Lane Kenworthy, The Good Society
June 2023

Capitalism has been the engine of unparalleled increases not only in economic output but also in living standards for ordinary people. Even when the rich get a disproportionate share of the proceeds of economic growth, there has tended to be enough “trickle down” that households in the broad middle class and at the bottom see their well-being rise significantly.¹ This happens partly via government transfers and services but also partly — and for many people mainly — through more jobs and rising wages.²

However, capitalism creates incentives for firms to try to minimize labor costs.³ And since the late 1970s, a variety of developments — computers and robots, globalization, heightened product market competition, looser labor markets, and the turn to a “shareholder value” orientation in corporate governance — have increased firms’ incentive to resist wage increases and enhanced their leverage vis-à-vis workers. In some of the affluent democratic nations, especially the United States, the result has been stagnant wages and limited increases in household incomes despite healthy economic growth.⁴

One helpful response is to increase the expansiveness and/or generosity of public goods, services, and insurance programs — from early education to active labor market policy to unemployment insurance to eldercare and much more. In most rich democratic countries, government pays partially or fully for these things via tax revenues. Many households utilize them, and some rely quite heavily upon them. Increasing the number of such programs, boosting their quality, and reducing their cost is therefore likely to improve the lives of many ordinary citizens. We know from the experience of the Nordic countries and others that this is good for living standards, economic security, equality of opportunity, and probably happiness — and it is doable without harming the economy or creating heavy government debt.⁵

We also want incomes to rise. In a good society, the incomes of most people, regardless of their position on the income ladder, should increase at roughly the same pace as the economy (GDP per capita).

Figure 1 shows that this is what happened in the United States from the mid-1940s through the late 1970s, the so-called “golden age” of post-World War Two capitalism. However, since the late 1970s the incomes of ordinary households have risen much more slowly than the economy.

The chief proximate cause of this slow growth of incomes is slow growth of earnings. Figure 2 shows median pay in the US. (These data are available only since the early 1970s.) There has been little increase.

Figure 3 shows the best picture we have of wage developments in all of the rich democracies in recent decades. It covers the period from 1995 to 2013 (the earliest and latest years of available data). On the horizontal axis is each country’s economic growth rate, and on the vertical axis is its growth rate of median compensation.⁶ The dashed line in the chart is a 45-degree line; a country will be on the line if median compensation has grown at the same pace as the economy. Again, that’s what we would hope for in a good society: ordinary people see their wages rise in sync with the economy’s rate of growth. But it’s what has actually happened only in some these countries. Many of them, including the United States, lie well below the line, with compensation increases lagging behind economic growth.

What, if anything, can be done to fix this problem?

Skip to:

• Strong labor unions would help but may not be feasible
• Six strategies that may not help much
• Four promising strategies
• Summary

STRONG LABOR UNIONS WOULD HELP BUT MAY NOT BE FEASIBLE

Across the rich democratic nations, the degree to which compensation growth has kept up with economic growth is largely a function of labor unions and collective bargaining. The Nordic nations and Belgium have high unionization rates, so it isn’t surprising to see them close to the line in figure 3 above. France and the Netherlands also are close to the line. In France the unionization rate is low, but the law requires extension of collectively bargained wage agreements to nonunionized workers. In the Netherlands this kind of extension of collectively bargained agreements isn’t legally mandated, but it is a strong norm. The countries with the lowest rates of unionization and no compensating mechanism, such as the United States, sit farthest below the line. According to one estimate, if labor unions today had the same power as they had in the early 1980s, compensation for ordinary American workers would be about 25% higher than it actually is.⁷

The fact that labor unions have been critical for wage growth in contemporary rich democratic countries is a problem, because unions have weakened over the past half century.

