Search

We measure the marginal effects of two crucial dimensions of the Investment Game design of Gneezy and Potters (1997, GP97) and its replications: time frequency and time horizon. To this aim, we randomize between subjects five different time frequencies: 1 round (“High Frequency” in GP97), 3 rounds (“Low Frequency”), but also 6, 9 and 12 rounds. As for time horizon, we compare the baseline GP97 horizon (9 rounds) with smaller (3 and 6 rounds), but also higher (12 rounds) alternatives. We find that, holding the time horizon constant, subjects invest more when they evaluate their investments less frequently, but this result is significant only when time horizon is sufficiently long. We also find that lower frequencies increase the marginal investment at a decreasing rate. As for time horizon, we find that, holding time frequency constant, the higher the horizon, the lower the investment (although, as in the previous case, investment lowers at a decreasing rate). Our reduced form results also show that subjects’ behavior is sensitive to the endowment stock provided along the experiment, independently of time frequency. Financial support from the Spanish Ministry of Science and Innovation (Grant PID2019-107081GB-I00) and Generalitat Valenciana (Research Projects Groups 3/086) is gratefully acknowledged.

We examine whether different phenomena of narrow bracketing can be traced back to some common characteristic and whether and how different phenomena are related. We find that making dominated lottery choices or ignoring the endowment when making risky choices are related phenomena and are both associated with low levels of cognitive reflection. In contrast, the phenomena of setting narrow goals or narrow mental budgets seem not to reflect choice errors due to low cognitive reflection, but are tools to overcome self-control problems. Buying small scale insurance is associated with having narrow mental budgets - suggesting that people buy such insurance to insure themselves against the consequences of their own self-control strategy.

We study the role of cognitive ability and financial literacy for inter-temporal decision-making using an adapted version of the Convex Time Budget Protocol. We document substantial heterogeneity in choice-patterns and estimated parameters at the individual-level: We find that subjects with higher cognitive ability and domain specific-knowledge are more likely to make patient inter-temporal choices, to allocate the entire budget to a single payment-date, and to allocate the entire budget to corner choices as interest rates increase. At the same time, domain specific knowledge is uncorrelated with choice consistency and estimated individual error parameters, suggesting these results are not driven by a reduction in random noise among high ability respondents. These results serve as suggestive evidence for inter-temporal arbitrage among high ability respondents, thereby revealing a potential confound in time-preference elicitation tasks relying on time-dated monetary rewards.

We study the role of financial literacy for inter-temporal decision-making using an adapted version of the Convex Time Budget Protocol (Andreoni and Sprenger 2012). While we find no evidence of dynamically inconsistent preferences in the aggregate, we document substantial heterogeneity in choice-patterns and estimated parameters at the individual-level: We find that subjects with higher levels of financial literacy are more likely to make patient inter-temporal choices, to allocate the entire budget to a single payment-date, allocate the entire budget to corner choices as interest rates increase, and to show individual discount factors which are in line with extra-experimental market rates. At the same time, financial literacy is uncorrelated with choice consistency and estimated individual error parameters. These results serve as suggestive evidence for inter-temporal arbitrage among financially literate respondents, thereby revealing a potential confound in time-preference elicitation tasks relying on time-dated monetary rewards.

We study theoretically and experimentally how the bracketing of non-binding goals in a repeated task affects the level of goals that people set for themselves, the actual effort provided, and the pattern of effort over time. In our model, a sophisticated or partially naive individual sets either daily or weekly goals to overcome a motivational problem caused by present-biased preferences. In an online, real-effort experiment, we randomly allocated subjects to treatments where they set either daily goals for how much to work over a one-week period or a single weekly goal. Consistent with the theoretical predictions, in the treatment with daily goals, the aggregate goal level for the week was higher and subjects provided more effort compared to the treatment with a weekly goal. The higher effort was driven by the higher aggregate goal level. Additional treatments complemented internal commitment through goals with an externally enforced minimum work requirement to get started working each day.

Gneezy and Potters (1997) designed an investment game experiment and found that, consistent with Myopic Loss Aversion (MLA), individuals are more willing to take risks when they evaluate the results of their investments less frequently. We formally prove that these findings can be accommodated by a standard CRRA functional, once we assume narrow bracketing, a specific form of myopia by which subjects take each decision in isolation and neglect asset integration. We build a compound dataset containing the evidence of Gneezy and Potters (1997) and other replications and structurally estimate both “CRRA plus narrow bracketing” and MLA, together with a mixture model in which we estimate the probability with which each observation is generated by either model. Our results suggest that the CRRA model can also accommodate this experimental evidence, providing empirical support to our theoretical findings. Financial support from Ministerio de Economía y Competitividad (ECO2015-65820-P) and Instituto Valenciano de Investigaciones Económicas (IVIE) is gratefully acknowledged.

