Endogenous Peer Effects Research Articles

Disentangling peer effects in transportation mode choice: The example of active commuting

We investigate the role of peer effects in the workplace on individual active transportation mode choices. We collect original data through an online survey on networks and sustainable behaviors among 334 individuals working in ten research laboratories at the University of Grenoble Alps in February 2020. We apply linear and nonlinear models of peer effects on active modal choice, untangling the role of conformism and strategic complementarity in social influence. We show that given our data, a linear local-average specification is the preferred empirical model of peer effects, and we estimate strong and significant endogenous peer effects.

“Like with like” or “do like?” Modeling peer effects in the classroom

AbstractObjectiveThe authors discuss the role of peer networks in shaping the decision to enroll at university. Using panel data from Italy, they apply innovative statistical methods to study a sample of students as they complete high school and decide whether or not to attend university.MethodsThe authors use simultaneous autoregressive (SAR) models to analyze a four‐wave panel database of Italian students. They explore the role of endogenous, exogenous, and correlated peer effects in relation to the decision about whether or not to enroll at university.ResultsThe findings suggest that endogenous peer effects exert a significant influence on the probability of enrolling after controlling for homophilous preferences and a range of variables. Exogenous peer effects do not appear to influence this outcome. Sensitivity tests suggest that the results of the estimation are robust to selection.ConclusionsThis article contributes to an emerging body of literature on the use of SAR models to study peer effects, illustrating its considerable potential in the study of educational outcomes.

Peer effects on corporate R&D investment policies: A spatial panel model approach

Identifying peer effects has important policy implications because endogenous peer effects can amplify policy effects. Using a spatial panel model with network-structure peer groups and data on listed manufacturing firms in China, we investigate whether a firm’s R&D expenditure affects its peers’ R&D policies. Our results show: (1) peer firms’ R&D expenditures play an essential role in determining corporate R&D investment policies; (2) competition and information incentives are the two main underlying channels; and (3) R&D subsidies can indirectly increase a firm’s R&D via the endogenous effect. Using a novel identification strategy, this study provides additional empirical evidence of peer effects on corporate policies. The findings contribute to our understanding of the impacts of R&D subsidies and indicate that the overall policy effect can be leveraged by considering firms’ networks.

Endogenous peer effects in diverse friendship networks: Evidence from Swedish classrooms

Do students benefit differently from interacting with similar and diverse peers? Using register-linked survey data from a stratified sample of Swedish classrooms I test whether endogenous peer effects in student achievement are heterogeneous by immigrant status. My empirical strategy combines instrumental variables, classroom fixed effects, and a control function to identify the parameter of interest separately from contextual and correlated effects. In particular, by considering partially overlapping networks of peers and family members, I use peers’ parents’ education as instruments for peer behaviour. My findings suggest that endogenous effects are limited to interactions with native peers only, but benefit both native and migrant students. High-ability children of migrants appear to be particularly vulnerable to friendship segregation.

Identification and Estimation of Endogenous Peer Effects Using Partial Network Data from Multiple Reference Groups

There has been a considerable amount of interest in the empirical investigation of social influence in the marketing and economics literature in the last decade or so. Among the many different empirical models applied for such investigations, the most common class of model is the linear-in-means model. These models can be used to examine whether social influence is truly a result of agents affecting each other through their choices simultaneously (endogenous effect) or of having similar taste and characteristics (homophily). However, the two effects are not separately identified in general in the standard linear-in-means model unless data on all members of an individual’s network are available. With data on a sample of individuals from a network, these effects are not identified. In this research, we leverage a very natural aspect of social settings, namely that consumers are usually part of multiple—as opposed to single—networks. We discuss the sufficient conditions for identification when the standard linear-in-means model is extended to allow for multiple sources of social influence. We also show how the additional variation generated by more than one source of social influence actually allows estimation of endogenous effects with sample data. We demonstrate the feasibility of our approach via simulation and on the National Longitudinal Study on Adolescent Health data, which has been used in a number of studies examining social influence. Our approach is, therefore, likely to be useful in typical marketing settings. This paper was accepted by Matthew Shum, marketing.

