Research

Overview

I am an applied microeconomist, where most of my studies lay out the rational-actor model and then propose a behavioral model that provides an alternative prediction. I then test the predictions with data gathered mainly from experiments (both lab and field), surveys, and, more recently, observational sources. I have experience starting and maintaining a lab, working with standard and non-standard populations and organizational partners, and generating grants. I have worked on projects in many domains, including creativity and innovation, personnel economics, managerial economics, contests, behavioral economics, game theory, and health economics. The commonality between these lines of research is my desire to collaborate with others around me.

One of my original lines of research examines the productivity of remote workers. Even with the COVID-19 remote work experiment, managers are still reluctant to allow their employees to work remotely. One of the stated reasons for this hesitance is almost always tied to productivity concerns. We have shown where these concerns may or may not be justified and the specific aspects of worker behavior that productivity-enhancing policies can focus on. Ongoing work further identifies selection effects and pulls in my other research on innovation in remote work settings.

I am also actively working on developing a deeper understanding of the innovation process. My current studies focus on creativity and ideation, the first stage of innovation. The goal of this line is to open the black box of innovation and test if the traditional economic models of production also apply to innovative work. In this line, we have developed an empirically validated metric of diversity's impact on innovative ideation, shown how creative processes are learned and passed on to others, and causally identified the impact of different experiences on the creative process. Ongoing work further develops these themes and ties them to innovative production with a focus on entrepreneurship.

One of my new lines of research involves an NIH-funded study to better understand the decision-making process of transplant surgeons and nephrologists when accepting or rejecting deceased donor kidneys (DDK). Policy proposals meant to address the shortage of DDK transplants make assumptions about clinicians, but clinician-level data does not exist to support or refute these assumptions. We are putting this data together, linking administrative data to traditional surveys, providing insight into the decision-making of individual clinicians for the first time in the field of transplantation. Our focus is on what information clinicians have, what information they use, and how their individual preferences and biases may affect their choices.

Research Abstracts:

Surgeon and Center-Level Variation in Kidney Transplant Offer Decision-Making with Ellen P. Green, Darren Stewart, and Jesse Schold

Purpose: Kidney disease is a leading cause of death in the US and a major OPTN priority is to ensure that all patients on the waiting list for a deceased donor kidney (DDK) receive a successful transplant before they become too ill for the surgery. In Jan. 2023, approximately 88,901 patients were waiting for a DDK, yet only about 25,000 kidney transplants are performed each year and 25% of kidneys recovered for transplant are discarded, most often after being declined by all offered centers. Centers vary tremendously in kidney offer acceptance practices but individual, surgeon-level variation has not been studied.

Methods: Using a novel dataset constructed by linking surgeon on-call logs from 9 participating centers by offer response date/time to OPTN kidney offer data, we quantified variation in surgeon-level offer acceptance rates, both within and among centers. Rates were adjusted by computing – for each offer – the expected acceptance probability based on an SRTR offer acceptance model, and calculating the ratio of observed to expected (O/E) acceptances for participating surgeons.

Results: Figure 1 (available upon request) provides early evidence that clinicians vary substantially in what they deem an acceptable kidney. Here we plot the distribution of observed to expected (O/E) ratios of clinicians from 9 transplant centers. If acceptance thresholds were consistent within a clinic, we would expect narrow interquartile ranges; however, we see that some centers have wide O/E ranges. This indicates that even within a center–with the same resources and patient panels–clinicians differ in what they deem an acceptable kidney.

Conclusions: These initial results demonstrating significant doctor-level heterogeneity in acceptance rates within and among transplant centers suggest that tailored interventions attuned to individual and center-level differences may be necessary to reduce discards. Augmentation of the novel, surgeon level offer dataset via growing center participation is anticipated.

The data reported here have been supplied by UNOS as the contractor for the Organ Procurement and Transplantation Network (OPTN). The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy of or interpretation by the OPTN or the U.S. Government.

