Pricing Decisions and Over-Entry of Delhi Fruit Vendors

Many development economists have noted the puzzle that seemingly too many vendors operate side by side, duplicating one another's efforts, yet neither competition nor cooperation drives vendors out of the market. The reserachers here propose a theory that rationalises this phenomenon: first, despite high density of vendors, there is a lack of price competition, keeping profits per transaction high. Second, barriers to entry are low, leading to high entry until the volume of transactions per vendor (and therefore profits per vendor) are relatively low (equal to vendors’ outside options). These two features of low price competition and low barriers to entry naturally lead to an equilibrium with over-entry and inefficiently high prices. With census data of around 1,500 fruit vendors in South Delhi previously compiled by the reserachers, they find some support for this theory. In this project, they conduct a field experiment that would explore the mechanisms behind low price competition (despite high density) in the context of fruit vendors in Delhi.

The reserarchers plan to identify 350 spatially-dispersed vendors to participate in the study. The prices at which vendors sell at will be recorded on three days across two weeks. After the first day of measurement, vendors will be assigned to one of five treatments that provide incentives for them to either cut or raise their prices and that vary on whether their sales will be shared with nearby vendors or not. On the last day of measurement, the treatments are removed. The experiment tests for collusive norms by testing for whether vendors are less willing to cut prices when their price-cutting behaviour is made public to neighbouring vendors. The key prediction of collusive pricing is that the price-cut incentives should be less effective when the price cuts are publicised. In particular, we expect to find that prices charged by vendors assigned to the price-cut, public treatment are higher than those charged by vendors assigned to the price-cut, private treatment. The experiment tests for underexperimentation by testing for persistent effects of temporary incentives to change prices. It looks to discover whether those vendors assigned to treatments with incentives to raise or cut prices continue this pricing behaviour on the last day of measurement when the incentives have been removed. If prices are significantly lower on the last day among vendors assigned to the price-cut treatments than among control vendors, this would give support to the hypothesis that initial prices were sub-optimally high as a result of a lack of under-experimentation.

The oversupply of labour in low-skill, low barrier to entry markets is a first order policy concern across many developing countries, and this project aims to illuminate an underappreciated feature of these markets - low price competition - which exacerbates this phenomenon. By understanding the cause of low price competition, the authors may also shed light on natural policy responses, such as regulating prices, entry, and visibility of pricing decisions that could reduce collusion or induce experimentation. This would substantially reduce the need for labour in these markets and free up labour for more efficient uses. One simple such policy, for example, would be an explicit price ceiling, which might resolve both of the frictions we intend to explore.


Benjamin Roth

Massachusetts Institute of Technology

Matthew Lowe

Massachusetts Institute of Technology