Sage Journals: Discover world-class research

Abstract

We study the value of information in a supply chain where the manufacturer introduces a direct‐selling channel to encroach in the retailer's market. Based on different information structures regarding the direct‐selling cost for the manufacturer, we analyze three models: incomplete information model, asymmetric information model, and full information model. By comparing the firms' profits in these models, we derive the value of the manufacturer's information acquisition and the value of sharing that information with the retailer. We find that the manufacturer's information acquisition may hurt the manufacturer but benefit the retailer simultaneously. However, we also find a win–win outcome for the manufacturer and the retailer due to the manufacturer's learning of its direct‐selling cost. Further, in expectation, the manufacturer always benefits from sharing its private cost information with the retailer, while the retailer does not. We also show that the direct‐selling channel is always valuable to the manufacturer, while the retailer's preference will depend on the information structure. Finally, we consider a supply competition model where the manufacturer competes with a spot supply source to sell to the retailer. In contrast to the case of no supply competition, the manufacturer sells less using direct selling. However, direct selling can become more valuable as it may contribute higher profit for the manufacturer due to the reduced wholesale price charged to the retailer. We find that supply competition can completely reverse the manufacturer's disincentive into incentive for information acquisition and facilitate information sharing with the retailer.

Keywords

asymmetric information manufacturer encroachment supply competition value of information

INTRODUCTION

In addition to partnering with retailers to channel products to consumers, manufacturers have the option of creating independent, direct access to the market. For example, Nike, Hewlett Packard, IBM, and Apple have built and operated their online stores for direct selling (Guan et al., 2019). With direct selling, manufacturers hope to augment the existing retail channel with additional benefits that were previously beyond reach, such as increasing sales, tightening control of flow of goods, direct marketing, and improving customer experience before and after sales.

Direct selling, however, complicates the manufacturer–retailer relationship. Retailers generally perceive manufacturers' direct selling as competition and may react aggressively in selling and contracting with manufacturers, causing reduced trust and cooperation. For example, Home Depot issued a letter to its suppliers to put forth a threat of cutting off suppliers that practiced direct selling online (Brooker, 1999). Heightened strain in the manufacturer–retailer relationship may cause excessive friction in the channel and erode any benefits of direct selling to the manufacturer. The problem is further exacerbated when different levels of information are available to the channel partners. The focus of our research is to examine these information issues and its resulting impact on the supply chain partners.

It is well‐known that manufacturers have to carefully balance between the dual roles they play with their retailers: being a supply chain partner and being a competitor for end‐market demand. The trade‐off critically hinges on manufacturers' efficiency in selling relative to that of the retailer. When the manufacturer is a new player selling direct in the market, they often experience a steep learning curve with high costs of direct selling, which puts them in a disadvantageous position in competition with retailers. In addition, direct selling often occurs via online stores, which have different cost structures from traditional brick‐and‐mortar retailers. Kapner (2014) finds that the operational fulfillment costs of online stores are generally higher than the fulfillment costs of bricks‐and‐mortar retailers and the fulfillment costs of online stores can run as high as 25% of their sales. In some cases, shipping expenses for some online orders can even exceed the price of the product (Whelan, 2016). High selling costs erode manufacturers' profit margins, compromise their competitiveness in the market, and lead to more reliance on retailers. When manufacturers' costs are moderately high, they may find it beneficial to complement the retail channel with direct selling, when “manufacturer encroachment” occurs as so perceived by the retailer. Striking an optimistic note, Arya et al. (2007) use a game‐theoretical model to show that if the manufacturer's selling cost is moderately higher than the retailer's, manufacturer encroachment may increase their profits simultaneously. In the extreme case, if manufacturers' costs are very high, they may choose not to engage with direct selling. Sony, for example, closed its retail stores in the United States in 2014 “to further streamline costs and continue focus on existing partner relations” (Welch, 2014).

Often, manufacturers closely guard their sensitive cost information for competitive reasons. Information asymmetry reduces the retailers' visibility of manufacturers' competitiveness and weakens the retailers' ability to respond. Consider, for example, a scenario where a manufacturer chooses a wholesale price for a retailer. The retailer may not be able to identify the manufacturer's motivation in price selection as it could be an attempt to undermine the retailer's competitiveness in protection of direct selling or it may be due to extracting higher profits from the retailer. As such, information asymmetry may cause unintended consequence, such as stirring up the retailer's aggressive moves in choosing its order quantity which can put extra stress on the manufacturer–retailer relationship.

We have four research objectives in this paper. First, we determine the value of information acquisition by the manufacturer. Specifically, when the manufacturer gains more accurate cost information, how does it affect profits for both players? Our second objective is to study the effect of information asymmetry regarding the manufacturer's direct‐selling cost. We analyze how information asymmetry affects the manufacturer's distribution strategy in terms of its use of direct‐selling channel versus selling through the retailer. Specifically, does information asymmetry intensify or mitigate competition between the manufacturer and the retailer? Also, would sharing its cost information with the retailer improve profits for the manufacturer and the retailer?

Our third objective is to understand the value of the direct‐selling channel in relation to information asymmetry in the supply chain. Specifically, does information asymmetry make the direct‐selling channel more or less valuable to the manufacturer and the retailer? Although previous research has examined the benefits of direct selling by the manufacturer, we contribute to the literature by studying the impact of different states of cost information (incomplete, asymmetric, and full) on the manufacturer's optimal distribution strategy and on profits for both players.

Finally, our fourth research objective is to study the effect of supply competition. We are interested in analyzing how supply competition affects the manufacturer's distribution strategy and the incentive of the manufacturer and the retailer for information acquisition. Previous research on multichannel distribution has focused on these issues with a single supplier. As such, our research contributes to the literature by studying the effect of supply competition on multichannel distribution and the value of information.

We model strategic interaction between the manufacturer and the retailer as a signaling game (i.e., the asymmetric information model). In this game, the manufacturer moves first to set the wholesale price, which sends a signal to the retailer. The retailer then updates its belief and chooses its order quantity. Finally, the manufacturer and the retailer sell to the market simultaneously. In sending the signal, the manufacturer faces a trade‐off between the need to communicate the true selling cost to the retailer and the need to remain competitive in the market. It is intuitive that lack of visibility into the manufacturer's selling cost would discourage the retailer from ordering more. To mitigate this, the manufacturer could downward‐distort the wholesale price in order to credibly signal its true cost. Conversely, the manufacturer would want to increase the wholesale price (which would lead the retailer to order less) and reduce competition for direct selling.

To answer the research questions, we analyze why it is not evident that getting better (more precise) information on its direct‐selling costs would always benefit the manufacturer. To study this, we compare the outcome of the asymmetric information model with the incomplete information model, in which both players only have a belief on the manufacturer's selling cost. Interestingly, we show that the manufacturer's information advantage may not always increase profits for the manufacturer even though it can benefit the retailer at the same time. However, there exist scenarios where better information (for the manufacturer) can lead to a win–win outcome as profits for both players can increase.

Next, we study the effect of information sharing by comparing the outcome of the asymmetric information model with the full information model in which both players have perfect information. We find that the manufacturer, in expectation, always benefits from sharing its private information with the retailer, while the retailer may be worse off. In the signaling game, the manufacturer has to downward distort its wholesale price that leads to a signaling loss. By sharing its cost information, the manufacturer avoids the loss from signaling. However, this can lead to lower profits for the retailer due to a higher wholesale price. A rather interesting implication of this finding is that information sharing may hurt the retailer (the receiver of information) but not the manufacturer (i.e., the provider of information). Further, we show that supply chain profits may increase due to information sharing, which makes it feasible for the manufacturer to share some of its benefit from information sharing with the retailer. Finally, we study the use of direct selling by the manufacturer and interestingly show that in comparison to the full information case, the use of direct selling is more likely to be a win–win for both players when the manufacturer has private information about its direct‐selling costs.

