Papers - Framed

Greer K Gosnell, John A List, Robert D Metcalfe

Cited by*: 8 Downloads*: 42

Understanding motivations in the workplace remains of utmost import as economies around the world rely on increases in labor productivity to foster sustainable economic growth. This study makes use of a unique opportunity to "look under the hood" of an organization that critically relies on worker effort and performance. By partnering with Virgin Atlantic Airways on a field experiment that includes over 40,000 unique flights covering an eight-month period, we explore how information and incentives affect captains' performance. Making use of more than 110,000 captain-level observations, we find that our set of treatments-which include performance information, personal targets, and prosocial incentives-induces captains to improve efficiency in all three key flight areas: pre-flight, in-flight, and post-flight. We estimate that our treatments saved between 266,000-704,000 kg of fuel for the airline over the eight-month experimental period. These savings led to between 838,000-2.22 million kg of CO2 abated at a marginal abatement cost of negative $250 per ton of CO2 (i.e. a $250 savings per ton abated) over the eight-month experimental period. Methodologically, our approach highlights the potential usefulness of moving beyond an experimental design that focuses on short-run substitution effects, and it also suggests a new way to combat firm-level externalities: target workers rather than the firm as a whole.

Steven D Levitt, John A List, Sally Sadoff

Cited by*: 8 Downloads*: 46

We test the effect of performance-based incentives on educational achievement in a low-performing school district using a randomized field experiment. High school freshmen were provided monthly financial incentives for meeting an achievement standard based on multiple measures of performance including attendance, behavior, grades and standardized test scores. Within the design, we compare the effectiveness of varying the recipient of the reward (students or parents) and the incentive structure (fixed rate or lottery). While the overall effects of the incentives are modest, the program has a large and significant impact among students on the threshold of meeting the achievement standard. These students continue to outperform their control group peers a year after the financial incentives end. However, the program effects fade in longer term follow up, highlighting the importance of longer term tracking of incentive programs.

Douglas Dyer, John H Kagel

Cited by*: 7 Downloads*: 29

Experienced construction industry executives suffer from a winner's curse in laboratory common value auction markets (Dyer et al. [Dyer, D., J. H. Kagel, D. Levin. 1989. A comparison of naive and experienced bidders in common value offer auctions: Laboratory analysis. Econom. J. 99 108-115.]). This paper identifies essential differences between field environments and the economic theory underlying the laboratory markets that account for the executives' success in the field and a winner's curse in the lab. These are (1) industry-specific mechanisms which enable contractors to escape the winner's curse even when they bid too low, (2) learned, industry-specific evaluative processes which enable experienced contractors to avoid the winner's curse in the first place, and (3) important private value elements that underlie bidding. Also identified are a number of industry-specific bidding characteristics whose evolution can be explained using modern auction theory. Lessons are drawn regarding the use of experimental methods in economics.

Richard Hofler, John A List

Cited by*: 7 Downloads*: 5

The lack of robust evidence showing that hypothetical behavior directly maps into real actions remains a major concern for proponents of stated preference nonmarket valuation techniques. This article explores a new statistical approach to link actual and hypothetical statements. Using willingness-topay field data on individual bids from sealed-bid auctions for a $350 baseball card, our results are quite promising. Estimating a stochastic frontier regression model that makes use of data that any contingent valuation survey would obtain, we derive a bid function that is not statistically different from the bid function obtained from subjects in an actual auction. If other data can be calibrated similarly, this method holds significant promise since an appropriate calibration scheme, ex ante or ex post, can be invaluable to the policy maker that desires more accurate estimates of use and nonuse values for nonmarket goods and services.

Peter Bohm

Cited by*: 7 Downloads*: 22

The main purpose of this article is to advance a set of conditions which demand-revealing mechanisms must pass in order to be politically acceptable for real-world applications and - to begin with - for real-world experiments. Without such non-laboratory experiments, real progress seems unlikely to take place in this field. So far, there are few indications that these conditions can be met with respect to the proposals made in the literature on public goods. One possible example of a mechanism that meets the ""acceptability"" conditions is given here. In addition, we present some comments as to why demand-revealing mechanisms constitute and important economic problem, a view which has recently been questioned.

