Today marks the 20^th anniversary of Wikipedia’s website going live. The online, collaboratively sourced encyclopedia is one of the internet’s biggest success stories, but one that, on the face of it, conventional economic analysis would suggest was the least plausible.

The free, volunteer‐​edited site today hosts 55 million articles in 300 languages, including 6.2 million individual content pages in English that have been subject to almost 100 million edits. The Wikipedia page about Wikipedia itself cites articles claiming it is the 13^th most popular site on the internet, with 1.7 billion unique visitors and 20 billion page views per month.

Some academics remain snooty about an encyclopedia that can be “edited by anyone,” but Wikipedia is often a go‐​to website for even established researchers looking for a quick overview of conventional wisdom or to double‐​check a fact cited elsewhere. The site has been adopted by companies such as Amazon and Apple to answer factual questions in voice assistants and smart speakers, and by social media companies such as Facebook to provide information links on posts.

There is good reason for this. A 2005 Nature study found that Wikipedia had decent accuracy compared to the expert‐​written Encyclopedia Britannica. Why? Aggregating information from a wide and diverse editor and reader base helps quickly correct obvious errors, especially on contentious and highly read topics. The articles that are most read then, over time, tend to be more accurate. Embedded within Wikipedia’s model is a market‐​style feedback mechanism that ensures resources head to improve product quality where demand is highest. And, of course, because it utilizes new, rapid internet technologies, this collaboratively sourced website can update much more quickly than the book‐​bound encyclopedias of yesteryear.

In a world of filter bubbles and media segregation, Wikipedia has also proven somewhat of a mediating resource. Analysis has found that a weak version of Linus’ Law holds for Wikipedia—that “Given enough eyeballs, all bugs are shallow.” On articles with plentiful contributors, such as political pages, the large number of editors combined with Wikipedia’s own evolving protocols helps achieve a high degree of neutrality.

This is an underappreciated success. Public debate focuses on the idea of internet gatekeepers “censoring” free speech. Wikipedia has largely avoided that accusation to date, despite being a highly read source with explicit safeguards to try to eliminate fake news and ill‐​sourced opinion. True, errors or slant are harder to police on less well‐​read pages. Wikipedia’s founder Jimmy Wales acknowledges that on certain niche topics, only those are who strong fans of the subject will tend to contribute, meaning the sentiment of articles will be biased. But in all Wikipedia is well‐​trusted and widely believed to be doing a reasonably decent job, especially compared to the media. Some research even posits that engaging in Wikipedia edits causes contributors on political topics to become less slanted over time. Imagine!

Of course, such an open model is made possible by Section 230’s intermediary liability protections. As a nonprofit enterprise, the Wikimedia Foundation cannot afford to litigate the decisions of its volunteer editors. Section 230 allows it to both avoid liability for user errors and rectify errors without provoking litigation from those unhappy with the changes. While discussions of the law often focus on the judgments of ‘Big Tech’ moderators, it’s important to remember it safeguards the ongoing editorial judgment of Wikipedia’s hivemind.

As a non‐​profit, Wikipedia is part of the rich tapestry of organizations that arise in a free economy. Often public debate overly focuses on “the market” versus “the government.” Wikipedia’s success as a free content, non‐​profit institution highlights how within free economic systems intermediate organizations arise that can develop safeguards and standards that achieve the desired ends willed on by people calling for heavy‐​handed government regulation, while obviating the need for that path which crushes new innovation.

Indeed, the site should be of particular interest to libertarians. Jimmy Wales is on record at Cato as highlighting that the inspiration behind it lay with his reading of Friedrich Hayek’s “The Use of Knowledge In Society.” That famous essay explained a key reason why central planning couldn’t match the efficiency of an open market. The market order is dictated by prices, which themselves reflect a host of locally embedded information that no single planner could ever comprehend or collect. Wales took the message to heart in regards to knowledge, thinking he could build something that could harness that localized, dispersed, and niche knowledge among individuals to produce a resource as globally comprehensive as possible.

Wikipedia is a useful case‐​study for economists on certain other policy issues.

