Russell Morton is a Job-Market Ph.D. Candidate at the University of Michigan. His research interests are in development economics, industrial organization, and health economics.
Click here for Russell’s job market paper Elevator Pitch.
As international supply chains generate over two-thirds of international trade, how firms split the gains from trade these chains create has important welfare implications. The World Development Report 2020 highlights concerns that surplus may be inequitably distributed “across and within countries,” with low-income country producers receiving low prices and, therefore, a small share of surplus.
Because trade-supporting institutions are limited in many low-income countries (e.g., firms might not trust courts to enforce contracts with suppliers effectively), supply chains involving low-income country firms often rely on relational contracts rather than official legal arrangements. Relational contracts are informal agreements that facilitate cooperation and prevent self-interested behavior. Cooperation creates surplus for both parties compared to transacting in the market where firms maximize their own surplus. Relational contracts facilitate cooperation by punishing self-interested behavior through ending the relationship, which blocks access to future surplus. Therefore, understanding the distributional consequences of these supply chains requires analyzing how firms bargain over surplus in relational contracts.
I propose that firms can increase their surplus share by vertically integrating. Consider an unintegrated buyer bargaining with an external supplier. The buyer evaluates contract terms relative to what other external suppliers offer. Now, suppose the buyer adds some integrated capacity, reducing quantities sourced externally.[1] Then, it evaluates contract terms relative to the cost of in-house production. Integration, therefore, changes the buyer’s threat point– its best option if negotiations with a supplier dissolve. When trade-supporting institutions make it costly and difficult to work with external suppliers, the buyer prefers sourcing in-house, meaning integration improves the buyer’s threat point. The buyer can leverage this threat point change to better its bargaining position vis-a-vis external suppliers.
In the context of relational contracts in Indian garment manufacturing, my Job Market Paper defines and quantifies the threat point effect—the price change caused by the change in threat point due to integration (i.e., the increased discount the buyer receives from relational contract suppliers post-integration). This quantification is important because of two reasons. First, the threat point effect shapes how firms share gains from trade, and little is known about surplus sharing in relational contracts. After finding that the threat point effect disproportionately reduces prices for small suppliers, I analyze policies aimed at reallocating surplus to them: increasing downstream buyer competition and enabling suppliers to insure profits against demand shocks. Second, understanding the motivations for and effects of vertical integration informs competition policy, which matters in low-income countries where markets are often fragmented and uncompetitive. Integration can be anticompetitive—integrated firms raise input prices for downstream rivals to reduce competition. Conversely, when firms integrate because market power makes input prices high, integration increases production and efficiency. If regulators ignore large threat point effects, they inaccurately evaluate how integration affects efficiency and inappropriately regulate it.
Because of challenges writing and enforcing contracts in the Indian fabric market, suppliers can charge high prices (see paper for details). The large Indian garment manufacturer (henceforth, buyer) uses relational contracts to obtain a discount relative to market prices in exchange for reducing suppliers’ quantity variability. Suppliers benefit from reduced quantity variability because it helps them control costs and stabilize profits.
I examine these relational contracts between the buyer and its fabric suppliers—a setting well-suited to analyze bargaining motivations for vertical integration for two reasons. First, the buyer adds integrated capacity—a fabric mill with more capacity than any external supplier but only ~25% of all fabric needed. Second, the buyer’s large size (~$230,000,000 of fabric purchased yearly) generates a comprehensive sample of 471 external suppliers, both relational and market. Therefore, I can analyze how contracts with external suppliers respond to integration.
Quantifying the Threat Point Effect
I build a model of the buyer and supplier bargaining over relational surplus to study the threat point effect. The model illustrates how the level of surplus and its allocation between the buyer and supplier shape the threat point effect.
The threat point effect is large when the pre-integration relational surplus level is high. This happens when suppliers highly value the reduction in quantity variability from the relational contract. Therefore, even with low relational prices, these suppliers benefit from the relationship. The buyer can then leverage its threat point change from integration to further reduce relational prices.
Conversely, the threat point effect is small when surplus is small pre-integration. Then, if the buyer switches its threat point to its integrated supplier, no surplus remains and the relational contract dissolves. However, the buyer does not want the relational contract to dissolve—it must still source some inputs externally, and dissolution requires the buyer to forfeit the pre-integration relational discount. Therefore, the buyer chooses to keep its pre-integration threat point— other suppliers—to preserve the relational contract. The threat point effect is also small when the buyer receives most surplus pre-integration. Then, the buyer cannot further improve its bargaining position post-integration, so relational prices are unaffected.
I estimate the model using rich transaction data on the universe of the buyer’s fabric purchases. Figure 1 depicts threat point effects, as a percent of pre-integration relational prices, for different levels and allocations of surplus. Surplus levels increase in supplier risk aversion, as risk-averse suppliers especially value the profit stability from the relational contract. Small firms could be risk-averse, for example, because they need consistent revenue to meet payroll obligations when underdeveloped financial markets create challenges to borrow and save. The buyer bargaining parameter defines surplus allocation—higher values allocate more surplus to the buyer.
I find that for risk-averse suppliers, the threat point effect reaches up to 6.7% of pre-integration relational prices—larger than Indian value-added tax (VAT) for apparel at the time.
I validate the model three ways, including a Difference-in-Differences (DiD) approach that directly estimates the threat point effect. I compare prices before and after the buyer adds integrated capacity, examining whether prices change differentially for suppliers who are more susceptible to displacement by in-house supply. Suppliers are more susceptible to displacement if they sell fabrics that can be produced in-house. I find that the model-implied threat point effect coincides with the DiD estimate.
Policy Implications
I assess two policies to reallocate surplus to small risk-averse suppliers. I focus on these firms for three reasons: the threat point effect disproportionately reduces prices they receive; their policy relevance as the majority of low-income country firms and employers; and the special attention they receive from both the World Bank and IMF.
First, I evaluate the effects of increased downstream buyer competition. This policy trivially impacts prices small suppliers receive—not even a 1% change. The threat point effect drives this result as omitting it leads to large overestimates of this policy’s benefit: a 67% relational price increase.
Second, I create a market for suppliers to insure profits against quantity variability. Figure 2 shows how allowing suppliers to purchase insurance affects the relational discount relative to market prices, as a percent of market prices. Once suppliers purchase insurance, the discount decreases and the threat point effect vanishes (i.e., the increased discount that integration causes disappears). Although governments might not offer insurance, policies can both improve access to financial markets so firms can self-insure and support trade associations and/or cooperatives providing revenue or price protection.
These results suggest that policies and future research should target market features beyond competition that lead small firms to provide discounts in relational contracts, such as missing insurance markets for the large levels of risk pervasive in low-income countries.
[1] Partial integration appears to be common in low-income countries. For example, 62% of integrated firms in Karnataka, India source some inputs externally.
