TRANSACTIONALMICROSTRUCTURE AND MACROECONOMIC PERFORMANCEt
The "Fundamental Transformation" in Macroeconomics
By RICARDO J. CABALLERO AND MOHAMAD L. HAMMOUR*
When factors enter into joint production, pensate the appropriatedfactors, providing a they often develop a degree of "specificity" highly inefficient macroeconomic "solution" with respect to each other. Their value in the to the unresolved microeconomic contracting context where they have come to produce ex- problem. Appropriabilityaffects major aspects ceeds their outside value. Committing a pro- of the macroeconomy. It results in misalloca- duction factor in such a relationship is a form tion, underutilization, and unemployment of of specific investment, whose sunk component the economy's productive factors; it hampers may have technological as well as institutional growth by depressing the incentives to replace origins. what is outdated and to utilize fully the econ- By itself, this "irreversibility" has impor- omy's resources; it disrupts macroeconomic tant and extensively studied implications for adjustment, by inducing a wedge between investment decisions. But, when combined timid creation and excessive destruction of the with contracting problems, it acquires an en- old system; and it exacerbates downturns by tirely new dimension. When two factors enter "elastifying" the cyclical response of inelastic into a joint-productionrelationship, specificity factors. creates ex post quasi-rents that need to be pro- tected through ex ante contracting. Unless L Microeconomics complete and enforceable contracts are avail- able, those quasi-rents may be appropriated The prototypical macroeconomic example by the other factor. This transformationfrom of specificity we choose to focus on is that of an ex ante competitive setting to an ex post the relationship between capital and labor. bilateral monopoly-what Oliver Williamson Other examples are also important and affect terms the "fundamental transformation"-is such diverse relations as that between an a central building block in the modem eco- entrepreneur and his external sources of nomic theory of institutions (e.g., Benjamin finance (e.g., Oliver Hart and John Moore, Klein et al., 1978; Williamson, 1979). 1994) or that between an upstream firm and The problem of appropriabilityspreads un- its downstream customers (as discussed ex- protected quasi-rents throughout the economy tensively in the microeconomic literature on and, in a partial-equilibrium setting, leads to vertical integration). underinvestment. In general equilibrium, the In the presence of specificity, the centralunit market system will adjust to help partly com- of analysis is the joint-productionrelation be- tween the factors. We embody this relation in the concept of a "productionunit" which, we assume, combines in fixed proportionsone unit of capital and one unit of labor. Because of t Discussants: Steven J. Davis, University of Chicago; specificity, productionunits cannotbe costlessly Mark Gertler, New York University; Abhijit Banerjee, reshuffled.It is thereforemeaningful to analyze Massachusetts Institute of Technology. * both the creation decision, by which factors Caballero: Departmentof Economics, Massachusetts choose to enter into a productionrelation, and Institute of Technology, 50 Memorial Drive, Cambridge, MA 02139, and NBER; Hammour: Capital Guidance, 32 the destructiondecision, by which factorswithin Avenue Hoche, 75008 Paris, France. We thank Olivier an existing productionunit choose whether to Blanchard and Steven Davis for helpful comments. continue producingjointly or not. 181 182 AEA PAPERS AND PROCEEDINGS MAY 1996
We outline a minimalist model to help struc- with respect to the relationship, so we repre- ture our discussion. For a more complete treat- sents both its ex ante and its ex post opportu- ment, we refer the reader to Caballero and nity cost. Hammour (1996a, b, c). The units of analysis that arise from the notion of specificity provide The Fundamental Transformation.-If cap- the natural setting for a Schumpeterian view ital and labor were able to precommit to their of a macroeconomy with ongoing creation and ex post terms of trade, they would follow the destruction (see e.g., Caballero and Hammour, efficient entry rule. But when complete con- 1994). In this short paper, however, we reduce tracts are costly to write and enforce, agents' our treatmentof dynamics to a minimum. We ability to precommit is limited, and a William- model destruction and creation decisions dur- sonian "fundamental transformation" arises: ing a single period. We assume that factors can the ex ante competitive relationship between either be part of a production unit in the capital and labor turns into an ex post bilateral "joint-production" sector, or remain in their monopoly. Assuming that no precommitment "autarky" sectors. The period starts with pre- is possible, the specific quasi-rents from a pro- existing production units as well as a supply duction unit is the difference between the of uncommitted units of capital and labor. unit's revenue and its factors' ex post oppor- First, the factors in all preexiting units decide tunity costs: whether to produce jointly in the coming pe- riod or to separate and join the uncommitted (2) sY -We. factors; second, all uncommitted factors either form new production units or remain in au- Following the Nash bargaining solution for tarky; finally, production of the economy's sharing the unit's revenue, we assume that unique consumption good takes place in the each factor gets its ex post opportunity cost joint-production and autarky sectors. plus half of the unit's bargaining-surplus s. Using the superscript j to denote payments Creation. -We denote by y the production to factors in new joint-production units, we revenue of a newly created production unit, have and by wk and we the reward of capital and labor in autarky (all in terms of the consump- - tion good). In partial equilibrium, the efficient (3) Wk 2 decision is to create a production unit as long as (4) = We + W5 2 (1) Y?Wk + We.