Finally, we study the effect of supply competition by extending our model to include a spot supply market as a second source for the retailer. The retailer's sourcing decision is price‐driven, and the manufacturer sets the wholesale price to compete with the spot supply. We find that supply competition discourages the manufacturer's use of direct selling, that is, it would now sell a lower quantity in the direct channel. However, interestingly, the contribution to overall profits from direct selling may be even higher because the manufacturer charges the retailer a lower wholesale price due to competition from spot supply. We also analyze the effect of supply competition on the value of information, and show that it may change the manufacturer's disincentive to acquire information of the direct‐selling cost. Further, supply competition may also change the retailer's disincentive to get information on the manufacturer's cost information and, as a result, supply competition facilitates information sharing between the manufacturer and the retailer.

The rest of this paper is organized as follows. We review the relevant literatures in Section 2. In Section 3, we describe the model setting and present the result of two benchmark models: the full information model and the incomplete information model. In Section 4, we analyze the asymmetric information model. In Sections 5 and 6, we examine the effect of the manufacturer's learning of its direct‐selling costs and information sharing with the retailer. In Section 7, we analyze the effect of supply competition. Finally, we conclude in Section 8.

LITERATURE REVIEW

Our work falls in the literature on channel management that explores strategic interaction among the channel players—manufacturers and retailers—in the existence of multiple channels. Although a stream of work in this literature focuses on understanding competition between brands when they share the same retailer (see Zheng et al., 2020 for a recent paper), our paper explores interaction between the manufacturer and the retailer, when the manufacturer uses multiple retail channels at the same time. Specifically, our work is closely related to the stream of work on manufacturer encroachment. The literature on manufacturer encroachment mainly focuses on the situation when the manufacturer actually sells through the direct‐selling channel, (e.g., see Arya et al., 2007; Cao et al., 2013; Huang et al., 2018; Li et al., 2013). One exception is Chiang et al. (2003) who use a Bertrand competition model in which the manufacturer does not explicitly use the direct‐selling channel but can achieve higher profit by using it as a threat. In their paper, the manufacturer will never sell through the direct‐selling channel in the equilibrium. We extend this stream of literature on manufacturer encroachment by considering a Cournot competition model as in Arya et al. (2007) but without restricting the manufacturer's direct‐selling strategy. With asymmetric information, the manufacturer encroachment model has been considered in Cao et al. (2013), Li et al. (2013, 2015), and Gao et al. (2021). Li et al. (2013) consider the situation when the retailer will use the order quantity to signal its private demand information. Cao et al. (2013) use nonlinear price contracts to screen the retailer's private selling cost information. Li et al. (2015) also consider nonlinear price contracts, but to screen the retailer's private demand information. Our paper differs from these papers by considering the manufacturer's private information of its direct‐selling cost.

A relevant paper to ours is Gao et al. (2021), who also consider the situation when the manufacturer can use the wholesale price to signal its direct‐selling cost. Our paper differentiates from their paper in several aspects. First, the main analysis and results of Gao et al. (2021) are based on the “dual selling” case, in which the manufacturer always sells a nonzero quantity through the direct‐selling channel. In our paper, we allow the manufacturer to optimally choose the distribution strategy—the direct‐selling quantity can be zero or positive. Our analysis and results are built upon full characterization of the optimal decision of the manufacturer. Second, the main analysis and results of Gao et al. (2021) are based on the separating equilibrium of the signaling game in the asymmetric information model. In comparison, we provide a full characterization of both separating and pooling equilibrium of the signaling game. This enables a comprehensive analysis of the value of information in our paper. Third, Gao et al. (2021) analyze the value of information sharing—information acquisition by the retailer. In comparison, we analyze both the value of information sharing and the value of information acquisition by the manufacturer. Moreover, we analyze the effect of supply competition on distribution and the value of information with manufacturer encroachment. As such, our paper complements Gao et al. (2021) and provides additional insights.

Our paper is also related to the literature on upstream information sharing. This stream of literature focuses on the situation in which the manufacturer has more information about demand than the downstream partner (e.g., see Chu, 1992; Desai, 2000; Guo & Iyer, 2010; Jiang et al., 2016; Lariviere & Padmanabhan, 1997). In Guo and Iyer (2010), the manufacturer can acquire information on the consumers' preference. They find that the manufacturer may not prefer accurate information even if information acquisition is free. Regarding the manufacturer's incentive to share information, it depends on the market parameters. In their paper, however, the retailer cannot rationally infer information from the manufacturer's wholesale price. When wholesale price is treated as a signal of the manufacturer's demand information, Jiang et al. (2016) show that the manufacturer may downward distort the wholesale price to convey a low demand forecast to the retailer. This wholesale price distortion may result in the manufacturer having lower profit but the retailer having higher profit, compared to the full information case. Other researchers (Dukes et al., 2011, 2017; He et al., 2008) also consider the situation when both the manufacturer and the retailer have private demand information. He et al. (2008) show information sharing may not always increase the channel profit. As in Guo and Iyer (2010), they assume that the retailer cannot derive any information from the manufacturer's wholesale price. When the wholesale price is used as a signal, Dukes et al. (2017) find that the manufacturer may downward distort the wholesale price to transmit a low demand forecast to the retailer, which causes information sharing may only benefit the manufacturer at the expense of the retailer and consumers. Finally, Gal‐Or et al. (2008) consider two competing retailers and show that the more informed manufacturer may only share information with the less informed retailer.

Our paper differs from the aforementioned literature in the following important ways. First, our model focuses on the manufacturer's private information about its direct‐selling cost. Second, in our model because the manufacturer can sell to the end market by direct selling, the manufacturer competes with the retailer. With dual channel, the manufacturer signal its direct‐selling cost information to the retailer through the wholesale price.

MODEL PRELIMINARIES

Model setting

We consider a supply chain with a manufacturer (supplier) and a retailer (buyer). The manufacturer sells a product through a retailer to consumers, and, in addition, the manufacturer has a direct‐selling channel for access to consumers. We assume that the products distributed via these two channels are perfect substitutes from the consumer's perspective.

The sequence of events is as follows. In stage 1, the manufacturer quotes a take‐it‐or‐leave‐it wholesale price, w, to the retailer. In stage 2, the retailer orders quantity,

q_{r}

from the manufacturer. In stage 3, the manufacturer and the retailer simultaneously decide their respective selling quantities,

q_{m}

and

{\tilde{q}}_{r}

(with

{\tilde{q}}_{r} \leq q_{r}

), and sell them to the downstream consumer market. We assume that the market‐clearing price, denoted as p, is linearly decreasing in the total quantity sold in the market:

p = α - q_{m} - {\tilde{q}}_{r}

, where α is the highest possible market price. We normalize the retailer's per‐unit selling cost

s_{r}

to be zero. The retailer's total profit is:

Π_{r} (w, q_{r}, {\tilde{q}}_{r}, q_{m}) = (α - q_{m} - {\tilde{q}}_{r}) {\tilde{q}}_{r} - w q_{r} .

1The manufacturer incurs a selling cost

s \geq 0

per unit. As such, we have

s \geq s_{r} = 0

. This inequality reflects the situation where it is more costly for the manufacturer to sell to the consumers than the retailer. This could happen, for example, when the manufacturer is a new entrant to the downstream market, is inexperienced in dealing with consumers or due to the cost of investing in e‐commerce (e.g., see The Economist, 2013; Kapner, 2014). We will discuss the implication of

s < 0

, that is, when the manufacturer has a cost advantage over the retailer, in Section 8. To further simplify the model, without loss of generality, we normalize the manufacturer's unit production cost to be zero. The manufacturer's total profit is:

Π_{m} (w, q_{r}, {\tilde{q}}_{r}, q_{m}) = (α - q_{m} - {\tilde{q}}_{r} - s) q_{m} + w q_{r} .

In this research, we consider three information states with respect to the manufacturer's direct‐selling cost s: incomplete information, asymmetric information, and full information, depending on how much information is available to both players. In the incomplete information model, neither the manufacturer nor the retailer has perfect information of the manufacturer's selling cost at the time of selling. This information structure occurs when the manufacturer is in early stage of operating the direct‐selling channel and does not have precise information on its selling costs. For example, the cost of a newly opened online channel with its associated fulfillment cost of orders may not be fully transparent until after a period of selling and data accumulation. In the asymmetric information model, the manufacturer has private information, but the retailer does not. Finally, in the full information model, both the manufacturer and the retailer have perfect information.

