Good article. Edit needed: “Imagine you're the sole buyer of apples and all the orchards from Washington to Virginia wait patently for your quoted price.”
Additional comment. Most search models have something like "firms post vacancies and successful match between one worker and each vacancy results in A units of output". In other words, no diminishing returns- hence the perfectly inelastic demand mentioned above. MPL=0 if job is vacant and MPL=A if it is filled. Additionally since there are costs of posting/keeping vacancies, any firm that pays marginal product, A, would obviously make negative profit and hence never open up the vacancy in the first place. So in such a model it makes no sense to talk about whether wages are at or below marginal product.
One of the most frustrating things about this debate is the refusal of some of the pro-minimum wage folks (like Dube, or Noah Smith) to recognize that the idea that "search costs cause monopsony power" has never actually been formalized in any rigorous model.
If you take search models like Shapiro-Stiglitz or the original Mortensen-Pissarides, then in those models a binding minimum wage has the same effect - decreased employment - as in the vanila labor supply/labor demand model. This is due to fewer vacancies being opened up by firms in response to minimum wage, resulting in higher equilibrium unemployment.
Then you have models where firm level labor demand is perfectly inelastic (one worker = constant level of output, additional workers require separate jobs) and the model also assumes a fixed number of vacancies. So overall labor demand is perfectly inelastic. And sure in such a model a minimum wage will not cause unemployment but that's not because of any "search costs" or "monopsony power" but simply because you assumed *perfectly inelastic labor demand*. One labor textbook even makes the claim while explicitly holding the number of vacancies constant.
Here is the other frustrating claim about this debate - the idea that, empirically, no effect of minimum wage on employment supports the monopsony model against the competitive model. This is of course nonsense. Zero effect rejects BOTH models. A monopsony model predicts that a binding minimum wage would INCREASE employment (moving up that labor supply curve). So the effect of 0 is a rejection of that. But AFAIK, no proponent of minimum wages and monopsony in labor markets has been willing to bite the bullet and make the argument that "no, minimum wages actually increase employment". Probably because that would be silly, and kind of expose the faulty logical foundations upon which the whole argument rests.
Good article. Edit needed: “Imagine you're the sole buyer of apples and all the orchards from Washington to Virginia wait patently for your quoted price.”
Additional comment. Most search models have something like "firms post vacancies and successful match between one worker and each vacancy results in A units of output". In other words, no diminishing returns- hence the perfectly inelastic demand mentioned above. MPL=0 if job is vacant and MPL=A if it is filled. Additionally since there are costs of posting/keeping vacancies, any firm that pays marginal product, A, would obviously make negative profit and hence never open up the vacancy in the first place. So in such a model it makes no sense to talk about whether wages are at or below marginal product.
One of the most frustrating things about this debate is the refusal of some of the pro-minimum wage folks (like Dube, or Noah Smith) to recognize that the idea that "search costs cause monopsony power" has never actually been formalized in any rigorous model.
If you take search models like Shapiro-Stiglitz or the original Mortensen-Pissarides, then in those models a binding minimum wage has the same effect - decreased employment - as in the vanila labor supply/labor demand model. This is due to fewer vacancies being opened up by firms in response to minimum wage, resulting in higher equilibrium unemployment.
Then you have models where firm level labor demand is perfectly inelastic (one worker = constant level of output, additional workers require separate jobs) and the model also assumes a fixed number of vacancies. So overall labor demand is perfectly inelastic. And sure in such a model a minimum wage will not cause unemployment but that's not because of any "search costs" or "monopsony power" but simply because you assumed *perfectly inelastic labor demand*. One labor textbook even makes the claim while explicitly holding the number of vacancies constant.
Here is the other frustrating claim about this debate - the idea that, empirically, no effect of minimum wage on employment supports the monopsony model against the competitive model. This is of course nonsense. Zero effect rejects BOTH models. A monopsony model predicts that a binding minimum wage would INCREASE employment (moving up that labor supply curve). So the effect of 0 is a rejection of that. But AFAIK, no proponent of minimum wages and monopsony in labor markets has been willing to bite the bullet and make the argument that "no, minimum wages actually increase employment". Probably because that would be silly, and kind of expose the faulty logical foundations upon which the whole argument rests.