Procurement Contracts Abstract

In this chapter we discuss supply contracts. We consider environments in which some type of uncertainty plays an important role. In general the main objective of supply contracts is to coordinate the supply chain and: 1. Ensure that buyers are able to meet their customers uncertain and changing demand. 2. roviding mechanisms so that suppliers !manufacturers" and buyer share systems ris#s and costs. $ore specifically ensure that suppliers dont assume the entire ris# of buffering raw materials and finished goods to meet the final demand for goods. %. rovide incentives for suppliers !manufacturers" to build enough capacity. We discuss several different supply contracts. &ll contracts balance between the fle'ibility of the buyer adapting to changing mar#et conditions and the stability of the production environment. (hese contracts differ in their assumptions and comple'ities. (he main differences between these supply contracts are that some are single period contracts the )uy )ac# *ontract is a good e'ample. +thers are multi,period contracts and they capture the dynamics of a long hori-on. (he eriodical *ommitments and .le'ibility contract is a good e'ample for the multi,period models. (he different contracts consider different mechanism to balance fle'ibility and stability: .i'ed delivery contracts ensure that the sellers deliver a fi'ed /uantity every period while the option contracts use options to achieve the same objective. (he eriodical *ommitments and .le'ibility contract provide buyers with limited fle'ibility to change previously made commitments. 1. Introduction.

(he problem of securing a supply of raw materials0 components0 subassemblies and finished goods at a stable and #nown price is not a new one. It is one of the main difficulties of every industrial !and not only industrial" organi-ation. (his problem is becoming even more acute due to the common practice of outsourcing and decentrali-ation. In a decentrali-ed environment0 independent entities0 each with its own plans and objectives0 are purchasing raw materials0 components0 subassemblies and finished products from each other. *learly0 each entity is interested in ensuring that it obtains the right /uantity of inputs0 of the right /uality0 at the right time0 while minimi-ing cost. *oordinating the supply chain is comple' and difficult. 1ome of the main reasons are: players in the supply chain are facing uncertainties regarding the supply and demand2 production lead times are long0 and thus procurement managers are forced to place orders long before they observe the actual demand2 their decisions are

based on forecasts that are usually not accurate2 economies of scale often force decision ma#ers to place large orders and thus prevent them from observing the actual demand and continuously adjusting to the changing mar#et conditions. .inally0 one of the most important difficulties is that each entity in the supply chain is interested in ma'imi-ing its own utility but not necessarily the utility of the entire supply chain. (he issue of designing supply contracts is not new0 but today it has became a crucial element of the procurement function. (his is due to the decentrali-ation of companies and to the fact that outsourcing of services and products is by now the common mode of operation. (his is in contrast to the not3so,distant past0 in which industrial companies were vertically integrated and coordinating the supply channel was much easier. In such environments0 a single decision ma#er made the decisions for all of the involved entities of the organi-ation0 thereby optimi-ing the performance of the entire organi-ation and not just the performance of each of the entities as a 4stand,alone5 unit. 6sually0 supply contracts are part of the process of procuring goods in manufacturing and retail industries. $ore recently0 it has been observed that such contracts are also an important part of health care delivery and especially the global health care delivery system. (he reason is that many of the patients live in low,income countries and are too poor to pay the full price of e'pensive medications. In such environments0 where mar#ets do not e'ist0 manufacturers are reluctant to invest in research and development of new drugs and in production capacity. We believe0 as we will describe in what follows0 that supply contracts are an important tool to ensure availability and affordability of drugs. In this chapter we concentrate in contracts the deal with ris# and uncertainty: in most cases we deal with the uncertainty in the demand for goods and services. )ut0 contracts are used also in other cases0 and especially in cases in which there is no uncertainty. & good e'ample is /uantity discount contracts. 7uantity discount contracts can be applied in a deterministic environment. (he main idea is to induce buyers and sellers to produce and purchase /uantities that generate operational efficiencies. 1uch contracts are common when economies of scale are presented. (ransportation systems are a good