In the United States and other affluent democracies, labor unions arose with the industrial revolution in the mid-1800s. American unions grew in size and strength through the first several decades of the 20th century. In 1935, the National Labor Relations Act guaranteed the right of workers in private-sector firms to form a union, required firms to negotiate with that union, and ensured protections for workers who go on strike. Over the ensuing twenty years union membership surged, reaching a peak of around 33% in the decade from 1945 to 1955. Since then, unionization in the United States has fallen steadily and sharply. Today, only 10% of employed Americans are union members.⁸

In the 1970s and 1980s, America’s union decline was widely viewed as exceptional.⁹ But developments in recent decades suggest a different conclusion. As figure 4 makes clear, unionization rates have been falling in most of the affluent democratic nations. Only five countries — Belgium, Denmark, Finland, Norway, and Sweden — still have a rate above 35%, and four of those five are helped by the fact that access to unemployment insurance is contingent on union membership. The average unionization rate in the other 15 countries dropped from 42% in the late 1970s to just 19% in the late 2010s.¹⁰

The situation isn’t as dire when it comes to the share of workers whose pay is determined by collective bargaining, because, as I just noted, extension practices in some countries mean that pay developments for workers who aren’t represented by a union are nevertheless determined by a collective agreement. Collective bargaining coverage has remained relatively high in some rich democratic countries despite low unionization. But in a number of these nations bargaining coverage is quite low, and it may decline in others, as it already has to a significant degree in the United Kingdom and Germany.¹¹

So while strong labor unions almost certainly would be of considerable help in boosting incomes, it’s unlikely that very many of the rich democracies will be able to make use of this solution.

SIX STRATEGIES THAT MAY NOT HELP MUCH

Faster economic growth

As I noted earlier, over the long run economic growth has tended to yield significant increases in living standards for ordinary citizens in the rich democracies. Economic growth has been slower since the late 1970s than it was during the post-WW2 “golden age.” Would a resumption of faster economic growth give us faster pay growth?

It might. However, over the short- and medium-run, there has tended to be little correlation between the pace of economic growth and pace of median compensation growth, as figure 3 above suggests.

Perhaps more important, when it comes to rich democratic capitalist nations, we have very little clue about what yields faster economic growth over the medium to long run.¹² So even if we could be confident that faster economic growth would solve the problem, we wouldn’t know how to achieve it.

Increased education

In their book The Race Between Education and Technology, Claudia Goldin and Lawrence Katz showed that since the late 1800s rising educational attainment has been a key contributor to rising pay in the United States.¹³ And education is associated with income growth in other rich democracies.¹⁴ Could further educational advances help improve pay growth?

Education has many benefits, so there are many reasons for policy makers to aim for a better-educated populace.¹⁵ But there is reason for skepticism about how much additional educational attainment will boost pay in the rich democracies.

Secondary-school completion has likely reached its ceiling. In the United States, for instance, more than 90% of each cohort complete high school.¹⁶

What about college? College completion in the US has increased at a fairly steady rate since around 1940, as we see in figure 5.¹⁷ There was no significant slowdown after the late 1970s that could account for the sharp slowdown in the rate of pay growth. Moreover, while the pay premium for college graduates increased in the 1980s and 1990s, it was stagnant in the 2000s and 2010s.¹⁸

Resurgence of manufacturing jobs

Jobs in the manufacturing sector were vital to broad pay growth in the rich democratic countries in the twentieth century. They required limited formal schooling, as most skills could be learned on the job. And productivity in manufacturing rose significantly, facilitating rising wages. Policy makers frequently propose to expand manufacturing employment as a way to rebuild the middle class. Will this work?

That’s unlikely. As we see in figure 6, manufacturing employment has been declining steadily in all of the affluent democracies since 1970. This owes to technological advance and to firms’ increased ability to shift production to low-cost countries. There is little reason to expect this trend to reverse.

Reduced trade and immigration

Trade increases domestic firms’ ability to outsource production of their parts and other inputs, and it enables them to threaten to close a plant or office and move its operations abroad. This puts downward pressure on pay. Immigration increases the supply of workers, and that too might be bad for domestic wages and salaries.