This rule would, in fact, be followed as long Underinvestment.-It is clear from equa- as factors of production can freely and cost- tions (3) and (4) that labor will willingly par- lessly move in and out of production units. ticipate in production as long as the surplus from joint production is nonnegative. Capital, Specificity.-Such freedom of movement the factor that is vulnerable to appropriation, rarely holds in reality. Let us take the extreme requires more. Its share of the ex post surplus irreversible-investmentview, where once cap- must be enough to justify forfeiting its ex ante ital is locked into a specific relationship, it is opportunity cost: lost if the relationship breaks up. Although it can also be due to such institutional factors as s - organized labor or legislated firing costs, we 2 Wk. view specificity here as a purely technological phenomenon. As a result, while the ex ante Replacing (2) into this expression yields the opportunity cost of capital is Wk,its ex post entry rule more explicitly: opportunitycost is zero. On the other extreme, we assume that labor develops no specificity (5) y 2 2Wk+ We VOL. 86 NO. 2 TRANSACTIONALMICROSTRUCTURE AND MACROECONOMICS 183 which states that a unit's revenue must not assume that preexisting production units are only cover its factors' opportunity costs, as in of infinitesimal mass. We assume that, in its the efficient condition (1), but must also gen- autarky sector, capital can be turned one-to- erate enough rents for capital to recover its ex one into the consumption good, so Wk = 1. ante opportunitycost after being appropriated. On the other hand, labor experiences decreas- This increased-revenue requirement generally ing returns in its autarky sector, resulting in gives rise to an underinvestment problem, a joint-production labor supply function: a result most clearly stated by Paul Grout (1984) in this context. (7) We = 2L"/77 r1> O
Destruction. -The other decision of inter- where L measures joint-production employ- est to us is that of a preexisting productionunit ment. The supply of capital into joint produc- with production revenue x that must decide tion is therefore perfectly elastic, while the whether to remain in operation or shut down supply of labor is partly inelastic. Finally, we and free its nonspecific resources. The pri- restrict our analysis to values of y E (2, 3), vately efficient rule is to exit if the opportunity where both the efficient and inefficient econ- cost of the two factors of production cannot be omies exhibit positive but not full employment covered. Since capital is fully specific, its op- in joint production. portunity cost is zero. On the other hand, the opportunity cost of labor in a preexisting unit A. The Efficient Economy is pwle + (1 - p)we, where p is the proba- bility of finding employment in joint produc- If precommitment with complete and en- tion (see below). The exit rule is therefore forceable contracts is available, equilibrium will be efficient. Capital and labor will receive (6) x
Depressed Creation. -In general equilib- the unit will exit if and only if its revenues are rium, and off corners, entry condition (5) must lower than pwie + (1 -p)wE. It is easy to hold with equality. As before, noting that Wk= see that 1 and taking labor supply (7) into account, the opportunitycost of labor and the level of joint- pwle + (1-p)We < We production employment are equal to since, by (10) and (12), we know that we < (10) w= y-2 < we we and wit = w*, and the probability of find- ing joint-production employment is p = L < 1. Comparing (6) and (9), this implies that = (11) L < L* the incomplete-contracts economy keeps in operation low-productivity units that would be With incomplete contracts, the "appropriat- scrapped in an efficient world. Not only is cre- ing" factor's opportunitycost and the creation ation depressed by the appropriability prob- of joint-production units are depressed in gen- lem, but also pressure on outdated units is eral equilibrium. In order to ensure a proper reduced by the equilibrium decline of the return to the "appropriated" factor, the eco- opportunity cost of labor. This combination nomic system must depress the opportunity makes for a powerful drag on growth. cost of labor so that, capital receives its ex ante requiredreturn. This is done by underutilizing Scrambling.