This study used data from the Organ Procurement and Transplantation Network (OPTN). The OPTN data system includes data on all donor, wait-listed candidates, and transplant recipients in the US, submitted by the members of the Organ Procurement and Transplantation Network (OPTN). The Health Resources and Services Administration (HRSA), U.S. Department of Health and Human Services provides oversight to the activities of the OPTN contractor.

Linking Doctors’ Hypothetical with Retrospective Choices on Deceased Donor Kidney Offers. with Ellen P. Green, Darren Stewart, and Jesse Schold

Purpose: In 2021, SRTR estimated 41,765 additions to the deceased-donor kidney waitlist, yet only 19,762 transplants were performed, and the nonuse rate rose to 24.6%, primarily for high KDPI offers. Various proposals aimed at decreasing the nonuse rate could work in theory but may have unintended consequences in practice. Implementing untested policies is thus risky, but testing these policies is equally tricky given the care that must be taken in pre-rollout phases at pre-selected centers. Hypothetical environments offer a promising alternative, yet there is no evidence that a hypothetical scenario can mimic clinicians' actual setting, especially for high KDPI offers.

Methods: The data comes from two sources. The first is SimUNet, a hypothetical tool developed by UNOS Labs that asks clinicians to make hypothetical choices where the information and layout match DonorNet. Eight out of ten offers in SimUNet have a KDPI greater than 0.70 to determine the decision-making process of clinicians facing higher-risk offers. The second is constructed by linking surgeon on-call logs from participating centers by offer response date/time to OPTN kidney offer data over four years. We compare choices made in the two scenarios to determine the degree of similarity.

Results: When comparing the relative ranking of risk taken by clinicians in SimUNet and DonorNet, we find very little correlation. A similar result is found using a random-effects logistic regression when all offers are pooled and a binary variable for SimUNet offers is included and interacted with KDPI. Specifically, more risks are taken, and KDPI is more strongly associated with choices, in SimUNet. Figure 1 (available upon request) shows why. This Figure shows the locally weighted smoothing plot of the relationship between KDPI and SRTR’s probability of acceptance. From this Figure, it is seen that this relationship is strongly negative for lower KDPI, but becomes flat for higher KDPI offers that make up much of our SimUNet data.

Conclusions: These results highlight how difficult choices are for high KDPI offers. For these offers, the risk of failure, as represented by KDPI, is no longer an important metric. A challenge in the future is to determine what may drive choices for high KDPI offers and if it is possible to include them in hypothetical scenarios.

The Formation of Risk Preferences Through Small-Scale Events. with Silvia Angerer, Daniela Glätzle-Rützler, Philipp Lergetporer and Matthias Sutter.

Large, macroeconomic shocks in the past have been shown to influence economic decisions in the present. We study in an experiment with 743 subjects whether small-scale, seemingly negligible, events also affect the formation of risk preferences. In line with a reinforcement learning model, we find that subjects who won a random lottery took significantly more risk in a second lottery almost a year later. The same pattern emerges in another experiment with 136 subjects where the second lottery was played more than three years after the first lottery. So, small-scale, random, events affect the formation of risk preferences significantly.

Remotely Creative? What happens when creative teams work apart? with Cortney Rodet

Creative teams are increasingly being asked to work apart in work-from-home arrangements; a practice exacerbated by the COVID-19 pandemic. Beyond the cost savings to the organization, working apart may decrease distractions and increase efficiency. However, working apart does not allow for synergies stemming from face-to-face interactions. This highlights a fundamental question about creative teams: are they better together or apart? We test teams' creative potential via the creative uses task in two extreme settings; team members working face-to-face vs. team members working apart. We find that teams working apart are able to produce a greater quantity of creative ideas. However, face-to-face teams are better able to take advantage of the team's diversity and a higher proportion of these team's output is unique. These results imply an organization must choose the option more pertinent to their goals with the knowledge that such a choice will entail a trade-off.