Fungibility of money is a central assumption in the theory of consumer choice: any unit of money is substitutable for another. This implies that the composition of income or wealth is irrelevant for consumption. We find that even in a simple, incentivized setup many subjects do not treat money as fungible. When a label is attached to a part of their budget, subjects change consumption according to the label. We first provide suggestive evidence from the field and then use a laboratory experiment to show this result in a tightly controlled environment. In the lab, subjects with lower cognitive abilities are more likely to violate fungibility. The findings lend support to behavioral models of narrow bracketing and mental accounting. One implication of our results is that in-kind benefits distort consumption more strongly than usually assumed. (JEL: C91, C93, D01, H31, I38)

We study in an online, real-effort experiment how the bracketing of non-binding goals affects performance in a work-leisure self-control problem. We externally induce the goal bracket - daily goals or a weekly goal - and within that bracket let subjects set goals for how much they want to work over a one-week period. Our theoretical model predicts (i) that weekly goals create incentives to compensate for a lower than desired performance today with the promise to work harder tomorrow, whereas daily goals exclude such excuses; (ii) that subjects with daily goals set higher goals in aggregate and work harder than those with weekly goals. Our data support these predictions. Surprisingly, however, when goals are combined with an externally enforced commitment that requires subjects to spend less than a minute each day on the task to get started working, performance deteriorates because of high dropout rates from the task.

We characterize optimal individual tax evasion and avoidance when taxpayers “narrow bracket” the joint avoidance/evasion decision by exhausting all gainful methods for legal avoidance before choosing whether or not also to evade illegally. We find that (1) evasion is an increasing function of the audit probability when the latter is low enough, yet tax avoidance is always decreasing in the probability of audit; (2) an analogous finding to the so-called Yitzhaki puzzle for evasion also holds for tax avoidance—an increase in the tax rate decreases the level of avoided income and the level of avoided tax; and (3) that, holding constant the expected return to evasion, it is not always the case that the combined loss of reported income due to avoidance and evasion can be stemmed by increasing the fine rate and decreasing the audit probability.

© 2015 Economic Science Association We examine experimentally how complexity affects decision-making, when individuals choose among different products with varying benefits and costs. We find that complexity in costs leads to choosing a high-benefit product, with high costs and overall lower payoffs. In contrast, when complexity is in the benefits of the product, we cannot reject the hypothesis of random mistakes. We also examine the role of heterogeneous complexity. We find that individuals still (mistakenly) choose the high-benefit but costly product, even if cheaper and simple products are available. Our results suggest that salience is a main driver of choices under different forms of complexity.

We examine experimentally the effect of complexity on individual decision making. We focus on credit choices, as they have been widely criticized for their complexity in recent years. In a first study, we find that complexity in benefits leads to random mistakes, while complexity in costs leads to a specific mistake: choosing a high-benefit loan, with very costly repayment schemes. In a second study, we show that individuals still (mistakenly) choose the high-benefit loan, even if cheaper and simple loans are available. This suggests that, when costs are complex, individuals bracket narrowly, focus on benefits and ignore costs, while they do not when benefits are complex. Hence, our results show that complexity and narrow bracketing may be deeply intertwined: complexity that makes narrow bracketing cognitively easier is likely to lead to myopic choices, such as choosing complex and expensive loans, despite the presence of simple and cheaper loans.

The calibration theorem by Rabin (2000) implies that seemingly plausible smallstake choices under risk imply implausible large-stake risk aversion. This theorem is derived based on the expected utility of wealth model. However, Cox and Sadiraj (2006) show that such implications do not follow from the expected utility of income model. One may then wonder about the implications for more applied consumption analysis. The present paper therefore expresses utility as a function of consumption in a standard life cycle model, and illustrates the implications of this model with experimental small- and intermediate-stake risk data from Holt and Laury (2002). The results suggest implausible risk aversion parameters as well as unreasonable implications for long term risky choices. Thus, the conventional intertemporal consumption model under risk appears to be inconsistent with the data.