e'ample. 1ellers offer price discount to induce buyer to purchase larger /uantities than those that they would have purchase otherwise. )y doing so sellers are able to ship full truc# loads and reduce the per unit shipping cost. (he interested reader is referred to the seminal paper by 8osenblatt and 9ee !1". & description and classification of /uantity discount contracts in $unson and 8osenblatt !2" .or obvious reasons we are unable to discuss here all types of procurement contracts. We describe shortly some of the most important contracts. & more inclusive and detailed review can be found in *achon !%". 1ection :.;.; in <1;= provides a very good discussion of many different aspects of contracts. 2. Flexibility and Stability &s mentioned above0 the main objective of supply contracts is to ensure the availability of the right /uantity and /uality of goods and services at the right time and at the lowest possible cost. In most of the contracts that we discuss here0 one of the main drivers of cost is ris#. $ore specifically0 all supply contracts are concerned with the allocation of ris# between the different entities in the supply chain. !8eferred to here as suppliers and buyers". (he /uestion that all supply contracts are addressing is the /uestion of who should bear the ris# of producing ahead of time and #eeping raw materials0 components and finished goods. 1uppliers li#e a stable environment in which all orders are firm0 and once a buyer places an order the buyer must ta#e and pay for the goods. +n the other hand0 buyers li#e the fle'ibility to place orders and then change these orders as they learn more about the actual demand. We argue that in many cases the right balance between stability and fle'ibility ma'imi-es the utility of the supply chain and can ensure a 4win, win5 situation. & good e'ample that illustrates this phenomenon was provided long ago by )lois <:=0 as shown in (able 1.

+ne can easily see how the monthly order is changed as we are approaching the delivery time. .or e'ample0 in $arch the scheduled delivery for >ecember was 220?@@ units. )ut in Aune the scheduled /uantity was reduced to 1:0;@@. It is not difficult to imagine the difficulties that the supplier is facing by such big changes in the orders and the resulting additional cost. (hus0 the role of a supply contract is not only to ensure demand !for the supplier"0 supply !for the buyer"0 and to improve the production planning process but also to create a fle'ible environment in which buyers can adjust to the changing conditions of the mar#et. &t the same time it is necessary to #eep the suppliers environment stable without imposing too much ris# and costs on the suppliers. (o further illustrate the tradeoff between stability and fle'ibility we use the graph provided by 8ietveld Bans <;=

.igure 1

(his graph illustrates several #ey issues crucial to the design of supply contracts. >espite the high demand for $alaria drugs !%@@$,;@@$ instances per year" and despite the available capacity !1@@ $ units"0 the manufacturer does not use all of its capacity. While there are several reasons for this0 one of the main reasons mentioned by the manufacturer !Covartis" is the uncertainty in the delivery /uantity. In 2@@;0 a large /uantity was ordered but only a relatively small /uantity was pic#ed and delivered. (he reason is that orders are placed long in advanced because of the long production lead time. When delivery time arrives0 the mar#et conditions may be different0 forcing the buyers to change their previously made orders and resulting in a large loss to the manufacturer. (his graph illustrates the general issue of balancing fle'ibility and stability. )uyers would li#e the fle'ibility to change previously made orders to better adjust to changing mar#et conditions. 1uppliers0 on the other hand0 would li#e to operate in a more stable environment in which the fle'ibility to change previously made orders is limited. If the 4right5 balance between fle'ibility and stability is not achieved sellers !buyers" may

produce !purchase" less than the desired /uantity and reduce their e'posure to ris#. Bow to balance fle'ibility and stability and allocate the ris# between buyers and manufacturers is e'actly the role of the procurement contracts that are discussed in this article. 3. Minimum Commitments