Banks’ interconnections and peer effects: Evidence from Chile

In this paper, we identify and quantify the importance of endogenous peer effects in the interbank market, allowing for varying degrees of intensity of these peer effects. We base our analysis on a unique dataset that includes all interbank loans that have taken place between 15 banks in the Chilean interbank market representing more than 95% of the market between 2009 and 2016. This approach contrasts sharply with the geographical definition of peers used by most of the literature. As an application of our model, we examine an episode of liquidity shortage experienced by one Chilean bank in the interbank market, with the lenses of our model. We show evidence consistent with a herding behavior of the lender banks which, according to our model, were peers of the stressed bank.

Measuring Social Interaction Effects When Instruments Are Weak

Studies that distinguish between exogenous and endogenous peer effects are relatively rare. To separate these effects, De Giorgi, Pellizzari, and Redaelli exploited partially overlapping peer groups where attributes of a student’s peers in one group provide instrumental variables (IV) for endogenous effects in another group. We apply this identification strategy to data from a period of transition at a Chinese university: dormitory roommate assignments were changed as students moved between campuses. This transition allows us to measure the endogenous effects between test scores of current roommates, but the traditional IV method suggests the potential for weak IV. We use weak-IV robust techniques to obtain properly sized tests. The S-test, K-test, and QCLR test all reject the null of zero endogenous effects with p-values between 0.01 and 0.05, as compared with 0.003 implied by the traditional IV estimator. The largest 95% confidence interval lower bound is 0.154 from the QCLR test, in contrast to 0.244 from traditional IV. Our findings offer unique evidence that endogenous peer effects influence academic outcomes at an empirically relevant magnitude, and an example where weak-IV robust tests are essential to quantify the relationship. Our results are robust to alternative model specifications.

Identification of Peer Effects using Panel Data

This paper provides new identification results for panel data models with contextual and endogenous peer effects, respectively operating through individual heterogeneity and outcomes. The results apply for general network structures governing peer interactions, and hinge on a conditional mean restriction requiring exogenous mobility of individuals between groups over time. Some networks preclude identification, in which case we propose additional conditional variance restrictions. We apply our method to surgeon-hospital-year data to study take-up of keyhole surgery, finding a positive effect of the average individual heterogeneity of peers. This effect is equally due to endogenous and contextual effects.

Identification and Estimation of Endogenous Peer Effects in the Presence of Multiple Reference Groups

There has been a considerable amount of interest in the empirical investigation of social influence in the marketing and economics literature in the last decade or so. Among the many different empirical models applied for such investigations, the most common class of model is the linear-in-means model. These models can be used to examine whether the social influence is truly due to agents affecting each other through their choices simultaneously (endogenous effect) or due to having similar taste and characteristics (homophily). However, the two effects are not separately identified in general in the standard linear-in-means model unless data on all members of an individual’s network are available. With data on a sample of individuals from a network, these effects are not identified. In this research, we leverage a very natural aspect of social settings viz. that consumers are usually part of multiple, as opposed to a single, networks. We discuss the sufficient conditions for identification when the standard linear-in-means models is extended to allow for multiple sources of social influence. We also show how the additional variation generated by more than one source of social influence actually allows estimation of endogenous effects with sample data. We demonstrate the feasibility of our approach via simulation and on the National Longitudinal Study on Adolescent Health data, which has been used in a number of studies examining social influence. Our approach is therefore likely to be useful in typical marketing settings.

