| Restricted transfers, or "Money follows people", are a policy instrument that combines public provision of private goods with competition between suppliers. Fee-for-service health care and school vouchers are examples. I find that restricted transfers have two unintended consequences, categorical inequality and competition attenuation, which threatens their promise to deliver the advantages of both government (categorical equality) and the market (competition). Most economists believe that competition (Smith, 1776; Hayek, 1968) drives innovation and thus productivity growth. Friedman (1955) famously argues that restricted transfers allow for higher quality at lower cost than government production.;The intuition behind the two unintended consequences is as follows: the quality a consumer receives depends not only on her direct spending (price paid), but also on her acquisition activities. Examples of acquisition activities are travel, search, information gathering, and bargaining. Holding price constant, the more a consumer is willing to engage in acquisition activities, the higher quality she will receive. But, as Southworth (1945) points out, restricted transfers distort consumption. The consumer can, and will, partially undo that distortion by engaging in less acquisition activities compared to what she would do, if the direct spending had been the result of her own volition. Categorical inequality (ch. 1) occurs since for a normal good the poorer a consumer is, the more she is distorted, and thus the less she is willing to engage in acquisition, which undoes categorical equality. Competition attenuation (ch. 2) occurs as the consumption distortion implies that consumers are less willing to engage in acquisition activities than under ordinary circumstances. This means that suppliers operate in a market where consumers are less willing to switch suppliers for that requires acquisition activities.;The result of chapter 1, the positive quality-income correlation, relies on quality being a normal good. Quality, unlike quantity, is merely ordinal. Normality is conventionally defined with respect to the demand function, i.e. linear budget sets. So the question arises what normality means when budget sets are not intrinsically linear. Chapter 3 untangles the definition of normality from linear budget sets, which allows chapter 4 to show that normality of a good is invariant to any order-preserving transformation of its dimension. Thus irrespective of any particular scaling of quality, and shape of the budget set, a consumer's preference is either normal in quality or not. |