Yet in practice, there is little evidence that reducing trade or immigration would help to increase pay.¹⁹ Both trade and immigration were rising in the rich democratic countries well before pay growth began to slow down in the 1980s. And pay increases haven’t been slower in nations with greater trade or immigration.

Even if it would help, I wouldn’t recommend that rich nations significantly reduce imports or immigration, because these are key sources of improved well-being for people in poorer nations.²⁰

Trade between poor nations and rich ones can help the incomes of people in poor countries grow faster. If producers in poor countries are able to sell their goods and services in rich countries, the size of the market expands enormously. There are more customers and those customers are, on average, able to pay more than customers in the poor nation. This enables increased production in the poor country, which can lead to more jobs and rising wages. Virtually every successful economic development story of the past half century — including South Korea, Taiwan, Hong Kong, Singapore, China, Brazil, Botswana, and Mauritius — has relied heavily on exports to rich countries.²¹

When economic growth increases in poor nations, some of the added income goes to wealthy owners in those countries or to executives and shareholders of multinational corporations. But some of it goes to ordinary workers in the poor nations. Between 2000 and 2012, China’s share of world manufacturing exports increased from 5% to 17%, and during that decade more than 200 million Chinese moved up into the global middle class.²² More broadly, the period of rising trade since 1970 has coincided with, and almost certainly has been a key contributor to, the most rapid decrease in extreme poverty in human history.²³

An even more effective way to help the world’s least well-off is to allow them to migrate to richer countries.²⁴ The pay of an unskilled worker who moves from Mexico to the United States goes up by 150% on average. For an unskilled worker who migrates from Nigeria to the United States, the pay rise is more than 1,000%.²⁵ This is partly because schools and other skill-development institutions are more widely available and more effective in rich countries and partly because the economic system is more productive.

Reduced monopsony power

Competition among employers for workers is one of economic processes that generates pay increases. In some industries in the United States, the number of employers is relatively small, or a small number of employers account for a large share of the jobs. Also, some American firms have pressured employees to sign a noncompete clause as a condition of employment, which prevents them from leaving the firm to work for a competitor. These two developments have reduced employees’ exit options, and hence their pay leverage.²⁶

How much has this mattered for pay growth in the United States? It’s difficult to tell. One reason to be skeptical that it has had much impact is that this is posited as a phenomenon of the 2000s and 2010s, and the pay trend during this period doesn’t look different from the trend in the 1980s and 1990s, as figure 2 above suggests. Moreover, some analyses suggest that apart from some rural areas and small towns, few local labor markets actually have high levels of concentration.²⁷

Employee board-level representation

Unions negotiate with management about wages. An alternative form of employee voice would give workers a say in electing the people who make far-reaching decisions about the firm’s direction and about who gets to be the management. The most common way this is done is via board-level employee representation (sometimes called codetermination), whereby employees elect a portion of their company’s board of directors.

In Germany, workers have been able to elect 50% of the directors in firms with 2,000 or more employees since the early 1950s and 33% of the directors in firms with 500 to 2,000 employees since the mid-1970s. As of the mid-2010s, similar rules existed in Austria, Denmark, Finland, France, Ireland, the Netherlands, Norway, and Sweden. Board-level employee representation is rare or nonexistent in Australia, Belgium, Canada, Italy, Japan, Korea, New Zealand, Portugal, Spain, Switzerland, the United Kingdom, and the United States.²⁸

Critics of board-level employee representation often suggest it will weaken firms’ performance. However, it appears to have had no such adverse effect in the European countries where large firms operate under codetermination requirements.²⁹

Board-level representation tends to have limited reach. About one-quarter of German workers are employed in firms that have it.³⁰ In 2018, Democratic Party lawmakers in the United States proposed the Accountable Capitalism Act, which would require employee election of 40% of the board of directors in large US corporations — those with annual revenues of $1 billion or more. The roughly 1,300 firms that meet this criterion employ approximately 45 million Americans, or about one-third of all workers.³¹

Can employee board-level employee representation increase pay growth, as some proponents hypothesize?³² Available evidence is fairly thin, but the evidence we have isn’t especially supportive.

