Counting Ghosts: The 19th-Century Circulation Fraud That Invented the Fake Engagement Economy

In 1914, the Audit Bureau of Circulations opened its doors in Chicago with a mission so obvious it bordered on embarrassment: it would simply verify whether magazines and newspapers actually reached the readers publishers claimed they did. The bureau's founding was not a proactive reform. It was a response to decades of systematic, industrialized lying—a reckoning that came only after the advertising industry had grown so weary of being defrauded that it built its own oversight institution rather than trust the sellers of attention to count their own audiences honestly.

The question worth asking, a century later, is why that institution had to be invented at all. And why, once invented, the same fraud reappeared almost immediately in every new medium that followed.

The Architecture of the Phantom Subscriber

By the 1880s, the American magazine industry had undergone a structural transformation. Subscriptions were no longer the primary revenue source. Advertising was. And advertising rates were set by circulation figures—numbers that publishers reported themselves, with no external verification, no audit trail, and no meaningful penalty for exaggeration.

The incentive structure was not subtle. A publisher who claimed 50,000 subscribers could charge significantly more per page than one who honestly reported 31,000. The difference between those two numbers was, in many cases, the difference between solvency and failure. So publishers inflated. They counted lapsed subscriptions, unsold copies distributed free to barbers and hotel lobbies, sample issues mailed unrequested to strangers, and in some documented cases, names drawn more or less at random from city directories.

None of this was secret, exactly. Advertisers suspected the numbers. Publishers knew they were fabricating them. What kept the system running was a collective tolerance for imprecision—a gentleman's agreement that everyone would pretend the figures were approximate rather than fictional, because the alternative meant confronting a market built on air.

This tolerance had a name, though no one called it that at the time. Behavioral economists would eventually label it "motivated reasoning." Advertisers wanted to believe the audiences existed because they needed the medium to work. Publishers needed them to believe it because their revenue depended on it. The fraud was, in a real sense, collaborative.

Exposure Without Consequence

The fraud was exposed repeatedly before anyone did anything about it. Trade publications ran investigations. Disgruntled former employees gave interviews. Competing publishers occasionally ratted each other out. None of it produced lasting reform.

What made the exposures ineffective was not ignorance but the absence of an alternative. An advertiser who stopped buying pages in inflated-circulation magazines did not thereby reach a better-measured audience. The entire industry operated on the same terms. Walking away from the fraud meant walking away from print advertising entirely, which was not a realistic option for companies that needed to reach consumers at scale.

This dynamic—where a known fraud persists because the alternatives are equally opaque or less convenient—is not a historical curiosity. It is the operating condition of every digital advertising market that has ever existed.

The Bureau That Couldn't Stop the Pattern

The Audit Bureau of Circulations did work, in the narrow sense. Publishers who submitted to its audits were, in fact, reporting more accurate numbers. Advertisers who bought only from audited publications were getting closer to what they paid for.

But the bureau could not audit radio. It could not audit television. It could not audit the web. Each new medium arrived with its own measurement problem, its own self-reported metrics, and its own version of the phantom subscriber. Radio networks in the 1930s cited listener estimates produced by their own research arms. Television ratings were, for decades, extrapolated from a sample of households small enough that the margin of error was larger than many shows' actual audiences.

When the internet arrived, it brought with it a measurement apparatus that looked, superficially, more rigorous than anything that had preceded it. Clicks were counted. Impressions were logged. Every interaction left a trace. What the medium also brought, inevitably, was a new class of phantom: the bot.

Bots Are Subscribers Who Never Unsubscribe

The bot follower is the 19th-century phantom subscriber with better infrastructure. Both are units of apparent attention that exist only as entries in a ledger. Both inflate the apparent value of an audience to attract payments from advertisers or platform partners. Both are sold with sufficient plausible deniability that the buyer cannot easily prove the fraud without building, as the advertising industry eventually did in 1914, an entirely new verification institution.

What the historical record adds to this analysis is a prediction about what happens next. After the Audit Bureau was established, publishers did not stop inflating their audiences. They shifted the inflation to metrics the bureau did not measure: newsstand pickup rates, pass-along readership, demographic quality claims that could not be audited because they were not strictly quantitative. The fraud did not end. It migrated.

Every platform crackdown on fake engagement has produced the same result. When Facebook purged bot accounts in 2018, the industry responded not by abandoning artificial inflation but by developing more sophisticated methods of generating engagement that looked organic. Click farms replaced simple bots. Coordinated inauthentic behavior replaced crude follower purchases. The underlying psychology—that buyers of attention will accept the metrics they are given if the alternative is abandoning the market—remained constant.

What the Pattern Tells Us

The durability of circulation fraud across five distinct media generations is not a story about bad actors or insufficient regulation, though both have played their roles. It is a story about the structural conditions that emerge whenever attention is commodified and measurement is delegated to the seller.

Human beings are not well-equipped to resist motivated reasoning when the alternative is costly. Advertisers in 1890 who audited their own media buys would have discovered that their campaigns were reaching far smaller audiences than they had paid for—and would have had to either absorb that loss or restructure their entire marketing approach. Most chose not to look too closely. Their successors, managing programmatic budgets in 2024, are making the same calculation in the same cognitive register.

History does not repeat because the actors are lazy or corrupt, though they sometimes are. It repeats because the psychological architecture that makes a particular fraud viable does not change between centuries. The medium changes. The incentives change. The human tendency to believe convenient numbers, and to punish those who insist on inconvenient ones, does not.

The Audit Bureau of Circulations is still operating, more than a century after its founding. It has expanded its scope to include digital metrics. Whether that expansion has meaningfully reduced the gap between claimed and actual audiences is, perhaps appropriately, a matter of some dispute.