In our recent work (Couture et al., 2018), we combine an experiment that we implement across Chinese villages with a new collection of survey and administrative microdata to provide evidence on the potential of e-commerce integration to foster economic development in the countryside. We also explore the underlying channels and the distribution of the gains from e-commerce across households and villages.
In this context, the Chinese government recently announced the expansion of e-commerce to the countryside as a national policy priority to foster rural economic development and reduce the rural-urban economic divide. Other developing countries, such as Egypt, India, and Vietnam, have recently announced similar policies to invest in the expansion of e-commerce trading into rural areas, where the majority of their populations reside (See Note 1).
To date, these policies have been motivated mainly by a number of prominent case studies of highly successful “e-commerce villages” that have experienced rapid output growth by selling both agricultural and non-agricultural products to urban markets via e-commerce. For example, by the end of 2017, China’s largest e-commerce platform, Taobao, had branded more than 2,000 rural markets in China as so-called “Taobao villages” based on their concentration of online sellers and high sales volumes on the firm’s platform (AliResearch, 2017). Inspired by these success stories, much of the current policy focus has been on rural producers. By lowering trade and information costs to urban markets, the arrival of e-commerce is meant to increase rural incomes through higher demand for local production and stronger incentives for rural entrepreneurship. There has been much less emphasis on the potential benefits to rural consumers. However, recent descriptive evidence from urban China suggests that e-commerce demand is strongest in smaller and more remote cities, pointing to potentially significant consumer gains in rural areas.
Despite the fast growth of e-commerce, we currently have limited empirical evidence on the economic consequences of access to e-commerce trading in developing countries. The recent growth of a number of “e-commerce villages” in China has captured the imagination of policy-makers and the general public, but important questions remain about whether market integration through online trading platforms have a broad and significant impact on rural development. To answer these questions, our recent paper (Couture et al., 2018) combines a randomized control trial (RCT) that we implement across villages in collaboration with a large Chinese e-commerce firm, with a new collection of household and store price survey microdata and the universe of transaction records from the firm’s internal database. The RCT takes place in eight counties that are located in three provinces (Anhui, Henan, and Guizhou), where we collect baseline and endline data for about 3800 households in 100 villages.
E-commerce is the ability to buy and sell products through online transactions coupled with transport logistics for local parcel delivery and pickup from the producer. Bringing e-commerce to the countryside in developing countries requires more than internet access. As in many emerging countries, the internet has spread rapidly to many parts of the Chinese countryside due to both smartphones and expanding broadband access. Instead, there are two currently binding barriers to e-commerce trading in the Chinese countryside, which we refer to as the logistical and the transactional barriers. First, the logistical barrier relates to the lack of modern commercial parcel delivery services. These distribution networks have started operating in Chinese cities, but have not started servicing large parts of the countryside. One well-known challenge to rural delivery is the so-called “last mile” between urban logistical hubs and relatively small pockets of populations in the countryside. Second, many rural residents potentially face a transactional barrier, due to lack of familiarity with navigating online platforms and lack of access to online payment methods. In addition, villagers may not trust transactions that occur without inspection of the product or without interacting with buyers in person.
To overcome these barriers, the Chinese government recently partnered with a large firm that operates a popular Chinese e-commerce platform. The program aims to invest in the necessary transport logistics to offer e-commerce in rural villages at the same price, convenience, and service quality that buyers and producers face in their county’s main urban center. To this end, the e-commerce firm builds warehouses as logistical nodes for rural parcel delivery near the urban center and fully subsidizes transport between the county’s urban center to and from the participating villages. To address the transactional barrier, the program installs an e-commerce terminal in a central village location. A terminal manager, who is employed by the firm, assists villagers in buying and selling products through the firm’s e-commerce platform and villagers can pay in cash at the terminal location upon receipt of their products or get paid upon pickup of their shipments.
Theoretically, we rationalize the program as a reduction in trade and information costs between participating villages and the urban areas that are already connected to e-commerce. An advantage of this setting is that we can study the reduction in trade frictions through e-commerce without confounding the counterfactual with the effects of first-time internet access more broadly. The participating villages were already connected to the internet and the program makes no changes on this front.