-Efficiently, the scrapping rule labor and, throughthis mechanism, weakening is based on a strict productivity ranking of the bargaining position of "insider" labor. preexisting production units, with the least productive units exiting, and others surviving. MarketSegmentation and Unemployment. The ranking that forms the basis for the In equilibrium, the economy must generate a scrapping rule could be "scrambled" in the positive joint-production surplus s to compen- presence of appropriabilityand bargaining in- sate capital ex post for its ex ante opportunity efficiencies. Suppose we had introduced pri- cost [see equation (3)]. This implies that vate inefficiency in separations of some form. joint-production workers, who share in that Some high-productivity units could be de- surplus, must earn more than those who re- stroyed because of private inefficiency, even main in autarky [see equation (4)]. The labor though they would survive in the efficient econ- market is therefore "segmented" between omy; other, much less productive units, if they those two sectors, and workers who remain suffer less from private inefficiency, may sur- in autarky are, in this sense, involuntarily vive because of "sclerosis," even though they "unemployed." would be destroyed in the efficient economy. In fact, because capital was assumed to be infinitely elastic, the full cost of inefficiency C. Macroeconomic Response to Shocks is borne by the unemployed. The autarky "wage" we is depressed in equilibrium so as Specificity together with incomplete con- to reduce the opportunity cost of labor and tracting not only affects levels, but also the help compensate capital for appropriationby cyclical features of the economy. We now turn insider labor. The latter, on the other hand, to the implications of appropriability for the suffer no welfare loss in the joint-production economy's response to aggregate shocks. sector, and receive Excessive Destruction.-Paradoxically, al- though scrapping is insufficient when com- (12) we = we + = W-Y-- e - 2 pared with the efficient economy, it is excessive given the level of creation. To see Technological Sclerosis.-We now turn to this, note that the social shadow wage in this the exit decision of a preexisting production economy is WE,while the shadow wage used unit with gross revenue x. From exit rule (6) in the private destruction condition (6) is in the microeconomics section, we know that pwJe + p)we > We (since wJe > We) VOL. 86 NO. 2 TRANSACTIONALMICROSTRUCTURE AND MACROECONOMICS 185
Excessive destructioncaptures an important and output declines during major adjustment aspect of the employment crises that are feared episodes. during major macroeconomic adjustment epi- sodes (see Caballero and Hammour, 1996b). III. Policy Considerations The coexistence of sclerosis with excessive destructionalso uncovers the fallacious nature, At a very general level, the multidimen- in our context, of the "liquidationist" argu- sional macroeconomic problem we have de- ment, which sees in a contraction a way to scribed in the previous section stems from improve welfare by reducing sclerosis. ill-defined property rights. These have long been of central concern to development econ- Elastification.-Consider the economy's omists, whether they affect the employment response to an aggregate shock to gross reve- relation stressed here, or others, such as exter- nues y. Differentiating the equilibrium em- nal finance. The design of institutions with the ployment expressions (8) and (11), one can ability to safeguard proper contract enforce- easily show that ment is a main ingredient of a successful development strategy, an old idea that has ac- dL/L dL *IL* quired heightened relevance in the recent East- dyly dyly ern European reform episodes. However, there is no need to go to such ex- In other words, the cyclical response of the treme episodes. The disruptive consequences incomplete-contracts economy is more elas- of ill-defined property rights are present even tic than that of the efficient economy.' A de- in economies with the most developed judicial cline in profitability requires a decline in the and regulatory systems. There are aspects of cost of labor to ensure capital of its required relationships that are simply too difficult to return. This is more difficult to achieve in the regulate contractually and to verify so as to presence of appropriable quasi-rents and re- fully protect the investing parties. Under those quires a proportionally larger decline in em- circumstances, and as a complement to insti- ployment. More generally, in Caballero and tutional reforms, there is a need to design Hammour (1996c) we show that, when the aggregate policies intended expressly to rem- unemployed factor is the less elastic of the edy the macroeconomic