What happens when citizens befriend government officials? The literature offers two competing explanations. One body of research suggests that the formation of such “relational contracts” could lead to collusion and costly misallocation of government resources. On the other hand, scholars have posited that these relationships may also bolster efficiency of corruption by reducing the transaction costs of corrupt exchange. Since friendships rarely form randomly, quantifying the impacts of friendships between citizens and state officials has proven a challenge.
In this paper, the authors induce randomness in whether public transit drivers have social relationships with police officers along their route in the Democratic Republic of the Congo’s (DRC) capital, Kinshasa. Drivers in Kinshasa are frequently stopped by police officers, leading to lengthy negotiations with the police over bribes. Here is the problem for the drivers: spending time negotiating with passengers on travel price or trying to attract new passengers increases the risk of detection by the police, potentially leading to lengthy bribe negotiations, the loss of passengers, and significant revenue loss. In this context, the authors use data covering bribes and negotiations to document the following:
- Negotiating a bribe can cause significant time waste— in some cases up to one hour—and can result in large payments—up to $27 USD for a single bribe.
- Drivers typically opt for a once-daily payment in lieu of continual negotiations with the officers along their route. To be allowed to participate this unofficial toll system, called the “Mbote ya Likasu,” drivers must first undergo a lengthy negotiation and make a large payment to a team of officers. Upon establishing this relationship, drivers can proceed through the unofficial toll by making daily payments of $0.53 cents. Such interactions typically take less than 40 seconds.
- Of all interactions with the police, 84.1% consist of the toll fee described above, while only 2.7% are bribes that involve negotiation. The remaining 13.1% are amicable conversations without payments.
Next, the authors randomize the likelihood of social ties between drivers and officers by experimentally re-routing drivers to lines on which they do not have a relationship with officers. They find the following:
- As one might expect, re-routing is associated with lower prevalence of relationships between the drivers and the officers, especially friendship. While the drivers and the police officers with whom they interact have social ties in 56.2% of interactions on home lines, they only do in 29.9% of interactions when re-routed.
- Re-routing also decreases the surplus from driving (revenue minus repairs and gas) by 43.3%.
- To uncover what drives this impact on surplus, the authors provide third-party protection to drivers on randomly selected days, drawing on commonly used practices whereby individuals high in the police hierarchy sometimes offer protection to drivers against harassment by individuals lower in the police hierarchy (those working in the street). They find that the effect of re-routing on driving surplus is driven by days on which drivers are not assigned third-party protection, suggesting that the effect of re-routing on the surplus is explained by the properties of the interaction the driver can expect to have with the police officers on the street, and not by other factors related to re-routing.
- The authors also experimentally vary the duration of time that drivers spend on unfamiliar routes, with drivers rerouted for a single day or a period of three days. They find that short-term re-routing decreases surplus by decreasing the surplus per trip, through passenger prices, and halves bribe costs. These results suggest that drivers in short-term re-routing are less patient when negotiating prices with passengers because time negotiating exposes them to the risk of being detained by the police.
- Long-term re-routing decreases surplus by decreasing trips per day, but increases bribe costs by 60%. This suggests that, while they could choose to reduce negotiations with passengers to avoid police detection and hence raise the surplus in the same day, drivers on foreign lines who expect to drive there in the long run instead invest in creating new ties when the horizon is long (at the expense of surplus today). Indeed, within three days of re-routing, drivers have built new ties at rates comparable to those on home lines, and the number of trips per day and profits catch up to those in the line in which they have social ties with the police.
- Finally, the authors combine these results to quantify the value of relationships between drivers and officers. Short-term re-routing, which captures the destruction of relationships because drivers do not take costly actions to create new ones, decreases the surplus from driving from $24 to $9 USD when drivers are not assigned third-party protection, amounting to a reduction of 280% in minibus revenue. Notice that these costs are about the surplus of driving, and occur independently on how such surplus is allocated between the driver and the police officer through bribe payments.
The upshot is that transaction costs arising from negotiating corrupt transactions can be huge, and that relationships form relatively quickly to sustain arrangements that mitigate these costs. Policy-wise, this implies that, if the government’s objective is to reduce opportunities for collusion, officials may need to be rotated more frequently than is traditionally assumed by governments attempting to design a bureaucracy. Over and above that, given the large transaction costs documented here, attempting to delay the creation of relationships between corrupt enforcers and citizens not only may be hopeless, but also socially costly. For example, if the elasticity of demand for driving is sufficiently low, the transaction costs which they document, huge, may well outweigh the social costs arising from the mis-allocation corruption induces.