Equilibrium Computation in Discrete Network Games

Counterfactual policy evaluation often requires computation of game-theoretic equilibria. We provide new algorithms for computing pure-strategy Nash equilibria of games on networks with finite action spaces. The algorithms exploit the fact that many agents may be endowed with types such that a particular action is a dominant strategy. These agents can be used to partition the network into smaller subgames whose equilibrium sets may be more feasible to compute. We provide bounds on the complexity of our algorithms for models obeying certain restrictions on the strength of strategic interactions. These restrictions are analogous to the assumption in the widely used linear-in-means model of social interactions that the magnitude of the endogenous peer effect is bounded below one. For these models, our algorithms have complexity O_p(n^c), where the randomness is with respect to the data-generating process, n is the number of agents, and c depends on the strength of strategic interactions. We also provide algorithms for computing pairwise stable and directed Nash stable networks in network formation games.

Judicial arbitration of unfair dismissal cases: The role of peer effects

The paper analyses how peer effects affect judicial decisions over dismissal disputes in Australian labour courts. Using a panel of judges and data over 2000 unfair dismissal decisions we first test for selection effects and for randomized matching of judges to cases. We then use spatial econometrics methods to test whether individual judges’ decisions depend on the behaviour and/or characteristics of their peer group. We do not find evidence of endogenous (behavioural) peer effects, but we find significant exogenous (characteristics) peer effects.

Endogenous peer effects and level of informality: some evidence from micro and small firms in Cameroon

Standard economic theory assumes that individuals’ preferences are independent of their social environment. However, this basic assumption seems partly unrealistic because individual utility can be affected by a variety of social interactions. This paper assesses the role of peer effects on the informality of Micro and Small firms. We use the instrumental variable approach with fixed effects on survey data in the informal sector in Cameroon. Our results show a positive impact of informal behavior of peers of the firm on its level of informality. Thus, we find a social multiplier of 9.43 and 4.65 according to the nature of the reference group. These results show that, in reality, a policy leading one firm to formalize will lead at least nine (or four depending on the reference group) others to do so due to peer effects.

Exploring peer effects in education in Latin America and the Caribbean

This paper assesses peer group influence on academic performance of primary school students in Latin America and the Caribbean. Based on TERCE data set, we investigate peer effects in mathematics, language and sciences tests outcomes among sixth grade students. We apply a social interaction model which allows to identify endogenous and exogenous peer effects while controlling for group-level fixed effects. We explore some heterogeneities related to the school type (private, public or rural). The estimates suggest the existence of endogenous peer effects but their magnitude and significance depend on subject and school type.

Equilibrium computation in discrete network games

Counterfactual policy evaluation often requires computation of game‐theoretic equilibria. We provide new algorithms for computing pure‐strategy Nash equilibria of games on networks with finite action spaces. The algorithms exploit the fact that many agents may be endowed with types such that a particular action is a dominant strategy. These agents can be used to partition the network into smaller subgames whose equilibrium sets may be more feasible to compute. We provide bounds on the complexity of our algorithms for models obeying certain restrictions on the strength of strategic interactions. These restrictions are analogous to the assumption in the widely used linear‐in‐means model of social interactions that the magnitude of the endogenous peer effect is bounded below one. For these models, our algorithms have complexityOp(nc), where the randomness is with respect to the data‐generating process,nis the number of agents, andcdepends on the strength of strategic interactions. We also provide algorithms for computing pairwise stable and directed Nash stable networks in network formation games.

The role of peer effects in firms’ usage of R&D tax exemptions

Survey evidence suggests that firms are insufficiently aware of newly introduced R&D support measures due to the complexity of the public support landscape. As a result, adoption is slow and incomplete, which implies that eligible firms leave money on the table. We hypothesize that a key coping mechanism involves firms relying on their peers’ behaviour to inform their own adoption decision. We test this hypothesis by analysing firms' first use of a newly-introduced R&D tax exemption scheme in Belgium. We identify endogenous peer effects in industry- and location-based peer groups by exploiting the intransitivity in firms’ networks as well as variation in peer group size. The results show that firms’ decisions to use R&D tax exemptions are influenced by the choices of their peers, primarily in the time window immediately following their introduction. The findings suggest that the efficacy of R&D policy can be improved by accounting for the structure of firm networks in the communication of new support initiatives.

