1. The document describes an experiment that examines herd behavior among fund managers in a laboratory financial market under different performance evaluation mechanisms and pricing mechanisms.
2. The experiment finds that the frequency of herd behavior is lower when asset prices are flexible rather than fixed, and lower under relative performance evaluation compared to absolute performance evaluation.
3. Analysis of the experimental results supports predictions from theoretical models that incentives under different price mechanisms affect traders' herd behavior.
1. The document describes an experiment that examines herd behavior among fund managers in a laboratory financial market under different performance evaluation mechanisms and pricing mechanisms.
2. The experiment finds that the frequency of herd behavior is lower when asset prices are flexible rather than fixed, and lower under relative performance evaluation compared to absolute performance evaluation.
3. Analysis of the experimental results supports predictions from theoretical models that incentives under different price mechanisms affect traders' herd behavior.
1. The document describes an experiment that examines herd behavior among fund managers in a laboratory financial market under different performance evaluation mechanisms and pricing mechanisms.
2. The experiment finds that the frequency of herd behavior is lower when asset prices are flexible rather than fixed, and lower under relative performance evaluation compared to absolute performance evaluation.
3. Analysis of the experimental results supports predictions from theoretical models that incentives under different price mechanisms affect traders' herd behavior.
performance evaluation programs 2. Performance evaluation and herd behavior in a laboratory financial market
NURMA WAHYUNI (1501103010119)
EFITAMALASARI (16011030100 15) ABSTRACT During the past decades numerous environmental performance evaluation programs have been developed and implemented on different geographic scales. This paper develops a taxonomy of environmental management behavioral patterns in order to provide a practical comparison tool for environmental performance evaluation programs. Ten such programs purposively selected are mapped against the identified four behavioral patterns in the form of diagnosis, negotiation, learning, and socialization and learning. Overall, we found that schemes which serve to diagnose environmental abnormalities are mainly externally imposed and have been developed as a result of technical debates concerning data sources, methodology and ranking criteria. Introduction
Since the 1970s, environmental
protection has become one of the core tasks of the government. Global, regional, and national policy makers have declared their ambition to improve the environment and pursue sustainable development driven by universally agreed principles and goals at the UN Conferences related to the environment, the Millennium Development Goals (especially goal No. 7 concerning environmental sustainability) and based thereon various implementations of the Agenda 21 cascaded down to national and local level Categorization of environmental management behaviors
• Addressing environmental challenges
requires both scientific knowledge and collective actions. Usually, it involves multiple stakeholders in problem identification and solution finding for pollution control, eco- conservation, and green transformation • Efforts in addressing environmental challenges may take different forms. It can be scientific research conducted in labs, by a group of scientists who share common goals. Data and Method For analyzing behavioral patterns by design of environmental performance evaluation programs, we selected ten quantitative evaluation schemes with global, regional, and/or national coverage. We mainly adopt a desktop research method of data collection, complemented with primary data collected from the first author's participation in OECD environmental performance review for China and the assessment of provincial environmental performance in China by the YCELP and its partners. Design of and roles played by the ten selected quantitative evaluation schemes
The allocation of the ten selected
quantitative evaluation schemes to the four roles of performance evaluation is: 1. Diagnosing 2. Learning 3. Influencing and bargaining 4. Social and learning Conclusion ALTHOUGH THIS PAPER HAS BEEN ABLE TO BROADLY CLASSIFY EACH OF THE SCHEMES, BY ITS FUNCTION ACCORDING TO A FRAMEWORK OF FOUR MANAGEMENT MECHANISMS IT HAS NOT BEEN POSSIBLE AT THIS STAGE TO SUPPORT A DEFINITE CONCLUSION ON WHICH SCHEME IS DESIRABLE OR UNDESIRABLE FOR ADDRESSING ENVIRONMENTAL CHALLENGES. THUS, THIS STUDY SUGGESTS A SYSTEM APPROACH TO ENVIRONMENTAL MANAGERS AND POLICYMAKERS IN ADOPTING EVALUATION PROGRAMS FOR ADDRESSING WICKED ENVIRONMENTAL CHALLENGES We consider the effect of performance evaluation on the herd behavior of fund managers in a laboratory financial market. Subjects acting as fund managers receive imperfect private information concerning the fundamental value of a stock, which they then trade in sequence with a market