Endogenous Identity in a Social Network

arXiv:2406.10972 · econ.TH · Submitted 2024-06-16 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Economics. The day in economics: theory, mechanisms and the evidence beneath them.

Ruth: We're looking at the paper: "Endogenous Identity in a Social Network".

Warren: Here's the gist of the paper. Interaction with individuals from other socioeconomic classes has been shown to be a main driver for social mobility <ref:2406.10972#pg2> The paper employs tools of social identity theory and network analysis to show how exposure to individuals of different social identities can lead to interactions with them, and an adoption of their identity, creating social mobility <ref:2406.10972#pg2> The authors find that even if all individuals have the same ability, they may endogenously choose different identities, leading to different classes and actions <ref:2406.10972#pg4> In particular, we derive a sufficient condition for such an equilibrium to exist, which equates to a novel measure of cohesion <ref:2406.10972#pg4> Furthermore, we show that the most socially mobile individuals (changing their identity) are those who either have few connections or a more heterogeneous mix of identities in their connections <ref:2406.10972#pg4> Finally, we show that upward social mobility increases action levels in society, but not necessarily welfare <ref:2406.10972#pg4>.

Ruth: Model Framework and Identity Utility The model is based on social identity theory where belonging to a certain identity provides utility to individuals through three channels <ref:2406.10972#pg9> These channels include a cognitive factor, an emotional factor, and how many ties one has with others of a given identity <ref:2406.10972#pg9>.

Warren: Following Shayo (2009) the economic literature has adopted a dual formulation where associating with a particular identity confers both a certain status to the individual and prescribes a specific action level.

Ruth: Individuals prefer identities with a higher status, but incur a cost whenever their actual action differs from the prescribed action of the identity.

Paper summary: Ruth: Building on what we just touched on, this paper "Endogenous Identity in a Social Network" really zeroes in on how social identity theory can explain the mechanisms behind social mobility. They argue that belonging to an identity gives people utility through three main channels: how many connections you have with people of your own group, conforming to the actions prescribed by your identity and your neighbors, and simply having a high social status attached to that identity >

Warren: It sets up a dual formulation where associating with an identity gives you both a certain status and prescribes a specific action level >

Iris: And there’s this cost involved too. If you actually act differently than what your identity tells you to do, it costs you utility >

Ruth: Right, and the model for the equilibrium in stage two shows that individuals choose their action level to maximize that utility equation, leading to a Nash equilibrium where actions are determined by the identity's value and its prescribed action >

Warren: The first order condition boils down to individual i choosing an action xi of one, which is equal to one plus gamma times vI plus alpha times the average action of all other individuals of type I >

Iris: So it predicts a Nash equilibrium for the actions for each identity, where the equilibrium action level is determined by that value and its prescribed action >

Ruth: Then they move on to coexistence and diffusion, showing that even if everyone has identical abilities initially, multiple identities can exist in the network >

Warren: They show that identity choice depends on three things: how much status the identity has, how far away from ability the prescribed action is, and how many neighbors of that type an individual actually has >

Iris: And they find an equilibrium where everyone chooses the intrinsically most valuable identity, which they call A >

Ruth: But they also show that if there's a certain group S where everyone in S has more connections than some threshold c, then multiple identities can coexist >

Warren: It’s a condition related to this cohesion measure that determines if these different identities can actually stick together in the network >

Conclusion: Iris: So looking at the whole paper "Endogenous Identity in a Social Network," the authors are showing us that social mobility isn't just about external factors; it’s deeply rooted in how we feel about our group and how those feelings shape our choices and actions >

Ruth: The title itself, "Endogenous Identity in a Social Network," highlights that these identities don't just exist outside the system; they are chosen within the network structure itself, which is a key insight for understanding mobility >

Warren: The main implication for me as someone who looks at strategy is that if you want to influence mobility, you might need to focus less on external opportunities and more on shaping these internal identity dynamics >

Iris: It’s important to remember the authors found that upward social mobility can actually be welfare-reducing, which is a significant caveat for policy discussions >

Ruth: Right. So while this model shows how interactions drive mobility, it also warns us that just moving up socially doesn't automatically translate into better outcomes for everyone involved >

Essex University · Ca’ Foscari University of Venice

econ.TH

Submitted: 2024-06-16

Updated: 2026-10-31

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 77/100

The gist: Interaction with individuals from other socioeconomic classes has been shown to be a main driver for social mobility <ref:2406.10972#pg2> The paper employs tools of social identity theory and network…

Key concepts

Social Identity Theory
This theory suggests that belonging to a certain group or identity provides utility to an individual through cognitive, emotional, and social factors. In this model, an individual's choice of identity is driven by these psychological benefits.
Identity Utility Channels
Utility from an identity comes from three main sources: a cognitive factor (how it changes thinking), an emotional factor (how it makes one feel), and the number of connections one has with others sharing that same identity. These factors determine how attractive an identity is.
Nash Equilibrium Action Profile
This is the stable state where no individual benefits from changing their chosen action level, given what everyone else is doing. The model shows that for each social identity, there is a unique set of actions that represents this stable outcome.
Cohesion Measure
A novel measure derived in the paper that serves as a sufficient condition for an equilibrium to exist where multiple identities can coexist in the network. This measure relates to how interconnected and diverse the social structure is.