(he total,minimum,/uantity contract addresses two different issues0 as follows. !1" It gives the buyer some fle'ibility to change orders and to adjust to changing mar#et conditions0 but at the same time it limits the buyers fle'ibility to change order si-es drastically0 which helps to avoid increases in the production cost for the supplier. !2" It provides incentives to suppliers to build capacity even when the demand is uncertain0 and it shares the ris# of investing in capacity between the supplier and the buyers. )asso# and &nupindi <D= use the framewor# of a multi,period news,vendor model0 which assumes that the periodical !month0 wee#" demand is un#nown but that a forecaster is able to provide a probability distribution of the demand. &lso0 in this model it is assumed that the demand is stationary and independentEthat is0 the demand in each period follows the same probability distribution and demands across periods are not correlated. (he authors consider a finite hori-on problem that is a problem with a limited number of periods !for e'ample0 one year". (he optimal solution for this problem is well #nown: every period has a critical number 1t !the 4base stoc#5"0 so that in period t it is optimal to raise the available inventory !on,hand inventory" and order up to the base stoc#. (he reader will notice that in such a model the buyer firm has all the fle'ibility to order any /uantity it wishes0 which means that the supplier environment is far from stable. (his 4order,up,to5 policy obviously minimi-es the buyers cost0 but just as obviously it ignores its effects on the suppliers cost. (o mitigate the negative effects of demand uncertainty on the suppliers performance0 the buyer is re/uired to announce its minimum purchasing /uantity for the entire hori-on. .or ma#ing the minimum purchasing commitment0 the supplier offers a discounted

purchasing price per unit. )y ma#ing the 4minimum commitment05 the buyer limits its own fle'ibility. If0 for e'ample0 the total demand during the hori-on turns out to be lower than the total commitment0 then the buyer must still ta#e delivery of the committed /uantity and pay for it. If the total demand turns out to be higher than the total commitment0 then the buyer will have to purchase the difference between the total demand and the total commitment at a higher price. +n the other hand0 the buyer still has the fle'ibility to place periodical orders of any si-e0 as long as the sum of the periodical orders is not greater than the 4minimum commitment.5 (he supplier also benefits from such a contract because it #nows with certainty that during the hori-on it will have a buyer for the committed minimum /uantity0 which means the firm can plan its production in an efficient way. (he authors concentrate on the effects of the minimum commitment contract on the buyers performance. (hey are mainly interested in determining the optimal purchasing policy of the buyer assuming that it ma#es a minimum commitment to purchase F units during the hori-on. (hey prove that the optimal purchasing policy of the buyer has a very simple structure and that it is very easy to calculate the optimal periodical purchasing /uantity. (he cru' of the optimal policy is as follows. &t the beginning of every period0 if the sum of the buyers earlier purchases is still smaller than the minimum commitment0 then it is optimal for the buyer to raise its inventory to the base stoc# 1$0 where 1$ is the optimal base stoc# for a single,period news,vendor problem assuming that the purchasing cost is e/ual to -ero. If0 at the beginning of a period0 the sum of earlier purchases is larger than the minimum commitment0 then the buyer needs to solve a standard multi,period problem without any commitments. )y following such a simple policy0 the buyer minimi-es its cost. *leary0 the supplier is interested in the buyer ma#ing a large commitment0 and to encourage the buyer to do so0 the supplier is willing to provide the buyer with a per,unit discount. (he effect of the commitment on the buyers cost is illustrated by the following graph <D=.