Creative capital generation: the role of diverse teams, experience, and communication. with Cortney Rodet

The capacity to create is vital to many organizations' success, yet we know very little about how this capacity is generated or maintained. Our study uses a carefully controlled experiment to show that, relative to a baseline where only learning-by-doing is present, creative capital generation through communication alone is ineffective; however, face-to-face engagement on the task within a diverse team works quite well. Latent semantic analysis indicates that diversity is key because teammates' ideas tend to cross-fertilize during face-to-face engagement. Our study is the first to explore human capital determinants of creativity in economically meaningful settings and suggests newly discovered benefits of teamwork in this setting, and challenges generalizations made outside of the creative framework.

Do Competitive Incentives in Team Production Lead to More Effort or More Sabotage? (with Regine Oexl, Dmitry Ryvkin and Timothy Salmon)

There is an important tension in the incentive systems of many firms. While firms want to incentivize team production by providing rewards for overall output, they also want to incentivize individual effort with relative performance-based bonuses. When production relies on team cooperation though, these competitive incentives may harm production as team members may refrain from helping each other and could even begin sabotaging their colleagues in order to get ahead. We examine these issues via theory and experiments in the presence of individual heterogeneity to try to determine the manner in which individual effort provision as well as help/sabotage behavior may change as competitive incentives increase and as team composition changes. Our results help understand better how to balance out individual versus team rewards and how firms might want to structure teams when employees have heterogeneous abilities.

Which two heads are better than one? Uncovering the positive effects of diversity in creative teams (with Cortney Rodet) Journal of Economics and Management and Strategy

We examine how diversity affects teams' output in the creative uses task. Diversity among team members' experience or knowledge leads to greater creative output; however, diversity over observable characteristics has no measurable impact. Surprisingly, we nd no correlation between experiential diversity and diversity over observable characteristics in our sample. We propose that creative organizations can benet by also emphasizing experiential diversity when constructing diversity policy to foster effective teams.

Gender differences in discrimination emerge early in life: Evidence from primary school children in a bilingual city (with Silvia Angerer, Daniela Glätzle-Rützler, Philipp Lergetporer and Matthias Sutter) Economics Letters

We report the decisions of 824 young children who allocated a fixed endowment between two other children where only one spoke the same language as the decision maker. We show that boys are more likely to discriminate against the out-group.

Is "Real" Effort More Real? (with Krista Jabs Saral and Timothy C. Salmon)

In recent years, a growing number of studies have researchers opting to use real effort designs for laboratory experiments where subjects complete an actual task to exert effort rather than using what is perhaps a more traditional design of stylized effort where subjects simply choose an effort level from a pre-defined set. The commonly argued reason for real effort is that it makes the results more generalizable and field relevant. Some researchers go further and make a distinction between trivial and useful real effort, i.e. whether the task is only relevant for the experiment or if it leads to tangible production for some purpose outside of the experiment, and claim that the useful effort model is even more likely to be generalizable. We present an experiment designed to test whether these three modes of effort, stylized, trivial, and useful, have any impact on behavior in a public goods setting. We find that all three forms of effort lead to identical decision making and then discuss how these results help to inform us about the use of real effort in laboratory experiments.

Social dilemmas with public and private insurance against losses. (with Esther Blanco and Tobias Haller) Journal of Economic Behavior and Organization

This study presents empirical findings on the use of public and private insurance investments in the presence of a potential loss at the group-level. We introduce a novel experimental game where subjects face a group loss and have the choice to reduce the probability of the group loss (public insurance), reduce the magnitude of the individual loss (private insurance), or to increase own payoffs (private earnings, non-insurance fund). We explore subjects’ responses to three treatment conditions; high loss, low loss and asymmetric loss. Results show that, investments in group insurance are lower for those who face lower levels of potential loss, who also invest more in the private earnings fund. Investments in the private insurance are not significantly different for agents facing different potential losses. These results stand both for symmetric and asymmetric groups, as subjects respond primarily to their own marginal incentives. When examining dynamic behavior, we find that subjects use the public and private insurance as substitutes; that is, they contribute more to the private insurance when they expect lower aggregate investments in the public insurance. To better understand the robustness of responses to treatment variations, in some treatments the loss was probabilistic (either occurred or not based on a probabilistic outcome), while in others it was deterministic (losses were presented in expected terms). These results are robust to probabilistic and deterministic presentations of losses.

