Klein, Crawford, and Alchain, "Vertical Integration, Appropriable Rents, and the Competitive Contracting Process"

Big picture: Incomplete contracting and asset specificity can make vertical integration more efficient than competitive contracting.

Why are contracts incomplete? Costs of monitoring, constructing, enforcing contract, uncertainty.

What problem does this bring about? Appropriability of quasi-rents from specific asset.

Is this inefficient?

Even if investment in specific asset is made, potential for appropriability creates the prospect of inefficient behavior, behavior that does not create value but rather increases bargaining power.


Printing Press Example:

Suppose Party A supplies a printing press, potentially serves parties B, C, and D (book publishers).

Assume that in order to supply the printing press, party A has to go to the bank and borrow money. Assume that loan payments are $4000/day. Because this must be paid regardless of whether or how much the printing press is used, this is a fixed cost.

Assume that the printing press costs $1,500/day to operate. These are variable costs. Therefore, TC=AC=$5,500/day.

Assume also that if the press is not used for publishing books, it can be used to publish other things such as pamphlets, flyers, etc. In these other uses, it has a value of $1,000/day. That is, people who want to use the press for printing pamphlets, etc., would be willing to pay $1,000/day to rent the press.

Case 1:

Suppose a competitive market for printing. If this is the case, then A will receive the market price=MC=AC, which is $5,500/day. If, after agreeing to this price, the publisher tries to renegotiate the terms, the owner of the printing press can simply say no and receive $5,500 from another publisher.

In this case, there is no incentive problem, and no inefficiencies arise.

Case 2:

Suppose instead that party B needs specialized printing press -- for example, he or she needs it within the same builiding, or wants to use special inks.

Suppose also that if once it is bolted into B's building or is contaminated with B's inks, it is not worth as much to other book publishers. (It is either very inconvenient for them to deal with a press that is in B's building, or the contamination makes it unsuitable to publish other books.) Suppose for now that once the press is bolted in, it becomes completely unsuitable to other book publishers.

Party A solicits bids, and B wins. Suppose the price is $5,500/day. The machine is then bolted into the building and the press uses these special inks.

Now party A comes in to collect payment...uh oh!

B says the deal is off -- unless A agrees to the new rate of $2,500/day. (opportunism!)

What does A do?

A's next best alternative is renting it to the pamphleteers. If it does this, its variables profits are $1,000/day.

If A gives in to B's demand, A's variable profits are $1,000/day ($2,500-$1,500 operating costs).

Without another alternative, he can do no better than accept the offer! Loses the entire quasi-rent.

Calculating quasi-rents and specialized quasi-rents in this example.

Quasi-rents are payments in excess of the amount necessary to keep an asset in its current use. The printing press' next best use is printing pamphlets. Party A's reaps $1,000 in variable profits if the press is put to use that way. Therefore, $2,500 is all that is necessary to keep the press in its current use -- publishing books. The QR is $5,500-$2,500=$3,000.

Specialized quasi-rents are payments in excess of the amount necessary to keep the asset with its current user. The printing press' next best user is someone printing pamphlets. Therefore, SQR=QR=$3,000. The entire QR is specialized, and therefore potentially appropriable.

Case 3:

Change case 2 so that one of the publishers, party C, is willing to pay at most $3,500/day for the press, even after it has been bolted down at party B's building and has been contaminated with party B's ink. Now what are QRs and SQRs?

QR's are the same as before. The value in its next best use -- printing pamphlets -- has not changed. QR=$3,000.

SQR's are not the same. It would take at least $3,500/day to ensure that the printing press be used to serve party B. SQR=$5,500-$3,500=$2,000.

Therefore, not all QR's are specialized. The maximum party B could appropriate is $2,000.

Why is this a problem?

The problem arises because if party A, the supplier of the printing press anticipates this happening, he or she will not supply the printing press (or at least he or she will not offer to specialize it for party B) because he or she will not be able to cover fixed costs. He or she has to pay the bank $4,000/day. Variable profits are $1,000/day in case 2 and $2,000/day in case 3. In neither case does the variable profits cover fixed costs.

