The domed marble banking hall of the former Buffalo Savings Bank in Buffalo, New York, with a gilded ceiling and a bronze fountain
Unit 15 · Advanced + Behavioural Economics · 60 min

Present Bias, Self-Control, and Nudges

Why plans slip, and how defaults and commitment change what we do
Image: Andre Carrotflower, via Wikimedia Commons (CC BY-SA 4.0)

People plan to save, to exercise, and to finish on time, and then put it off. This unit sets out the standard model of choice over time, shows the evidence that real choices reverse as the date approaches, and builds the beta-delta model that captures it. It then follows the idea into practice: commitment devices that people buy to bind themselves, defaults that moved millions into pensions, and nudges, with an honest account of how large their effects really are.

Preview: the beginning of the unit

In three American health clubs, the members who paid a flat monthly fee of a little over 70 US dollars came to the gym 4.3 times a month on average, which works out at more than 17 dollars a visit, although the same clubs sold a pass of ten visits for 10 dollars a visit. The members were not confused about the prices. When they signed the contract they expected to come more often, and month after month they did not. This unit is about the gap between the plans people make and the things they do: why it exists, how economists model it, and how banks, employers, and governments have learned to design around it, sometimes to help people and sometimes to profit from them.

1 Choices over time in the standard model

Almost every important economic decision compares costs and benefits that fall at different dates. A pension contribution costs money this month and pays out in forty years, a loan does the opposite, and a diet trades dinner tonight for health in a decade. Unit 9 compared sums of money at different dates with the present value, which divides a payment due in \(t\) years by \((1+i)^t\), where \(i\) is the interest rate. The standard model of choice over time applies the same idea to satisfaction rather than money. In 1937 Paul Samuelson (1915-2009), later awarded the 1970 Nobel prize in economic sciences, proposed in a short note in the Review of Economic Studies that a person values a plan of consumption, that is, the amounts \(c_0, c_1, \ldots, c_T\) to be consumed at the dates \(0, 1, \ldots, T\), by adding up the utility of each date, with the utility of later dates shrunk by a constant factor for every step into the future. This is the model of discounted utility, and the factor is the discount factor \(\delta\), a number between 0 and 1.

A \(\delta\) of 0.95 per year means that a given amount of utility next year counts for 0.95 of the same amount today, and in ten years for \(0.95^{10} = 0.60\) of it. Economists often write \(\delta = 1/(1+\rho)\) and call \(\rho\) the rate of time preference, so that \(\delta\) plays the same role for utility as \(1/(1+i)\) plays for money in Unit 9. Textbooks often teach the case with only two dates, today and tomorrow, under the name “two-period model”.

The model has one consequence that matters more than any other. Because the weight of date \(t\) relative to date \(t+1\) is always the same factor \(\delta\), no matter where the two dates lie, a plan made today for a choice between two future dates is the same plan the person makes when those dates arrive. Economists call this property time consistency. A time-consistent person who decides in January to start saving in March still wants to start saving when March comes, and may change her mind only if she learns something new. Samuelson did not claim that people are built this way; he presented the model as a convenient hypothesis and warned in the same note that its assumptions were arbitrary. The warning was forgotten for forty years.