آموزش کاهش هزینه ها: درباره پویایی تنظیم بهینه محصولات تجربه ای

We study the introduction of new technologies when their costs are subject to idiosyncratic uncertainty and can only be fully learned through individual experience. We set up a dynamic model of clean experience goods that replace old polluting consumption options and show how optimal regulation evolves over time. In our base setting where social and private learning incentives coincide, the optimal tax on the polluting consumption is increasing over time. We show, however, that if social and private learning incentives diverge because the private discount rate exceeds the social discount rate, it may be optimal to temporarily increase the tax rate beyond net marginal external damages to induce more learning before reducing the tax rate to the steady state level. Alternatively, one could complement the tax with subsidies for first-time users which can be phased out over time. Similar results apply if consumers have biased expectations. We therefore give a rationale for introductory subsidies on new, clean technologies and non-monotonic tax paths from a perspective of consumer learning.

Many environmental regulation schemes involve taxes or subsidies that change over time. In this paper, we study the dynamics of environmental regulation to control the adoption of a socially beneficial experience good. That is, by trying out a new, less polluting consumption choice, consumers may learn the personal net costs they incur from its use, i.e. their personal abatement costs. Examples are widespread: Car users are often only partially informed about the specific costs and benefits of using public transport (or other means of transportation). Ecologically produced food and clothing may have attributes unknown to the inexperienced user, including taste, durability, social acceptance and the like. Household or office appliances are often advertised in terms of their higher energy efficiency relative to the older versions they seek to replace, but consumers only have limited knowledge about the operating costs and convenience associated with these new products. Learning by experience can also be linked to motivational factors such as the warm-glow effect or social reputation derived from using environmentally friendly products [2] and [4]. These examples share two features: that the new product reduces an externality and that consumers are uncertain about their personal benefits and costs before trying out the new product. We study the optimal regulation of an experience good, explore rationales for initial subsidies for a new product, and demonstrate how optimal regulation levels change over time. We focus on the dynamics of a government intervention that is driven exclusively by consumers learning the net costs that they personally incur from adopting a new technology (hereafter denominated ‘learning’). That is, we abstract from learning spillovers among consumers as well as from supply-side arguments such as cost reductions through an intensified use of the new technology caused by technological spillovers among firms. The study of experience goods is novel in the environmental economics domain. Our paper is, however, related to studies in industrial organization that focus on the supply of experience goods under imperfect competition. Bergemann and Välimäki [5] examine monopolistic pricing of experience goods in a dynamic model. They show that price dynamics crucially depend on a simple dichotomy between mass and niche markets.1 While prices in mass markets are declining over time, they may initially be low but subsequently increase in niche markets. The low initial prices are set with a policy focus aimed at increasing market penetration, whereas the higher prices in the steady state (where no more learning takes place) maximize monopoly profits. The literature on dynamic pricing of experience goods goes back to Shapiro [14] who considers learning in a simple two-period model. The pricing of experience goods has also been discussed by Cremer [7], Farrell [10], and Milgrom and Roberts [13]. We set up a dynamic model of a new experience good that replaces a preexisting consumption choice which generates an environmental externality. We consider an infinite-horizon, discrete-time model with a continuum of consumers, who have (at most) unit demand per period for the new product. Alternatively, they consume the traditional, more polluting product. Consumers differ in their expected net costs of using the new technology and face an ex-ante unknown cost component. The latter is subject to idiosyncratic uncertainty and can only be learned through individual experience. We assume that consumers only learn their true costs, i.e. their valuation relative to the old technology, once they have used the new technology for one time period.2 We consider two different regulatory regimes: We start by analyzing the first-best case in which the regulator, at each point in time, determines both the number of inexperienced consumers who are exposed to the new technology for the first-time, and the set of experienced consumers whose continued use of the technology is optimal. Second, we consider a setting where the regulator needs to rely solely on subsidies or taxes. The subsidy in the given period then determines both the behavior of the experienced consumers as well as the inexperienced consumers' decision to try the new technology. We show how optimal regulation evolves over time. As long as consumers' and social discount rates coincide and expectations about the net costs of the new technology are unbiased, the first-best case can be decentralized by simply relying on a corrective tax levied on the polluting technology, with the optimal tax rate increasing over time. If the private discount rate exceeds the social discount rate or if consumers' cost expectations are biased, the tax alone does not suffice. The first-best solution can then be implemented by complementing the tax with a subsidy for first-time users. This subsidy will be non-increasing over time. If the regulator cannot discriminate between first-time and experienced users, the second-best taxation scheme may involve a non-monotonic path: tax rates are initially increasing with a policy focused on reaching the optimal amount of knowledge (i.e. the number of consumers that have learned their net personal costs), before being reduced to the level that reflects the marginal social costs of using the polluting alternative. Our results are qualitatively similar to those derived by Bergemann and Välimäki [5] in the context of the optimal monopoly pricing of experience goods. During the approach path, the government (or monopolist) sets the tax with a focus on inducing optimal learning. In the steady state, however, no more learning takes place, and the tax (price) is chosen so as to maximize social welfare (monopoly profits). We believe that our results indicate an important reason for a slow introduction of pollution taxes that is motivated by the fact that consumers are uncertain about their personal costs or benefits from using a new and cleaner technology. Only by trying it will they learn about the personal fit. We show that such a setting not only motivates taxes that are increasing over time, but that it may also require introductory subsidies for first-time users that are phased out over time. Our paper is structured as follows. Section 2 presents our basic model, and Section 3 discusses the social optimum. Section 4 considers the case of first-best regulation, while we turn to a second-best regulation that solely relies on usage taxes in Section 5. Section 6 concludes.

