I’ll do this tomorrow, or will I?

The phrase ‘rational decision-maker’ is perhaps one of the most fundamental assumptions in a variety of economic models. But are decision-makers always rational and consistent in their choices? This is indeed a question which has been explored in a multitude of disciplines – from philosophy and psychology to economics and statistics. The answer, perhaps unsurprisingly, is that we are not always rational in our choices and can often make decisions which are inconsistent with our own preferences. One of these inconsistencies in decision-making is known as present bias.

What is ‘present bias’?

Present bias is used to describe the tendency of individuals to attach higher weights to rewards now versus rewards in the future or have a higher tendency to defer costs to a later date. This concept is certainly not new and has been observed and debated since ancient times. For example, Greek philosophers argued against immediate gratification in favour of a greater reward in the future. In other words, they advocated being consistent and opting for greater rewards at a later date as opposed to accepting meagre outcomes now out of ease.

Returning to the domain of economics, we can refer to Ted O’Donoghue and Matthew Rabin who comprehensively discuss and summarise present bias as a model of discounting future benefits and costs [1] in their 2015 paper titled ‘Present Bias: Lessons Learned and To Be Learned’. This paper gives insight into how individuals make decisions under the present bias model and how researchers have tested and calibrated this model to match real world evidence. Furthermore, the authors also highlight the importance – and challenges – of conducting welfare analysis of individuals who have time-inconsistent preferences. Interestingly, the paper does not only collate and restate results from existing literature but also asks questions which would lead to new insights in the study of present bias.

Having presented the concept of time-inconsistent preferences and present bias, we can consider an example to disentangle the abstract concept into a more concrete idea. Say, Person A is offered a choice of $10 today or $15 in one month, it is likely that A will opt to immediately receive $10. Now, we change the choice set to $10 after 6 months or $15 after 7 months. In other words, we have deferred the benefits by 6 months while maintaining the time gap between them. In this situation, Person A is more likely to choose the $15 since waiting one more month for five extra dollars ‘seems easier’ after waiting 6 months. Therefore, A is present-biased when it comes to choosing preferences.

Let us consider another example – one closer to home – concerning discounting of costs. This time Student B has two options at the start of the semester: (1) complete a 3-page essay by week 8 of the semester or (2) complete a 5-page report by week 12. Of course, here Student B analyses the situation carefully and opts for option 1 since it requires less effort. However, if the same choices are presented in week 6 of the semester, Student B is likely to choose option 2 despite the larger workload – assuming the window to complete both is sufficient. This change in choices is called ‘preference reversal’ and highlights the inconsistency in an individual’s preferences.

How does present bias lead to preference reversals?

Preference reversals occur when individuals are rational in the long-term and consider the present value of future costs and benefits with a consistent discounting factor. Conversely, when the time horizon is reduced – the due dates come closer – the individual becomes impulsive, and the costs today appear exceptionally large while the costs in future are discounted heavily. Similarly, benefits today appear more attractive compared to benefits at a future date [2]. These preferences are represented through a ‘hyperbolic discounting model’ (see Fig. 1) which aims to explain why the value of future benefits or incurring of future costs is discounted so heavily compared to outcomes today. 

The above figure highlights how over time the hyperbolic discounting intersects the exponential – time consistent – discounting curve highlighting graphically where preference reversals occur. This systematic treatment, specifically in an economic context, was introduced by David Laibson in his 1994 doctoral dissertation submitted as part of the PhD requirement [3].

Naivety, sophistication and commitment devices

A question that may naturally arise from this discussion is: what if we know and understand that we are present-biased individuals? The answer to this leads us to the definition of naïve versus sophisticated individuals. As the adjectives imply, naïve individuals are unaware of their bias and repeatedly choose and switch their preferences as the time horizon changes. Whereas sophisticated individuals are aware of their present bias and are more likely to adopt commitment devices which would force them to keep their choices consistent over time. 

Having made this distinction, it is not necessarily a negative trait to be naïve as O’Donoghue and Rabin claim that giving higher weight to nearer costs and benefits may be inherent risk aversion from individuals from an uncertain future. However, a large-scale experiment in the Philippines using restricted bank accounts as commitment devices did identify that that some customers opted to have the withdrawal restriction on their savings and were able to grow their savings more than those who opted out [4]. This does provide some evidence towards sophisticated selectors being more successful over the long run as they are able to maintain consistency and commit to better rewards. Further details on the experiment and its results can be found in the paper ‘Tying Odysseus to the mast: Evidence from a commitment savings product in the Philippines’ by Ashraf, Karlan and Yin.

To conclude, we return to a more philosophical tone and end on the words of Roman emperor and Stoic philosopher Marcus Aurelius, ‘You would rather become good tomorrow than be good today.’

References:

[1] O’Donoghue, T., & Rabin, M. (2015). Present bias: Lessons learned and to be learned. American Economic Review105(5), 273-79.

[2] O’Donoghue, T., & Rabin, M. (1999). Doing it now or later. American economic review89(1), 103-124.

[3] Laibson, D. I. (1994). Hyperbolic discounting and consumption (Doctoral dissertation, Massachusetts Institute of Technology).

[4] Ashraf, N., Karlan, D., & Yin, W. (2006). Tying Odysseus to the mast: Evidence from a commitment savings product in the Philippines. The Quarterly Journal of Economics121(2), 635-672.