maker. When prices are flexible, subjects regard their private information and herd less frequently than when the price is constant. When price is fixed, subjects evaluated by relative performance tend to "go with the flow" to reduce any price deviation from their peers, but when the price is flexible, herd behavior almost disappears under relative performance evaluation as the increasing trading costs squeeze the net profit of follow-up imitators. Overall, the likelihood of rational decisions under relative performance evaluation is higher than under absolute performance evaluation Extensive empirical evidence suggests that herding is quite common among fund Managers. Indicate that the herd behavior of career-concerned fund managers could be induced by analyst recommendation revisions, especially under a consensus analyst downgrade, and this could destabilize stock prices. Overall, they conclude that an informational cascade is impossible when new information is fully reflected in prices. in a laboratory financial market by comparing herd behavior between markets with a flexible price mechanism (where a market maker updates asset prices according to trade history) and markets with a fixed price mechanism (where asset prices are constant). Their results reveal that the frequency of herding under a flexible price mechanism is lower than is that under a fixed price mechanism, but unlike the predictions of the theoretical models, the flexible price does not completely prevent herding. Experimental design
WE DESIGNED OUR EXPERIMENT TO OBSERVE
HERD BEHAVIOR UNDER TWO DIFFERENT PERFORMANCE EVALUATION MECHANISMS AND TWO ALTERNATIVE PRICING MECHANISMS. THE EXPERIMENTAL ANALYSIS IS BASED ON THE INFORMATION CASCADE MODEL IN BIKHCHANDNI, HIRSHLEIFER, AND WELCH. The trading market in each session consisted of six participants acting as fund managers with the computer acting as the market maker. The procedure was as follows: 1. Each session commenced with a public reading of the instructions, followed by 10 trading periods. 2. At the beginning of each period, the computer assigned a random number between 1 and 6 to each subject without replacement. Then subjects were asked to trade one after another, following the assigned order from 1 to 6. 3. Before deciding whether to sell or buy a stock at a given price proposed by the market maker in each trading period, each trader received a private signal on the value of the stock. Subjects also observed the history of past trades. 4. At the end of the trading period, we randomly chose one of the 10 periods to determine the real payoff for the subjects. 5. At the end of the sessions, we introduced the Holt and Laury (2002) risk preference task to measure individual risk aversion. We use the crossover point, where the subject switches from the safe option to the risky option, to infer the degree of risk aversion. When price is fixed and there is absolute performance evaluation.
the trading price is fixed and equals the asset’s
unconditional expected value, i.e., formula (2). As in the example in Table 2, to maximize the expected value of payoff, it is obvious that the trader should ignore his/her private signal and buy once the excess demand is either higher than or equal to 2, and similarly ignore his/her private signal and sell once the excess demand is either less than or equal to –2. This brings about herd behavior. Analysis of herding
• Analysis of rational behavior
There is an important assumption applied in deducing our models, which is that subjects are rational traders, making full use of any available information and acting in a Bayesian way. During our experiments, a rational subject would join the majority (herd) under a fixed price, and follow his/her own signal (not herd) under a flexible price when potential information cascades existed. However, subjects are not always as rational as we expect. Empirical analysis
We employ two logit models to examine the effect of
incentives on herd behavior under the fixed and flexible price mechanisms. The dependent variable is Herd, which is a dummy variable, and which equals one if herding occurs when there is a potential informationcascade, and otherwise zero. We set a dummy variable Relative-Evaluation, to equal one for relative performance evaluation and otherwise zero. |z| and Risk-Attitude, respectively, represent the absolute value of the excess demand and subjects’ risk attitude, as measured by the numbers of safe options chosen in the lottery choice test. We also include period as a variable to control for any possible learning effect. Conclusion
IN THIS PAPER, WE EXAMINE WHETHER THE
INCENTIVES UNDER TWO DIFFERENT PRICE MECHANISMS AFFECTS TRADERS’ HERD BEHAVIOR. WE DESIGN FOUR TREATMENTS WITH RELATIVE/ABSOLUTE PERFORMANCE EVALUATION UNDER FLEXIBLE/FIXED PRICE MECHANISMS. FOLLOWING THE BASIC MODELS IN AVERY AND ZEMSKY (1998), BIKHCHANDNI, HIRSHLEIFER, AND WELCH (1992), AND CIPRIANI AND GUARINO (2005), WE PROPOSE SEVERAL PREDICTIONS ABOUT HERD BEHAVIOR, GIVEN THESE FOUR ALTERNATIVE TREATMENTS. THE EXPERIMENTAL RESULTS STRONGLY SUPPORT OUR PREDICTIONS ABOUT HERD BEHAVIOR, AS DERIVED FROM THE THEORETIC MODELS.