.igure 2. Effect of rice discount

(he graph shows us that as the total commitment is increasing the total saving is decreasing. (his is e'pected because the larger the commitment is0 the larger is the buyers ris#. (he graph is useful in determining the optimal commitments. .or e'ample0 it is better for the buyer to commit for ?;@ units with a ;G discount than for H@@ units with 1@G discount. (he minimum,commitment contract can be used to encourage companies to build production capacity. (his is especially true in environments in which there are no mar#ets for the produced goods. & good e'ample is the 1ub,1ahara countries0 in which most of the population is unable to purchase life,saving drugs. rivate and public organi-ations are willing to subsidi-e the drugs and to ensure that they are available and affordable to the needy0 but usually the si-e and the duration of subsidy is not well defined and is uncertain. &s a result0 drug manufacturers are unwilling to ta#e the ris# and build capacity to produces drugs for low,income countries. (o solve this problem0 several researchers have suggested that the private and public organi-ations ma#e clear and well,defined commitments to subsidi-e certain drugs !for e'ample0 malaria drugs". (he total amount and the duration of such a commitment should be advertised ahead of time and ensured by a third party. .inally0 a neutral organi-ation !e.g.0 the World Bealth +rgani-ation" will have to provide a list of drugs that should be

subsidi-ed by the private and public organi-ations. )y providing a total minimum, /uantity commitment0 the private and public donors are able to mimic mar#et conditions and provide a more stable environment for the manufacturers0 who will now be certain that they will sell at least the minimum,commitment /uantity. *learly0 this is an incentive for pharmaceutical companies to enter the mar#et and invest in production capacity. 4. Periodical Commitments and Flexibility

(he total,minimum,/uantity commitment solves several problems0 most especially ensuring that suppliers will sell at least a minimum /uantity during the hori-on. )ut it does not deal with the changing periodical orders. )uyers still have the fle'ibility to change their previously made orders to better meet the actual demand. It is clear that buyers would li#e the fle'ibility to change previously made orders at any time0 but it is also evident that suppliers would li#e to limit this fle'ibility0 ensure stability0 and force the buyer to commit and ma#e firm periodical orders. )asso# et. al <?=0 &nupindi and )asso# <I= and (say and 9ovejoy <H= are all addressing the issue of balancing fle'ibility and stability. *ommon to these studies are several concepts. (he buyer0 before the hori-on starts0 ma#es periodical commitments /@010/@020J/@0C where /i0j is the commitment made at period i for period j. &t the beginning of period 1 the buyer may change the order /uantity to be delivered during this period /101 to be: !1 l @ "q @01 q101 !1 + u o " q@ 01 . (hus0 the buyer has the downward fle'ibility of l @ and the upward fle'ibility of u o . (hus0 the buyer does not have the infinite fle'ibility to purchase any /uantity it wishes but rather must stay within the limits determined by its original commitment and the downward and upward fle'ibility limits. In addition0 the buyer may change its future commitment in a restricted way:
!1 l1 "q @0 2 q10 2 !1 + u1 " q @0 2 !1 l 2 " q @ 0% q10% !1 + u 2 " q @ 0%

!1 l N 1 " q @0 N q10 N !1 + u N 1 " q @0 N 0

where l1 l 2 ...l N 1 and u1 u 2 ...u N 1 . &nupindi and )asso# <I= describe graphically the dynamics of this contract in the following way:

Figure 3:

!e dynamics o" t!e #eriodical commitment and "lexibility contracts.

(he current period is represented by a diamond and the future periods by a rectangle. 1uppose that the commitment for period % made in period @ was 12@ units. In period 1 this commitment can be adjusted upward or downward by 2@G. (hus0 in period 1 the buyer can adjust the commitment in period % to any /uantity between <HD01::=. 1ay that the buyer decides0 at the beginning of period 10 to commit to purchase 11@ units. In period 20 the commitment made for period % can be adjusted upward or downward only by 1@GEthat is0 the adjusted commitment could be between <HH0121=. Cow let us say that the buyer chooses 1@@ units. In period % the buyer has ;G upward and downward fle'ibility and can purchase any /uantity between <H;01@@=

