“Theory and practice should not contradict one another.” George J. Stigler

Sometimes researchers make claims that can grab headlines, but upon further investigation, their theories don’t line up with reality.

A recent paper and accompanying blog follow this pattern. They conclude that merger policy in tech should block more mergers, including mergers that would make consumers better off, because doing so increases innovation. That’s pretty surprising! Even if each approved merger makes consumers better off, the overall effect is consumer harm? How could that be?

The paper develops a theory in which blocking even good mergers makes more consumers drop their old services in favor of those offered by new entrants, leading to more successful startups. More startups mean more innovation because incumbents never innovate, according to this paper’s theory.

To show that the theory works in practice, the paper looks at nine tech acquisitions and concludes that, true to the theory, investment in startups declined once an acquisition occurred. Less investment means fewer startups, which means less innovation, according to the paper.

But does the theory really align with the cases as the paper and blog claim? No.

The theory makes a few key assumptions that drive its results. These include:

1. The entrant offers the same type of product as the incumbent, but better.
2. Early adopters have to spend money to learn to use the entrant’s new product, but only if there is no merger. There is no cost to learning about the upgraded product if there is a merger.
3. Customers can choose either the entrant’s product or the incumbent’s product, but not both (i.e., no multi-homing).
4. Customers don’t know how good the entrant’s product is before trying it and once they try it, they cannot switch back to the incumbent if they are disappointed. And if customers don’t leave the incumbent at their first opportunity, they never leave.
5. The incumbent never innovates.

These assumptions (and others) are pretty questionable. But let’s see if they align with the cases. The cases involve either Google (whose core products are search and associated advertising) or Facebook (whose core products are social media and associated advertising).

Of the nine acquisitions considered, none of the five key assumptions apply to:

• Google’s acquisition of YouTube (core product was video sharing)
• Google’s acquisition of Postini (core product was filtering email for spam and malware) [Note: It is unclear that Google was doing much in this space before the acquisition]
• Google’s acquisition of Waze (core product was crowdsourced mapping service) [Note: Customers appear to be divided on whether pre-Google Waze was better than Google Maps, but both continue to exist]
• Facebook’s acquisition of Instagram (core product was sharing photos linked into stories) [Note: Customers appear to be divided on their preferences for Facebook or Instagram, but both continue to exist]
• Facebook’s acquisition of WhatsApp (core product was texting using the internet)

Only the first assumption, and none of the others, applies to:

• Google’s acquisition of DoubleClick (core product was online advertising)
• Google’s acquisition of AdMob (core product was mobile advertising)
• Google’s acquisition of ITA Software (core product was searching for airline prices) [Note: Google’s general search might have been useable for finding prices, but probably not.]
• Google’s acquisition of Apigee (core product was web programming interfaces and analytics)

So four of the five important assumptions listed above apply to none of the cases. And the first assumption applies to only four of the nine cases. So the “theory” in the paper isn’t even a theory. As economist George Stigler said, a real theory aligns with evidence. Otherwise, what is offered as a theory doesn’t even rise to the level of fable.

There are other reasons to be skeptical of how the paper and blog use these cases. For example, Instagram has leapt in value since the Facebook acquisition, implying that Facebook improved the product (or at least its business model). And innovation is more than just an idea: Innovation requires an idea, turning the idea into a real product, developing a viable business plan, and executing the business plan.

Sometimes incumbents can do a better job with the last two steps than entrants, making an acquisition a good thing for all affected parties, including consumers. Microsoft’s acquisition of DOS is a clear example.

What’s the bottom line? As I have explained before, profit potential drives startups. And options have value to startups, including options of selling an idea or product to an incumbent that has proven business acumen.

(Disclosure statement: Mark Jamison provided consulting for Google in 2012 regarding whether Google should be considered a public utility.)

This article by Mark Jamison first appeared at the American Enterprise Institute.