A recent study uses within-country variation in Germany to assess the impact of board-level employee representation on wage increases. Beginning in the early 1950s, “stock companies” in Germany — companies not fully owned by a single family — were required to have one-third of their board elected by workers regardless of how many employees they had. In 1994, the German government eliminated this requirement for newly-created stock firms with 500 or fewer employees. The authors of the study compare wage developments in stock firms created just before this change with those in stock firms created just after the change. The firms with employee board-level representation didn’t have faster wage growth.³³ A study of Norwegian firms reaches a similar conclusion — employee board-level representation doesn’t seem to have contributed to higher wages or faster growth in wages.³⁴

We also can compare across countries. To do so, it would help to have a country in which labor unions aren’t especially strong but board-level representation is. Germany fits the bill. While German unions and collective bargaining remain powerful in some manufacturing industries, they have weakened considerably in much of the rest of the economy. But board-level employee representation has remained quite prevalent (despite the 1994 change mentioned in the previous paragraph). So has Germany had healthy wage growth? No, it hasn’t. In fact, Germany’s record has been similar to that of the United States: growth of median compensation has been much slower than growth of the economy, and it has lagged well behind compensation growth in most other affluent democratic countries (figure 3 above).³⁵ Germany’s slow wage growth owes partly to its reunification with the former East Germany in 1990 and its intentional creation of a low-wage (“mini-jobs”) segment of the labor market in the early 2000s. Still, its wage performance gives us little cause for optimism about board-level employee representation’s ability to boost wages.

FOUR PROMISING STRATEGIES

Tight labor markets

When employers can benefit from hiring more workers but find it difficult to do so, they are more likely to increase wages. The key indicator here is the unemployment rate. A low unemployment rate suggests there are relatively few people who don’t have a job but want one. In this situation, employers will be willing to offer a higher wage in order to attract additional workers and keep the ones they have. Wages will tend to rise.

The US experience bears this out. Over time, there is a strong association in the United States between the unemployment rate and the rate of pay increase. That’s true both for the country as a whole and within states.³⁶

Once the unemployment rate dips below a level that begins to push up wages — around 4.5% in the US in recent decades — what matters most is the duration of the tight labor market rather than its degree of tightness. The longer the job market remains tight, the larger the bump in wages and the longer that bump persists for workers.³⁷

To create and sustain a tight labor market, the key policy lever is monetary policy. Central banks in the affluent democratic nations are charged with maintaining both price stability and low unemployment. The longer the labor market can remain tight, the stronger the pressure on employers to increase wages and salaries. To achieve this, the central bank needs to be willing to resist raising interest rates when the unemployment rate gets low enough to potentially cause a jump in inflation.

In the United States, the central bank’s tendency since the inflationary period of the late 1970s and early 1980s has been to increase interest rates quickly and sharply when the unemployment rate falls to 5% or 6%. There have been two notable exceptions: 1997-2000 and 2016-19. These are the only post-1979 periods in which the Federal Reserve allowed the unemployment rate to stay below 5% for multiple years. As we see in figure 7, both periods saw significant pay increases. Preliminary evidence from comparison across states suggests that the low unemployment rates of the post-Covid period have had the same impact.³⁸

There are two limits to the efficacy of tight labor markets as a driver of pay growth. First, periods of low unemployment tend to be brief. Even if the central bank does a good job of allowing a low jobless rate to persist, tight labor markets may obtain in a relatively small number of years, and that may not be enough to secure sizable pay increases over the long run.

Second, the evidence for this strategy’s effectiveness comes mainly from the United States. It might not work everywhere.