In these models, an uninformed firm faces uncertainty about the manufacturer's direct‐selling cost. We assume that the direct‐selling cost is either high or low. We denote the high‐ and low‐cost type of the manufacturer with the subscript

i = h

and l, respectively. The direct‐selling cost

s = s_{h}

if it is high and

s = s_{l}

if the cost is low, with

0 \leq s_{l} < s_{h} \leq α

. Without loss of generality, we further normalize

s_{l} = 0

.¹ The uninformed firm holds a prior belief about the value of s:

s = s_{h}

with probability λ, and

s = s_{l}

with probability

1 - λ

. All other information is common knowledge.

The timing of events is illustrated in Figure 1. To begin, nature reveals the direct‐selling cost to neither firm, the manufacturer only, or both firms, depending on the information state of the supply chain. Then, the manufacturer sets the wholesale price. In sequel, the retailer orders from the manufacturer. Finally, both firms sell to the market at the same time.

FIGURE 1

Timing of events

We analyze the full information model in Subsection 3.2 and the incomplete information model in Subsection 3.3, and will use them later as benchmarks for the asymmetric information model, which we analyze in Section 4.

The full information model

In this section, we analyze the model in which the manufacturer's direct‐selling cost per unit s is public information. This model setting is identical to that of Arya et al. (2007) after swapping the notation of a, c and b in Arya et al. (2007) with α, s and 1 in our model², with one important extension. Arya et al. (2007) present the results under the condition

s \leq \frac{3 α}{5}

, whereas, we consider all

s \geq 0

, and so our model is applicable over a wider range that allows us to discuss new insights.

We use backward induction to derive the sub‐game perfect Nash equilibrium for this perfect information model. The equilibrium results are presented in Proposition 1. All proofs are in EC.2.1 in the Supporting Information. Proposition 1

In the full information model, in equilibrium the wholesale price,

w^{*}

, the retailer's order quantity,

q_{r}^{*}

, the manufacturer's direct‐selling quantity,

q_{m}^{*}

, the manufacturer's and the retailer's profits,

Π_{m}^{*}

and

Π_{r}^{*}

, and the consumer surplus, Γ, are given in Table 1, in which the Consumer Surplus is computed as

Γ = \frac{{(q_{m}^{*} + q_{r}^{*})}^{2}}{2}

TABLE 1

The equilibrium result for the full information model

	$0 \leq s < \frac{3 α}{5}$	$\frac{3 α}{5} \leq s < \frac{5 α}{6}$	$\frac{5 α}{6} \leq s \leq α$
$w^{*}$	$\frac{3 α - s}{6}$	$\frac{3 s - α}{2}$	$\frac{α}{2}$
$q_{r}^{*}$	$\frac{2 s}{3}$	$α - s$	$\frac{α}{4}$
$q_{m}^{*}$	$\frac{3 α - 5 s}{6}$	0	0
$Π_{m}^{*}$	$\frac{3 α^{2} + 7 s^{2} - 6 α s}{12}$	$\frac{(3 s - α) (α - s)}{2}$	$\frac{α^{2}}{8}$
$Π_{r}^{*}$	$\frac{2 s^{2}}{9}$	$\frac{{(α - s)}^{2}}{2}$	$\frac{α^{2}}{16}$
Γ	$\frac{{(3 α - s)}^{2}}{72}$	$\frac{{(α - s)}^{2}}{2}$	$\frac{α^{2}}{32}$

Proposition 1 shows that as the manufacturer's selling cost s increases, not surprisingly, the manufacturer direct‐sells less to the market and relies more on the retailer for selling. At

s = 0

, the manufacturer is as efficient as the retailer in selling to consumers. The retailer orders

q_{r} = 0

, and the manufacturer is the solo seller in the market. When the selling cost is low with

0 < s < \frac{3 α}{5}

, as Arya et al. (2007) find, the manufacture sells directly to the market as well as via the retailer, encroaching on the retailer's market share. When the selling cost is high with

s \geq \frac{3 α}{5}

, the manufacturer direct‐sells

q_{m}^{*} = 0

, fully relying on the retailer for selling to the market. There are, however, two cases to consider.

\frac{5 α}{6} \leq s \leq α

, the full information model converges to a model without the direct‐selling channel, that is, a conventional dyadic supply chain with

w = α / 2

and

q_{r} = α / 4

. But when

\frac{3 α}{5} \leq s < \frac{5 α}{6}

, the manufacturer direct‐sells

q_{m}^{*} = 0

, and the retailer orders and sells

q_{r}^{*} = α - s

, which differs from the retailer's order quantity without the direct‐selling channel,

\frac{α}{4}

. In other words, although the manufacturer does not actually exercise the direct‐selling channel, its mere presence distorts the ordering behavior of the retailer. The presence of the direct‐selling channel creates a credible threat of competition to the retailer, and so it may respond to the wholesale price with a larger order, that is,

α - s > α / 4

s < 3 α / 4

. However, if

3 α / 4 < s < 5 α / 6

, the threat may reduce the retailer's order size. At a high direct‐selling cost, the threat from direct selling is weak. Intuitively, to forsake the manufacturer's direct selling to the end‐market, the retailer only needs to order a small quantity.

In summary, the analysis of the full information model reveals three modes of the manufacturer's use of the direct‐selling channel—“encroachment” under low s, “threat” under moderate s, and “retailer only” under high s—with one unified framework. This enables us to construct a full view of the effect of the direct‐selling cost. In contrast, “threat” is absent from Arya et al. (2007).

The incomplete information model

In the incomplete information model, neither the manufacturer nor the retailer has perfect information of the manufacturer's selling cost s, but holds the prior belief as defined in Subsection 3.1. The manufacturer and the retailer make the contracting and selling decisions based on the expectation of the cost of direct selling,

λ s_{h} + (1 - λ) s_{l} = λ s_{h}

. As we assume that the manufacturer is risk‐neutral and has unit direct‐selling costs, we can obtain the result for the incomplete information model by substituting s in the full information model in Subsection 3.2 with

λ s_{h}

. For the analysis of this model, please refer to EC.2.2 in the Supporting Information.

THE ASYMMETRIC INFORMATION MODEL

In this section, we first analyze the asymmetric information model where the manufacturer has private information of its direct‐selling cost but the retailer holds a prior belief only. The manufacturer and the retailer play a signaling game. Nature moves first to reveal the manufacturer's cost type

i = h

or l to the manufacturer, but not to the retailer. The informed manufacturer of i‐type announces the wholesale price

w_{i}

to the retailer, for

i = h

and l. Observing the wholesale price, the retailer forms a posterior belief regarding the direct‐selling cost s and chooses its order quantity from both types of the manufacturer,

q_{r | i}

for

i = h

and l. Note that the retailer makes rational inference about the true value of s from the observation of the wholesale price. Finally, the manufacturer and the retailer choose their respective selling quantities

q_{i}

and

{\tilde{q}}_{r | i}

in the market for

i = h

and l.

The essence of the signaling game is the low‐cost type manufacturer's incentive to mimic the high‐cost type. In general, the retailer tends to order more from the manufacturer with a higher selling cost, because the retailer sees less competition in the market. By mimicking the high‐cost type, the low‐cost type can earn extra profit from the supply contract while maintaining a strong direct presence in the market. Being unable to discern the true cost type of the manufacturer, the retailer may distort its order quantity which would lower profits for high‐cost type. As such, the high‐cost manufacturer has an incentive to credibly distinguish itself from the low‐cost type in contracting with the retailer. To do so, it downward‐distorts the wholesale price w to make the supply contract unattractive to the low‐cost type. This results in signaling cost for the high‐cost type. When the signaling cost is low, the high‐cost type can afford to distinguish itself from the low‐cost type; the game ends in a separating equilibrium. Otherwise, a pooling equilibrium occurs.