Don't hate the player, hate the game: Uncovering the Foundations of Cheating in Contests. (with Daniela Glätzle-Rützler and Dmitry Ryvkin)

It is commonly believed that contests - environments in which rewards are based on relative performance - attract and identify the best performers in the population. However, contests also incentivize cheating which may reduce the efficiency of the contest mechanism. We outline a theoretical model of dishonesty in contests with and without selection and test the basic predictions via a controlled experiment. In doing so, we map out the foundations of lying in contests and test the effectiveness of the contest mechanism to attract the best performers when lying is possible. Contrary to common wisdom, we find no difference in the ability (perceived or actual) of those who selected into the contest versus those who selected into a comparable noncompetitive pay scheme. This is due to the finding that high-ability subjects who should have entered, chose to avoid the contest since they believed the amount of lying in the contest would be quite high while many low-ability subjects who should have stayed out, selected into the contest because they believed others would lie less than them. Risk preferences also play a vital role in selection, with less risk-averse subjects being more likely to select into the contest. Lying is negatively correlated with perceived own relative ability and risk-aversion and positively correlated with perceived lying by others. These findings have direct implications for contest designers, and inform on the welfare effects of contests in selecting the ``best.'' In particular, our results have implications for the design of probabilistic monitoring and punishment mechanisms.

Improving Environmental Quality Through Aid: An Experimental Analysis of Aid Structures With Heterogeneous Agents, (with Paul Chambers and Mark Isaac) Ecological Economics

In the pursuit of aid, potential recipient groups may engage in societally damaging behavior. This unintended consequence arises since aid may be based on the severity of the problem. However, given that engaging in damaging behavior is personally costly and aid is given to a group, the process can be modelled as a collective good where the agents in the model ignore the broader social costs. We experimentally investigate the level of unintended consequences of two major aid allocation mechanisms with heterogeneous agents. With the first mechanism, aid is allocated in a more continuous manner based on the perceived need of each recipient with the potential of multiple recipients. In the second mechanism, aid is allocated only to a single recipient with the greatest perceived need resulting in competition between groups for the aid. We find that both allocation mechanisms result in a certain level of negative effects however the degree of damaging behavior depends on which mechanism is used, the opportunity cost of engaging in damaging behavior and how often the aid is given. Our most surprising result is that the average level of damaging behavior is lower in a competitive setting than in the non-competitive setting. Even though average levels of damaging behavior are lower in the competitive setting, this mechanism elicits to the highest variance of damaging behavior. These findings have direct implications on which mechanism the granting agency should use dependent upon their specific goals.

Strive to be first or avoid being last: An experiment on relative performance incentives. (with Loukas Balafoutas, Florian Lindner, Dmitry Ryvkin and Matthias Sutter) Games and Economic Behavior

Managers often use tournaments which motivate workers to compete for the top, compete to avoid the bottom, or both. In this paper we compare the effectiveness and efficiency of the corresponding incentive schemes. To do so, we utilize optimal contracts in a principal-agent setting, using a Lazear-Rosen type model that predicts equal effort and efficiency levels for the three mechanisms with the appropriate distribution of prizes. We test the model's predictions in a laboratory experiment and find that a mechanism which incorporates both competition for the top and away from the bottom produces the highest effort from agents, especially in contests of a relatively larger size. Avoiding being last is shown to produce the lowest variance of effort, be more effective and, in larger contests, more efficient than competing for the top. Finally, we show that behavior in all mechanisms is consistent with basic directional and reinforcement learning.

Remote Work and Team Productivity (with Krista Jabs Saral)

Remote work policies remain controversial mainly because of productivity concerns. The existing literature highlights how the remote setting affects individual productivity yet little is known about how the remote setting affects work in teams - where productivity losses are potentially higher given the additional role of beliefs over partner productivity. Our study closes this gap by examining the effort of individuals randomly assigned to work in either a remote or office setting with partners who are remote and office based. We find that remote workers contribute more effort to the team than office workers, with no differences based on the location of their partners. Office workers incorrectly believe their remote teammates’ contributions will be lower and respond by contributing less effort to the team when paired with remote partners versus office partners. Hence, productivity issues in remote teams are driven by the biased beliefs of office workers rather than true productivity differences, which suggests that managerial policies should focus on correcting these incorrect beliefs rather than limiting remote work.