Statutory minimum wages

In the middle of the twentieth century, during the era of strong labor unions, few of the affluent democratic nations had a statutory minimum wage. Unions often didn’t want one, fearing that this type of government intervention would weaken their position vis-à-vis employers. But this has begun to change, and 14 of the 21 rich democracies now have a minimum wage.

Figure 8 shows the inflation-adjusted value of the statutory minimum wage in these 14 countries. In most of them it currently sits at between $9 and $12 per hour.

A statutory minimum wage allows policy makers to increase pay via a simple political decision. In some of the rich democracies that’s what they have tended to do in recent decades. In France, for example, the minimum wage was increased from $3 per hour in 1960 to $7 in 1980, and then to $11.50 as of 2018. The United Kingdom enacted a minimum wage in 2000 at $6.50 and by 2018 had raised it to $9.60. Ireland too instituted a statutory minimum in 2000, setting it at $7.50, and then increased it steadily to $9.60. Canada’s minimum wage also was $9.60 as of 2018, up from $7.60 in 1980. New Zealand’s jumped from $6 in 1980 to $10 in 2018.

In other nations, policy makers have decided to limit the degree of increase in the minimum wage, often in an attempt to bring wages more into line with competitor nations. In the Netherlands, the minimum wage was increased steadily from $6.80 in 1961 to $12.60 in 1979. But policy makers then allowed it to fall to $10.50 during the 1980s and have kept it at that level since.³⁹ Belgium, too, had a comparatively high minimum wage around 1980 but then held it flat in subsequent decades.

The same is true in the United States, at least at the federal government level. The US federal minimum wage was increased steadily and sharply from its inception in the late 1930s until the late 1960s. It then dropped a bit in the 1970s, due mainly to higher inflation. Since 1980 it has stayed essentially flat. In the late 1990s, however, some states and cities began adopting a statutory minimum wage above the level of the federal minimum, sometimes also indexing their minimum wage to prices or increasing it regularly. As a result, the average minimum wage across the country has risen, and the rise has been particularly sharp since 2014.

In all of the nations that have a minimum wage, its level is set by elected policy makers, though sometimes on the advice of a group of experts (such as the United Kingdom’s Low Pay Commission).⁴⁰ In principle, this means it can be increased as much as the populace would like it to be. In practice, opponents of a high wage floor, such as business organizations, may be able to exert more influence on policy makers than ordinary citizens and thereby keep the minimum wage below what the majority wants.

Another constraint on the ability of a statutory minimum wage to ensure wage growth is the fact that it will tend to apply to a relatively small number of people. In the United States, for example, the federal minimum wage applies directly to around 2% of workers, and the effects of increasing it tend to fade out by around the 20th percentile of the wage ladder.⁴¹ A rise in the minimum wage doesn’t, therefore, guarantee that wages of most ordinary workers will increase.

A way to address this problem is via sector-specific or occupation-specific minimum wages. This enables policy makers to directly affect the wages of a much larger share of the workforce. Australia illustrates how this can work. Each year a Fair Work Commission sets minimum wages for more than 100 different sectors and occupations, from “Aboriginal Controlled Health Services” to “Wool Storage, Sampling, and Testing,” as well as for various pay grades within these categories.⁴² These minimum wages (“wage awards”) are based on characteristics of the work and required skills. They directly determine the pay of about 20% of Australian employees, and indirectly of many more.⁴³

The United States has experience with something similar. In the 1940s, “wage boards” determined pay levels for particular occupations and sectors. They exist today in a few states, including California and New York, though they play a small role in overall wage setting.⁴⁴