In general, in a signaling game, multiple equilibria may co‐exist. To eliminate equilibria with unreasonable beliefs, we apply the refinement of “undefeated” equilibrium proposed by Mailath et al. (1993). This refinement stipulates that one should retain only the equilibrium that yields the highest profit for the type of the message sender that has incentive to signal its true type. In our setting, only the high‐cost manufacturer has such incentive, and so we select the equilibrium with highest profit for it as the unique equilibrium of the signaling game.³ We present the equilibrium result in Proposition 2. Proposition 2

In the asymmetric information model, the unique “undefeated” equilibrium is as follows.

At each

(λ, s_{h}) \in S E

, there exists a unique “undefeated” separating equilibrium. Specifically, set SE is divided into three subsets: SE1, SE2, and SE3. In SE1, the high type of the manufacturer uses the “encroachment” strategy; in SE2, the high type uses the “threat” strategy; in SE3, the high type uses the “retailer only” strategy. In all SE, the low type of the manufacturer uses only direct selling, bypassing the retailer entirely. The sets SE and

S E i, i \in {1, 2, 3}

, are defined in EC.1.1 in the Supporting Information, and the detailed equilibrium results are presented in Table EC.1 in the Supporting Information thereof.

At each

(λ, s_{h}) \in P E

, there exists a unique “undefeated” pooling equilibrium. Specifically, set PE is divided into two subsets: PE1 and PE2. In PE1, the high type of the manufacturer uses the “threat” strategy; in PE2 the high type uses the “retailer only” strategy. In all PE, the low type of the manufacturer uses the “encroachment” strategy. The sets PE and

P E i, i \in {1, 2}

, are defined in EC.1.1 in the Supporting Information, and the detailed equilibrium results are presented in Table EC.2 in the Supporting Information thereof.

To aid the understanding of the results in Proposition 2, we illustrate SE1, SE2, SE3, PE1, and PE2 in Figure 2.

FIGURE 2

Signaling equilibrium in the

(λ, s_{h})

space

As shown in Figure 2, in general the signaling game has pooling equilibrium only if the high type's cost

s_{h}

and the probability of being the high type λ are both large. This can be explained as follows. Note that if

s_{l} > 0

, due to the inefficiency in selling in the downstream market, the low‐cost manufacturer always has an incentive to mimic the high‐cost type. Then, in order to signal its high‐cost information, the high‐cost manufacturer needs to downward distort its wholesale price even with small

s_{h}

, that is, in region SE1. However, when the probability of the direct‐selling cost being high is large, the uninformed retailer behaves mostly as if the manufacturer is of the high‐cost type. The high type is then less pressured to signal its true cost to reduce wholesale price concession. Therefore, with a large

s_{h}

, pooling equilibrium occurs if λ is large enough.

When the high‐type manufacturer's cost is small or medium (i.e.,

s_{h} < 3 α / 5

), the manufacturer is competitive, and the retailer orders a small quantity from the high‐type manufacturer. Anticipating a small order for the high‐cost type, the low‐cost type has no incentive to mimic the high‐cost type. Consequently, separating equilibrium occurs.

When the high‐type manufacturer's cost is moderately large (i.e.,

s_{h} \geq 3 α / 5

), the retailer orders a large quantity from the high‐type manufacturer. The low‐type manufacturer then has a strong incentive to mimic the high type. When there is a small or medium probability of the manufacturer being the high‐cost type, the uninformed retailer expects the manufacturer's direct‐selling cost to be low. Anticipating overall intense market competition against the manufacturer, the uninformed retailer would order much less than what the high type would normally expect. Therefore, the high‐type manufacturer has a strong incentive to signal its true cost type, and separating equilibrium occurs. When there is a high probability of the direct‐selling cost being high, the uninformed retailer behaves mostly as if the manufacturer is of the high‐cost type. Then, the high‐type manufacturer is less pressured to signal its true cost, and pooling equilibrium occurs.

MANUFACTURER'S ACQUISITION OF COST INFORMATION

When the manufacturer introduces direct selling as part of its distribution strategy, it may not have perfect information about its selling cost, and instead it relies on expected costs only. As the manufacturer acquires information about its selling costs, it could enable better informed decisions due to the informational advantage over the retailer. However, it is not evident whether this would necessarily benefit the manufacturer or the supply chain as a whole, depending on the retailer's response to it. To study this, we analyze the effect of the manufacturer's learning on the two firms' profits by comparing the equilibrium outcome of the incomplete information model with the asymmetric information model. The effect of the manufacturer's learning on its expected profit is summarized in Proposition 3. Proposition 3

The manufacturer's expected profit under asymmetric information may be greater or less than that under incomplete information. Specifically, only when

(λ, s_{h}) \in (S E 3 \cap {{\hat{s}}_{1}^{m} < s_{h} < {\hat{s}}_{2}^{m}}) \cup (P E \cap {{\hat{s}}_{3}^{m} < s_{h} < {\hat{s}}_{4}^{m}})

, the manufacturer's expected profit under asymmetric (private) information is strictly lower than that under incomplete information, that is, better information may hurt the manufacturer's profits. Here,

{\hat{s}}_{i}^{m}, i = 1, 2, 3, 4

are defined in EC.2.4 in the Supporting Information.

In Figure 3, we illustrate the region where the manufacturer's expected profit is greater under incomplete information than under asymmetric information. In this region, where the expected direct‐selling cost

λ s_{h}

is moderately large or large (i.e., roughly from

3 α / 5

5 α / 6

), the manufacturer is hurt by learning of its private cost information. As such, Proposition 3 suggests that while, in general, it makes sense for the manufacturer to learn its direct‐selling cost information, the manufacturer could be worse off with learning when the expected cost of direct selling is moderately large or higher.

FIGURE 3

Effect of the manufacturer's learning on the two firms' expected profit

To understand this result, we break down the effect of the manufacturer's learning on its profit by cost type. We can show that, not surprisingly, when the direct‐selling cost is low type, the manufacturer always benefits from learning.⁴ Knowing that its selling cost is as low as the retailer's, the manufacturer has the option of replacing the retailer with the direct‐selling channel and eliminate any inefficiency from double marginalization. When the direct‐selling cost is high type, however, the manufacturer may suffer from the learning when the expected cost of direct selling is moderately large or higher. By learning its cost to be high type, the manufacturer loses the power of exerting vertical control on the retailer and extracting its profit. With a sufficiently large probability of drawing the high‐cost type, in expectation, the benefit of learning for the low‐cost type manufacturer is outweighed by the loss of the high‐cost type.

We also find that the manufacturer's acquisition of cost information may or may not benefit the retailer. We present this finding in Proposition 4 and illustrate it in Figure 3. Proposition 4

The retailer's expected profit under asymmetric information may be greater or less than that under incomplete information. Specifically, only when

(λ, s_{h}) \in S E 1 \cup S E 2 \cup (S E 3 \cap ({s_{h} < {\hat{s}}_{1}^{r}} \cup {{\hat{s}}_{2}^{r} < s_{h}})) \cup (P E 2 \cap {{\hat{s}}_{3}^{r} < s_{h} < {\hat{s}}_{4}^{r}})

, the retailer's expected profit under asymmetric information is strictly greater than that under incomplete information. Here,

{\hat{s}}_{i}^{r}, i = 1, 2, 3, 4

are defined in EC.2.4 in the Supporting Information.