consequences of en- two, its cyclical responsiveness will inherit trenched appropriabilityproblems. some of the elasticity of the other factor in The first and most direct impact of appro- the presence of appropriability, a phenome- priability is underinvestment.Thus a key ele- non we term "elastification." ment in an effective policy package is some sort of creation incentive. But, in isolation, the Decoupling. -In tenns of the joint behavior benefits of such a policy may come at a sig- of creationand destructionflows over the cycle, nificant cost. Because of the wage pressures of excessive destructionleads to a "decoupling" increased hiring, a creation incentive may ex- of creation and destruction,characterized by a acerbate the excessive-destruction problem. destructionof jobs that is not accompaniedby a Thus, excessive destruction calls for another simultaneoussurge in job creation.In Caballero type of policy aimed at protecting existing pro- and Hammour( 1996a), we arguethat this man- duction units. But, in isolation, such a policy ifestation of inefficiency is broadly consistent does not address the need for more creation with what we know aboutjob flows; in Caballero incentives and, depending on the form it takes, and Hammour (1996b), we argue for its rel- may actually further depress creation. evance in accounting for massive employment The problem of these two approaches con- sidered in isolation is that, while partially cor- recting a problem along one margin, they fail to address-and often exacerbate-the prob- ' If 3 < A < 4, the elastification result is more stark. The efficient economy exhibits full joint-production em- lem on the other. A well-designed policy pack- ployment in that case, while the inefficient economy does age must not only raise the level of creation, not. but must also succeed in "synchronizing" de- 186 AEA PAPERS AND PROCEEDINGS MAY 1996 structionwith it. Only a well-balancedcom- . "On the Timing and Efficiency of binationof creationincentives and protection Creative Destruction." Quarterly Journal of the existingstructure can succeedin achiev- of Economics, 1996a (forthcoming). ing thosegoals. In fact, it is straightforwardto . "On the Ills of Adjustment." Journal show in our simplemodel thatboth problems of Development Economics, 1996b (forth- are solved by simultaneously implementing an coming). investmentsubsidy equal to wk, and a job- . "The Macroeconomics of Specific- protectionsubsidy L*s */2 to preexistingunits ity." Mimeo, Massachusetts Institute of if they remainin operation. Technology, March 1996c. This two-prongedapproach offers a useful Grout,Paul A. "Investment and Wages in the frameworkto discuss macroeconomicpolicy Absence of Binding Contracts:A Nash Bar- in a varietyof contexts.On adjustmentstrat- gaining Approach." Econometrica, March egies, it highlights the limitations of the 1984, 52(2), pp. 449-60. "gradualismversus cold-turkey" debate, which, Hail, RobertE. and Lazear,Edward P. "The Ex- in its purestforn, is a debateconfined within the cess Sensitivity of Layoffs and Quits to De- set of single-instrumentpolicies aimed at slow- mand." Journal of Labor Economics, April ing destruction.In Caballeroand Hammour 1984, 2(2), pp. 233-57. ( 1996b), we arguethat Korea provides a good Hart, Oliver and Moore, John. "A Theory of example of the managed-adjustmentpolicies Debt Based on the Inalienability of Hu- that stem from our framework,where invest- man Capital." Quarterly Journal of Eco- mentincentives through the creditmarket were nomics, November 1994, 109(4), pp. 841- complementedwith protection of domesticsec- 80. tors, as well as with a direct interventionin Klein, Benjamin; Crawford, Robert G. and labor marketsto keep wage demandscom- Alchian, Armen A. "Vertical Integration, mensuratewith productivitygains. Over the Appropriable Rents, and the Competitive business cycle, our approachspeaks on the Contracting Process." Journal of Law and soundnessof countercyclicalpolicies aimed Economics, October 1978,21(2), pp. 297- simultaneouslyat encouraginginvestment and 326. protectingemployment during downturns. Myerson,Roger B. and Satterthwaite,Mark A. "Efficient Mechanisms for Bilateral Trad- REFERENCES ing." Journal of Economic Theory, April 1983, 29(2), pp. 265-81. Caballero,Ricardo J. and Hammour,Mohamad Williamson,Oliver E. "Transaction-Cost Eco- L. "The Cleansing Effect of Recessions." nomics: The Governance of ContractualRe- American Economic Review, December lations." Journal of Law and Economics, 1994, 84(5), pp. 1350-68. October 1979, 22(2), pp. 233-61.