Maria De Paola, Francesca Gioia, Vincenzo Scoppa

Cited by*: 7 Downloads*: 12

We conducted a field experiment involving 720 Italian undergraduate students to investigate the existence of gender differences in performance in competitive settings and whether performance is affected by one's opponent gender. The experimental design was aimed at disentangling gender differences in taste for competition from other differences in psychological attitudes, such as self-confidence and risk aversion. Students were invited to undertake a midterm exam under a tournament scheme having as a prize some bonus points to add to the final grade. Students competed in pairs of equal predicted ability but different gender composition. We find that females are as likely as males to take part in the competition and to obtain a good performance. The gender of one's competitor does not play any role in shaping students' behavior. Men and women perform similarly both in the competitive and in the non-competitive environment.

Esther Duflo, William Gale, Jeffrey Liebman, Peter Orszag, Emmanuel Saez

Cited by*: 7 Downloads*: 7

This paper analyzes the effects of a large randomized field experiment carried out with H&R Block, offering matching incentives for IRA contributions at the time of tax preparation. About 14,000 H&R Block clients, across 60 offices in predominantly low- and middle-income neighborhoods in St. Louis, were randomly offered a 20 percent match on IRA contributions, a 50 percent match, or no match (the control group). The evaluation generates two main findings. First, higher match rates significantly raise IRA participation and contributions. Take-up rates were 3 percent for the control group, 8 percent in the 20 percent match group, and 14 percent in the 50 percent match group. Average IRA contributions (including non-contributors, excluding the match) for the 20 percent and 50 percent match groups were 4 and 7 times higher than in the control group, respectively. Second, several additional findings are inconsistent with the full information, rational-saver model. In particular, we find much more modest effects on take-up and amounts contributed from the existing Saver's Credit, which provides an effective match for retirement saving contributions through the tax code; we suspect that the differences may reflect the complexity of the Saver's Credit as enacted, and the way in which its effective match is presented. Taken together, our results suggest that the combination of a clear and understandable match for saving, easily accessible savings vehicles, the opportunity to use part of an income tax refund to save, and professional assistance could generate a significant increase in contributions to retirement accounts, including among middle- and low-income households.

John A Fox, Jayson L Lusk

Cited by*: 6 Downloads*: 13

An increasing number of studies have begun conducting economic experiments in field, rather than laboratory settings. We directly compare results from laboratory and field valuation experiments. After controlling for unengaged bidders, results indicate field valuations were greater than laboratory valuations. Results are discussed in the context of recent literature on commitment costs.

Steffen Andersen, Alec Brandon, Uri Gneezy, John A List

Cited by*: 6 Downloads*: 51

Perhaps the most powerful form of framing arises through reference dependence, wherein choices are made recognizing the starting point or a goal. In labor economics, for example, a form of reference dependence, income targeting, has been argued to represent a serious challenge to traditional economic models. We design a field experiment linked tightly to three popular economic models of labor supply-two behavioral variants and one simple neoclassical model--to deepen our understanding of the positive implications of our major theories. Consistent with neoclassical theory and reference--dependent preferences with endogenous reference points, workers (vendors in open air markets) supply more hours when presented with an expected transitory increase in hourly wages. In contrast with the prediction of behavioral models, however, when vendors earn an unexpected windfall early in the day, their labor supply does not respond. A key feature of our market in terms of parsing the theories is that vendors do not post prices rather they haggle with customers. In this way, our data also speak to the possibility of reference-dependent preferences over other dimensions. Our investigation again yields results that are in line with neoclassical theory, as bargaining patterns are unaffected by the unexpected windfall.

Thomas S Dee

Cited by*: 5 Downloads*: 7

Wisconsin's influential Learnfare initiative is a conditional cash penalty program that sanctions a family's welfare grant when covered teens fail to meet school attendance targets. In the presence of reference-dependent preferences, Learnfare provides uniquely powerful financial incentives for student performance. However, a 10-county random-assignment evaluation suggested that Learnfare had no sustained effects on school enrollment and attendance. This study evaluates the data from this randomized field experiment. In Milwaukee County, the Learnfare procedures were poorly implemented and the random-assignment process failed to produce balanced baseline traits. However, in the nine remaining counties, Learnfare increased school enrollment by 3.7 percent (effect size = 0.08) and attendance by 4.5 percent (effect size = 0.10). The hypothesis of a common treatment effect sustained throughout the six-semester study period could not be rejected. These effects were larger among subgroups at risk for dropping out of school (e.g., baseline dropouts, those over age for grade). For example, these heterogeneous treatment effects imply that Learnfare closed the enrollment gap between baseline dropouts and school attendees by 41 percent. These results suggest that well-designed financial incentives can be an effective mechanism for improving the school persistence of at-risk students at scale.