Interestingly, Proposition 4 shows that the retailer may also benefit from the manufacturer's learning. This result suggests that while the manufacturer's learning leads to relative information disadvantage for the retailer, such information disadvantage needs not translate into a lower profit for the retailer. From Figure 3, we further observe that the retailer benefits from learning mainly in regions where the manufacturer's learning leads to separating equilibrium (i.e., regions SE1, SE2, and left and right portions of SE3). To understand this result, we first note that if the uninformed manufacturer learns its cost type to be low (with

s_{l} = 0

) and the retailer is aware of this (via credible signaling), the manufacturer stops selling to the retailer. So, the low‐type manufacturer's learning hurts the retailer⁵, and the retailer benefits from the manufacturer's learning only if the direct‐selling cost is high type. In a separating equilibrium, the retailer benefits from the high‐type manufacturer's learning for two reasons. On one hand, the high‐type manufacturer sees a high cost of direct selling and loses some competitive power to the retailer. On the other hand, due to information asymmetry the high‐type manufacturer lowers its wholesale price to credibly signal its true cost type. With these two factors combined, the high‐type informed manufacturer's ability to extract profit from the retailer is lower than that of the uninformed manufacturer.

There are also regions where the manufacturer's learning hurts the retailer. In the middle part of the separating equilibrium region SE3 where the expected direct‐selling cost is moderately large, the retailer suffers from the manufacturer's learning. This can be explained as follows. When the manufacturer is uninformed and sees a moderately large expected direct‐selling cost, the retailer enjoys the largest expected profit as the uninformed manufacturer decreases its wholesale price to boost the retailer's order. When the manufacturer learns its direct‐selling cost to be low type (with

s_{l} = 0

), the manufacturer does not sell through the retailer leading to zero profit for the retailer.

In addition, the retailer suffers from the manufacturer's learning of its cost in region PE1 and most of region PE2, where pooling equilibrium occurs. Because the retailer cannot resolve its information disadvantage in a pooling equilibrium, the retailer is vulnerable to exploitation by the informed manufacturer. More specifically, at pooling equilibrium the retailer remains uninformed and its order quantity is independent of the manufacturer's cost type. Both types of the informed manufacturer can choose their respective selling strategy to take advantage of the retailer's under‐order or over‐order, causing the retailer's profit to suffer.

Finally, Figure 3 suggests that there are regions in which both the manufacturer's and the retailer's expected profit are greater under asymmetric information than under incomplete information, that is, the manufacturer's learning of its direct‐selling cost is a win–win for both firms at the same time. We formalize this observation in Proposition 5 and delineate the win–win regions in Figure 4. Proposition 5

When

(λ, s_{h}) \in S E 1 \cup S E 2 \cup (S E 3 \cap ({s_{h} < {\hat{s}}_{1}^{r}} \cup {{\hat{s}}_{2}^{m} < s_{h}})) \cup (P E 2 \cap {{\hat{s}}_{4}^{m} < s_{h} < {\hat{s}}_{4}^{r}})

, both the manufacturer's and the retailer's expected profits under asymmetric information are strictly greater than those under incomplete information at the same time, respectively.

FIGURE 4

Simultaneous effect of the manufacturer's learning on both firms' expected profit

Proposition 5 and Figure 4 together show that the win–win outcome occurs mostly in regions where separating equilibrium occurs under asymmetric information. Because the informed manufacturer can credibly signal its true cost, the manufacturer benefits from tailoring its wholesale price and selling quantity to its cost type. The retailer benefits from loosened vertical control by the high‐type manufacturer, and from a reduction in the wholesale price due to the high‐type manufacturer's need to send a credible signal, as noted in the discussion following Proposition 4. This leads to a win–win outcome in expectation for both players.

INFORMATION SHARING

As we have shown in Sections 4 and 5, information asymmetry distorts the manufacturer's and the retailer's actions and affects their profits. In this section, we study information sharing by the manufacturer with the retailer which brings added visibility to the supply chain. The question then is whether the two firms can benefit from more visibility. We explore the effect of information asymmetry on the manufacturer's distribution strategy and discuss the value of information sharing for the two firms. To do this, we compare the equilibrium outcome of the asymmetric information model with the full information model.

We compare the results in Propositions 1 and 2 to identify the directional changes in the manufacturer's and the retailer's selling qualities and profits caused by information asymmetry. Table 2 summarizes the directional changes in the selling quantities of the high‐ and low‐cost manufacturer,

q_{h}

and

q_{l}

, the retailer's order (and selling) quantity from them,

q_{r | h}

and

q_{r | l}

, the profits of the two types of the manufacturer,

Π_{h}

and

Π_{l}

, the retailer's ex post profits,

Π_{r | h}

and

Π_{r | l}

, and the expected profit of the manufacturer, the retailer and the channel,

E [Π_{m}]

E [Π_{r}]

, and

E [Π_{C}]

. Not surprisingly, information asymmetry has no effect in region SE1 (i.e.,

s_{h} < 3 α / 5

; see Figure 2), because the low‐cost manufacturer has no incentive to mimic the high‐cost type. We focus on the other four regions in the discussion.

TABLE 2

The directional changes in the manufacturer's and the retailer's selling quantities and profits caused by information asymmetry

						PE2
		SE1	SE2	SE3	PE1	$s_{h} < \frac{5 α}{6}$	$s_{h} \geq \frac{5 α}{6}$
Manufacturer	$q_{h}$	‐	‐	‐	‐	‐	‐
	$q_{l}$	‐	‐	‐	↓	↓	↓
Retailer	$q_{r \| h}$	‐	‐	↑	‐	↑	‐
	$q_{r \| l}$	‐	‐	‐	↑	↑	↑
Manufacturer	$Π_{h}$	‐	↓	↓	↓	↓	↓
	$Π_{l}$	‐	‐	‐	↑	↑	↑
Retailer	$Π_{r \| h}$	‐	↑	↑	↑	↑	↑
	$Π_{r \| l}$	‐	‐	‐	↓	↓	↓
	$E [Π_{m}]$	‐	↓	↓	↓	↓	↓
	$E [Π_{r}]$	‐	↑	↑	↑	↑	↑
	$E [Π_{C}]$	‐	‐	↑	↓	↑	↓

Note: “‐” indicates no change, “↑” indicates an increase, and “↓” indicates a decrease.

Effect of information asymmetry on distribution strategy

We first study how information asymmetry affects the manufacturer's distribution strategy by examining the directional changes in the two firms' selling quantities (that is,

q_{h}

and

q_{l}

for the manufacturer, and

q_{r | i}

i \in {h, l}

, for the retailer). From Table 2, we note that overall information asymmetry tends to discourage the manufacturer from direct selling and encourages more selling via the retailer. To understand this result, we consider the two cases in which direct selling is high‐cost and low‐cost, respectively.

When direct selling is of the high‐cost type, information asymmetry has no effect on the manufacturer's direct‐selling quantity

q_{h}

and pushes up the retailer's order quantity

q_{r | h}

in regions SE3 and PE2

\cap {s_{h} < \frac{5 α}{6}}

. In these regions, because direct selling is costly (i.e.,

s_{h} \geq 5 α / 7

), the manufacturer does not sell direct under both full and asymmetric information. The explanations for the effect of information asymmetry on the retailer's order quantity differ in these two regions. In region SE3, separating equilibrium occurs under asymmetric information. To credibly assure the retailer that its direct selling is not of low‐cost type, the high‐cost manufacturer downward distorts its wholesale price, inducing a larger order from the retailer. In region PE2

\cap {s_{h} < \frac{5 α}{6}}

, because direct selling is moderately costly, the direct channel imposes a “threat” under full information, which depresses the retailer's incentive to order. Under asymmetric information, pooling equilibrium occurs, and the retailer orders according to the expected direct‐selling cost. Note that the expected cost is smaller than the high type's cost. Facing a smaller direct cost in expectation, the retailer perceives stronger threat and competition so it orders more to deter the manufacturer's incentive in selling via the direct channel. As such, the retailer orders more under asymmetric information than under full information.

When direct selling is of the low‐cost type, information asymmetry causes the manufacturer to cut back on its direct‐selling quantity in regions PE1 and PE2 in exchange for a larger order quantity of the retailer. Under full information, the retailer stops ordering from the manufacturer with low direct‐selling cost, leaving the manufacturer to be a monopolist in the market. However, under asymmetric information, pooling equilibrium occurs in regions PE1 and PE2, and the retailer orders from the manufacturer based on the expected costs regardless of the manufacturer's cost type. Facing competition, the low‐cost manufacturer sells less than a monopolist. As such, information asymmetry causes the manufacturer to cede some market share to the retailer.

