In the last year, the most important question for U.S. economists and economic journalists has changed from Where are the jobs? to Where are the wages?

It's a problem best summed up by Matthew O'Brien in the Washington Post. As the labor market approaches full employment, there should be more pressure on wages to rise. In the graph below, that would look like a trend-line pointing up and to the left. Instead, as you can see in a half-a-second glance, the trend-line is a blob and it's certainly not pointing up. The unemployment rate has fallen below 6 percent, and earnings growth is flat.

Hourly Earnings vs. Unemployment Rate by Month

Matthew O'Brien/Washington Post

Here are three theories for why.

1. Wage growth and job growth are happening in different places.

When economists and writers say "wages aren't growing," we're making a blanket statement that hides the fact that some wages are growing somewhere. Mining and energy jobs have had a fantastic few years, while retail and food service wage growth has been awful. The problem is that there are far more retail and food service workers than mining and manufacturing employees.

We're adding lots of jobs in industries with stagnant wages, and a few jobs in industries with rising wages, according to new research out of the Cleveland Fed. "It may seem counterintuitive that wages and salaries are growing the slowest in industries where jobs are growing the fastest, but it actually is not," writes LaVaughn M. Henry, vice president of the bank's Cincinnati branch. We're adding few jobs in goods-producing industries like manufacturing, which have the highest overall post-recession wage growth, and lots of jobs in service-producing industries (e.g. health care, leisure and hospitality, and education), which have the lowest real wage growth.

Real Annual Wage Growth, by Industry

Cleveland Fed

2. The rise of invisible unemployment is too large to ignore.

What is "invisible unemployment"? It's discouraged workers and part-timers who want more hours. The official unemployment rate doesn't consider them unemployed. So when we talk about the official unemployment rate—now at a lowish 5.8 percent—we're ignoring these workers. They're statistically invisible.

Here's a picture of invisible unemployment (in blue) vs. official unemployment (in red). Since early 2010, the number of unemployed Americans has declined by twice as fast as the number of discouraged/part-timers (42 percent vs. 21 percent).

Invisible Unemployment vs. Official Unemployment


In 2002, official unemployment swamped invisible unemployment. The official unemployment rate was an accurate description of the labor force. But the spread between invisible and official unemployment is shrinking. In the last 20 years, the six months with the smallest gaps between official and invisible unemployment were all in 2014. That means the official unemployment rate is getting worse and worse at describing the real conditions facing American workers.

Invisible unemployment is hurting the participation rate even more than economists predicted with an aging work force. The entire developed world is getting older. But U.S. participation fell faster in the years after the recession that just about any other country.

2009-2012 Change in Participation Rates


3. The rise of invisible work is too large to ignore.

By "invisible work," I mean work done by American companies that isn't done by Americans workers. Globalization and technology is allowing corporations to expand productivity, which shows up in earnings reports and stock prices and other metrics that analysts typically associate with a healthy economy. But globalization and technology don't always show up in U.S. wage growth because they often represent alternatives to U.S.-based jobs. Corporations have used the recession and the recovery to increase profits by expanding abroad, hiring abroad, and controlling labor costs at home. It's a brilliant strategy to please investors. But it's an awful way to contribute to domestic wage growth.