Third, governments in the post-war era expanded access to education, which meant that more workers had the skills that were in demand. In the US, for example, the federal government made higher education and vocational training available to millions of citizens with the G.I. Bill, Pell Grants, support for research, and other measures. Yes, all of this investment in innovation and education required higher tax revenues. But moderately higher tax rates and economic growth itself were sufficient to make up the difference.

A similar institutional architecture took hold across much of the industrialized world during the postwar era. For example, in Scandinavia, shared prosperity was achieved not through redistribution, as is commonly assumed, but as a result of government policies and collective bargaining, which created an environment conducive to the creation of high-wage jobs.

This is not to suggest that the 1950s and 1960s were perfect. In the US, discrimination against African-Americans and women remained deeply ingrained, and educational opportunities were not equally distributed. Still, in many other ways, economic conditions then were better than they are now, particularly when it comes to the availability of high-wage jobs.

Having averaged about 2.5% per year between 1947 and 1987, private-sector wage growth in the US decelerated sharply after 1987, and ceased altogether after 2000 – a full seven years before the Great Recession. This stall has coincided with a period of weak productivity growth and a reorientation of investment toward automation, and away from the creation of new, higher-productivity tasks for humans. As a result, high-wage jobs have dried up, wages have stopped growing, and a larger share of prime-age adults have dropped out of the labor force.

More broadly, the institutional architecture that once underpinned job creation crumbled during this period. Protections for workers were steadily weakened, market concentration in many sectors increased, and the government abandoned its previous support for innovation. By 2015, federally funded research and development had fallen to 0.7% of GDP, down from 1.9% in the 1960s.

Many regard the falloff in the creation of high-wage jobs as the inevitable result of advances in artificial intelligence and robotics. It isn’t. Technology can be used either to displace labor or to enhance worker productivity. The choice is ours. But to ensure that such decisions benefit workers, governments need to coax the private sector away from its singular focus on automation.

In the US, this might start with fixing tax policy, which is far too favorable toward capital income. Because firms can reduce their tax burden by deploying machines, they often have an incentive to pursue automation, even in cases where hired workers would actually be more productive. Government also needs to get back into the business of supporting technological innovation, in order to provide a counterweight to the big tech firms, whose business models are overwhelmingly geared toward automation at the expense of job creation. And, of course, expanding educational opportunities across the board is essential.

As in the postwar era, this institutional architecture will require higher tax revenues, especially in the US, where annual tax revenue relative to GDP is around 27%, well below the OECD average (34%). In raising that figure, the point should not be to punish the rich, but to remove distortions such as the favorable treatment of capital. That means broadening the tax base and increasing tax rates modestly (to avoid discouraging investment and innovation).

A society powered by shared prosperity is not out of reach. But getting there will require urgent action to align technology with workers’ needs, prevent monopolization, and fix the tax code so that we can fund the investments we need. Restoring the post-war institutional architecture will be a job for humans only.