Effect of information sharing on profits

The difference in a firm's profit in the full information model versus the asymmetric information model gives us the value of information for the firm and its incentive for information sharing. Table 2 contains the directional changes in the profits due to information asymmetry versus full information. A decrease in profit caused by information asymmetry indicates an opportunity of profit improvement by information sharing, and vice versa.

To shed light on the manufacturer's and the retailer's incentive for information sharing, we explain the effect of information sharing on the two firms' profits

Π_{m}

and

Π_{r | i}

with

i \in {h, l}

for the high‐ and low‐cost types. When the manufacturer's direct‐selling cost is of high type, the manufacturer's profit is higher under full information than under asymmetric information and the retailer's profit is lower under full information. In other words, information sharing benefits the manufacturer but not the retailer. To understand this, we note that under asymmetric information, the high‐type manufacturer bears the brunt of signaling cost, as the low‐cost type has an incentive to mimic the high‐cost type. The high‐type manufacturer has to then lower the wholesale price that lowers its profits (in comparison with the full information case).

When the manufacturer's direct‐selling cost is of low type, we observe the opposite effect—the manufacturer's profit is lower under full information than under asymmetric information, and the retailer's profit is larger under full information. In other words, information sharing benefits the retailer but not the manufacturer. Note that this result occurs only in regions PE1 and PE2, in which pooling equilibrium occurs under asymmetric information. In the pooling equilibrium, the retailer cannot distinguish between the cost type of the manufacturer and orders in expectation. The manufacturer with low direct‐selling cost enjoys higher profit from the retailer's supply contract under asymmetric information, which is forgone under full information.

In short, when the direct‐selling cost is high type, only the manufacturer has an incentive to share information resulting in a loss for the retailer. Conversely, when the direct‐selling cost is low type, only the retailer has an incentive to get information. This conflict of interest may possibly be overcome with an ex ante commitment by both firms to information sharing. To understand the two firms' incentive for such commitment, we analyze the effect of information sharing on their expected profits. Table 2 shows that information sharing increases the manufacturer's expected profit

E [Π_{m}]

and decreases the retailer's expected profit

E [Π_{r}]

. In other words, better information for the retailer hurts itself and is beneficial to the manufacturer—the owner of private information. As such, the retailer would be reluctant to accept the manufacturer's offer to share information. This is in contrast with conventional wisdom on information sharing in the supply chain—the owner of private information is typically worse off by sharing information, and the receiver is typically better off from more information. As such, our finding contributes a possible explanation for why information sharing may not be observed in distribution channels.

The analysis above suggests that without additional arrangement, the manufacturer and the retailer may not form a consensus on committing to information sharing, because the retailer could be worse off. Such an arrangement can take the form of an enforceable transfer payment between the manufacturer and the retailer, for example. Table 2 shows that information sharing increases total channel profits

E [Π_{C}]

in regions PE1 and PE2

\cap {s_{h} \geq \frac{5 α}{6}}

. Hence, in these regions, it is feasible for the manufacturer to ex ante commit to sharing information and compensate the retailer for its reduced profit.

Value of the direct‐selling channel

Finally, we discuss the value of the direct‐selling channel when perfect information on selling cost may not be available to the players. In the full information model, Arya et al. (2007) highlight the possibility of a win–win outcome when the manufacturer encroaches into the retailer's market under certain conditions. We now explore if a win–win outcome can still occur under asymmetric information. We measure the value of the direct‐selling channel for the firms as the difference in their expected profit with and without the direct‐selling channel. The results are summarized in Proposition 6. Proposition 6

Under asymmetric information, the manufacturer's value of the direct‐selling channel is positive for all

(λ, s_{h})

. The retailer's value may, however, be positive or negative. Specifically, only when

(λ, s_{h}) \in ((S E 1 \cup S E 2) \cap {{\tilde{s}}_{1} < s_{h}, λ > \frac{3}{4}}) \cup (S E 3 \cap {\frac{49}{64} < λ}) \cup (P E 1 \cap {s_{h} < {\tilde{s}}_{2}})

, the retailer's value is strictly positive. Here,

{\tilde{s}}_{1}

and

{\tilde{s}}_{2}

are defined in EC.2.4 in the Supporting Information. Moreover, the set for the retailer to have positive value is greater than that under full information.

It follows from Proposition 6 that the creation of the direct‐selling channel is a win–win for both firms under asymmetric information whenever the retailer's value is positive. More interestingly, under asymmetric information the direct‐selling channel is more likely to be a win–win outcome than under full information. We illustrate, respectively, the set of

(λ, s_{h})

that is win–win under asymmetric information and full information in Figure 5.⁶ To understand this result, recall that as we have shown in Table 2, information asymmetry benefits the retailer in expectation for all

(λ, s_{h})

. Thus, the value of the direct‐selling channel for the retailer under asymmetric information is greater than that under full information. As such, there exists a broader range of situations under asymmetric information where the direct‐selling channel benefits the retailer and is win–win than under full information.

FIGURE 5

The regions in which the direct‐selling channel is win–win under asymmetric information and full information

SUPPLY COMPETITION

The manufacturer faces supply competition when the retailer has the option to order from multiple sources. In this section, we explore how supply competition affects the manufacturer's channel and selling decisions and the firms' incentive of learning and information sharing.

To model supply competition, we focus on the outcome of supply competition—downward distortion of the manufacturer's wholesale price, but not strategic interactions between the competitors that shape the wholesale price decision. Specifically, we model a scenario in which the retailer can source from a spot supply market as well as the manufacturer. We assume that the products from the manufacturer and the spot market are perfect substitutes in the end market, and so the retailer's sourcing decision is purely price‐driven. We assume the unit price of the spot supply is constant, denoted by

w_{0} > 0

. Under

w > w_{0}

the retailer sources from the spot market only, and the manufacturer makes zero profit. To win the supply contract, the manufacturer must set

w \leq w_{0}

. Note that the manufacturer's production cost is normalized to be

c = 0

, so setting w to be

c \leq w \leq w_{0}

is feasible. The manufacturer can always compete with the spot supply by setting

w \leq w_{0}

and earn a positive profit from supplying the retailer. In other words, in our model, the presence of the spot market caps the manufacturer's wholesale price decision. To obtain the model with supply competition, it suffices to append the constraint of

w \leq w_{0}

to the manufacturer's wholesale price decision in the original model described in Section 3.

We fully characterize the equilibrium of the model with supply competition for all

w_{0} > 0

under all three information structures. When the spot price w ₀ is high, intuitively, the spot market presents no concrete threat to the manufacturer, and the model with supply competition degenerates to the model without supply competition in Section 3. Supply competition materializes only if the spot price is not excessively large:

w_{0} < \frac{3}{4} α

. We delegate the analysis to EC.4 in the Supporting Information and the equilibrium outcomes to EC.1.2 in the Supporting Information and discuss their managerial implications in this section, with focus on

w_{0} < \frac{3}{4} α

Effect of supply competition on channel strategy

In this section, we examine how supply competition affects the manufacturer's channel decision. Generally speaking, the retailer's access to a second source would reduce the retailer's propensity of sourcing from the manufacturer. For this reason, one may expect that the presence of the spot supply market pushes the manufacturer to be more aggressive with direct selling to the end‐market. We compare the equilibrium results of the models with and without supply competition under all three information structures, and surprisingly show that while the manufacturer sells a lower quantity through the direct channel, the contribution of direct selling to the manufacturer's total profit may be even higher. This is due to the need to reduce the wholesale price to sell to the retailer (due to competition from the spot supply), which in turn could lead to higher profit contribution from direct selling. We present these findings in Propositions 7 and 8. Proposition 7

Under all three information structures, the manufacturer direct‐sells a smaller quantity under supply competition than under no supply competition.

Proposition 7 shows that supply competition discourages the manufacturer from direct selling. To explain this result, note that the spot price caps the manufacturer's wholesale price. This enables the retailer to order more from the manufacturer, which reduces the incentive for the manufacturer to use direct selling.

