The Effects of Price Discrimination in Village Economies

EGC Research Summary, February 2021

As tools to address poverty, cash transfer programs have had much success in developing countries over the past 20 years. World Bank researchers have documented a dramatic expansion of these programs to address the Covid-19 crisis, and argued that they should continue to be scaled up and improved after the pandemic ends. 

The conditional version of these programs – Conditional Cash Transfer programs, or CCTs – offer cash payments, typically paid by the government, to eligible families so long as they satisfy certain requirements that support development goals, such as immunizing children or keeping them in school. 

Despite their strong track record, a recent study finds that CCT programs may exacerbate consumption distortions and ultimately lead to higher unit prices for the poorest households. This is a risk in settings in which sellers have market power – meaning they are able to raise prices with little risk of being undersold – and exercise it by discriminating among consumers through quantity discounts, that is, by charging higher unit prices for small quantities than for large quantities. Using high-quality data from Mexico’s large-scale Progresa program, the researchers find that CCTs have countervailing effects: on the one hand, they can lead to higher food prices for the poor, regardless of whether they receive transfers or not, but on the other hand, they ensure that poor households are not completely priced out of markets. 

Nonlinear Pricing in Village Economies

Orazio Attanasio and Elena Pastorino, Econometrica, Volume 88, Issue 1, 2020, Pages 207-263.

Results at a Glance

  • Price discrimination in the form of quantity discounts is common in rural Mexico for basic food staples like rice, beans, and sugar.

  • Mexico’s CCT program did not have a significant impact on the average unit prices of these food items.

  • Yet, the program lead to greater price discrimination for each of these goods in the form of higher quantity discounts with poorer households facing higher unit prices for small-quantity purchases and richer households facing lower unit prices for large-quantity purchases.

  • Indeed, the slope of the schedule of unit prices for each good steepened significantly as a result of the program: it increased from -0.320 to -0.328 for rice, from -0.188 to -0.196 for beans, and from -0.198 to -0.210 for sugar.

  • In general, some degree of price discrimination could benefit poor consumers by allowing for greater market participation and consumption than under a linear pricing scheme.

Understanding the Price Effects of CCT Programs 

CCTs seek to alleviate the short- and long-term effects of poverty by providing additional income to households while also incentivizing parents to invest in the health and education of their children. Since their inception, though, policymakers have been concerned about the possibility that the positive impact of cash transfers would be reduced if consumers' higher spending power led to increases in prices. Intuitively, when consumers receive extra income, sellers have an incentive to adjust offered quantities and prices in response to consumers’ higher ability to pay. In the context of Progresa, a celebrated CCT program in rural Mexico, early research showed that the program did not have any impact on average unit prices.

A fruit vendor in a rural market
A market in Uman, Yucatan. KrisNM, Flickr

But these studies did not assess CCTs’ impact on prices taking into account the possibility of price discrimination through nonlinear pricing, or quantity discounts. The phenomenon works like this: when consumers differ in their preferences for a good, a seller has an incentive to sell different quantities of a good at different unit prices in order to induce consumers who like the good relatively more to purchase more of it. This is the usual logic of quantity discounts, that is, of lower unit prices for larger quantities. In developing countries, indeed, sellers often charge higher unit prices for small quantities than for large quantities of common goods. For this reason, nonlinear pricing is usually thought to disproportionately hurt the poor, who have fewer resources and so are more likely to buy goods in smaller quantities that are sold at higher unit prices. 

New research by Orazio Attanasio, Cowles professor of Economics at Yale University and an EGC affiliate, and Elena Pastorino, research fellow at the Hoover Institution at Stanford University and research scholar at the Stanford Institute of Economic Policy Research, takes nonlinear pricing into account. Specifically, Attanasio and Pastorino develop a model of price discrimination that allows them to measure the effects of CCTs on consumption and prices in the presence of nonlinear pricing. Building on previous work on economic models of price discrimination, their study explicitly incorporates differences across households in budgets, incomes, and preferences for consumption, which imply that price discrimination can lead to levels of consumption below or, surprisingly, above the efficient level. The authors then apply their model to data from Progresa and find that the price effects of the program might have been more subtle than often thought, as the program has affected the degree of price discrimination faced by consumers, which is common in these contexts.

Questions on post-Covid public policy renew the focus on income transfers and therefore call for a thorough assessment of their impact so as to make them work best for the poor."

- Orazio Attanasio and Elena Pastorino, study authors

Analyzing Price Discrimination in the Context of Mexico’s Progresa Program

Progresa is a national CCT program started by the Mexican government in 1997 to improve education, health, and nutrition in poor rural villages. During the first phase of the program, randomized among 506 poor rural villages, Progresa conducted extensive evaluation surveys that resulted in a rich dataset containing both the quantity bought of a number of commodities and the price paid for them by each household in each village. Using this information, the authors characterized the demand structure faced by sellers in each village. 