(his process repeats itself every period. (he idea is to enable buyers to change their previously made commitments in a restricted way. When the time between ma#ing a new commitment !period i" and the actual delivery time !period j" is short0 the fle'ibility is very limited. )ut when the time between ma#ing the commitment and the actual delivery is long0 the fle'ibility is large. (his is achieved by l j i and u j i increasing with j,I0 and the number of updates that the buyer is able to ma#e. (he reason behind the structure of fle'ibility contract downward and upward constraints is that changes in the order /uantity0 with a very short notice0 are costly for the supplier0 while changes with a long notice are much cheaper0 and in this case changes in the previously made commitment can be relatively large. &ll three studies aim to identify the optimal initial periodical commitments and the optimal periodical updates. >etermining the optimal policy is very difficult0 and all studies resort to appro'imation and heuristic. )asso# et al <?=. develop a lower and upper bound on the total costs. (he lower bound is achieved by assuming that the fle'ibility is very largeEi.e.0 assuming that changing previously made commitments is not limited. In this case0 the structure of the optimal policy is well #nown: there are critical numbers !base stoc#s" 110 120J01C so that at the beginning of each period it is optimal to raise the inventory up to the base stoc#. (he idea behind the upper bound on the total cost is somewhat more involved. (he upper bound policy is a feasible policy !changes in the commitment /uantities are limited and satisfy the fle'ibility constraints". (he changes are made so that the probability of reaching the base stoc#s 110120J01C is ma'imi-ed. (he authors show that the gap between the upper and lower bounds is not too large. .or e'ample0 when the coefficient of variation of the demand is smaller than @.; the ma'imum gap is only ?G !assuming upper and lower fle'ibility of only ;G". 1ince the gap is small0 the authors use the upper bound to calculate the optimal initial commitments and the updates of the periodical commitments. (he ability to determine close,to,optimal commitments and then to update them provides a very important tool in the negotiation process between buyers and suppliers. (he

suppliers may provide different levels of upward and downward fle'ibility0 but for a higher fle'ibility the supplier demands a higher per,unit cost. (his relationship is illustrated in the following graph.

Figure 4: (rea)%e'en #urc!asing cost 's. "lexibility under stationary demands *c +4,- ! + 1- # + 1,,- + 12- + 1,,,. /!ere c is t!e #urc!asing cost #er unit- ! is t!e !olding cost #er unit- # is t!e s!ortage cost #er unit- is t!e mean demand and is t!e standard de'iation o" t!e demand

1ource: )asso# et. al. <?= . (he reader can see that when the coefficient of variation is @.?; the buyer should pay no more than :1.? per unit for @.2 upward and downward fle'ibility.

$. Fixed%&eli'ery Contract & contract that has similar features to the periodical orders and fle'ibility is the one presented by $oin-adeh and Cahmias <1@=. (hese authors consider a fi'ed,delivery contract0 and they assume a finite,hori-on model with bac#orders and independent and

stationary demand. (he buyer is committed to purchasing the same /uantity 7 at a cost per unit of c in every period. (he buyers have upward fle'ibility to order more than 70 but they have to pay a premium for the additional /uantity0 and the cost per unit is cB !where cB K c". In addition0 every time the buyer deviates from the committed /uantity 7 it must pay a fi'ed cost F. (he fi'ed delivery contract captures the fact that in certain industries it is very costly to change the production /uantity. (hus0 suppliers prefer to produce the same /uantity in every period. +n the other hand0 the buyer prefers to have fle'ibility to adapt to the uncertain demand and update the fi'ed,delivery /uantity. (he fi'ed,delivery contract offers limited fle'ibility because the buyer may purchase more than the fi'ed delivery /uantity but not less. &t first glance0 it seems that the fi'ed,delivery contract is mostly advantageous to the supplier0 who is able to reduce the variability in the orders. Bowever0 this observation is misleading. (he benefit to the buyer depends upon the price discount that the supplier offers0 c0 and the premium price cB. (here is always a large,enough discount so that the buyer also benefits from the fi'ed discount contract. 0. (ac)u# Contracts Eppen and Iyer <11= present a bac#up contract that has some of the flavors of the commitment,and,fle'ibility contract presented above. (hey consider a model that is very common in the apparel industry and especially between the manufacturer and a catalog !retailer". (hey suggest a two,period model. )efore the first period0 the buyer must ma#e a purchasing decision0 /0 for the two periods. Let0 they ta#e delivery of only 7!1," units. +nce they observe the demand0 at the first period0 they ta#e a delivery of up to 7 units0 but not more than that. While the cost per unit is c0 the cost per unit that the buyer does not ta#e !at the most 7 unit" is b. (his contract is similar to the periodical, commitment,with,fle'ibility contract. Cotice that the buyer !catalog" ma#es a commitment to purchase 7 units. (he upward fle'ibility is -ero and the downward fle'ibility is . (he main differences between the