Reduced direct selling due to supply competition suggests that the manufacturer becomes more reliant on, and less competitive toward, the retailer for reaching the end market. In other words, supply competition mitigates demand competition between the retailer and the manufacturer. The manufacturer's use of direct selling in the end‐market promotes the retailer to source from the spot market, which can push the manufacturer out of the end‐market. To preempt this, the manufacturer appeases the retailer by imposing self‐restraint on direct selling.

The preceding discussion may suggest that supply competition diminishes the value of the direct‐selling channel for the manufacturer. We evaluate this conjecture with the full information model⁷ and summarize the result in Proposition 8. Proposition 8

Under full information, there exists a nonempty set of

(w_{0}, s)

, such that

w (s) < w_{0} < α / 2

and

0 < s < 2 α / 3

, at which the value of the direct‐selling channel for the manufacturer is greater under supply competition than without supply competition.⁸

Proposition 8 shows that when the direct‐selling cost is low, supply competition may enhance the manufacturer's incentive to create a direct‐selling channel, even if supply competition discourages the manufacturer's use of direct selling. With an efficient direct‐selling channel, the manufacturer encroaches into the retailer's market share. Although the manufacturer enjoys an extra stream of profit from direct selling, it comes with a downside—the manufacturer has to set a high wholesale price to restrict the retailer's order quantity. Under supply competition, the spot price automatically caps the manufacturer's wholesale price and thus mitigates the downside of direct selling. As a result, the net benefit of direct selling may become larger. In other words, profit contribution from direct selling may become more valuable for the manufacturer due to the reduced wholesale price in selling to the retailer (due to competition from spot supply).

Effect of supply competition on the value of information

In this section, we explore how supply competition affects the manufacturer's and the retailer's incentive to acquire information of the manufacturer's direct‐selling cost. We use the equilibrium outcome in EC.1.2 in the Supporting Information to compute the expected benefit of information acquisition for the manufacturer and the retailer under supply competition and compare the benefit with that under no supply competition.

The manufacturer's information acquisition

We first analyze how supply competition affects the manufacturer's incentive of information acquisition. Recall that Proposition 3 shows that in the absence of supply competition, the manufacturer's acquisition of private information of direct‐selling cost may increase or, interestingly, decrease its expected profit. In Proposition 7, we show that supply competition reduces the manufacturer's use of direct selling. From this, it appears to be reasonable to infer that supply competition discourages the manufacturer's incentive of information acquisition and exacerbates the manufacturer's disincentive for information acquisition. To investigate this conjecture, we compare the expected benefit for the manufacturer from its information acquisition under supply competition with that under no supply competition and summarize the finding as follows: Proposition 9

Supply competition may increase or decrease the manufacturer's expected benefit from learning its direct‐selling cost. Moreover, when

(λ, s_{h}) \in S E 3 \cap {{\tilde{s}}_{2}^{m} < s_{h} < {\hat{s}}_{2}^{m}, \frac{3 α}{4 λ} \leq s_{h} < \frac{5 α}{6 λ}}

and

\frac{α^{2}}{8 (α - λ s_{h})} \leq w_{0} < \frac{3 λ s_{h} - α}{2}

, the sign of the benefit goes from negative under no supply competition to positive under supply competition. Here,

{\tilde{s}}_{2}^{m}

and

{\hat{s}}_{2}^{m}

are defined in EC.2.4 in the Supporting Information.

Proposition 9 shows that supply competition may strengthen or weaken the manufacturer's incentive to learn its direct‐selling cost. Interestingly, supply competition can completely reverse the manufacturer's disincentive into incentive for information acquisition. As such, supply competition can reinforce the manufacturer's incentive to acquire its direct‐selling cost information, even though supply competition discourages the manufacturer's use of direct selling.

To understand the incentive‐strengthening effect of supply competition, note that the manufacturer's information acquisition causes information asymmetry and signaling—the high‐cost type downward distorts its wholesale price to prevent the low type from mimicking. Supply competition with a reasonably small spot price w ₀ alleviates the signaling burden on the high‐cost type manufacturer, because w ₀ caps the high type's wholesale price and restricts the region that the low type can exploit to mimic the high type. When there is a large probability that the high‐cost type will be burdened with a large signaling cost (i.e., with large λ and

s_{h}

), supply competition is most effective in alleviating the signaling burden and thus strengthening the manufacturer's incentive to learn.

Information sharing

We next analyze how supply competition affects information sharing between the two firms. In Subsection 6.2, we discovered that without supply competition the retailer does not have incentive to share the manufacturer's private information of its direct‐selling cost; so, information sharing is infeasible without additional arrangement. It is unclear a priori whether or not this conclusion remains valid if the manufacturer faces supply competition. To answer this question, we compare the benefit of information sharing for the manufacturer and the retailer with versus without supply competition. We summarize the finding in Proposition 10: Proposition 10

When (1)

s_{h} < \frac{α}{2}

, (2)

w_{0} < \min {w_{b}, \frac{α}{6}}

, and (3) a pooling equilibrium is selected, the benefit of information sharing for the retailer may be positive.

w_{b}

is defined in EC.4 in the Supporting Information. Moreover, the benefit of information sharing for the retailer and the manufacturer may be positive at the same time.

We find that supply competition can reverse the retailer's disincentive to learn the manufacturer's private cost information. Better yet, supply competition may enable information sharing between the manufacturer and the retailer without the aid of other arrangements such as a transfer payment, and so on. These findings are in contrast to the findings in the absence of supply competition in Subsection 6.2. As such, our finding contributes a new perspective on facilitating information sharing in distribution channels.

We now explain why the retailer can benefit from information sharing under supply competition. Recall that without supply competition, information sharing frees the manufacturer from the signaling burden of downward price distortion and affords the manufacturer freedom of exploiting the retailer by charging a high wholesale price. So, the retailer's profit can suffer from information sharing. With supply competition, the spot market price caps the manufacturer's wholesale price and limits the manufacturer's ability to exploit the retailer. This protects the retailer's profit margin and enables the retailer to earn more profit than without supply competition.

In summary, we find that supply competition mitigates the wholesale price distortion for the retailer. As a result, supply competition may strength the manufacturer's incentive for creating the direct‐selling channel. Moreover, supply competition relieves the manufacturer's signaling burden, and, as a result, it may strengthen the manufacturer's incentive for acquiring private information and facilitate information sharing between the players.

CONCLUSION

In this paper, we study a distribution channel where the manufacturer introduces a direct‐selling channel to encroach into the retailer's market. The manufacturer's cost of direct selling may be known (or unknown) to both players, or may be private information for the manufacturer. As such, a number of different informational structures are analyzed, and a number of interesting results are derived. First, we show that the manufacturer may prefer not to have private information on its selling cost, as the resulting profit loss due to a downward distortion in the wholesale price may make it worse off. However, we show that the manufacturer always benefits from sharing its cost information with the retailer, who may interestingly be worse off with more information. Finally, we show that the region where the bright side of encroachment, that is, where both players benefit from direct selling by the manufacturer, is enlarged when there is information asymmetry between the players in comparison to the full information case.

We also extend the models to include supply competition. We assume that the retailer can also source from a spot supply market for a perfectly substitutable product. As such, the manufacturer faces competition with the spot supply. Comparing this model with the standard single‐supplier model, we analyze the effect of supply competition on the manufacturer's distribution strategy and on the value of information. Our analysis there yields a number of interesting findings. First, we find that supply competition may make the direct‐selling channel even more valuable for the manufacturer, even though supply competition always discourages the manufacturer's use of direct selling. Second, supply competition may fully reverse the manufacturer's disincentive into incentive to acquire information. Finally, supply competition facilitates information sharing between the two firms.

The paper can be extended in several directions. One may consider variant information structures in the supply chain. In general, both supply chain players may have private information of different variables of the supply chain at the same time. For example, the retailer may have private information of the market demand potential and the manufacturer may have information about its direct‐selling cost. Depending on which player makes the first move and has more contracting power, the issue of double signaling or informed principle may arise. It remains to be explored how an additional layer of information asymmetry may alter the findings in this paper.

