**2. Literature Review**

Our research is connected to two streams of the operations management literature: (1) circular operations management and (2) operational flexibility. One of the fundamental problems in the circular operations management literature is how to transform the linear "take-make-dispose" operational model to a circular structure, so that products can stay in the market after their lifetime to minimize waste on the consumption side [13,14]. The phenomenon of circular operations management is also known as closed-loop supply chain management (CLSC). There are three different layers of CLSC, which aim to minimize product waste on the consumption side. We depict these layers in Figure 1.

**Figure 1.** Closing the loop in supply chains.

The first layer of the CLSC is *reusing*, which focuses on strategies to extend the consumption length of products [14–16]. If a product is damaged or loses its functionality, it must be repaired to increase the length of the consumption period. When a customer loses interest in using a product, it must be sold in the secondary market or shared with other people. Therefore, the ease of repairing, resharing, and selling in secondary markets are key elements of the first layer [14,16]. One of the successful applications of reuse is the online marketplace of Patagonia, an outdoor apparel firm, where customers can exchange their clothes when they lose interest in them [17]. Another example is the product ownership program of Xerox whereby the printing company retains ownership of the printers and leases them to customers [14]. When a customer terminates its contract, Xerox leases the product to a new customer, so the company's products are shared over their lifetime.

The second layer of the CLSC is *remanufacturing*, whereby a set of refurbished and new components are used to manufacture products [18–20]. There are two main challenges regarding the implementation of remanufacturing. The first is uncertainty about the flow of used components, which will later be refurbished for use in the manufacturing process. The second challenge is the cannibalization of the original items because introducing the remanufactured products to the market would result in lower sales of the original ones, leading to lower profits. Ref. [19] address the first challenge by developing a queuingtheory model that dynamically estimates the flow of used products and then applying an aggregate base-stock policy to optimize the inventory policy. To address the second challenge, [20] develop a diffusion model and categorize products depending on their market diffusion and purchase frequency. The authors outline a decision typology that shows the product categories with the maximum potential for remanufacturing.

The last layer of the CLSC is *recycling*, whereby products at the end of their life go through a series of operations to manufacture new items. Well-known examples of recycling are paper and plastic recycling, which are observed in the recycling centers of municipalities of big cities. Yet, the most important challenge of recycling remains the collection of products from households. Ref. [14] give the example of Norway, where

the recycling rate for plastic bottles is impressively high—97%. Norwegians achieve this by providing government funds to support retail stores in collecting plastics via reverse vending machines (the same system can be observed in other western European countries such as the Netherlands). Another approach to increasing the recycling rate is to mandate manufacturers to develop collection and recycling mechanisms for their products, which is popularly known as extended producers responsibility (EPR) [21]. EPR has been popularized in the electronics industry, with an example being the Minnesota Electronics Recycling Act [22]. According to this act, the state of Minnesota imposes strict collection and recycling targets on producers as a percentage of their total sales volume [22].

Studies in the extant literature successfully address the most important problems related to improving sustainability on the consumer side. Once a product reaches the market, keeping it in the loop of the CLSC has certain environmental benefits. However, the extant literature does not quantify the environmental impact of overproduction nor develop remedies for that problem. We contribute to the literature by filling this gap.

Our research is also related to a second stream of literature that prices the value of operational flexibility. Companies establish operational flexibility in different ways, such as lead-time reduction [2,8], quantity-flexibility contracts [10], and multiple sourcing [9,23]. These operational-flexibility strategies make it possible for buyers to determine order quantities after the partial or full resolution of demand uncertainty, helping them to better match supply with uncertain demand. One of the challenges in the extant literature is related to demand modeling because the demand model should involve the time element in order to quantify the benefits of delaying the ordering decision. In practice, manufacturers often employ demand planning teams that collect credible information from customers and update demand forecasts over time. Thus, demand forecasts are improved over time as a result of such efforts. For this reason, the modeling approaches used in the operationalflexibility literature incorporate the evolutionary dynamics of demand forecasts in order to price the value of operational flexibility. Ref. [8] use a multiplicative demand process to price the value of lead-time reduction. Ref. [2] later extend the multiplicative demand model by incorporating sudden changes in the demand forecasts and show that the value of lead-time reduction increases with positive jumps in the demand forecasts. Ref. [10] use a multiplicative demand model to price the value of quantity flexibility and show that the value is jointly affected by the order-adjustment flexibility and the time when the order adjustments are made. Ref. [23] develop a tailored capacity model, which is analogous to the multiple-sourcing model that we consider in this research. In their model, a buyer utilizes a speculative capacity under demand uncertainty, but also reserves a reactive capacity that can be utilized once the demand is known. Ref. [9] extends [23] by using the extreme-value theory, so the tailored capacity model can be applied to a wider selection of product categories.

Our contribution to the operational-flexibility literature is that we quantify the environmental benefits of operational flexibility that appear in the form of reduced product waste during the sourcing process. The studies in the extant literature are based on a profit-oriented view of the firm such that a reduction in the supply–demand mismatch costs determines the value of operational flexibility. We hopefully expect that companies will be less concerned about increasing their profits in the future, focusing rather on understanding the environmental impact of their operations. Our research aims to fill this gap in the literature by showing how operational flexibility can help minimize product waste during the sourcing process.