Regulation often involves introductory taxes or subsidies that may later be reduced. Usually this is motivated by supply-side considerations such as decreasing production costs or technology spillovers. In this article, we discuss a different rationale for introductory subsidies that is motivated by the demand-side: If consumers are uncertain about their tastes regarding a new technology, they may learn by trying. A prominent and environmentally relevant example involves the introduction of new public transport options. Whereas consumers usually know the total private costs and benefits of their status quo alternative (commuting by car), at least a part of the true opportunity costs of using public transport need to be experienced before they can be known. In this paper, we demonstrate how an optimal regulation should incorporate dynamic features that initiate from this “learning-by-trying” in an intertemporal setting. Any regulation needs to simultaneously account for two dimensions: First, experienced consumers will use the new technology if their private opportunity costs are outweighed by a tax on the old technology; second, the policy in its introductory phase needs to control the optimal number of new consumers. We demonstrate that the optimal transition path to a steady state involves increasing regulation levels as long as social and private learning incentives coincide. In this case, the first-best path can be decentralized by taxing the dirty alternative. Along the optimal path, the tax rates are increasing, which corresponds to a slow introduction of taxes. With zero or low fixed costs to expand the capacity of the new technology, the fraction of consumers using the clean alternative actually falls over time, since initially many consumers try the alternative, but then may (temporarily) go back to the polluting option. If the capacity expansion is costly, both the optimal tax and usage rates increase over time. The qualitative features of the optimal policy significantly change if the private discount rate exceeds the social discount rate. Due to the divergence between private and social learning incentives, a tax on the dirty alternative alone cannot implement the first-best solution, but complementing the tax with a subsidy for first-time users can. This subsidy is (weakly) decreasing over time. If such a special treatment of first-time users is not feasible, the regulator's second-best tax path also needs to take the different learning incentives into account. A tax path that is first increasing, but will decrease at one point in time before being constant, can improve upon a policy that relies on a monotonically increasing tax path. Although we derive our results in the context of differing private and social discount rates, they also apply to situations where consumers have biased expectations about the full costs or benefits of a new technology. Similar to the inefficiencies resulting from the difference between private and social discount rates, biased expectations result in socially inefficient learning that has to be corrected by means of subsidies or taxes for first-time users in order to obtain the first-best solution (where the subsidy is equal to the bias), or by a non-monotonic regulation path if the first-best solution is not feasible. More generally, our results suggest that if private learning incentives lead to a rate of exposure to a new experience good that lies below the social optimum, introductory subsidies can be justified not only with decreasing production costs, but also when consumers learn by experience.