A few years ago, dealers begged auto makers to stop shipping so many vehicles because they were jamming up their lots and straining their floorplanning budgets.

Now, dealers are asking auto makers to send more inventory on the working premise that it's hard to sell and deliver vehicles if they're not in stock.

Some people are scratching their heads. They wonder if dealers are schizoid, because first they complain about too much inventory and now they beef about not enough. Why the polar-opposite positions?

The answer is simple. Dealers want plenty of cars that sell. Conversely, they don't want a glut of vehicles that will linger on the lots unless prices are slashed and incentives slapped on the hoods.

Auto makers that once notoriously overproduced vehicles and shoved them down dealers' throats now have gone to the other extreme. They haven't just cut back on making slow-sellers. They've reduced production schedules on virtually everything, including vehicles people want to buy.

That makes little sense to dealers. They want a balanced inventory, a smart mix that is low on slow-sellers and high on fast. Right now, that's not the case.

Once the paragons of overproduction, American auto makers in these post-recession and post-bankruptcy times now proceed cautiously. Build rates and inventory levels across the board are the lowest in years.

Auto makers say they will build more cars when the economy improves. Likewise, some tight-fisted banks say they will make more auto loans when things get better. But the recovery will come sooner if those two groups build enough in-demand vehicles and make enough suitable loans. Their wait-and-see positions create a vicious cycle.

Risk management is fine. We've seen what happens when risk managers get lax. That's when the soft-boiled eggs hit the fan. But you can risk-manage your way into paralysis. You definitely want risk managers at the party, but not hosting it. They lock up the booze and turn off the music too early.

The great economic storm that some auto makers barely survived came with a sterling-silver lining.

It forced them and many of their dealers to right-size their operations. Best of all, it killed the push-production system, an industry bane. It went this way: Auto makers produced too many vehicles, forced them on the market that wasn't asking for them and then had to lure people to buy them with incentive ploys.

Now we have a much more preferable pull system, whereby market demands dictate how many particular vehicles are built.

But for the pull system to work, there must be a vehicle at the other end of the rope that consumers are pulling.

For example, the Chevrolet Equinox is a hot-seller and a great cross/utility vehicle with an engine that was a Ward's 10 Best this year. But General Motors is struggling to meet the demand for that vehicle. The last thing GM needs is impatient customers unduly waiting around for a popular vehicle to get shipped to the dealership.

The solution is flexible manufacturing, the nimble ability to quickly increase and decrease vehicle production, depending on demand.

That requires a versatile production process that can build a variety of vehicles off a few common platforms and build different platforms at the same plant.

The old antithesis of flex is one plant, one model, and pray that vehicle sells or you are in big trouble.

The Japanese, as pioneers of flex manufacturing, enjoy a head start. Domestic auto makers are playing catch-up. The industry as a whole isn't there yet but is headed in the right direction.

Pull plus flex equals healthy profits, even if the industry isn't selling as many cars as those heady 17-million unit years of a decade ago.