FOREIGN KNOW-HOW, FIRM CONTROL,ANDTHEINCOMEOFDEVELOPINGCOUNTRIES∗Ariel T. Burstein and Alexander Monge-NaranjoJune 17, 2008Management know-how shapes the productivity of firms, and can be reallocated acrosscountries as managers acquire control of factors of pro duction abroad. We constructa quantitative mo del to investigate the aggregate consequences of the internationalreallocation of management know-how . Using aggregate data, we infer the relativescarcity of this form of know-how in a sample of developing countries. We find thatdeveloping countries gain, on average, 12% in output and 5% in welfare (with widevariation across countries) when they eliminate policy barriers to foreign control ofdomestic factors of production.I. IntroductionThe diffusion of productive knowledge has a prominent place in the literatureon cross-coun try income differences. Much of the attention has been on flows of∗We thank Manuel A m ad or, Pol Antràs, A nd y Atkeson, R obert B arro , W ilb ur Jo hn Cole-man, Pete Klenow, Éva Nagypál, G iorgio Primiceri, A ndrés R o dríguez-Clare, E steban R ossi-Hansberg,MartinSchneider,AlehTsyvinski,andfive referees for use ful comments a nd sug-gestions. G ilb erto A rce provided sup erb research assistance.1know ledge that is embedded in patents and internationally traded goods, andon the role of cross-country spillovers in the diffusion of ideas. There has beenless focus on the transfer of productive know-how that takes place when firmmanagemen t crosses borders and directly controls factors of production in aforeign country.1These transfers appear to have gained importance with theexpanding multinational activity of recent years.2In this paper, we investigate the aggregate consequences of reallocating firmmanagemen t know-how across countries. In a world economy in which firmscan reallocate and produce abroad, it is important to distinguish between thefixed productivity components of a country and those that are internationallymobile. We refer to the immobile components of productivity, which impact allfirms operating in a country (suc h as infrastructure, regulations, and naturalamenities), as “country-embedded productivity.” In contrast, we refer to the in-ternationally mobile components, which generate productivity differences acrossfirms operating within the same location, as “firm-embedded productivity.”At the aggregate level, separating country- from firm-embedded productivityis not straightforward. For given levels of capital and labor, a combination ofhigh country- and low firm-embedded productivity can lead to the same observedoutput level as a combination of low country- and high firm-embedded produc-tivity. However, if firm-embedded productivity flows from countries where itis relatively abundant to countries where it is scarce, then country- and firm-embedded productivities can be separated. A high share of capital and laborcontrolled by foreign firms indicates that the domestic level of firm-embeddedproductivit y is lo w relative to country-embedded productivity.1. Barro and Sala-i-M artin (1998), Keller (2004), and K lenow and Ro driguez-Clare (2005)review th e literature on t echnology d iffusio n a n d its i mplic a tio n s o n cross-c o u ntry in c o medifferences.2. See chapter one of Barba Navaretti and Venables (2004) and references therein for anaccount of the impressive growth of multinational firms d uring the last two decades of the20th century.2We develop this logic in a quantitative model of the world allocation offirm-em bedded productivity. We use the model and aggregate data to disentan-gle country- and firm-embedded productivities in a set of developing countries.Then, we conduct counterfactuals on policies that restrict the control by foreignfirms of local factors of production.Our model follows a long tradition in the literature that links firm-embeddedproductivity to the management know-how and skills of individuals leading thefirm.3Management know-how is similar to codified technological knowledge, asit can be reallocated across sectors, regions, and, albeit imperfectly, across coun-tries. But management know-how differs from codified technological knowledge,as it is to a large extent a rival factor requiring the holder’s direct involvementwhen making the critical decisions facing a firm.4Since time and attention arelimited, the use of know-how in one task or location entails an opportunity costin another.5The reallocation of firm-embedded productivity across countries in our model,is consistent with the observation that multinational firms in developing coun-tries rely largely on expatriates from the source country for “senior managementpositions and key technical and engineering jobs to execute sophisticated or spe-cialized production tasks” (UNCTAD 1994, p. 238). Bloom, Sadun and VanReenen (2007) and Bloom and Van Reenen (2007) show that multinational firmsshape their productivity abroad by transplanting their organization structuresand management practices.63. See, for example, Kaldor (1934), Lucas (1978), Rosen (1982), Oi (1983), and Garicano(2000). F irm -emb edd ed pro d uctivity is also related to “organ ization capital” (Prescott andVisscher 1980), which also includes th e firm-sp ecific knowledge of non-managem ent workers.4. Firm -le a d in g skill s and m a n a ge ment k n ow -h ow are to a larg e ext ent a “ ta cit kn owing”,as defined by Polanyi (1967), that lies within the individuals m aking the critical choices ofthe firm. This know -how is costly —if not imp ossible— to describe and teach to others. Kaldor(1934) highlights the limited span of control in the co ordinating role of m anagement.5. The rival nature of firm-embe d d ed prod u c tiv ity is at the core of a larg e litera tu re on thefirm-s iz e dis trib ution (e .g . Luc a s [19 78 ], an d Atkeso n an d Keh oe [20 0 5 ]).6. Chari, Ouimet, and Tesar (2007) find that firm s from developed countries obtain positivereturns from a cqu iring fir ms in emerg in g m a rket s only when the y acq u ir e a c o ntro llin g sta ke.3We extend a standard neoclassical model by introducing management know-how as an additional factor of production. As in Lucas (1978), firms are teamsof managers, workers and capital. Each country has domestic supplies of la-bor and of management skills, both of which can change over time with theoccupation choice of individuals. Managers can reallocate their skills and leadfirms in foreign