A Muddle on Vertical Integration
21 August 2006 at 8:58 pm Peter G. Klein 4 comments
| Peter Klein |
Slate’s Daniel Gross offers a hopeless muddle on vertical integration in his latest Moneybox column. Citing several examples of forwards integration mentioned in a recent WSJ article, Gross proclaims “the return of one of the great industrial developments of the late 19th century: vertical integration.” But the reasoning is a mess.
First, there’s no systematic evidence that vertical integration ever went away. We have plenty of anecdotal information about outsourcing, increased use of networks and alliances, “refocusing,” and the like, but little systematic evidence for a comprehensive, economy-wide trend toward vertical dis-integration.
Second, it’s hard to believe that Gross produced an article on vertical integration without consulting a single IO economist or intermediate textbook. He alludes to scale and scope as explanations for the rise of vertically integrated enterprises in the early twentieth century, but is seemingly unaware of the transaction cost or property-rights rationales for vertical integration (let alone double-marginalization). We’re told that firms integrate “to ensure they have adequate supplies of raw materials and parts.” But supply uncertainty can be reduced by contract. Why incur the additional transaction costs of hierarchy simply to ensure supplies of off-the-shelf commodities?
Next, we’re told that during the 1980s and 1990s, “[t]he tenets of re-engineering and focusing on core competencies led firms to focus on performing only those functions they could carry out profitably and to outsource everything else.” Globalization also gets a nod. Again, however, systematic empirical evidence on refocusing is sketchy, and invoking fuzzy notions of core competencies hardly constitutes an explanation.
One doesn’t expect theoretical and empirical rigor in a magazine article. But surely Slate can do better than this.
Update: bobvis didn’t like it either.
Entry filed under: - Klein -, Business/Economic History, Strategic Management, Theory of the Firm.
Organizations and Markets 
1.
Bob V | 22 August 2006 at 6:40 pm
Ajit Kambil tied investments in information technology to vertical integration. This was using data from the Bureau of Economic Analysis. Since IT investments have been increasing, we can guess that there has been a shift away from vertical integration. (Caveat: this isn’t economy-wide; he showed it only for the manufacturing sector.)
Cite:
Guerin-Calvert, M. E. and Wildman, S. S. Electronic Services Networks: a business and public policy challenge, Praeger, 1991.
Chapter by Ajit Kambil: Information technology and vertical integration: evidence from the manufacturing sector, pages 22–38.
2.
Peter Klein | 22 August 2006 at 8:18 pm
Thanks for the cite, I hadn’t seen this paper.
I’m not sure I see why increased IT investments are an indicator of dis-integration, however. Do we know that this IT was for managing relations with external suppliers or customers? Couldn’t it also be investments in improved production technology, internal ERP systems, and other intra-firm processes?
3.
Ajit Kambil | 5 September 2006 at 10:59 pm
Vertical Integration.
The work cited above was based on the preliminary analysis for my Masters thesis in 1989 – chaired by Tom Malone at MIT. Vertical integration was measured using a value added over sales indicator at the 2 digit SIC level in manufacturing. Similarly the IT investment data was at a high level of aggregation. We were exploring the hypotheses of an increased shift toward markets enabled by IT reducing transaction costs. The vertical integration as well as the IT data was thus at a very aggregate level. Firm size and other dependent variables were explored by Erik Brynjolfsson in his thesis.
Does Information Technology Lead to Smaller Firms?
EJ Brynjolfsson, T Malone, V Gurbaxani, A Kambil – Mgt Sci 1994 reports on the broader data.
The original thesis was that IT would lead to more electronic markets and thus less vertical integration as firms could contract more efficiently in the market, and benefit from the economies of scale of producers who aggregate demand from multiple buyers.
I am not a great fan of econometric analysis of such broadly aggregate data. We may learn something directional but after my masters thesis I shifted to look at institutional arrangements in more detail through rich cases. However, since we have more data since 1989 – it may be a good doctoral student exercise to plot and review trends in vertical integration using the crude measures such as Adelman’s value added over sales or other more refined metrics. Doing this in manufacturing may be a good starting point.
I do think IT enabled more efficient coordination, contracting and monitoring – reducing transaction costs and encouraging greater outsourcing across the value chain. IT has been vital to the growth of business process outsourcing, contract manufacturing etc. This in some measure has contributed to the vertical disintegration of firms, and the focus of firms on specific aspects of the value chain. But as you point out its hard to measure vertical integration across the broad economy and to establish a definitive causal relationship.
I also think the transaction cost approach while powerful should be complemented by other perspectives. I have long felt it is generally a static equilibrium model and does not consider the costs of transitioning from one form of organization to another. This implications of transition costs are illustrated in some case examples from my last book analyzing the success and failures of various electronic markets (www.makingmarkets.org)
regards
Ajit
4.
Peter Klein | 5 September 2006 at 11:12 pm
Ajit, thanks very much for the comments. I don’t at all disagree that IT has lowered the cost of transacting between firms. My point (on IT) was only that IT may also lower the costs of internal organization, such that the net effect on firm size and scope is ambiguous. And I agree that TCE (and the property-rights approach) offer plausible comparative-static explanations of the relevant trends, but not very rich dynamic ones.
Finally, I agree strongly that exploring the more recent data would make an excellent PhD thesis. Any takers out there? Or are there already some more recent empirical studies that address the question satisfactorily?