H.L. Menken supposedly once said, "there is always a well-known solution to every human problem - neat, plausible, and wrong." Robert Frank's explanation of market failure was described as "simple," but the former three adjectives would suit the purpose better. His explanation rests on the central premise that, by competing, workers find themselves coerced into a situation where they end up seeking sub-optimal outcomes to secure their relative position. Thus, life is a zero sum game (almost exactly like hockey!) where a portion of the participants win only at equal expense to the losers. Were relative measures of wealth the only major factor in an individuals career decisions, Frank would have a point. However, life is much more complex than the Frankian model of the human psyche.
Frank uses the example of workplace safety as an area where market failures occur because of the strong incentive to forgo safety in exchange for increased relative wealth. Because we all do it, no one ends up seeing a gain in their relative status, and we all are resultantly worse off. Frank's argument rests on assumptions that he never properly addresses. For example, he assumes that absolute income doesn't matter. It doesn't matter, he argues, because the most important goods are allocated in proportion to one's relative income. The example he features most prominently is that of school districts. The best school districts are generally in the wealthiest areas. Because of this, only the upper echelon of workers have the opportunity to give their children a decent education. Any worker who cares enough about his or her child would certainly, then, take a riskier job to secure a better salary and ensure a better education for their child. To the extent that this happens, and to the possibly greater extent that it would happen in the absence of worksplace safety regulations, this is somewhat unfortunate. However, such a situation is not the product of the market. Instead, it is a product of the public education monopoly. If we stopped determining school attendence by geography, much of the problem would be alleviated because the quality of a child's education wouldn't be determined by income of their zip code.
Barring such government failures, there are few area's in life where relative income means more than absolute income. Houses, contrary to what Frank implied, are not a fixed quantity. Instead, when absolute income rises we see many middle-income Americans buying second homes. In some other countries, middle income people would be lucky to have a house at all. Even though they share the same relative income as their American counterparts, their poor on every measure that matters. This is because life, it turns out, is not a zero-sum game. The absolute gains that we make by increasing efficiency and even accepting higher risk are real. Trade-offs are are reality of economics and everday life, we shouldn't step in and bar individuals from making them because those same actors chose to compete.