John A List

Cited by*: 5 Downloads*: 5

Several experimental studies have recently provided strong evidence that the basic independence assumption, which is used in most theoretical and applied economic models to assess the operation of markets, is rarely appropriate. These results, which clearly contradict closely held economic doctrines, have led some influential commentators to call for an entirely new economic paradigm to displace conventional neoclassical theory. This paper refutes the generality of these experimental findings by going to a well-functioning marketplace and examining more than 350 individual decisions across various incentive compatible elicitation mechanisms. The data suggest that individuals with significant marketlike experience behave largely in accordance with neoclassical predictions: any observed WTA/WTP disparity amongst this group can be explained by neoclassical arguments. In light of these findings, I believe that we have discarded neoclassical explanations of the value disparity too quickly. More narrowly, these empirical results have important implications for stated valuation methods, such as contingent valuation.

John A List

Cited by*: 5 Downloads*: 6

This paper pits neoclassical theory against prospect theory by investigating several clean tests of the competing hypotheses. Consistent with previous work, the field experimental data suggest that prospect theory adequately organizes behavior among inexperienced consumers, whereas consumers with intense market experience behave largely in accordance with neoclassical predictions. The data indicate that the convergence in values occurs entirely because of lower Hicksian equivalent surplus values.

Michael H Birnbaum

Cited by*: 5 Downloads*: 15

No abstract available

Esther Duflo, Rema Hanna

Cited by*: 5 Downloads*: 17

In the rural areas of developing countries, teacher absence is a widespread problem. This paper tests whether a simple incentive program based on teacher presence can reduce teacher absence, and whether it has the potential to lead to more teaching activities and better learning. In 60 informal one-teacher schools in rural India, randomly chosen out of 120 (the treatment schools), a financial incentive program was initiated to reduce absenteeism. Teachers were given a camera with a tamper-proof date and time function, along with instructions to have one of the children photograph the teacher and other students at the beginning and end of the school day. The time and date stamps on the photographs were used to track teacher attendance. A teacher's salary was a direct function of his attendance. The remaining 60 schools served as comparison schools. The introduction of the program resulted in an immediate decline in teacher absence. The absence rate (measured using unannounced visits both in treatment and comparison schools) changed from an average of 42 percent in the comparison schools to 22 percent in the treatment schools. When the schools were open, teachers were as likely to be teaching in both types of schools, and the number of students present was roughly the same. The program positively affected child achievement levels: a year after the start of the program, test scores in program schools were 0.17 standard deviations higher than in the comparison schools and children were 40 percent more likely to be admitted into regular schools.

Travis Lybbert

Cited by*: 4 Downloads*: 12

Potential poverty traps among the rural poor suggest a need to reduce poor farmers' vulnerability by stabilizing crop yields and limiting yield losses. Advances in biotechnology will help address this need directly with crops that tolerate climate fluctuation or resist biotic stresses. Evaluating ex ante how farmers will value these "poor" seeds is important for delivery design, but also challenging. This paper describes an experimental economic approach to understanding farmers' valuation of such seeds. Using data from a survey and experiment, I assess Indian farmers' valuation of changes in the mean, variance, and skewness of payoff distributions. These farmers value increases in expected value, but seem indifferent about higher moment changes in payoff distributions. Farmer traits such as wealth and risk exposure affect their valuation of these changes only mildly. While various limitations to the experimental approach must qualify practical implications of these findings, the experiment demonstrates the viability of conducting valuation experiments with open-ended questions in developing countries.