First, its existence shows how services with public good‐​like characteristics can be provided in a free economy. Knowledge is something that is non‐​rivalrous (me knowing something doesn’t “use it up” and so prevent you from knowing it) and non‐​excludable (you can try to raise barriers to its acquisition but it can still spread relatively freely). Introductory economics textbooks would say these characteristics mean such a good or service would be underprovided in a free economy, requiring some state subsidies or provision. Indeed, some Wikipedia employees have been known to joke “Thank God our little enterprise works in practice, because it could never work in theory”.

Yet it turns out donors worldwide are willing to fund the knowledge venture because they find it useful or buy into the vision, while enough editors volunteer to participate because they find satisfaction and usefulness themselves from the pastime. Mastering Wikipedia’s rules and guidelines make this a bit like Terence Kealey’s “contribution good” idea.

Second, the website reminds us that network effects, other things given, enhance consumer welfare, rather than diminish it. In the Big Tech antitrust debates it’s common to read that competition is inherently stifled in social media or the search engine sectors because users find services more useful when large numbers of other people are using them. This is said to constitute a major “barrier to entry” for competitors. But from the consumers’ perspective, it’s good that certain companies have high usership rates, as this improves the quality of the product.

In the case of Wikipedia, it should be obvious that the readers benefit from large numbers identifying errors and removing biases. Yet nobody says “Wikipedia is unfairly monopolizing the online encyclopedia market,” because we recognize those benefits arise from an open, competitive process. So while there might be other anti‐​competitive conduct charges against Big Tech companies, the existence of “network effects” should be separated out from other issues and not talked about as if they are a “bad.”

Finally, Wikipedia is a great example of how internet‐​based products have enhanced human welfare in ways not picked up in conventional GDP statistics. The decline in the purchase of physical encyclopedias would show up as a decline in measured market activity. But because of innovations such as Wikipedia, we now have access to more information than they provided at a zero out‐​of‐​pocket cost. Research from two years ago estimated U.S. consumers valued Wikipedia then at $150 per year. That’s some consumer surplus. For context, the cost of Encyclopedia Britannica’s 32‐​volume, five‐​yard‐​long set in 2012 was $1,400.

Sure, one can quibble that the way Wikipedia operates entrenches consensus positions and treats certain controversial ideas harshly. One can find examples of mistakes on the site, or bias or slant. But the counterfactual is the imperfect world we live in, just without Wikipedia. Human beings themselves exhibit biases every day. At least in‐​built into the editorial process is a means of correction. Even the Wikipedia page about Wikipedia is pretty critical!

Wikipedia itself will have to evolve as more internet‐​activity moves away from the desktop or laptop onto smart technologies. But for now, we should see sites such as Wikipedia as a testament to the open, collaborative opportunities a free economy allows.

2020 was a challenging year for the restaurant industry. The National Restaurant Association reports that 17 percent of restaurants in America have closed, with many more on the brink of failure. Restrictions on capacity and indoor service during the coronavirus pandemic forced many restaurants to pivot to a carry‐​out or delivery model.

For restaurants without a pre‐​existing, self‐​run delivery system, delivery systems like DoorDash and GrubHub were critical to reaching customers. These services publicize restaurants, process orders, and deliver food via independent couriers. These companies remain profitable by charging restaurants and customers fees for this convenience.

However, lawmakers around the country have begun to crack down on what they consider unreasonable fees. Los Angeles, New York City, Seattle, Chicago, and Washington, DC are among the cities that have imposed caps on the fees that delivery services can charge. These caps are misguided.

Introducing price controls will disincentivize delivery services from partnering with low‐​volume or remote restaurants, from increasing contractor pay, or from improving service. To make up for the revenue shortfall from restaurant fee caps, DoorDash and GrubHub will likely increase delivery and service fees for customers, which will then reduce demand for these services and ultimately harm restaurants.

Trying to protect restaurants while increasing costs to consumers is thus short‐​sighted. Many restaurants have failed, and many more will. Fee‐​limiting policies targeting the delivery middlemen will at best prolong the inevitable and more likely hasten that outcome.