Terminology

Summary

Interaction with individuals from other socioeconomic classes has been shown to be a main driver for social mobility <ref:2406. p. 010972#pg2> The paper employs tools of social identity theory and network analysis to show how exposure to individuals of different social identities can lead to interactions with them, and an adoption of their identity, creating social mobility <ref:2406. p. 010972#pg2> The authors find that even if all individuals have the same ability, they may endogenously choose different identities, leading to different classes and actions <ref:2406. p. 010972#pg4> In particular, we derive a sufficient condition for such an equilibrium to exist, which equates to a novel measure of cohesion <ref:2406. p. 010972#pg4> Furthermore, we show that the most socially mobile individuals (changing their identity) are those who either have few connections or a more heterogeneous mix of identities in their connections <ref:2406. p. 010972#pg4> Finally, we show that upward social mobility increases action levels in society, but not necessarily welfare <ref:2406. p. 010972#pg4>.

Model Framework and Identity Utility

The model is based on social identity theory where belonging to a certain identity provides utility to individuals through three channels <ref:2406.10972#pg9> These channels include a cognitive factor, an emotional factor, and how many ties one has with others of a given identity <ref:2406. p. 610972#pg9>. Following Shayo (2009) the economic literature has adopted a dual formulation where associating with a particular identity confers both a certain status to the individual and prescribes a specific action level. p. 7 Individuals prefer identities with a higher status, but incur a cost whenever their actual action differs from the prescribed action of the identity. p. 7 The utility of individual i, choosing action xi,I and identity I, when faced with network gI and the vector of actions x−i,I of all other individuals of type I is given by equation (1). p. 9

Equilibrium Conditions for Identity Choice

The equilibrium in stage 2 involves individuals choosing their action level to maximize equation (1). The first order condition of equation (1) with respect to xi,I yields: xi,I = bi [1 + γvI + αx¯i,I]. p. 9 This leads to the Nash equilibrium action profile of individuals of identity I being x∗ I = [1 + γvI]HbI. p. 9 Lemma 1 states that there always exists a unique Nash equilibrium in action levels for each identity. p. 2

Coexistence and Diffusion of Identities

Under the assumption that all individuals are ex ante identical in abilities, we show that multiplicity of identities may arise even in such a homogeneous population. Proposition 3 highlights that the choice of identity will depend on its status µI, how far away from ability w is the prescribed action vI, and - for each individual i - how many neighbors of type I they have. p. 8 Proposition 4 states that there always exists an equilibrium in which all individuals choose the intrinsically more valuable identity (here, A). p. 2 Furthermore, there exists an equilibrium in which multiple identities coexist if there is at least one group S, such that all i within S have ki(S) > c. p. 2

Implications for Social Mobility and Welfare

The paper analyzes a stylized scenario to inform on the policy of a common space in schools highlighted in Chetty et al. (2022b). The analysis shows that in Scenario 1, all students choose the SES of their parents as their own identity if and only if c ∈ (−d, d). p. 0 In Scenario 2, all students choose the same SES. p. 0 The analysis also shows that upward social mobility might be welfare-reducing and provides illustrative examples of welfare-enhancing and welfare-reducing upward mobility. p. 0

Conclusion

While the continuing lack of socioeconomic mobility remains somewhat of a puzzle, interactions across socioeconomic classes have been shown to be a leading factor in overcoming it. The present paper utilizes the tools of social identity theory and network analysis to shed light on possible factors affecting such interactions. p. 0 The authors conclude by showing that upward social mobility might be welfare-reducing. p. 0 The next step is to consider heterogeneous individuals, which will introduce a negative effect of being connected to others of your own identity. p. 5

Proofs

A.1 Proof of Lemma 1 states that the existence and uniqueness of the equilibrium in equation (4) follows the arguments put forward in Ushchev and Zenou (2020). p. 3 A.2 Proof of Proposition 1 shows that a Nash equilibrium always exists because each individual has a finite set of strategies available in the game. p. 2 A.3 Proof of Proposition 2 derives equation (7) and then equation (8). p. 9 A.4 Proof of Proposition 3 follows immediately from equation (8) in the main text. p. 0 A.5 Proof of Proposition 4 shows that the existence of a second equilibrium, in which all individuals choose the intrinsically less attractive identity, depends on the difference between the intrinsic values, c. p. 2 A.6 Proof of Proposition 6 provides a sufficient condition for A to diffuse through the entire network as that is the stated condition. p. 0

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Ability to predict individual identity choices based on network structure: The improved system can determine that Individuals with fewer neighbors are more likely to change their identity than those with more neighbors when abilities are homogeneous, as derived from Proposition 5.

  2. Ability to model social mobility dynamics through identity diffusion: The AI can simulate how an intrinsically more valuable identity diffuses through a network by calculating the necessary conditions for its complete adoption, such as the sufficient condition that "max di - 2 <= c'" for full diffusion (Proposition 6).

  3. Ability to identify welfare implications of social mobility: The system can assess whether upward social mobility is welfare-enhancing or reducing by comparing two illustrative examples, determining if welfare-enhancing upward mobility may be blocked by the importance to individuals of being 'similar' to their social connections (Welfare Example 1) or if coordination on identity B would be welfare maximizing (Welfare Example 2).

  4. Ability to predict equilibrium identity configurations in homogeneous populations: The system can determine that all individuals of the same identity choose the same action level under homogeneity, and find conditions for multiple identities to coexist based on the difference between intrinsic values and network structure, such as "there is at least one group S, such that all i within S have ki(S) > c" (Proposition 4).

  5. Ability to predict diffusion thresholds for identity change: The AI can calculate the threshold for an individual to change their identity from B to A based on the relative attractiveness of identities and their neighborhood structure, specifically using the condition di,A − di,B ≥ c (Equation 9).

Related papers