Business lore and business school classes are full of stories about vertical integration. They range from the classic tales of Carnegie Steel Co., Ford Motor Co.’s River Rouge complex, and General Motors Corp.’s 1926 acquisition of Fisher Body, to Boeing Co.’s recent decision to bring production of the 787 Dreamliner in-house after costly delays at its suppliers.

What these stories share is the notion that companies vertically integrate so they can ensure ready supplies of key inputs.

Yet a closer look at the data shows that, even among vertically structured firms, this sort of self-service is the exception rather than the rule. In a recent study, we tracked the flow of products through the economy using data on millions of individual shipments of goods from tens of thousands of U.S. firms.

This detailed information comes from the Commodity Flow Survey. It lets us see who makes what; how much it is worth; and where it is sent. In particular, it shows whether the upstream units of vertically integrated companies did, in fact, send their output to their downstream units for further processing.

Upstream Units

We were surprised to find that instead of keeping those potential inputs inside the company, the upstream units sell the vast majority of what they make to others. It wasn’t just that the vertically integrated companies make more inputs than they need and sell the remainder. Those selling their upstream production would simultaneously buy similar products from others to use as inputs for their downstream operations.

For example, some construction-products companies made both cement and concrete products (cement is an input in concrete production). Yet those companies sent most of the cement they made to other concrete producers while purchasing much of their own cement needs from other firms.

We found that vertically integrated companies sell more than 80 percent of what they make by way of inputs to other firms, rather than using those products themselves. Even this number is skewed by a few large Boeing-like exceptions.

About 40 percent of vertically integrated firms use none of their own upstream production, choosing instead to procure their inputs from outsiders.

If most vertically integrated companies aren’t supplying themselves, then why do they own production chains? It is harder to get a clear answer to this question from the data, but there are some clues to an explanation. Trying to understand why companies own production chains by looking at the way goods flow along the chain could be misleading. Instead, it might be the things we can’t see -- managerial oversight, marketing know-how, customer contacts, intellectual property, and other information-based capital -- that drive most vertical integration.

By integrating, companies can spread these types of capital over the production chain. Integrated firms appear to let the market and contracted suppliers handle most of the flow of tangible goods along the chain, while using control through ownership to apply the necessary intangible capital, which is by nature harder to write contracts over.

We found some evidence for this sort of spreading of intangible capital. When ownership of a business establishment (a manufacturing plant, say) changes hands, the plant reorients its activity toward its new owner. It sells less to its old customers and more to those that its acquirer had already been serving.

Acquirer’s Contacts

While that’s not completely surprising, it must mean that the plant’s acquirer transfers some sort of firm-specific sales contacts to the plant. Perhaps more notable is that the acquired plant doesn’t just serve different customers, it often changes what it produces. It stops making the products that were the focus of its old firm and shifts toward the types of products its new firm specializes in. This sort of transformation of the very purpose of the plant almost surely requires some sort of intangible capital inputs from the acquirer.

If this explanation for integration is correct, companies’ expansions along production chains may not be altogether different than their horizontal expansions, which occur when firms start operating in markets related to their current lines of business with the hope that their capital can be applied profitably to the new markets.

Vertical expansions may operate under the very same principle. After all, industries that supply a firm’s inputs are clearly related lines of business. Some companies may expand into them in the same way they would enter a new product market.

In other words, companies choose to grow in all sorts of directions. Some happen to do it along supply chains, but they are essentially doing it for the same reason as all the others.

We shouldn’t expect that the classic “make or buy” input decision is always part of the consideration. And the data show that, indeed, it usually isn’t.

(Ali Hortacsu is professor of economics at the University of Chicago. Chad Syverson, a professor of economics at the University of Chicago’s Booth School of Business, is a contributor to Business Class. The opinions expressed are their own.)

Read more opinion online from Bloomberg View.

To contact the writers of this article: Ali Hortacsu at hortacsu@chicago.edu; Chad Syverson at chad.syverson@chicagobooth.edu

To contact the editor responsible for this article: Max Berley at mberley@bloomberg.net