News reports indicate that COVID-19 vaccinations in the United States are happening more slowly than officials promised and in comparison to other countries. Some health care providers are racing — and some are failing — to administer their stock vaccines before they expire. In one case, more than 500 doses spoiled when an employee removed them from a freezer, allegedly on purpose.

These episodes highlight just some of problems inherent to using government rationing rather than market prices to distribute vaccines.

If manufacturers and retailers could charge, and consumers could pay, whatever they like, distribution of COVID-19 vaccines likely would be more rapid than today and likely none of the existing stock would expire.

Anecdotal reports indicate that some consumers are willing to pay $25,000 to receive the vaccine. If retailers could make thousands or even hundreds of dollars in profit from every dose they sell, they would have the incentive and the ability to invest greater resources in securing the vaccines and distributing them quickly. They would take greater steps to protect the vaccines from sabotage by deranged employees. (The law could also do so, if penalties for destroying the vaccines rose with the market valuation of each dose.) They would hire more personnel (e.g., nurses) at higher‐​than‐​usual wages to organize distribution and/​or to administer the vaccines. They could even train more personnel to administer vaccines — and form a lobby to demand that states suspend government regulations that prevent them from doing so.

If manufacturers could make thousands or even hundreds of dollars in profit from every dose they sell, they would have greater incentive and ability to expand production and produce more vaccines faster.

Would market prices guarantee that vaccines would go to the highest‐​value recipients first? Not at all. But market allocation doesn’t have to be perfect. It just has to outperform the alternative of government rationing.

In an earlier post on government vs. market distribution of COVID-19 vaccines, I wrote, “If the government could allocate vaccines in a way that gets more of them to the highest‐​value recipients than market forces would,” then government distribution would be defensible. But, “To improve on market forces, government must actually know who the highest‐​value recipients are[,] actually be able to allocate vaccines on that basis, [and] not detract from whatever good market forces would do on their own, or…diminish the incentives for pharmaceutical companies to boost production.” That does not appear to be happening.

Government rationing is detracting from the good that market forces would do, and slowing the distribution of COVID-19 vaccines, by diminishing the incentives for speed and security on the part of manufacturers and retailers. In many cases, it is resulting in low‐​value recipients receiving vaccinations before high‐​valued recipients do. It is diminishing the incentives for manufacturers to accelerate production. And in some, cases it is costing lives by allowing vaccines to spoil.

Rachel Handler has a delightful piece at New York magazine’s food and restaurant blog Grub Street on how Big Pasta is using government regulation to punish competitors and consumers. The result is that the U.S. Food and Drug Administration, in addition to causing a shortage of COVID-19 diagnostic tests and vaccines, is basically causing a nationwide shortage of bucatini.

On March 30, at the beginning of a pandemic whose supply shocks were making everything from toilet paper to pasta harder to get, the FDA blocked imports of De Cecco bucatini. The FDA found the iron content of the Italian company’s bucatini to be—brace yourself—10.9 milligrams per pound rather than the 13 milligrams per pound the FDA requires. The product in question is perfectly safe. It presents no threat to the public. It is legal to sell throughout the European Union. But since the FDA alleges it does not meet the agency’s arbitrary standard, the agency turned a temporary shortage of bucatini into a…less-temporary one. Handler surmises the FDA took the action at the behest of one of De Cecco’s competitors.

You might think it implausible that the FDA would seize one manufacturer’s inventory at the behest of a competitor. If so, you would be wrong. The Great Bucatini Shortage of 2020 isn’t even the first time the FDA told Italians how to make Italian food. In a similar episode, the FDA once told a native Sicilian he didn’t know what tomato sauce is.

Rosario Raspanti was born in Palermo, Sicily, where his father ran a tomato‐​sauce cannery. After learning the trade from his father, in 1913 the younger Raspanti brought that knowledge to the United States. He established a canning factory in Mississippi, which he claimed made him the first canner of tomato sauce in America. Raspanti used practically the same process and sold practically the same product his father did. By 1942, Raspanti estimated he had sold some 100 million cans of tomato sauce to satisfied customers in Arkansas, Louisiana, Mississippi, and Western Tennessee.

