Many of our consumption decisions are influenced by related decisions made by our friends (Moretti 2011, De Giorgi et al. 2016). These peer effects have important implications for firms and policymakers. One example is that, in the presence of peer effects, the aggregate elasticity of demand is lower than the elasticity of individual demand, because individuals who are persuaded to purchase by a price reduction might encourage their friends to buy too. By reducing aggregate price elasticities, peer effects therefore lead to lower markups, and higher consumer surplus. From a macro perspective, peer effects in consumption imply that the effects of stimulus policies on aggregate demand are larger than those we would estimate if we looked at directly affected individuals.
Although peer effects have important implications, we have limited empirical evidence on their exact nature. For example, peer effects may lead an individual to buy a new phone after her friend gets a new phone, but the effect of this purchase on firm behaviour would depend on whether the second purchase represented incremental demand or the retiming of an already-planned purchase. Peer effects influence overall aggregate demand elasticities if they bring forward a purchase, but not if they create incremental demand. The implications also depend on whether the changes in demand are restricted to the precise model purchased by her friend, or whether there are positive or negative demand spillovers to competing brands.
Using Facebook data
Using de-identified data from Facebook, we explored the nature of peer effects in the US mobile phone market (Bailey et al. 2019). Facebook has more than 242 million active users in the US and Canada, and more than 2.3 billion users globally. In this data set, individuals’ social networks are represented by their Facebook friends, which we know provide a fair representation of real-world US friendship networks (Bailey et al. 2018a, b). For mobile active users, we could also capture data on the device model used to log into Facebook accounts, and so we could identify when users bought a new phone. By combining these data sets, we could explore how phone purchases by a user’s friends influence the user’s own phone-purchasing behaviour.
We exploited quasi-random variation in friends’ phone purchasing behaviours to separate peer effects from common shocks or common preferences within friendship groups. Specifically, we used the number of friends who broke or lost their phones in a given week to instrument for the number of friends who purchased a new phone in that week. The identifying assumption is that how many friends break or lose their phones in a given week is conditionally random, and is not related to a user’s own propensity to buy a new phone in that week.
We identified individuals who randomly broke or lost their phones by applying natural language processing and machine learning techniques to public posts on Facebook. A post such as: "Phone broken... ordered a new one but if anyone needs me urgently, call Joe," is a signal of random phone loss by a peer, and people are substantially more likely to obtain a new phone in the week after posting such messages.
Size and duration of peer effect
If a friend purchases a new phone, an individual’s probability of buying a new phone in the following week rises by 0.04 percentage points. This estimated effect is large, as the weekly probability of buying a new phone is about 1%.
Peer effects are long-lasting and generate substantial incremental demand. If an individual randomly loses her phone, there is a positive effect on the total number of phones purchased by her friends in the following 10 months. Figure 1 shows the effects of having a friend purchase a new phone in week 0 (induced by that friend's random phone loss in week 0) on own-purchasing probabilities over four-week periods afterwards.
Figure 1 The peer effect of a friend's new phone purchase 10 months
Source: Bailey et al. (2019), using de-identified Facebook data.
Therefore peer effects cause an increase in the total number of phone purchases, at least over intermediate horizons. If one extra friend purchases a new phone, an individual’s own probability of purchasing a new phone over the 16 weeks that follow increases by 0.6 percentage points, relative to a baseline probability that person would buying a new phone in the period of about 14.6%.
An important implication of this creation of new demand is that the value to a firm of acquiring new customers is greater than the direct effect of these customers on the revenue and profitability of the firm.
Heterogenous effects and cronut sales
The peer effects are very different in different demographic groups. Panel A of Figure 2 shows the per-friend peer effects exerted by individuals in different groups, while Panel B shows the overall peer effect – it adjusts for the fact that individuals with different demographic characteristics have differentially many friends.
Figure 2 Differences in size of peer effect in different demographic groups
Source: Bailey et al. (2019), using de-identified Facebook data.
The evidence shows, for example, that younger and less-educated individuals have the largest effects on their friends’ purchasing behaviours. This has important implications for the design of viral or seed-marketing campaigns that target a small set of early adopters, who generate follow-on demand through peer effects.
Individuals who exert larger peer effects are generally more price sensitive, measured as the effect of a price cut for a phone model on the probability of purchasing that model. This suggests that the difference between the elasticities of aggregate and individual demand induced by peer effects is even larger than implied by the average peer effect. This higher price elasticity faced by firms leads to lower optimal markups.
The positive correlation between price sensitivity and peer influence way explain why markets with supply constraints clear through queuing, rather than through price increases. If price increases disproportionately reduce demand from individuals who have large peer effects on their friends, then optimal dynamic pricing for the supplier would be a willingness to accept lower revenue today, in return for additional sales generated through peer effects in future.
While demand at those lower prices sometimes exceeds supply, for example when a new iPhone is launched, assignment via queuing disproportionately selects the enthusiastic individuals who exert the largest peer effects. Similar mechanisms might be at work in other settings in which queuing selects individuals who exert large peer effects (new sneaker designs, restaurants, Broadway shows and, of course, Cronuts).
Competitive brand effects
Are these peer effects are limited to the brand purchased by the peer, or are there demand spillovers to other brands? We analysed three brands – iPhone, Samsung Galaxy, and 'other' – and found three main takeaways:
- Positive peer effects are largest for phones in the same category as that purchased by the peer, for all three categories.
- Same-brand peer effects are largest for less-well-known but cheaper 'other' phones, and smallest for expensive, well-known iPhones. This suggests that social learning is an important part of the explanation for these peer effects. Social learning should be more important for lesser-known brands, while keeping-up effects should be more important for expensive brands that signal high status. (Our results do not allow us to rule out that keeping-up effects also contribute to the observed peer effects, but our findings suggest that they cannot be the whole story, and that a large part of the peer effects come from social learning.
- When a friend purchases a new phone, a person's propensity to purchase a phone from a competing brand on the same operating system increases. The propensity to purchase a phone from a competing brand on a different operating system decreases. In other words, some of the positive same-brand peer effects are from entirely new purchases (consistent with the evidence in Figure 1), but others come from pulling demand away from rival firms with competing operating systems. Note these demand spillovers across operating systems could easily be positive. For example, a user who buys a Samsung Galaxy might have caused her friends to desire more expensive phones – of any type, including iPhones – through a keeping-up effect. These peer effects have important competitive implications: losing a customer to a competitor does not only mean missing out on positive peer effects that this customer could have had, but may also lead to future losses of other customers through competitive peer effects.
Bailey, M, R Cao, T Kuchler, J Stroebel, and A Wong (2018a), "Social Connectedness: Measurement, Determinants, and Effects", Journal of Economic Perspectives 32(3): 259–80.
Bailey, M, R Cao, T Kuchler, and J Stroebel (2018b), "The economic effects of social networks: Evidence from the housing market", Journal of Political Economy 126(6): 2224–2276.
Bailey, M, D M Johnston, T Kuchler, J Stroebel, and A Wong (2019), "Peer effects in product adoption", NBER working paper 25843.
De Giorgi, G, A Frederiksen, and L Pistaferri (2016), "Consumption network effects", NBER working paper 22357.
Moretti, E (2011), "Social learning and peer effects in consumption: Evidence from movie sales", The Review of Economic Studies 78(1): 356–393.