According to the authors’ model, a seller sets prices to discriminate among consumers based on the distribution of consumer preferences and resources in each village. The distribution of consumers’ characteristics in a village, in turn, is reflected in the distribution of quantities they purchase. Together with the price schedule, the distribution of quantities purchased can then be used to recover the determinants of prices and consumption in a village, in particular the distribution of consumers’ preferences. 

Attanasio and Pastorino estimated their model using the Progresa dataset to analyze the impact of price discrimination both on consumption (the amount customers buy) and the unit prices of rice, beans, and sugar – three basic food items that are widely consumed by Mexican households of all income levels, and that households typically consume more of when they have more income to spend. The authors found that sellers have market power in the villages they studied, and exercise it by price discriminating across consumers through distortionary quantity discounts. Interestingly, according to the authors’ estimates of consumer preferences and seller costs, a substantial fraction of consumers, including those of small quantities, consume more than the efficient level of each good rather than, as usually argued, less than it. Hence, nonlinear pricing in practice can support high levels of consumption and need not exacerbate, but may actually reduce, consumption inequality. 

Key Finding: CCTs Can Increase Prices for Poor Households

The authors also show that Progresa had a substantial impact on unit prices. Intuitively, since transfers are tailored to households’ needs, they generally lead to differential increases in households’ ability to pay depending on their characteristics. Correspondingly, cash transfers provide an incentive for sellers to differentially modify their unit prices for the different quantities they provide. The authors found that Progresa led in fact to significant changes in the price schedules of rice, beans, and sugar. With the program’s introduction, the schedules of unit prices of these basic goods became significantly steeper – that is, the magnitude of quantity discounts has increased in villages receiving Progresa transfers, with sellers charging higher unit prices for small quantities and offering instead higher discounts for large quantities. “Since poor households typically buy in smaller quantities, they have ended up facing higher unit prices – whereas less poor households ended up facing lower unit prices after the introduction of the program,” Attanasio and Pastorino explained in an email interview.

Crucially, the authors show that, consistently with their model, the changes in the distribution of expenditures and consumption in each village induced by the program can explain, in a precise statistical sense, the tilting of price schedules they document.

Key Finding: Price Discrimination Can Support Market Inclusion

The authors also used their model to explore the hypothetical effects of removing price discrimination entirely in order to illustrate the impact of price discrimination on consumption and market participation. In developing contexts like rural Mexico, poor households face significant subsistence constraints but also have alternative options to purchasing basic food items in regular markets, including directly farming them and buying them from heavily subsidized government stores. Under linear pricing, sellers have to charge all consumers the same unit price regardless of the quantity purchased. To attract poor consumers under linear pricing, sellers may then have to lower their unit prices for all consumers so much that their profits would substantially decrease. As a result, it may not be in a seller’s interest to cater to the poorest consumers under linear pricing, leading to their exclusion from a market. Depending on the good considered, the authors estimated that between 20% to 70% of households in the bottom 5% of the distribution of quantities purchased would be excluded from the corresponding market under linear pricing. 

Attanasio and Pastorino said these findings imply “both that price discrimination is less detrimental for consumers than often presumed, and that the appeal of identical unit prices for all consumers as a measure of equality of opportunity for market access may be overstated if not altogether misleading.” That is, whenever sellers can charge different unit prices for different quantities, they may have an incentive to serve those consumers who would be unprofitable under linear pricing, thus supporting market inclusion.

Implications for Public Policy

These findings have important implications for how economists and policymakers should evaluate the impact of CCT programs on poverty. Attanasio and Pastorino’s study is the first to show that although CCTs may not have any significant effect on average unit prices, they may have sizable effects on price discrimination and can lead to poor households paying more for essential items. Nonetheless, the authors also find that eliminating price discrimination altogether would likely cause some of the poorest households to be excluded from key markets. According to the authors, “It is important to think about the implications of income transfers for both sides of a market to assess their overall impact on consumers. In general, effective interventions of the scale CCTs contemplate involve considerations of seller competition and market structure. Questions on post-Covid public policy renew the focus on income transfers and therefore call for a thorough assessment of their impact so as to make them work best for the poor.”

As CCTs are increasingly used as poverty alleviation tools in Latin America and globally, economists and policymakers should strive to better understand – and plan for – their potential negative side effects associated with price discrimination in order to maximize their positive net impact on poor households. One possibility would be to foster greater competition among sellers. This might be accomplished through lowering barriers to seller entry in these markets, for instance, by reducing production and distribution costs through better infrastructure. If more sellers can get to market, increased competition could be good for the poor and reduce the negative price impact of cash transfers.
 

Research summary by Sarah Guan