Related Event

Defending Our Right to Test: How the FDA Restricts Direct‐​to‐​Consumer At‐​Home Testing

When the FDA announced in November 2020 it approved the first at‐​home COVID test, there was a catch: patients were required to get a prescription to purchase and use it. Not for the first time, the FDA is obstructing Americans from obtaining information about their own health. Does this regulatory content exceed the FDA’s authority?

In 1942, however, the FDA seized 36,144 cans of his tomato sauce. The agency furnished no evidence the sauce was harmful to consumers. Its entire justification for seizing the items was that, in the FDA’s opinion, Raspanti’s sauce was thinner than tomato sauce should be and didn’t have enough seasoning.

At trial, Raspanti testified, reasonably, that his customers preferred tomato sauce that was both unseasoned (so they could then season it to taste) and thinner than his competitors’ (allowing them to consume it as‐​is or reduce it to whatever consistency they prefer). Others testified in Raspanti’s defense. Two food brokers with a combined 42 years of experience testified that wholesalers, retailers, and consumers alike all accepted Raspanti’s sauce as tomato sauce; that the trade accepted it over competing products by a ratio of 10 to 1; that one retailer said it enjoyed as much consumer acceptance as Arm & Hammer Baking Soda; that none complained the sauce was too thin or lacked seasoning; that consumers bought it because they preferred a thinner, unspiced sauce; and that they (the brokers) could easily sell a large amount of this established product if it were available. One of Raspanti’s competitors, another Sicilian who also operated a tomato‐​sauce cannery in Mississippi, testified Raspanti knew perfectly well what tomato sauce is and more important (for legal reasons) so did Raspanti’s customers.

I can’t recall if the FDA’s action stemmed from a complaint by one of Raspanti’s other competitors, but I believe that was the case. Either way, Raspanti’s other competitors helped the FDA convince Judge Harry Jacob Lemley, himself a native of rural Virginia and member of the U.S. District Court for the Eastern District of Arkansas, that Raspanti didn’t know how to make tomato sauce. At trial, witnesses for the government included “chemists employed by FDA and competitors, a plant manager employed by a competitor, a buyer and sales manager of a food wholesaler, a housewife, a chef, [and] a restaurant manager.” Several government witnesses testified, uniformly and with stunning precision, that true tomato sauce contains no less than 8.37 percent tomato solids. Raspanti’s crime was to produce a tomato sauce—the faint of heart should stop reading here—that contained only 6.5 percent tomato solids. Budding tomato‐​sauce expert Judge Lemley personally and thoroughly assessed these claims at trial: “A can of [Raspanti’s sauce] and certain cans of other brands were opened and exhibited to the Court, by whom they were tested by pouring and tasting.”

In ruling for the government, Judge Lemley conceded, “It is true that in…Louisiana, Arkansas, Mississippi, and Western Tennessee, the claimant’s product has been accepted by the consuming public as tomato sauce over a long period of time.” Lemley nevertheless concluded, with similar stunning precision, “There seems to be no question but that dealers in, and consumers of, tomato products generally throughout the United States consider tomato sauce to be a spiced product containing not less than 8.37% of salt‐​free tomato solids.” If Raspanti wanted to sell his tomato sauce in the United States, he would most likely have to relabel it a beverage in accordance with the assessment of one of the witnesses, who was “an expert on beverages, being in charge of the Beverage Section of the Food Division of the Food and Drug Administration.” Lemley ordered the government to re‐​label and sell the 36,144 seized cans of tomato sauce to the government’s benefit or, in the alternative, to release the items to Raspanti provided his company both paid the costs of the seizure proceedings and posted a bond conditioned on the company re‐​labeling the cans prior to sale.

“Government,” Barney Frank reportedly said, “is simply the name we give to the things we choose to do together.” Like tell Sicilians how to make tomato sauce.

