I magine that a sailor sees clouds forming on the horizon, and estimates that it won’t be a major storm. She thus decides to stay on course. The next morning, she’s lashed by an hurricane.
This problem is clearly one of imperfect information, not of discounting, i.e. of time preferences.
Now imagine that a family is getting the services of a tax preparation agency. They are told that they are owed a refund of $1000 from the Treasury and offered to take a refund-anticipation-loan of $900 on that day, instead of waiting four weeks for the refund. They decide to take the loan. Is this discounting? Or is this excessive skepticism about the reliability of the government refund?
For the last century, economists have assumed that agents have ‘deep’ time preferences. Traditional theories of discounting posit that individuals care less about the future than they care about the present. More recent theories of present bias and hyperbolic discounting have formalized a large amount of evidence on the inconsistencies in the trade-offs that people make between utility experienced closer or farther to the present. In particular, individuals appear to discount the future at a much higher rate in the short than in the long term.
For instance, say that I ask you today to choose whether you would like to consume fruit or chocolate one week from today. What would you choose? About 74% of subjects asked to make this decision in an experiment chose fruit. Now imagine that I ask you today whether you would like to consume fruit or chocolate today. What would you choose? 70% of subjects in the same experiment chose chocolate (Read and van Leeuwen,1998).
Another well-known example of this type of behavior is given by Read, Loewenstein & Kalyanaraman (1999): in their study, subjects were asked to choose among 24 films. Some of the films were ‘low brow’, e.g. Four weddings and a funeral; while some were ‘high brow’, e.g. Shindler’s list. When asked to pick a film to watch on the same night, 66% of subjects chose low brow, while this figure dropped to 37% when picking a film to watch the following week.
These examples tell us that people want immediate gratification in the short term, but switch to wanting things that are good for them in the long term. There is, therefore, inconsistency in individual decision making, and it affects important realms of one’s life, from savings to health prevention, to possibly environmental conservation. A large share of behavioral economics has devoted itself to modelling this inconsistency as one affecting individual time preferences, and at identifying and testing policy interventions to help people overcome these inconsistency problems, ranging from commitment devices, to defaults, to simplification of decision environments.
As the examples above illustrate, though, preferences may not tell the whole story. David Laibson is one of the most important contributors to the literature on hyperbolic discounting and time inconsistency. In a recent talk, he presented an alternative model, that formalizes a different, but established, perspective on intertemporal trade-offs based on imperfect information.
A central concept in this theory is that of myopia, i.e. of a lack of foresight or, as Pigou (1920) put it ‘a defective telescopic faculty’ that makes us see future (and past) pleasures on a diminished scale. When individuals consider alternative options, they estimate the utility they will derive from each. The further into the future an event, the more imprecisely the agent will estimate the utility she will derive from it. As the future gets closer, this model predicts reversals of preferences similar to the ones that predicted by theories of present bias.
The model is appealing to an empirical economist like me because it rings true. It also explains puzzles and regularities that models of hyperbolic discounting struggle with.
A first puzzle that faces perspectives based on incoherent time preference is that we would expect time inconsistent individuals to value the ability to commit their future selves to choosing what is best for them, instead of immediate gratification. Yet, we observe little demand for commitment, both in real world markets (think of rare exceptions such as personal trainers or web-surfing blockers) and in experimental studies, where subjects are typically not willing to pay for commitment (Augenblick, Niederle and Sprenger, 2015). The myopia model explains this by making preference reversals the result of inference problems, not of self-control problems.
Second, there is a relationship between learning and discount rates that does not follow naturally from the hyperbolic discounting model, but is instead easily explained by the myopia one. A body of empirical evidence shows a negative relationship between discount rates and age, IQ test scores, and current cognitive loads. It is not clear why present bias should decrease in these variables, whereas it is pretty intuitive that greater experience with intertemporal decisions, the ability to understand and predict complex future situations, and being able to devote cognitive capacity to future utility predictions should all reduce the imprecision of future utility simulations, and thus lead agents to weigh the future more.
Finally, the myopia model explains findings of domain-specific impatience- for instance that framing intertemporal choices as investment problems leads to more patient choices (Read, Frederick and Sholten, 2013)- better than the alternative framework. Individuals find it easier to estimate future utilities in domains where they have more experience, and thus weigh future outcomes more in these domains than in others, where their simulations are more uncertain.
What are the implications of the myopia model for pro-environmental behavior? First, this theory explains why individuals appear extremely short-sighted when their decisions have consequences on climate or the environment: the uncertainty linked to the future utility from undertaking pro-environmental behaviors is very high, thus the future gets very little weight in this domain. And this is not because individuals do not care about the future, of the planet and their children, but because they cannot estimate what utility the behavior will bring them.
Imagine that I am considering the purchase of fridge. I care about saving money and preserving the environment in the future, but I do not know how much money my current inefficient fridge is costing me (and the disutility resulting from this expenditure), nor how much the efficient fridge will cost me, nor the value for me, in terms of future utility, of the savings from moving to a more efficient one. Therefore, when it comes to buying, I may choose an inefficient fridge.
A second implication of this model is, therefore, the role that education, information and experience can have in fostering pro-environmental behaviors. If I could understand better the relationship between energy use and cost, this would increase my ability to simulate the future utility from replacing my fridge with an efficient one: this implies that interventions aimed at making information on energy use more transparent should affect behavior. Moreover, if I knew how much my fridge was costing me in terms of energy bill, I would probably be better able to predict the future utility cost of an alternative fridge: this opens exciting opportunities to make the technological developments in the internet-of-things sector a driver of energy efficient behaviors. Finally, educating individuals about these trade-offs when purchasing items that are frequently replaced, such as light bulbs, may generate spillovers to more economically and environmentally relevant, but less frequent, decisions, such as the purchase of big appliances. These are all exciting areas of research that we are keen to explore within COBHAM.