Resumen
A key prediction of transaction cost economics is that the presence of relationship-specific assets increases the likelihood of vertical integration whenever contracts are incomplete. I explore a firm level data set on Eastern European and Central Asian firms, the BEEPS 2005 Survey provided by the EBRD and World Bank, to test this prediction. I measure lock-in by supplier substitution, and vertical integration by the presence of sales to the parent firm, and find the TCE prediction confirmed in the data: At the extensive margin, a firm whose customers are locked-in at medium or high levels is about 5 to 6 percent more likely to be vertically integrated than a firm whose customers are not locked-in. At the intensive margin, I find that high lock-in raises intra-group sales by about 2 percentage points. Being a large firm raises the probability of being vertically integrated significantly in itself, but does not alter the impact of lock-in on the probability of carrying out intra-group sales. Instead, operating in a non-manufacturing industry significantly reduces the probability of vertical integration, and also reduces the impact of high lock-in on the probability of having positive sales with a parent.