two contracts are !1" the periodical,commitment,with,fle'ibility contract is a multi, period contract0 while the bac#up contract is limited to only two periods2 !2" in the periodical,commitment,with,fle'ibility contract0 the demands in the different periods are assumed to be independent0 while the demands in the bac#up contracts are assumed to be correlated2 and !%" in the bac#up contract there is an e'plicit penalty per unit0 b0 for not purchasing units that were actually ordered. In the periodical,commitment,with, fle'ibility contract0 the cost of fle'ibility is captured by the cost per unit of the productE the greater the fle'ibility0 the higher the cost per unit. )ut there is no e'plicit cost for units that were ordered but not purchased. &s the value of b !the penalty per unit for not purchasing the committed /uantity" is increasing0 the value of the bac#up contract is decreasing. (his is /uite intuitive because when b is very large the buyer will tend to purchase all of the committed /uantity and thus the bac#up option has -ero value. 1imilarly0 as l is increasing and the buyer has the opportunity to purchase a smaller fraction of the commitment0 the total cost for the buyer is increasing. erhaps more interesting is the fact that the bac#up contracts provide the buyer with an incentive to commit to a larger /uantity than the /uantity that it would have purchased without the bac#up contract. (his is a very important observation. (he bac#up contract0 provides the buyer with fle'ibility and thus reduces the buyers ris# and induces the buyer to commit for more units. (he supplier0 who assumes a larger share of the ris# benefits by selling a larger /uantity. (he magnitude in this difference can be large0 and the authors provide an e'ample in which the committed /uantity in the bac#up contact is larger by D%G than the purchasing /uantity without a buybac# agreement. It is interesting to observe that for certain parameters both the supplier and the buyer benefit from the bac#,up contract. (hus0 one can as# what the parameters of the bac#up contracts that minimi-e the sum of the cost of both the buyer and the supplier are0 and when it is possible to coordinate the channel. We will address this issue shortly.

1.

2#tions Contracts

(he main idea of the above contracts is that they provide the buyer with fle'ibility to adjust to changing mar#et conditions0 and at the same time they provide the supplier with information that helps the supplier to create a stable environment and better plan its production and reduce cost. &ll of the above contracts can be viewed as special cases of the following option contract. In a two,period option contract0 the buyer ma#es periodical orders 71 and 720 at a price c per unit. (hese orders are firm orders and cant be changed. In addition0 the buyer purchases $ options0 at the option price co per unit. (hese options may be e'ercised at the second period at the e'ercise price ce. &s is the case in all of the previous contracts0 the buyer would li#e to have fle'ibility to change its previously made orders0 while the supplier would li#e to limit the fle'ibility and encourage the buyer to limit the changes to the previously made orders. *learly0 if co M ce is large enough0 then the buyer will tend not to purchase any options and will have -ero fle'ibility. +n the other hand0 if co M ce N c0 the buyer will not ma#e any commitment for the second period and will only purchase options. *hoosing the right levels of co and ce ma#es it possible to stri#e a balance between the interests of the buyer and the supplier and perhaps even to reduce !increase" the cost !profits" of both. (he reader may notice that the option contract is a generali-ation of the two,period fle'ibility contract and the bac#up contracts. (he table below demonstrates that these contracts are special cases of the option contract.

