David Kreps on Dynamic Choice

While researching and preparing for my applications for graduate studies in Economics, I came across a highly inspiring lecture by David Kreps who is the Adams Distinguished Professor of Management as well as Professor of Economics at Stanford. Under the heading “Choice, Dynamic Choice, and Behavioral Economics” Kreps contrasts how the discipline of economics models choice and, in particular, how the discipline models dynamic choice with more realistic models which draw on behavioural economics. Thereby Kreps’ research program is at the forefront of this with Kreps’ mission being:

To provide economists with “better” (more realistic, more consistent with observed behaviour) models of dynamic choice.

As an introduction to the lecture, Kreps first discusses how the discipline of economics works. Kreps argues that the bedrock of standard economics is that: (1) people make consistent choices and (2) these choices are equilibrated in markets and other institutional settings. Note here that some economists say that people make ‘rational choices’ rather than ‘consistent choices’. However, Kreps reverts to the wording ‘consistent choices’ to circumvent calling choices ‘irrational’ which do not conform with the standard economic model. Having laid out the basics, Kreps exemplifies what consistent choice in economics means in reality. It prohibits any situation like the following:

David Kreps.png
(Kreps, 2015, 2:37-3:12)

The situation in the roadside diner is the typical example of inconsistent choice. Why? Because the customer prefers the apple pie initially when having the choice between apple and peach pie. We could say that the individual maximises his utility by purchasing the apple pie. However, when confronted with a seemingly irrelevant third alternative, that is banana crème, his preference changes. He now chooses peach pie over both apple pie and banana crème. In the world of standard economics this cannot happen because economists assume that the customer will stick to his preferences no matter what. If he chooses apple pie and not peach pie if both options are available to him, then he will never choose peach pie over apple pie when both options are available to him. In short: He will always choose apple because he prefers apple pie. Besides consistent choice theory being the bedrock for economics, this assumption is also necessary for maximising utility. We need this independence of irrelevant alternatives/ preference stability in the choice of an economic agent to conclude that individuals are utility maximisers. Otherwise we cannot model the decision-making process as one of maximising utility with a simple numerical function as commonly done in utility theory, expected utility theory and subjective expected utility theory.

After having set the scene what consistent choice is in economics, Kreps looks at empirical evidence. He argues that empirically the assumption that people make consistent choices does not hold. There is watertight evidence that the environment and other seemingly irrelevant alternatives available to you change the ‘frame’ of the choice. One of the examples is marketing which relies heavily on manipulating the frame of a decision, inducing individuals to buy their products. Hence, compared to economists, marketers have long recognised that people fall prey to inconsistent choices, for example depending on which products they advertise simultaneously. In contrast, economists still go with ‘consistent choice’ because otherwise their decision theory breaks down. Conventional economists tend to ignore the ‘frame’ of the decision.

Next Kreps looks at dynamic choice. The situation of the apple and peach pie is an example of static choice. Kreps, however, is more interested in dynamic choice; that is, choices that have implications for the future and influence future choices like education or savings decisions. Again, standard economics looks at dynamic choice problems through the lens of utility maximisation. It assumes that economic agents act ‘as if’ they could figure out the set of strategies and the outcomes of each strategy in a dynamic choice problem and hence could plainly act ‘as if’ they maximised their utility subject to their ‘budget constraint’. According to Kreps, utility maximisation and the ‘as if’ assumption are not a good model of dynamic choice due to the following three reasons:

  1. Tastes change.
  2. People know that the future is highly uncertain and often unpredictable and they take this into account in the decisions they make today.
  3. People rely on heuristics when it comes to complex dynamic choice problems.

In the remainder of the lecture Kreps looks at all three reasons in more detail. In terms of changing tastes, he argues that people have three basic options. On the one hand, individuals can acknowledge that tastes changes but do nothing. On the other hand, if they wanted to take some action, they could either constrain themselves via a commitment device or deliberately choose to remain flexible and choose not constrain themselves.

In relation to the second reason, Kreps takes a closer look at what motivates people. Conventional economics models individuals as homo economicus; that is consistent (or rational) and selfish. Under these assumptions incentives are guaranteed to work. Economics has a complete theory of incentives. An example is the simple principal agent model: In order to incentivise the agent, the principal can pay the agent a monetary bonus. This is sufficient for compensating the agent for the risk of putting in high effort but just being unlucky and ending up with an unsuccessful project. On the other hand, social psychologists take a rather different stand on motivation. Kreps points towards the concepts of self-determination and self-perception in social psychology. He sees both as superior to the standard economics stories and argues that social psychology needs to be incorporated into economics if we want to build theories approximating real-world behaviour.

In terms of the third reason for why the ‘as if’ model is unlikely to be a good one for dynamic choice problems, one needs to acknowledge that most of the dynamic decision problems like how much to save for retirement, which school to go to or which career to pursue are far too complex. What is more, in the real world individuals tend to choose by trial-and-error. The only way to find out which option is the best one is often merely to try or alternatively simplify the problem and then use some rule of thumbs, i.e. certain heuristics.

The bottom line of David Kreps’ lecture is that dynamic choice problems matter. They are among the most important ones in terms of their impact on individual’s economic well-being, the economic well-being of society and in aggregate the macroeconomic outlook. However, standard economics, by modeling individuals as utility maximisers, abstracts from how actual individuals behave in our economy. In contrast, what economics really should do is modeling actual human choice using proper assumptions which necessarily move away from “rational decision making” because our world is complex, risky and uncertain. Acknowledging that people do the best they can, given their cognitive limitations, paves the way to a superior theory of choice and, in particular, theory of dynamic choice.

If my short discussion of David Kreps caught your interest; then you should check out the full lecture available from here! Kreps is an amazing speaker and his way of presenting makes the topic – which is fascinating in its own right – even more captivating!

Thanks for reading,




Kreps, D. (2015). Choice, Dynamic Choice, and Behavioral Economists [Video]. Available at: http://www.gsb.stanford.edu/insights/david-kreps-choice-dynamic-choice-behavioral-economics [Accessed 22 October 2016].


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s