I’m a fan of capitalism because history proves it works. Humanity has grown both populous and rich in lockstep with the rise and sometimes the decline of market economies and market freedom. Capitalism works mostly because it is more like a natural law than a human institution or design. Natural laws don’t play favorites and don’t change the rules to benefit themselves. Human planners make mistakes while natural laws just keep on ticking.
But does capitalism always work? Being a libertarian, I’d certainly like to think so. Imagine my surprise on seeing a possible flaw, a case where capitalism appears to blow up, what physicists call a singularity. One example of a singularity is a black hole, a point so heavy that it collapses all nearby matter. No one likes natural laws with singularities, just as no one wants to live near a black hole or travel among land mines. I’m no economist, so I may have this wrong. But it seems to me that Google could blow up the economy, at least in theory, just by succeeding in the free market. I think Google exposes a potential singularity in capitalism.
As you know, Google is an Internet search engine for finding web sites, sort of a modern-day TV Guide. Google does a lot of things right, but its key success was discovering a way to measure web site quality. Actually, Google’s Page Rank algorithm for grading web sites does not measure real quality, which after all is in the eye of the beholder. Instead it determines a site’s popularity with other web sites, counting others that link to it. More links imply web authors find the site useful, and while that may not be quite the same as usefulness to web users, it is a reasonable approximation.
Page Rank and other factors led to Google.com becoming immensely popular, itself having the highest ranking of all. A large majority of people search the Internet using Google. But what if you have a web site or web store that isn’t highly ranked on Google search results? In the Internet era, there may be almost no other way for people to find you. Google’s Page Rank algorithm assumes that quality and popularity are related, but of course nothing is popular at first until people learn about it.
Fortunately, Google solves this for us by including paid advertisements next to its search results. Google prices these ads using an auction, the oldest and purest form of capitalism. If you bid more than other advertisers, your ad appears. If you don’t pay as much as others, then you don’ get seen unless you are already popular enough to be in the non-advertising results.
Currently, Google is virtually the only game in town. Other search engines exist, and in the last few years they have duplicated most of Google’s Page Rank algorithm and so can provide similar results, but none come anywhere close to Google’s search volume. There is no good reason for Google users to switch, since the others do basically the same thing, just maybe a little worse due to Google’s advantages of scale and reputation. Right now it does not appear Google’s dominance is fading with increasing competition.
Also, the Internet is increasingly the only game in town. Newspapers and magazines are in decline. Even television is loosing audience to the Internet. TV channels are proliferating and fragmenting to the point that they may eventually be just a type of Internet video. If you want people to find you, increasingly your only recourse is a web site and Internet advertising.
Capitalism Blows Up
Here’s the hitch. If Google and the Internet are dominant, then Google advertising is also the only game in town and can be priced accordingly. Given Google advertising is efficiently auctioned, your competitors have an incentive to bid higher until you are no longer making any money. Your competitors probably similar products and costs (and if not, they will eventually in a free market). So if you are making money and bidding enough for customers to find you, then your competitor can win those customers instead with just a bit less profit per sale by bidding a bit higher.
In an efficient market (and an Internet auction is about as efficient as things get), the eventual result is that all competitors bid high enough that their profits are zero. In theory, all profits go to Google. And you wondered where Google gets the money to give away so many great free Internet services! So Internet advertising vacuums all profits out of the economy and into Google’s coffers, at least in theory. That is the “Google Singularity.”
Of course there are some exceptions. For one thing, people can find a local business without the Internet, usually because they drive by your location. Also, once a customer finds you, their repeat business can be yours without paid advertising. Maybe you make no money on the first sale, but when your new customer comes back for more, they can find you without Google. Most companies get repeat business because they sell consumables, and maybe they are fine. Though I wonder whether an advertising auction with enough market efficiency might eventually result in bidders tapping the expected profits of future repeat sales, so they pay all future profits just to find a new customer.
I am particularly sensitive to this potential problem, since I sell software, the ultimate non-consumable that never wears out. New software versions may be better, but they usually also fix old problems and I am reluctant to charge old customers to correct my mistakes. Plus you don’t sell software from a factory-direct storefront, so the above exceptions need not apply to package software.
