investments, I do not see a problem. A higher capital stock today allows for higher income and consumption in the future which is good for everyone.
Now I realise that this post was far too long and that I have been rambling more that usual. But before I leave it, I will give you some graphs showing developments in France. There has been a lot of turmoil there lately with outcries of increasing inequality and injustice. As the figure below shows, whatever the people in yellow vests tell you, any “injustice” is not due to capital owners taking a larger share of the pie (GDP), c.f. Figure 7
I am not prolonging this post by plunging into French statistics but I am sure that things are not as bad as the yellow vests and Russian Tv-stations like us to think.
Annex. Income concepts
[1] According to economic theory, real wages should be equal to the marginal productivity of labour while Figure 1 shows developments in average productivity of labour. With a Cobb-Douglas production function marginal productivity can be shown to be equal to the labour’s share multiplied by average productivity. This means that average and marginal developments over time are very similar.
[2] For estimates of household wealth in Sweden 2000-2012, see Lundberg, J. & Waldenström, D. (2017). See Waldenström has also constructed a Swedish national Wealth Database which can be found at http://www.uueconomics.se/danielw/SNWD.html
[3] Yes. They are also in different units. I was too lazy to do anything about that now I realised it.
[4] The study analysed effects up to 2013 which means that the effects of the large immigration waves are not included.
[5] Look here for yourself! https://ec.europa.eu/eurostat/data/database