The paper can also be extended by considering a more complicated supply chain structure. The issue of common retailer arises when a retailer sells substitutable products of multiple manufacturers. In such setting, a manufacturer's direct selling not only introduces competition with the retailer, but also aggravates its competition with the other manufacturers. It is interesting to study how direct selling of a manufacturer reshapes the other manufacturers' distribution decisions and their interaction with the retailer.

In our model, we assumed that

s \geq 0

. One may extend the model by setting

s < 0

, that is, the manufacturer has cost advantage over the retailer in direct selling. When there is no supply competition, the lack of double marginalization and cost advantage would obviate the need to sell through the retailer, and the manufacturer would only rely direct selling. Without supply competition, indeed, the retailer becomes irrelevant and the model degenerates to the manufacturer's centralized decision‐making problem. However, with supply competition, the manufacturer would have to engage with the retailer even if it is more costly than the direct‐selling channel, for the purpose of preempting the spot supply from reaching the end market via the retailer. As such, the manufacturer's distribution strategy is driven by eliminating supply competition and not purely by cost alone. When there is supply competition and asymmetric information, an interesting question arises—how does the manufacturer's signaling burden interfere with its need to engage with the retailer for countering supply competition? We reserve this extension for future research.

Footnotes

ACKNOWLEDGMENTS

We thank the Department Editor Asoo Vakharia, Senior Editor, and two anonymous reviewers for their constructive comments and suggestions that have improved the paper. Huiqi Guan's research is supported in part by the National Natural Science Foundation of China (Grant Numbers: 72102048, 72131004, and 71971065). Haresh Gurnani's research is partially supported by the Thomas H. Davis Endowed Chair at Wake Forest University.

1

The analysis can be extended to the case when the low‐cost manufacturer's cost is greater than zero. Details are in the EC.3 in the Supporting Information.

2

Arya et al. (2007) assume that the retailer will sell all supply to the market, which in our notation is

q ∼ r ≡ q r $\tilde{q}_r \equiv q_r$

. Our model allows the retailer to withhold some supply from selling to the market, that is,

q ∼ r ≤ q r $\tilde{q}_r \le q_r$

. We prove that in equilibrium the retailer does sell

q ∼ r = q r $\tilde{q}_r = q_r$

. However, buyer withholding can happen in practice and we refer to Yang et al. (

) for examples.

3

In case of a tie between two equilibria, we select the equilibrium with the highest profit for the low‐cost type.

4

The details of the proof are in EC.2.4 in the Supporting Information.

5

The details of the proof are in EC.2.4 in the Supporting Information.

6

The value of direct selling for the manufacturer is computed without accounting for the fixed costs of creating the direct‐selling channel. Strictly positive fixed costs will reduce the value of direct selling for the manufacturer and negatively affect its decision to open the direct‐selling channel. As a result, the win–win region will reduce or may even vanish. However, as long as the fixed costs are not prohibitively high, the win–win regions persist.

7

The result thereof can be qualitatively extended to the asymmetric information and incomplete information models.

8

w ( s ) $w(s)$

is a piece‐wise function of s.

ORCID iD

Haresh Gurnani

References

Arya

Mittendorf

Sappington

D. E.

(2007). The bright side of supplier encroachment. Marketing Science, 26(5), 651–659.

Brooker

(1999). E‐rivals seem to have home depot awfully nervous. Fortune, 140(4), 28–29.

Cao

Wan

Lai

(2013). Contracting with asymmetric cost information in a dual‐channel supply chain. Operations Research Letters, 41(4), 410–414.

Chiang

W. K.

Chhajed

Hess

J. D.

(2003). Direct marketing, indirect profits: A strategic analysis of dual‐channel supply‐chain design. Management Science, 49(1), 1–20.

Chu

(1992). Demand signalling and screening in channels of distribution. Marketing Science, 11(4), 327–347.

Desai

P. S.

(2000). Multiple messages to retain retailers: Signaling new product demand. Marketing Science, 19(4), 381–389.

Dukes

Gal‐Or

Geylani

(2011). Who benefits from bilateral information exchange in a retail channel? Economics Letters, 112(2), 210–212.

Dukes

Gal‐Or

Geylani

(2017). Bilateral information sharing and pricing incentives in a retail channel. In Ha

A. Y.

Tang

C. S.

(Eds.), Handbook of information exchange in supply chain management (pp. 343–367). Springer.

Gal‐Or

Geylani

Dukes

A. J.

(2008). Information sharing in a channel with partially informed retailers. Marketing Science, 27(4), 642–658.

10.

Gao

Guo

Orsdemir

(2021). Dual‐channel distribution: The case for cost information asymmetry. Production and Operations Management, 30(2), 494–521.

11.

Guan

Gurnani

Geng

Luo

(2019). Strategic inventory and supplier encroachment. Manufacturing & Service Operations Management, 21(3), 536–555.

12.

Guo

Iyer

(2010). Information acquisition and sharing in a vertical relationship. Marketing Science, 29(3), 483–506.

13.

Marklund

Vossen

(2008). Research note—Vertical information sharing in a volatile market. Marketing Science, 27(3), 513–530.

14.

Huang

Guan

Chen

(2018). Retailer information sharing with supplier encroachment. Production and Operations Management, 27(6), 1133–1147.

15.

Jiang

Tian

Zhang

(2016). To share or not to share: Demand forecast sharing in a distribution channel. Marketing Science, 35(5), 800–809.

16.

Kapner, S. (2014, December 1). How the web drags on some retailers. Wall Street Journal.

17.

Lariviere

M. A.

Padmanabhan

(1997). Slotting allowances and new product introductions. Marketing Science, 16(2), 112–128.

18.

Gilbert

S. M.

Lai

(2013). Supplier encroachment under asymmetric information. Management Science, 60(2), 449–462.

19.

Gilbert

S. M.

Lai

(2015). Supplier encroachment as an enhancement or a hindrance to nonlinear pricing. Production and Operations Management, 24(1), 89–109.

20.

Mailath

G. J.

Okuno‐Fujiwara

Postlewaite

(1993). Belief‐based refinements in signalling games. Journal of Economic Theory, 60(2), 241–276.

21.

The Economist. (2013). The emporium strikes back. http://strategy.sjsu.edu/www.stable/B290/reading/Dixons/The%20emporium%20strikes%20back%20‐%20Economist%202013.pdf (2022, September 1)

22.

Welch, C. (2014). Sony to close majority of US retail stores by year's end. https://www.theverge.com/2014/2/26/5451012/sony‐closing‐majority‐of‐us‐retail‐stores‐by‐end‐2014

23.

Whelan, R. (2016, March 28). High cost of keeping chocolate cool tempers online sales. Wall Street Journal.

24.

Yang

Gurnani

Guan

(2018). Multichannel distribution strategy: Selling to a competing buyer with limited supplier capacity. Management Science, 64(5), 2199–2218.

25.

Zheng

Pan

X. A.

Vakharia

A. J.

(2020). Common retailer channel revisited: The role of supply network size. Production and Operations Management, 29(9), 2175–2181.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.09 MB

Value of information with manufacturer encroachment

Abstract

Keywords

INTRODUCTION

LITERATURE REVIEW

MODEL PRELIMINARIES

Model setting

The full information model

The incomplete information model

THE ASYMMETRIC INFORMATION MODEL

MANUFACTURER'S ACQUISITION OF COST INFORMATION

INFORMATION SHARING

Effect of information asymmetry on distribution strategy

Effect of information sharing on profits

Value of the direct‐selling channel

SUPPLY COMPETITION

Effect of supply competition on channel strategy

Effect of supply competition on the value of information

The manufacturer's information acquisition

Information sharing

CONCLUSION

Footnotes

ACKNOWLEDGMENTS

1

2

3

4

5

6

7

8

ORCID iD

References

Supplementary Material