Mona Zhang has a good piece on marijuana and political corruption at Politico. She writes:

In the past decade, 15 states have legalized a regulated marijuana market for adults over 21, and another 17 have legalized medical marijuana. But in their rush to limit the numbers of licensed vendors and give local municipalities control of where to locate dispensaries, they created something else: A market for local corruption.

Zhang describes how the mayor of Fall River, Massachusetts, allegedly tried to extort $600,000 from cannabis companies in exchange for granting them sales licenses. She discusses numerous other cases in a “rash of cannabis‐​related corruption across the nation, from Massachusetts to California to Arkansas and beyond.”

The problem is not marijuana, but rather that politicians are overregulating and micromanaging the market. Zhang notes:

Almost all the states that legalized pot either require the approval of local officials — as in Massachusetts — or impose a statewide limit on the number of licenses, chosen by a politically appointed oversight board, or both. These practices effectively put million‐​dollar decisions in the hands of relatively small‐​time political figures — the mayors and councilors of small towns and cities, along with the friends and supporters of politicians who appoint them to boards … They have also created a culture in which would‐​be cannabis entrepreneurs feel obliged to make large campaign contributions or hire politically connected lobbyists.

It is sickening that people in “public service” are not satisfied with their fat salaries and pensions and seek illegal ways to further line their pockets. But that is what excess regulation often leads to, particularly when governments put artificial limits on valuable items and then dole out access in a discretionary manner. Chicago has been plagued by corruption partly because the government requires businesses to obtain masses of permits, licenses, and other approvals, and then gives individual city council members discretionary power over whether to grant them.

A similar problem exists with the Low‐​Income Housing Tax Credit (LIHTC), as discussed here. The federal government assigns each state a fixed amount of valuable tax credits, which the states dole out to local governments who distribute them to their favored developers. That setup has led to LIHTC corruption scandals in Dallas, Los Angeles, and other cities.

For the marijuana market, Zhang finishes her piece touching on the solution to the corruption problem:

States that have largely avoided corruption controversies either do not have license caps — like Colorado or Oklahoma — or dole out a limited number of licenses through a lottery rather than scoring the applicants by merit — like Arizona. Many entrepreneurs, particularly those who lost out on license applications, believe the government shouldn’t be in the business of picking winners and losers and should just let the free market do its job.

The transit industry will get $14 billion of the $900 billion coronavirus relief package passed by Congress on Tuesday. That’s less than half of what transit agencies wanted but enough to tide them over for five months or so by which time (the agencies hope) the next Congress will have a chance to pass another and even bigger relief bill. The $14 billion is on top of the $13 billion that Congress gave to transit as a part of its normal annual funding bill.

For those who care, the TransitCenter has posted a spreadsheet showing its estimate of how the $14 billion in relief funds will be distributed among the nation’s major urban areas. The New York urban area will get $5.5 billion of which $3.9 billion goes to New York, $1.4 billion goes to New Jersey, and $0.2 billion to Connecticut.

Los Angeles gets nearly a billion, San Francisco‐​Oakland $800 million, and Chicago and Seattle around half a billion. Curiously, what regions get isn’t closely related to how many transit riders they carry: Seattle apparently will get 16 percent more than Chicago even though Chicago transit carries more than twice as many riders as Seattle’s.

It’s even stranger that urban transit, which doesn’t even carry 1 percent of passenger travel in the United States and whose fare revenues represented less than 0.08 percent of the economy in 2019, gets more than 1.5 percent of the money that is supposed to help the entire economy. This is testimony to transit’s successful effort to portray itself as essential to urban living even though, outside of New York, it is actually pretty irrelevant. It is also testimony to the fact that transit, like someone whose job has been outmoded by automation and who refuses to learn new skills, is pathetically depended on public relief in order to survive.

“Do we need to save transit?” asked Cambridge Systematics transit consultant Herbert Higginbotham in a presentation earlier this month to the Transportation Research Forum in Washington DC. He pointed out that telecommuting was likely to remain high, eating into transit ridership, long after the pandemic was over.

