Analysis: What's wrong here? Unemployment is below 4 percent, but wage growth is still lousy

The unemployment rate has dropped below 4 percent for the first time in 18 years, and the truth of the matter is we have no idea how much further it can go.

That's the silver lining for what is, if not a dark cloud, at least a surprisingly off-color one: Even though unemployment is a mere 3.9 percent, wages don't seem to be rising much faster than they have for the last couple years, and at nowhere near the rate they were before the crash. That's not what's supposed to happen when unemployment gets as low as it can ... which tells us that unemployment must not have bottomed out yet.

If this sounds familiar, well, that's because it is. We've been going through different versions of this for a few years now, as unemployment first hit a 10-, then a 16-, and finally an 18-year low. Each time, this was better than economists had thought was possible not long before; each time, wage pressure was still muted to the point of being almost completely quiescent; and each time, there weren't any signs that this was about to stop, as the economy was still chugging along at the same good-but-not-great pace it had been for most of the recovery.

The details, of course, have been different every month, but this big picture has not. In April, for example, the economy added a slightly worse-than-expected 164,000 jobs, but over the last three months it's still averaging a robust 208,000. Similarly, the labor force just shrank by a disappointing 236,000 - the unemployment rate fell for this bad reason that fewer people were looking for work rather than finding it - but, again, this is more than in line with what we thought would happen if you look at the last three months and see that it's up 412,000 over that time. And wages, as we already said, were up the same 2.6 percent that they have been for most of the last three years.

The simple story, then, is that the economy is adding somewhere between 150,000 and 200,000 jobs a month, just enough people are entering the workforce to offset all the baby boomers' retirements, and wage growth is stuck at a level that suggests the recovery still has a ways to go.

In other words, the natural rate of unemployment must be much, much lower than we thought.

That, you see, is the point at which unemployment has fallen so far that inflation starts rising. The idea is that lower unemployment means companies will have to start paying people higher wages as they compete over the dwindling supply of workers, which will then eat into their profits enough that they have to raise prices. Economists used to think this happened at around 5 or 6 percent unemployment for no other reason than that sounded pretty low, but, well, it didn't. Nor did it when joblessness fell to 4.5 or even 4 percent. The best you can say is there are signs that it might happen soon. Alternative measures of wage growth have shown it growing at a slow and steady pace, and inflation has already in fact risen back to the Federal Reserve's 2 percent target. But if you look at the average hourly earnings for production and nonsupervisory workers - the only numbers we have that go back to the last time unemployment was this low - you see a different story right now: stagnant wage growth no matter how low unemployment is.

Consider this: Between 1998 and 2008, the unemployment rate explained about 58 percent of the raises that these workers were getting. Since then, though, it's only 2 percent. Now, a large part of that is because companies were reluctant to cut wages during the depths of the crisis because of how much that would have hurt morale, so they kept a lid on wages even after things started to improve. But even if you look at the last three years, the relationship is still a lot weaker than it used to be - the unemployment rate accounts for only about 30 percent of wage growth over that time.

What's going on? Well, as economist Adam Ozimek points out, this might not be as mysterious as it seems. Despite our low unemployment rate, there's still a smaller share of 25-to-54-year-olds, who, for the most part, should be too old to still be in school but too young to be retired, working today than there was before the recession hit in 2007, let alone the last time unemployment was this low in 2000. Once you control for that, wage growth is right about where you would expect it to be. Which is to say that we need a lower unemployment rate than we used to for workers to have the bargaining power they did before.

It might be that 4 percent unemployment is the new 5 percent unemployment: Good, but not good enough.