Amanda Kowalski

Cited by*: 4 Downloads*: 6

I examine treatment effect heterogeneity within an experiment to inform external validity. The local average treatment effect (LATE) gives an average treatment effect for compliers. I bound and estimate average treatment effects for always takers and never takers by extending marginal treatment effect methods. I use these methods to separate selection from treatment effect heterogeneity, generalizing the comparison of OLS to LATE. Applying these methods to the Oregon Health Insurance Experiment, I find that the treatment effect of insurance on emergency room utilization decreases from always takers to compliers to never takers. Previous utilization explains a large share of the treatment effect heterogeneity. Extrapolations show that other expansions could increase or decrease utilization.

Abigail Barr, Jose Garcia-Montalvo, Magnus Lindelow, Pieter Serneels

Cited by*: 4 Downloads*: 11

We explore the value of the strategy method to field experimentalists. Specifically, we demonstrate that, while the method may lead to reductions in subject understanding, it also generates valuable insights. We played the Third Party Punishment Game and the Generalized Trust Game with Ethiopian medical and nursing students applying the strategy method to the responding role in each case. Then, making use of two proxy measures for the students` cognitive abilities, we investigate the relationship between strategy-type choices and subject understanding. Thus, we find support for the assertion that apparently random and internally inconsistent strategies are symptomatic of problems of cognition. We also find support for the often, implicitly made assumption that, in BDM-type trust games, the ratio of what is returned to what is sent is an appropriate focus for comparative analyses of responder behaviour. Finally, we find evidence that an observed difference in third party punishing behaviour between Swiss and Ethiopian students is due, not to misunderstanding, but to variations in what is perceived as punishable. Our results lead us to conclude that the strategy method is of considerable value in Third Party Punishment Games, but need not be routinely applied in BDM-type trust games.

Stefano DellaVigna, John A List, Ulrike Malmendier, Gautam Rao

Cited by*: 4 Downloads*: 61

Why do people vote? We design a field experiment to estimate a model of voting 'because others will ask'. The expectation of being asked motivates turnout if individuals derive pride from telling others that they voted, or feel shame from admitting that they did not vote, provided that lying is costly. In a door-to-door survey about election turnout, we experimentally vary (i) the informational content and use of a flyer pre-announcing the survey, (ii) the duration and payment for the survey, and (iii) the incentives to lie about past voting. The experimental results indicate significant social image concerns. For the 2010 Congressional election, we estimate a value of voting 'to tell others' of about $15, contributing 2 percentage points to turnout. Lastly, we evaluate a get-out-the-vote intervention in which we tell potential voters that we will ask if they voted.

Lars Hultkrantz, Gunnar Lindberg, Jan-Eric Nilsson, Fridtjof Thomas

Cited by*: 4 Downloads*: 58

Around one million people are killed world wide every year in road-traffic accidents. The risks and consequences of accidents increase progressively with speed, which ultimately is determined by the individual driver. The behaviour of the motorist thus affects both her own and other peoples safety. Internalisation of external costs of road transport has hitherto been focused on distance-based taxes or insurance premiums. While these means, as they are designed today, may affect driven distance, they have no influence on driving behaviour. This paper argues that by linking on-board positioning systems to insurance premiums it is possible to reward careful driving and get drivers to self select into different risk categories depending on their compliance to speed limits. We report two economic field experiments that have tested ways to induce car-owners to have technical platforms installed in their vehicle in order to affect the extent of speeding. It is demonstrated that a bonus to remunerate those that have the device installed, tantamount to a lower insurance premium, increases drivers?propensity to accept the technical devices. In a second experiment the size of the bonus is made dependent on the actual frequency of speeding. We find that this is a second way to discipline users to drive at legal speeds.

Michael J. Seiler

Cited by*: 4 Downloads*: 0

Inequity Aversion has long been applied in a game theoretic setting to explain that individuals are willing to sacrifice personal wealth in order to financially penalize players they perceive to be acting selfishly or unfairly. I apply inequity aversion to strategic mortgage default decisions and find that individual homeowners (as well as a second sample of professional mortgage lenders) have a differential stated willingness to walk away from their mortgage based on the perceived characteristics of their lender. Importantly, these significant differences can be removed even with extremely modest loan modifications. Finally, I document that regular homeowners and even professional lenders do a poor job differentiating between the owner of their loan and the servicer of their loan. This is particularly troubling given the extreme misconception of their bank's true character. As a result, much of their willingness to penalize is misplaced resulting in an unnecessary number of strategic mortgage defaults.