The optimal allocation of prizes in tournaments of heterogeneous agents (with Loukas Balafoutas, Florian Lindner and Dmitry Ryvkin) Economic Inquiry

Tournaments are widely used in organizations, explicitly or implicitly, to reward the best-performing employees, e.g., through promotion or bonuses, and to punish the worst-performing employees, e.g., through firing or unfavorable job assignments. We use a principal-agent model to compare the efficiency of two tournament incentive schemes, reward tournament and punishment tournament, which, respectively, reward the best performer and punish the worst performer. We show that while the two schemes are equivalent when agents are symmetric in their ability, the equivalence is broken in the presence of heterogeneity. Specifically, punishment tournaments lead to higher profits of the firm. The reason is that low-ability agents are discouraged less in punishment tournaments than in reward tournaments, and hence can be compensated less to meet their participation constraints. Hence, our results predict that firms using punishment tournament contracts will perform better.

Observed Punishment Spillover Effects: A Laboratory Investigation of Behavior in a Social Dilemma, (with David L. Dickinson and Cortney S. Rodet) Experimental Economics

Punishment has been shown to be an effective reinforcement mechanism. Intentional or not, punishment will likely generate spillover effects that extend beyond one’s immediate decision environment, and these spillovers are not as well understood. We seek to understand these secondary spillover effects in a controlled lab setting using a standard social dilemma: the voluntary contributions mechanism. We find that spillovers occur when others observe punishment outside their own social dilemma. However, the direction of the spillover effect depends crucially on personal punishment history and whether one is personally exempt from punishment or not.

Productivity effects of discrimination in the workplace. An experiment on identity, favoritism and work effort.

Discrimination is common in many work settings, yet the general mechanism for how employees internalize, and react to, discriminatory behavior via changes in effort provisions is not well understood. To address this question, a controlled laboratory experiment is used where subjects participate in three-person games - two subjects are always in the role of an employee competing for a prize given out by a third in the role of a manager. Males do not change their behavior regardless of discrimination shown to/against them. Discriminated-against females, however, compete less relative to a setting when discrimination was not possible, while females who were favored compete much more relative to a setting when a manager could not favor them. This leads to the disappearance of an identified gender-performance gap when females are favored but is widened when they are not. Because of these offsetting effects, and the absence of a response by males, a manager’s payoff is not affected by the degree of favoritism shown.

The effects of telecommuting on productivity: An experimental examination. The role of dull and creative tasks Journal of Economic Behavior and Organization

In an effort to cut costs and improve worker morale, corporations are increasingly turning to telecommuting. Conflicting reports exist though on the effects that working outside the office has on productivity which directly affects a company’s bottom line. This study explores these controversies using an experimental approach. Creative and dull individual tasks were used to mimic two extreme work climates. Results of this study indicate that the telecommuting environmental effects may have positive implications on productivity of creative tasks but negative implications on productivity of dull tasks.

A Metastudy of the Dynamics of Responder Behavior in the Ultimatum Game, (with David J. Cooper) Experimental Economics

Using data aggregated from six papers that study repeated play in standard ultimatum games with either stranger or absolute stranger matching, we show that the behavior of responders changes with experience. High offers are more likely to be accepted with experience and low offers are more likely to be rejected. At the individual level, there is a negative relationship between the likelihood that a given offer is accepted and the size of the preceding offer. We compare the results with predictions generated by static models of distributed preferences (i.e. Bolton and Ockenfels, 2000; Fehr and Schmidt, 1999), implicitly dynamic models of preferences with reciprocity (ie. Charness and Rabin, 2002), and explicitly dynamic models of adaptive learning (Roth and Erev, 1995; Gale, Binmore, and Samuelson, 1995). The data is consistent with models of preferences with reciprocity.

Department of Economics, Ohio University, Athens, OH 45701

Email: dutcherg@ohio.edu