Figure $: S#ecial cases.

Source:

)arnes,1chuster et. al. <12=

.or e'ample it is clear that the bac#up contract is a special case of the option contract. In both contracts0 the buyer commits to purchase a certain /uantity. (he actions in the second period are different. In the bac#up contract0 the buyer does not commit to any /uantity0 while in the options contract the buyer may commit to any /uantity0 72. 1econd0 the bac#up contract limits the number of options that the buyer may purchase2 it is a function of the buyers first,period commitment. +n the other hand0 the option contract provides the buyer with the fle'ibility to purchase any number of options. )arnes,1chuster et. al. <12= provide a detailed two,period option model. In addition to studying the behavior of the buyer0 they also study in detail the behavior of the supplier and the entire channel. (he se/uence of action assumed in this study is provided below:

Figure 0: A time line o" buyer%su##lier decisions.

(his model assumes that the supplier may produce ahead of time up to 71 M 72 M $ units. )y producing ahead of time with a long production lead time0 the buyer is utili-ing a less,costly technology0 but the supplier is incurring holding cost0 and perhaps more important it is assuming the ris# of having0 at the end of the last period0 obsolete inventory. &lternatively0 the supplier can produce a smaller /uantity0 wait until the buyer decides how many options to e'ercise0 and only then produce the goods. In this case0 the production cost per unit is high due to the short production lead time. *learly0 such a policy seems to be costly due to the high production cost. Let0 it reduces the inventory holding cost and the ris# of obsolete goods that were produced but cant be sold.

1uch contracts are common practice in the apparel industry. (he difficulty in this industry is the long lead time in producing the raw material. (he #nitting and weaving0 dying and printing operations have a long production lead time. (he sewing itself can be done relatively /uic#ly. 1uppliers !manufacturers"0 in order to minimi-e ris#0 tend to #eep low inventories of raw materialsEjust enough to produce the ordered /uantity. (his results in one production run that is performed long before the season start0 which creates a situation in which it is retailers who are unable to place a second order during the season when they have better information about the actual demand. (he option contract reduces the suppliers ris# of #eeping raw materials because the buyer shares this ris# by paying the option price. )y #eeping raw materials in stoc#0 the suppliers are able to produce /uic#ly and0 if necessary0 to produce and deliver during the season. (hus0 it provides the buyer with the needed fle'ibility. (his detailed model not only determines the optimal commitments0 the number of options0 and the options and e'ercise prices0 but it is also instrumental in identifying the optimal production policy for the supplierEin other words0 how much should the supplier produce ahead of time at the low production cost0 and which fraction of the total should be produced at a later time only and at a higher production cost. (he fact that it is possible to consider that the supplier production policy leads the authors to as# whether it is possible to coordinate the channel. We remind the reader that we say that the channel is coordinated when the sum of the costs !profits" of the supplier and the buyer are the same as the costs !profits" in a centrali-ed system with a single decision ma#er who optimi-es the performance of the entire supply chain. In addition0 if it is possible to coordinate the channel it is interesting to find out whether it is possible to allocate the cost !profit" between the buyer and the supplier in any arbitrary way. If the answer to both /uestions is positive then it is possible to find options and e'ercise prices so that both players improve their costs !profits" as compared with other contracts. & detailer discussion of channel coordination is found in chapter :.;.;.1.

