Do Emotions Help or Hurt Decisionmaking?: A Hedgefoxian Perspective

12. Affect and Cognition as a Source of Motivation: A New Model and Evidence from Natural Experiments

• 12 • Affect and Cognition as a Source of Motivation: A New Model and Evidence from Natural Experiments LORENZ GOETTE AND DAVID HUFFMAN WHEN WORKING toward completion of a long-term project, individuals make effort choices. Many important stages in life involve working on such long-term projects; examples include completing an education, working toward a promotion, working out to lose weight, or working to generate the necessary income to pay for an important future expense. All of these examples share the property that completion of the project requires effort exerted over a sustained period of time, sometimes for many years. Progress toward completion is steady, but each day’s effort is only a small step towards completion of the overall project. Until recently, decision research assumed that the primary source of human motivation in these kinds of situations was cognitive. In the purely cognitive framework, motivation to exert effort is modeled as the outcome of a conscious calculation in which the individual chooses the course of action with the highest net benefit. By contrast, new evidence points to the importance of affect as a source of motivation.1 Experiments show that, at every instance, humans (and other animals) tend to evaluate performance on a task relative to narrowly defined goals or “mileposts” along the way. They experience affect as they make progress, or fail to make progress, toward these more 267 narrowly defined goals. Narrowly defined goals seem to be pervasive for the types of long-term projects we consider; for instance, individuals might set a target on how much progress to make on a dissertation per day, how many calories to lose per workout, or how much money to earn per day in a piece-rate job. The affective reaction triggered by progress relative to these narrow goals has an impact on behavior. In particular, affect apparently explains loss aversion, which refers to a strong preference for not falling short of a reference point or goal and acts as a psychological incentive to exert effort as long as the individual is below the goal. The tendency for affect to become increasingly intense as distance from a goal decreases can explain the so-called goal gradient effect, the tendency for humans, monkeys, and other animals to increase effort as a goal draws nearer. This is in contrast to what the standard economic model would predict: broadly speaking, the standard economic model says that exerting effort at a constant pace over every day is optimal when working on a longterm project. An exception is that effort should increase on days when a random change in the environment makes the marginal productivity of effort high, and less effort should be exerted on days when effort does not translate into significant progress on the project. But the model contends that variations in effort over time across or within days, unless they are a response to such external shocks, are not optimal. Because the standard model in economics is purely cognitive, an alternative model is necessary to incorporate affect as an additional source of motivation. The key feature of this model is that affect is aroused by performance relative to one or more narrowly defined goals which must be passed in order to complete the long-term project (for example, daily page goals as part of working towards a dissertation). This affective motivation can override the priorities assigned by cognitive decision making and distort the individual’s effort profile on the way to completion of the project. Affect is assumed to respond to the immediacy of a goal or reward, increasing in intensity and creating a stronger motivation to exert effort as one of the narrowly defined goal draws near. We formalize this tension between affect and cognition in similar way to George Loewenstein and Ted O’Donoghue (2005), assuming a two-part objective function for the individual; one part corresponds to the preferences of the forward-looking, cognitive self, and the other part to the more myopic process that drives affective impulses. This alternative model generates a psychological incentive to not fall short of a goal, consistent with experimental evidence on loss aversion, and predicts an increasing effort profile leading up to a goal, consistent with experimental evidence on the goal gradient effect. The model also predicts that a temporary shock to productivity (for example, a day on 268 Do Emotions Help or Hurt Decision Making? which it is particularly easy...