The chief worry about a rising minimum wage is that it may reduce employment. However, the best available evidence suggests that modest increases in the statutory minimum in the past haven’t done so. The best test, because it is closest to an experimental design, is a “difference in differences” approach.⁴⁵ The fact that many of the US states have set minimum wages higher than the federal minimum, in varying degrees and at different times, is helpful for analytical purposes. In the early 1990s David Card and Alan Krueger compared employment changes in fast food restaurants on either side of the New Jersey-Pennsylvania border after one state increased its minimum wage while the other didn’t. Arindrajit Dube and colleagues pursued this strategy for every pair of adjacent counties straddling state borders in which one increased its minimum wage between 1990 and 2006. They, like Card and Krueger, found no adverse employment effect of minimum wage increases.⁴⁶

Profit sharing

Another way to boost earnings is profit sharing, whereby employees receive part of their compensation in the form of a portion of the firm’s profit rather than as a guaranteed wage or salary. For owners, the advantage is that when the firm is struggling, for example during a recession, its labor costs will fall, because part of the reduction in profits will be absorbed by workers via reduced take-home pay. For workers, the advantage is that if profits rise, their pay automatically will too. Over time, their pay will be higher than it would have been without profit sharing.⁴⁷

There is, as just noted, a risk for employees: they will bear part of the cost of falling profits during bad economic times. Then again, workers will tend to have greater employment security, as the enhanced flexibility in labor costs makes it less likely that firms will need to fire employees during rough times.⁴⁸

Profit sharing isn’t common in the rich democratic nations. But it probably could be if it were subsidized. Hillary Clinton’s 2016 presidential campaign in the United States proposed offering firms that implement profit sharing a two-year tax credit equal to 15% of the amount they share (higher for small businesses). The credit would apply to shared profits up to 10% of a worker’s salary or wage. For instance, if a new profit share program in a firm added $5,000 to the pay of someone making $50,000 a year, the firm would receive a subsidy of $750.⁴⁹

Earnings subsidy

Cash transfers and tax credits to people in paid work but with low earnings are another mechanism for boosting incomes in a context of weak labor unions. The United States and the United Kingdom began using employment-conditional earnings subsidies in the 1970s, and in recent decades many other rich democratic countries have adopted some version of them. These programs have proven effective at raising the incomes of households who struggle in the labor market while also encouraging employment.⁵⁰

The dashed lines in figure 9 show the structure of the US Earned Income Tax Credit (EITC). The EITC subsidizes earnings by as much as 45%, providing up to $6,300 per year for a household with three or more children, though the average amount recipient households get is $2,300. The benefit level increased sharply between 1987 and 1996, but since then it has been flat.

The solid line in figure 9 shows a possible alternative version of the EITC, modeled on Sweden’s earnings subsidy.⁵¹ The alternative version would be paid to individuals rather than households, thereby enhancing work incentives for second earners in households. It also would include workers further up the distribution. The current EITC starts to taper off once earnings reach a certain level and disappears altogether at household earnings of $55,000. The alternative version would give every person who earns at least $10,000 the same amount, say $4,000, and raise the earnings cutoff so that most employed persons qualify. These changes would enhance the degree to which the EITC boosts household incomes.

To enable the EITC to assist with not only income levels but also income growth, it could be indexed to GDP per capita (the current EITC is indexed to prices). This would allow the EITC to rise over time in sync with the economy.

Some worry that an employment-conditional earnings subsidy will cause wage levels to fall. In the presence of the earnings subsidy, employers may offer a lower wage than they otherwise would, and workers may be willing to accept a lower wage. Also, the earnings subsidy might increase the supply of less-educated people seeking jobs, and without an increase in employer demand for such workers, this rise in supply could push wages down. Studies suggest that the EITC may indeed reduce wages somewhat, but the evidence is thin and the effect is likely fairly small.⁵² The best way to address this danger is with a moderate to high minimum wage.

SUMMARY

Since the late 1970s, ensuring that incomes rise in sync with the economy has become more difficult in the rich democratic nations. Strong labor unions probably would be very helpful, but they’ve been weakening in nearly all of these countries and there is no compelling reason to expect this trend to reverse. The most effective alternatives, in my estimation, are tight labor markets, statutory minimum wages, profit sharing, and earnings subsidies.