A Swedish TV documentary argues that the share of income taken by capital has increased significantly at the expense of the Labor Share. Obviously that would be a problem, since the middle class gets most of their income from wages and owns little capital.
The conclusion looks reasonable on the surface. However if we look deeper you will see that the result is driven by an oversimplified way of measuring the Labor Share and not taking Taxes into account.
The National Income Accounts is the part of National Accounts that estimates what share of the total GDP ends up as wage or capital income. In Sweden, this is calculated by SCB.
In the simplest analysis you divide GDP between wages and capital income. A common theoretical result in macroeconomics is that the Labor-Share is proportional to the marginal product of labor and constant over time.
In the 1950s economists realized that in the modern world, across countries, the Labor-share was surprisingly stable. Economists believe that 1/3 of total income goes to capital and 2/3 goes to labor. The empirical observation that the Labor-share is constant over time is one of Kaldor’s stylized facts.
Unlike what the some claim this is not just a theoretical assumption, it is a robust empirical observation over a long period of time in many countries in the modern era. (During early industrialization the capital share was closer to 1/2).
Reality is more complex than this simple division. GDP is not just divided between Capital and Labor in income accounts. GDP is often divided into even more segments, including some Taxes, which complicates the analysis.
Here are Income-account of Swedish GDP in 2011 in billions of kronor:
1. Gross Domestic Product: 3500
2. Consumption of fixed capital (depreciation of capital): -455
3. Net National Product (minus destroyed capital): 3045
4. Compensation of Employees (inc. compensations): 1837
5. Net Operating Surplus (of firms and self-employed): 616
6. Taxes on Production and Imports minus Subsidies: 552
If you just look at the compensation of employees as a share of Net National Product, the number has indeed fallen significantly from 70% of national income in 1980 to 60% in 2011.
It is however important to realize that this has not happened because capital is getting a bigger share. The surplus of business was 19.6% of national income in 1980 and 20.2% of national income in 2011.
What is happening is that the share of income allocated to the category “Taxes on Production and Imports – Subsidies” have risen sharply in Sweden. In 1980 this was 10.6% of GDP, and in 2011 it was 19.2% of GDP. This includes Moms, the Swedish VAT, which has risen significantly.
Obviously net wages going down because the government is taking in more taxes is not the same thing as capital taking a bigger share of the pie. Tax revenue is mostly spent on workers.
If we remove the other stuff and just divide income between workers and capital you see that no large shifts have occurred since 1980. The long-held belief by economists that the division between capital and labor is fairly stable holds in this period in Sweden once we adjust for taxes.
This conclusion is supported by Konjunkturinstitutet, who concluded that the Labor-Share has not declined and that the probability of Swedish industry has remained fairly constant between 1981-1997 and 1998-2011 (page 29). The report also has this graph:
It shows that wage increases between 1980-2011 were almost as fast as increases in worker’s productivity during the period. If the capital share had increased sizably at the expense of the labor share we would not observe this, we would observe worker’s wages rising significantly slower than their output.
Hasn’t income inequality in Sweden risen during this period? Yes, it has. However the fundamental issue in modern income inequality is no longer about capital vs. labor, it’s about highly educated employees vs. others. The overwhelming majority of high income earners in Sweden are high-wage employees, not passive capital owners. This includes people in finance, lawyers, doctors, consultants, executives and entrepreneurs who combine high wages with capital gains. Income inequality has risen because the earnings of this group has gone up more than for everyone else, not because capital returns are higher.
Here is an article by Tim Worstall in Forbes Magazine making a similar point about the U.K.
The producers of the program argue that I am wrong, and refer to wages as a share of GDP at factor prices. The producer is right that using factor prices solves the problem of VAT (moms). However, that is not the only adjustment I made. You also have to adjust for capital depriciation. Every year a lot of capital loses value, for example industrial machines and buildings that become less useful or equipment that is outdated. This is part of GDP, but it is not part of national income, since it cannot be consumed. To calculate capital income you must remove the cost of capital depreciation. Here is the labor share of GDP at factor prices, including capital depreciation.
As you see, the capital income share has not increased significantly. A falling wage share does not automatically mean that more income is flowing to capital. Not all of GDP is either capital income or labor income. In this case the falling wage share is due to rising production taxes and rising capital depreciation. The producers of the show do not appear to have adjusted for rising capital depreciation.
Konjunkturinstitutet also tracks this issue. Click on “BNP fördelad på vinster (driftsöverskott) och arbetskostnader”. You will see that ”Arbetskostnadsandel” was 63 percent in 1993 and 66 percent in 2012.
Here is a graph going back to 1980 in a factbook by Konjunkturinstitutet and Medlingsinstitutet showing no change in the profitshare. (The definitions are all a bit different but the results are the same).
Here is a graph from Riksbanken in 2003 with no change since 1970 in the profit share.
I see no evidence for the claim that the share of income that goes to capital rather than labor has increased in Sweden in recent decades.