That data too anecdotal for you? Sorry, but there aren't any good comprehensive statistics on how many companies should be hiring but aren't. It's too subjective. How many of the companies in McKinsey's survey, for example, are using the "no suitable candidate" explanation as an excuse to avoid hiring?

But there are estimates. Harry Holzer, an economist and public policy professor at Georgetown University, estimates that the unemployment rate would be 8% instead of 9.1% if existing job openings could be filled. The International Monetary Fund estimates that 25% of U.S. unemployment is structural.

If those estimates are anywhere near right, then most of U.S. unemployment is still cyclical. And it should be attacked with the available tools for creating demand in the economy. You know what those tools are, since we've tried most of them -- not always in the best or most efficient fashion, I'll grant you.

The Federal Reserve has tried to create demand (even for houses) by keeping interest rates low. Congress and the president have tried to create demand by spending on infrastructure (too little, I'd argue), tax cuts and investment tax credits to business, to mention a few approaches.

It's quite possible that these measures haven't worked because the Great Recession was the deepest economic downturn in the U.S. since the Great Depression and that cyclical downturns, when they're deep enough, are very difficult to reverse. Certainly the duration of the Great Depression suggests that.

But it's also possible that these cyclical measures have been ineffective because they've attacked only part of the unemployment problem. If companies can't find workers with the right skills, then creating demand won't reduce unemployment very quickly. Think of this mismatch of workers' skills and job openings as yet another force slowing down job creation. If a company has to train new hires to bring their skills up to speed, it has to be 100% convinced that the state of the economy makes the expenditure worthwhile. No company is going to spend six months training a worker only to discover that the orders it was counting on haven't come in the door.

How to speed up job creation

There are ways to attack this structural unemployment problem. The federal and state governments could pay for worker training through an expansion of community college systems, grants to workers (unemployed or employed) for training or subsidies to company- or industry-run apprenticeship programs like those in Germany.

Some of the existing programs in these areas -- community college funding, for example -- have been hit hard by state and local budget cuts. And there's certainly a spirit in the land (I've always wanted to use that phrase) that says we can't afford such frills.

Click here to become a fan of MSN Money on Facebook

Except that these aren't frills. And they wouldn't be even if the United States wasn't trying to figure out a way to speed up job creation after the Great Recession.

If you look around the world -- and take a point of view that stretches beyond the Great Recession -- you'll see that the great economic challenge of the next decades is competing on productivity and the skills of an economy's workforce. I don't think you can escape that competition even if you're China -- rising wages and incomes in that society are pushing the cheap-labor jobs to countries such as Indonesia, Bangladesh and Vietnam. Moving up the value chain is the next game every economy will have to to play.

And the United States doesn't seem to be half-trying.

Jim Jubak's column has run on MSN Money since 1997. He is the author of the book "The Jubak Picks," based on his market-beating Jubak's Picks portfolio; the writer of the Jubak's Picks blog; and the senior markets editor at Get a free 60-day trial subscription to JAM, his premium investment letter, by using this code: MSN60 when you register at the Jubak Asset Management website.

Click here to find Jubak's most recent articles, blog posts and stock picks.