Simultaneous Binary Choice Models with Social Interactions: An Empirical Application to Youth Delinquency

This study considers the estimation of simultaneous binary choice models with social interactions, with particular focus on a bivariate case. Assuming that individuals make binary decisions simultaneously to maximize their overall utility, defined as the sum of utilities from each binary outcome, under incomplete information, we provide a sufficient condition for unique equilibrium, and develop a copula-based maximum likelihood estimator based on the nested fixed-point algorithm. As an empirical analysis, we apply the proposed model and method to National Longitudinal Study of Adolescent Health (Add Health) data in order to investigate the determinants of adolescents’ smoking and drinking behaviors. Our estimation results suggest that smoking and drinking are complementary activities, and that endogenous peer effects are significantly present for smoking behavior, but not for drinking.

Public Expenditure Interactions of Chinese Local Governments: A Simultaneous Equations Network Approach

ABSTRACTFiscal competition, mimicking, and interactions among local governments exist as a complex network. To better model and estimate public expenditure interactions of Chinese local governments, we adopt a simultaneous equations network approach in which spatial dependence of multiple components of local governments’ spending and also the internal competition among different components of public expenditure within the same city are considered. Using Chinese county-level fiscal data, we find that the endogenous peer effect for capital construction expenditure is positive among Chinese local governments, capital construction expenditure and administrative expenditure are complements rather than substitutes, there is no fiscal mimicking of social welfare expenditure, and there is a strong crowd-out effect of administrative expenditure on social welfare spending.

Measuring Social Interaction Effects When Instruments Are Weak

Studies that can distinguish between exogenous and endogenous peer effects of social interactions are relatively rare. One recent identification strategy exploits partial overlapping groups of peers. If a student has two groups of separated peers, the peer choices are correlated through that specific student's choice, but one group's attributes are assumed to directly influence neither the other peer group's attributes nor the choices. In the context of academic performance in higher education, however, the evidence of peer effects on academic outcomes has been mixed, creating a potential for weak instruments. We utilize a period of transition when students were being reassigned to dormitories from a new campus to an old campus. Many groups of roommates were broken up at the end of freshman year, and then combined with other groups of students from the same school in the sophomore year. We find reduced-form evidence that information about a student's previous year roommates can explain the current test scores of their new roommates. However, due to weak instruments, the estimated endogenous effects appear unreasonably large. We draw on weak-IV robust tests, namely the Anderson-Rubin-type S-test (Stock and Wright, 2000) and Kleibergen's Lagrangian multiplier test (Kleibergen, 2005), to provide properly-sized tests for the endogenous effects between the test scores of current roommates and to calculate lower bounds of such effects. These tests strongly reject the null hypothesis of no endogenous effects.

Separating Gender Composition Effect from Peer Effects in Education

This paper aims to highlight the importance of considering endogenous peer effects, as defined by Manski (1993), in order to identify gender composition effect on education outcome appropriately. Using Manski (1993) linear-in-means model, this paper illustrates that the gender composition effect that is currently estimated in education function is the function of three parameters: social multiplier, gender differences in outcome and gender composition effect (known as a gender peer effect). The appropriate gender peer effect is identified after using Graham's variance restriction method to identify and rule out a social multiplier effect. The findings suggest that a social multiplier plays a crucial role in learning process for Italian secondary and US primary students, although a gender peer effect is not as important as highlighted in previous literatures (Hoxby, 2000; Whitmore, 2005; Lavy and Schlosser, 2011).

Separating gender composition effects from peer effects in education

ABSTRACTThis paper aims to demonstrate the importance of controlling for endogenous peer effects in estimating the influence of gender peer effects on educational outcomes. Using Manski's linear-in-means model, this paper illustrates that the estimation of gender peer effects is potentially biased in the presence of endogenous peer effect in education. The appropriate gender peer effect is estimated after identifying and controlling for the endogenous effect through the use of Graham's variance-restriction method.