Before the pandemic, he said, transit agencies were “prioritiz[ing] choice riders,” that is, people who ride transit even though they have a choice of driving a car. Light‐​rail and bus‐​rapid transit lines were specifically designed to attract such riders. With increased telecommuting, however, that market is likely to dry up.

As a result, Higginbotham suggested, many transit agencies after the pandemic will “prioritize” what he calls “equity” riders (otherwise known as transit‐​dependent riders), meaning low‐​income people without cars. The problem is that there aren’t that many transit‐​dependent people left. As the American Community Survey documents, only about 4 percent of workers (6.7 million people) live in households without cars, and only 40 percent of them (2.6 million people) take transit to work. That’s only about a third of transit’s pre‐​pandemic market (which is why transit has recently focused on choice riders). It may not be a coincidence that September and October transit ridership was barely more than a third of pre‐​pandemic numbers.

The pre‐​pandemic transit model, Higginbotham admitted, “was in danger of failure at worst, and stagnancy at best.” He cheerfully concluded (as any transit consultant must do) by arguing that the pandemic “gives us an opportunity to reimagine transit and mobility to be more flexible, equitable, and sustainable.” Will transit “accept the challenge?” he asks. The answer is probably no: Why bother to learn new skills when you can get paid even if you don’t produce anything of value?

In my imagination, a transportation system that is flexible, equitable, and sustainable is built around automobiles, not mass transportation. For now, transit is on relief, but if we don’t really need it anymore, why keep it?

Yesterday, the Centers for Disease Control and Prevention reported that the drug overdose death rate, already accelerating after a brief pause in 2018, increased at an alarming rate coinciding with the mobility restrictions and emotionally stress of the COVID-19 pandemic. There were more than 81,000 overdose deaths during the 12 months ending in May 2020—a new record. There were just over 71,000 deaths reported for the 12 months ending in December 2019.

To be sure, the isolation, loneliness, and anxiety of the pandemic and its management exacerbate mental health problems, including substance use disorder. Their effect on the overdose rate was noted by White House “drug czar” Jim Carroll in the late spring of this year.

But the most important feature of the latest report from the CDC is the fact that illicit fentanyl, made in clandestine labs in Asia and Mexico, was responsible for roughly 57 percent of all overdose deaths. Cocaine was found in roughly 22 percent of overdoses. And methamphetamines were found in 23 percent of all overdose deaths.

Yet the CDC maintains its focus on doctors prescribing opioids to their patients in pain in its list of recommendations for dealing with this dispiriting news. It tells health care providers to co‐​prescribe the opioid overdose antidote naloxone to their pain patients when they prescribe high doses of opioids—despite the fact that prescription opioids account for an ever‐​decreasing percentage of the drugs involved in overdose deaths, with overwhelming evidence that the overdoses are from illicit drugs obtained in the black market. As Josh Bloom and I recently pointed out, the crackdown on prescription opioids has fueled a “street drug epidemic” while it has harmed patients in pain.

The CDC also gives lip service to a few harm reduction strategies in its recommendations. For example, it recommends more widespread distribution of the overdose antidote naloxone, without urging the Food and Drug Administration to reclassify naloxone as an over‐​the‐​counter drug—an obvious way to increase its dissemination.

The CDC also encourages the expansion of medication assisted treatment (MAT) programs with drugs such as methadone and buprenorphine, which it recognizes as the most effective form of treatment for substance use disorder due to opioids. Yet it makes no mention of proposals to deregulate methadone and buprenorphine treatment programs by letting health care practitioners more easily prescribe these drugs to their patients, as they do in many developed countries. Congress can build upon the temporary relaxation of MAT regulations during the COVID pandemic and implement permanent reforms.

The tragic report by the CDC lacks the boldness to admit that restricting patients’ access to legal drugs (e.g., prescription opioids in the case of opioid‐​related deaths and pseudoephedrine in the case of methamphetamine‐​related deaths) has, if anything, inflamed the drug overdose crisis while harming patients. It failed to argue for reforming or repealing laws that obstruct harm reduction strategies. So it should come as no surprise that it failed to identify the ultimate source of the drug overdose crisis: drug prohibition and the dangerous black market it fuels.