The efficiency loss in this example lies in the fact that there is a potentially opportunity for trade that is lost. The press owner is willing to supply the press for the price of $5,500/day, and the publisher values the press at $5,500/day. Therefore, there are gains from trade. But because the press owner anticipates that, after the specialized investment is made, the publisher may appropriate some or all of the quasi-rents, he or she chooses not to supply the specialized press.

In general, efficiency losses arise from asset specificity include:

Competitive contracting in this case is costly.

Vertical integration is costly as well, because of additional costs of managing increased scope of economic activity.

Long-term contracts are another option. But they are detailed and costly to construct, and are they any easier to enforce than the short-term ones? Possibly...if parties are earning rents within this contract that would be lost if it were broken. Brand name serves as collateral.

Sometimes certain long-term contracts create new hold up problems, as we shall see. This may make it worthwhile to construct even more complicated ones.

Long-term contract is particularly an option when the asset in question is human capital, which cannot be owned by firms. Vertical integration is not an option.


Back to the question of vertical integration v. competitive contracting.

KCA assume that: as asset specificity increases, contracting costs increase by more than the costs of vertical integration.

Why? Because larger quasi-rents make reneging more attractive, in the competitive contracting situation, it will be optimal to expend more resources tinto writing more detailed contracts, monitoring more, enforcing more vigilantly. It is not obvious why managing economic activity is more expensive for specific assets than non-specific ones.

Therefore, greater specificity, more likely that vertical integration will take place.


Fisher Auto Body

Who owns the machine that stamp body parts? Assembler or outside supplier?

Originally, vehicles had individually-constructed open bodies.

By the late 1910s, production process was transformed to metal closed body. Production by specific machines which stamped the metal into its required shape.

Machine was specific to individual car-makers' design. This created the possibility of large hold up problems -- assembler could hold up the value of specific investment in stamping machine.

If rents they receive within this relationship and others are great enough, GM will not appropriate in order to protect its "reputation". But contracts can extend the states of the world in which appropriation does not occur by making more costly to do so.

GM enters into a exclusive dealing long-term contract (10 yr) with Fisher Auto Body. GM buys all of its closed bodies from Fisher. This resolves GM's ability to hold up Fisher -- exclusive dealing contract was easy for Fisher to enforce. Fisher secure about its specific investments in its machine.

However, GM was left very vulnerable. Large appropriable QRs.

Fisher could hold up GM by demanding monopoly prices rather than those stipulated in the contract, because stopping production was extremely costly to GM, who could not obtain auto bodies from other sources on short notice. Some of this was anticipated -- there were price provisions in the contract -- Fisher was paid variable costs plus 17 percent. Furthermore, Fisher could not charge GM more than they charged other car makers for similar bodies.

But Fisher could hold up GM through other means -- for example, by shirking on quality, slowing delivery, etc.

What it actually did is use inefficiently labor-intensive quantities of labor, because it could pass these variable costs along to GM. It also refused to locate its plant near GM plants -- disadvantageous when contract was renewed.

High cost of even a slight delay from litigation created possibility of hold up.

1926: GM buys out Fisher. This proved to be less costly than long-term contract.

costs from hold up potential:


Mine-mouth Coal Plants, Vertical Integration, and Contract Duration

About 15% of electric utility coal consumption is from "purchases" from integrated suppliers

70% in long-term contracts (specifying prices, quality, quantity, length).

15% in spot contracts.

What explains the contractual variation?

Joskow: Asset specificity.

Vertical integration more likely for "mine-mouth" plants -- those located literally next to a coal mine rather than next to its consumers.

Why? Non-integrated firm faces a hold-up of specific asset -- demand to renegotiate.

Conditional on contracting solution, longer (far longer -- 15 yrs) for mine-mouth plants.

Why? Holding up firm more costly during contract than at renegotiation time. Costly to bargain repeatedly. "Small numbers" bargaining problem.

Why aren't all mine-mouth plants vertically integrated? One reason is regulation --discouraged because of cost-shifting. Difficult to have rate-of-return or cost-based regulation.