It turns out that in general it is impossible to coordinate the channel with linear prices. (here are some instances !depending upon the cost parameters" that enable the coordination of the channel. )ut in these cases the prices are not individually rational for the supplier. Even return policies that are #nown to coordinate the channel in many environments fail to coordinate the channel in this case. +ne of the main difficulties in coordinating the channel is the fact that production cost is not linear. (o see this point0 consider a buyer that purchases $ options. (he supplier may produce ahead of time0 at a low cost0 some fraction of the $ options0 then wait until a later time to see the actual number of options that are e'ercised0 and finally0 if necessary0 produce an additional /uantity at a higher cost. (hus0 the actual production cost is not linear. (he option contract assumes that the supplier charges the buyer a linear price. &s a result0 the number of options bought and e'ercised0 assuming an option contract0 is not the same as in the centrali-ed system. (hus0 in general options contracts do not coordinate the channel. $ore on option contracts can be found in <1%= and <1:= 3. Conclusions

(he main issue in designing supply contracts is to ensure supplies and goods at the right /uantity0 /uality0 time0 and cost. (he tension is always between the wishes of buyers to have as much fle'ibility as possible and suppliers0 who benefit from a stable environment and early firm commitments. )alancing this with fle'ibility can be done in different ways0 but in all cases stability and fle'ibility have real costs that can be evaluated0 /uantified0 and used to calculate acceptable solutions for suppliers as well as buyers.

4.

5e"erences

<1= 8osenblatt0 $.A. and 9ee0 B.9. !1HI;" Improving rofitability with /uantity discount with fi'ed demand. IIE Transaction0 1?0 %II,%H;.

[2] $unson *. and 8osenblatt0 $.A. !1HHI". (heories and 8ealities of 7uantity >iscounts: &n E'ploratory 1tudy. Production and Operation Management. ?!:" <%= *achon0 O. !2@@%"0 1upply *hain *oordination with *ontracts. Bandboo#s in +perations 8esearch and $anagement 1cience0 11: 1upply *hain $anagement: >esign0 *oordination and +perations5. Edited by A.6. de 7o) and S.C. 6ra'es. Corth Bolland. <:= )lois0 F.A. !1H?;". 1upply contracts in the Oalbraithian planning system. The journal of industrial economics. 24( !. <;= 8ietveld Bans !2@@I" 4& Cew *lass of $alaria >rugs: (he *oartem )rea#through from Covartis and its *hinese artners5 Wor#shop on access and benefit sharing. )onn. <D= )asso#0 L. and 8. &nupindi. 1HH?. &nalysis of 1upply *ontracts with total minimum commitment. IIE Transactions. 2H!;". <?= )asso#0 L0 &. (ixby0 8. Srini'asan0 and B. 8. /iesel. 1HHH. >esign of component, supply contract with commitment,revision fle'ibility. I"M #ournal of $esearch and %e&elopment. :1!D" <I= &nupindi 0 8. and L. )asso#. !2@@I". &nalysis of supply contracts with commitments and fle'ibility. Na&el $esearch 'ogistics. ;;!2" <H= (say0 &. and W. 1. 9ovejoy. 1HHH. 7uantity fle'ibility contracts and supply chain performance. Manufacturing and (er&ice Operations Management. 1!2" <1@= $oin-adeh0 F. and 1. Cahmias. 2@@@. &djustment strategies for fi'ed delivery contracts. Management (cience. :I!%" <11= Eppen. >. O. and &. P. Iyer. !1HH?" )ac#up agreement in fashion buying,the value of upstream fle'ibility. Management (cience. :%!11". <12= )arnes,1chuster0 >.0 L. )asso# and 8. &nupindi. !2@@2". *oordination and fle'ibility in supply contracts with options. Manufacturing and (er&ice Operations Management. :. 1?1,2@?. [13] >onohue0 F. !2@@@". Efficient 1upply *ontracts for .ashion Ooods with .orecast6pdating and (wo roduction $odes. $anagement 1cience :D!11"0 pages1%H?,1:11 <1:= 1pinler0 1. !2@@%". *apacity 8eservation for *apital,Intensive (echnologies: &n +ptions &pproach. 9ecture notes in Economics and $athematical 1ystems. Pol. ;2;. ublished by 1pringer.

<1;= Wiley Encyclopedia of +perations 8esearch and $anagement 1cience.