The RCT and data analysis that we implement exploit this empirical setting to provide evidence on the local economic effects of e-commerce trading access on rural households. We find that the program leads to sizable gains in real incomes among the group of rural households who are induced to use the e-commerce terminal. These users represent 14 percent of the rural household sample and 13 percent of the village population after adjusting for sampling weights. For the average rural household, including non-users, these gains are statistically significant but more muted. Underlying these effects, we find strong heterogeneity across households and villages. Beneficiaries are on average significantly younger, richer, live in closer proximity to the e-commerce terminal, and in villages that are relatively more remote. Conditional on these characteristics, we do not find evidence that household education or the characteristics of the terminal managers affect the extent of household gains from e-commerce.
In terms of channels, we find significantly stronger gains among villages that were not previously serviced by commercial parcel delivery, suggesting that the program’s effects are mainly due to overcoming the logistical barrier rather than the transactional one. On the consumption side, we find that the e-commerce terminals offer lower prices, higher convenience, and increased product variety compared to the pre-existing local retail environment both within the village and in nearby towns. The gains in household purchasing power are strongest for durable product groups, such as electronics and appliances. We also find suggestive evidence that the program led to additional product variety in local stores, as their managers source some new products through e-commerce. On the other hand, we find no evidence of significant pro-competitive effects on local retail prices. On the production side, we find no evidence of significant effects on the local economy: online selling activity, purchases of production inputs, household incomes and entrepreneurship are not significantly affected by the arrival of e-commerce. Overall, we find that the gains from e-commerce are driven by a significant reduction in local household cost of living, which is mainly due to the direct gains from access to the new e-commerce shopping option. These gains are in the order of a 5 percent reduction in the cost of living for retail consumption among users and a 1 percent reduction for the average household living in participating villages. For durable consumption, the estimated reduction in the local cost of living is 17 percent among users and on average 3 percent among all households.
Using the firm’s database, we find little evidence suggesting that household adjustment on the consumption side takes longer than one year: the uptake materializes within 2–4 months of entry and then remains mostly constant over time. On the production side, we find evidence that village-level out-shipments significantly increase over time beyond the 12-month survey window. However, the effect on total out-shipments remains minor after more than two years post-program entry with a small upper-bound effect on local household incomes that is consistent with the statistical zero in the survey data.
Overall, our findings suggest that e-commerce trading access offers significant economic gains to certain groups of the rural population and locations, rather than being broad-based. Compared to the recent case studies highlighting the growing number of successful rural e-commerce production hubs, our analysis reveals that such transformative effects may not be representative more generally, even when focusing on positively-selected rural markets that were targeted by the program for successful e-commerce expansion outside of cities. In the absence of complementary interventions, such as business training, access to credit, or targeted online promotions, we show that large and significant production-side effects appear unlikely to materialize for the average rural marketplace in the short- to medium-run.
These findings serve to inform the current wave of policy interest in e-commerce expansions and future research in this area. In particular, a better understanding of the factors and potential complementary interventions that enable some markets to thrive on the production-side under e-commerce integration seems a promising agenda for future work in this area.
Note 1: Following this policy interest, UNCTAD recently announced the launch of a new platform, “eTrade For All: Unlocking the Potential of E-Commerce in Developing Countries”, to provide technical assistance and funding for e-commerce expansions in the developing world (UNCTAD, 2016b).
(Victor Couture, University of California, Berkeley; Benjamin Faber, University of California, Berkeley; Yizhen Gu, Jinan University; Lizhi Liu, Stanford University.)
AliResearch. 2017. China’s Taobao Villages. Research Report. Alibaba Research. Hangzhou. China.
Couture, Victor, Benjamin Faber, Yizhen Gu, and Lizhi Liu. 2018. “E-Commerce Integration and Economic Development: Evidence from China.” NBER Working Paper 24384.
PFSweb. (2016). “China’s e-commerce market 2015.” PFSweb Annual Market Report. Allen, Texas.
Statista. 2016. “Online Retail Statistics for China.” Market Statistics. Statistica. New York.
United Nations Conference on Trade and Development (UNCTAD). 2016a. UNCTAD B2C E-commerce Index 2016. United Nations Conference on Trade and Development: Geneva.
United Nations Conference on Trade and Development (UNCTAD). 2016b. “E-commerce Opens New Opportunities for Developing Countries.” United Nations Conference on Trade and Development: Geneva.
World Trade Organization (WTO). 2013. “E-commerce in Developing Countries: Opportunities and Challenges for Small and Medium-sized Enterprises.” World Trade Organization: Geneva.