There is another exception that does help even my business. Google’s advertising auction is intentionally inefficient. It may show an ad infrequently even with a low bid , because Google takes into account whether the ad is useful to searcher users and gets clicks. This may be a facet of Google’s “do no harm” philosophy or just a good business tactic. Either way, its Internet advertising cannot be totally efficient extracting customer profits. But in my experience with my kind of company, most profits really do go to Google. No other form of advertising works well and Google costs almost as much as we make no matter how we tinker with the ads.
The Google Singularity is the product of market efficiency, what I would call a “frictionless” market. Because the Internet connects so many people at such low cost, the market communication involved in signaling products and prices can become negligibly small. Does that endanger capitalism?
At first it appears that other markets could blow up in theory if they become too efficient. With frictionless communication, the best product should win and become a worldwide monopoly, hoovering all profits. Markets cannot work without competition, without product choices.
But frictionless communication should also allow arbitrarily narrow product categories. Cheap Internet communication increases the potential market size for a product, making it available to more consumers, so more targeted product categories can emerge. As the best product starts growing into a monopoly, related new offerings that better satisfy some subset of consumers can emerge, talking sales from the emerging monopoly. We end up with more widely sold products that are more narrowly targeted, instead of diverse products sold “regionally” where communications costs were low even before the Internet. There is still competition and consumers have an even wider choice. So as far as I can see, the Internet is not a theoretical doomsday machine, at least in general.
That still leaves my Google Singularity, the chance that Internet search could become the communist dream, a capitalist invention that destroys capitalism by becoming so profitable that no wealth is left for other productive endeavors. Without profitable products, even Google eventually implodes because no advertisers are left standing, and we’re all “Back in the U.S.S.R.”
As I see it, the real problem is that Google seems to be a natural monopoly without good competitive alternatives. And not just any monopoly, but one that is potentially the gatekeeper to our entire economy. When there are competitors, prices tend to fall toward the level of product costs. But without competition, prices rise to the level of the economic benefits, costs plus profits.
Of course, one solution is government action to correct the problem, possibly legislating away Google’s monopoly. I don’t believe Google breaks current anti-trust laws, since having a monopoly is legal if it is not wielded to harm consumers, but one could always dream up a new law. However, we libertarians don’t often favor government solutions. Government is an even more virulent monopoly, far better than even Google at vacuuming wealth.
A better solution would be to find Google a reasonable competitor, an alternative Internet search that offers users something Google does not. With competition, the market would reach a balance that keeps advertising costs in check rather than falling into a monopolistic black hole.
Improving Internet Search
Maybe the real problem is that Google’s competitors all rushed to do pretty much exactly what Google does. Google’s Page Rank approach was so good that competitors duplicated it rather than looking for something even better. The result is that Google clones offer no reason for users to switch. In a traditional market, a clone product can steal customers with a lower price. But Internet search is free and the users are not Google’s customers. Search competitors can lower advertising rates, but cannot offer users a cheaper solution than free, so Google clones are not really competition. People have no reason to switch and duplicating Google was a competitive dead end.
A viable alternative to Google needs to have a different kind of Internet search that helps some users more reliably find what they seek. Of course it has been tried, but so far without success. Given the possible Google Singularity, not to mention the immense potential advertising profits, I think we need to try harder.
My suggestion would be a blend of Amazon’s product reviews with Google’s automated search. Remember that the Page Rank algorithm only approximates site quality. It is really a sort of muted gauge of site popularity. What searchers really want to know is whether credible users like a site, not whether it is popular with other web sites. A better search would more closely estimate “quality” in the eye of the user.
Amazon’s great idea was to allow users to review products, and keep the reviews credible by giving reviewers a public identity and reputation established by user ratings of the reviewers. Amazon’s system works when comparing a few dozen competing products, but would bog down comparing thousands of sites relevant to an arbitrary search query. If I’m searching for dish washing jobs, it will not help to rank restaurant web sites using customer reviews, however credible they may be. I might want to see the worst rated restaurants because they might have more trouble finding good help. What I really want would be reviews by people like me, aspiring job applicants in this case.
Yet slicing and dicing thousands of reviewers is exactly the kind of thing that automated search servers might do. Maybe Internet search could be based on and aggregated Amazon-style reviews as long as credible reviewers are categorized in various ways, so that only those best matching the user’s interests and values apply to a given search. I think that might be a better Internet mouse trap that would give Google some real competition.