"Talent management" - the attempt to attract, retain and motivate the best employees - is one of the buzz phrases of bosses. But it might be distracting them from a bigger problem - that of bad workers.

Michael Housman and Dylan Minor estimate (pdf) that "toxic workers" cost companies twice as much as superstars benefit them. The rogue trader who costs a bank billions or the groper who attracts a big sexual harassment suit can do more damage to a firm than a decent employee does good. Companies should, they say, do more to screen out such workers.

But here's the thing. "Toxic workers" can look attractive; in fact they might be indistinguishable from the "talent" that firms are so keen to attract. For one thing, say Housman and Minor, such workers - those who would be later fired for fraud, violence or sexual harassment - claim to be honest: they are more likely than average workers to tell surveyors that it's important to always obey rules.

What's more, they are often overconfident about their abilities - and we know from experimental evidence that overconfidence is attractive to hirers because they often confound it with actual ability.

Such confounding, though, contains a germ of truth. Housman and Minor also found that toxic workers are more productive - in the sense of working quickly if not well - than others. This is consistent with two other findings. Francesca Gino and Dan Ariely show that creative people tend to be more dishonest (pdf) than others. And in a study of bank managers Douglas Frank and Tomasz Obloj have found (pdf) that those with higher firm-specific human capital are not only better at generating new business but also better at gaming incentive systems to get themselves big bonuses, with the result that such apparently skilled employees are actually a net cost to the bank.

For these reasons, "toxic workers" can not only stay in an organization for some time, but actually thrive - until they actually cause a big loss.

All this fits in with two other findings. One is from Nick Bloom ands John Van Reenen, who show that there is in most countries a long tail (pdf) of badly managed firms - which is what we'd expect if firms can't distinguish between good employees and bad ones. The other is from Luigi Zingales and colleagues who have estimated that one-in-seven US companies is defrauding its shareholders in some way. If you're wondering which firm will be the next VW, there are lots of candidates.

This sounds like a narrow management thesis. It's not. There are two broader points here.

First, there is a great deal of ruin not only in a nation but in companies. Market forces don't necessarily quickly and efficiently weed out bad hires. In fact, as I've said in a different context, they might sometimes actually select for them.

Secondly, what appears to be a narrow technical function is often in fact an expression of ideology and power. Firms' obsession with "talent management" might appear to be motivated by a desire to increase efficiency and "shareholder value." But if these are what matter, then firms should in fact be more concerned with managing "toxic workers" - many of whom will be in positions of influence. So why aren't they? It might be that "talent management" is in fact not (just) a means of increasing efficiency but a cover for rent-seeking - a justification of big salaries at the top of firms.