Counterfactuals matter

Alex Tabarrok at Marginal Revolution has a really great post on car safety in India, which illustrates a major problem with attempts by government to protect consumers by regulating unsafe products out of existence.

According to the Global New Car Assessment Program, 5 out of 7 new cars made for the Indian market lacked airbags and would fail basic UN crash tests. Tabarrok quotes Global NCAP Secretary General David Ward "Global NCAP strongly believes that no manufacturer anywhere in the world should be developing new models that are so clearly sub-standard".

But demonstrating that a product is unsafe is not sufficient to show that regulating it would make people safer. As Tabarrok shows, counterfactuals matter:

Let’s take a closer look. These cars are very inexpensive. A Renault Kwid, for example, can be had for under $4000. In the Indian market these cars are competing against motorcycles. Only 6 percent of Indian households own a car but 47% own a motorcycle. Overall, there are more than five times as many motorcycles as cars in India.
Motorcycles are also much more dangerous than cars.
The GNCAP worries that some Indian cars don’t have airbags but forgets that no Indian motorcycles have airbags. Even a zero-star car is much safer than a motorcycle. Air bags cost about $200-$400 (somewhat older estimates here a, b, c) and are not terribly effective. (Levitt and Porter, for example, calculated that air bags saved 550 lives in 1997 compared to 15,000 lives saved by seatbelts.) At $250, airbags would increase the cost of a $5,000 car by 5%. A higher price for automobiles would reduce the number of relatively safe automobiles and increase the number of relatively dangerous motorcycles and thus an air bag requirement could result in more traffic fatalities.

Substitution effects are basic economics, but regulators often ignore the fact that consumers responding to incentives will switch to cheaper and often (less safer) alternatives.

Take the European Union's recent Tobacco Products Directive:

Cartridges and tanks of e-liquid will be limited to 2ml, a incredibly small amount, and refill containers to 10ml, completely preventing bulk-buying and outlawing the products users find most convenient. The maximum strength will be 20mg, ruling out high strength varieties that most closely approximate cigarettes.

The regulators who drafted this law hope to see the public's health improve as vapers switch to safer, lower strength E-Cigarettes, but it could very likely have the opposite effect. As E-Cigarettes, which are 95-99% safer than traditional cigarettes, become weaker and less convenient, consumers have less of an incentive to switch from tobacco to e-liquids. 

Likewise, attempts to improve public safety by banning low-risk tobacco products like Snus have backfired spectacularly.

Or take for instance, Austin, Texas. Its City Council imposed onerous regulations on private ridesharing companies like Uber and Lyft with the stated aim of protecting the public from unsavoury drivers. One of the key sticking points was a finger-print background check requirement that went far beyond what was expected for standard Yellow Cabs.

Uber and Lyft responded by going Galt. They left Austin altogether. What happened next shouldn't surprise anyone. The Drive reports:

Though the city frightened voters with terrifying depictions of a plague of Uber-rape, it’s now come out that people with sexual assault convictions, and even drunk-driving arrests, are actually allowed to drive cabs in Austin, few questions asked. "If you have ever been convicted of theft at any point, you could never get a chauffeur's permit,” a city council member told a local news station. “That just seems like too much.”
Then there are the opportunists who are actually thriving. I talked to one driver who seemed to be enjoying himself in this bizarre ecosystem of random auto-barter. Because of his “local media presence and history of bartending in the city,” he immediately scored an under-the-table gig providing two to three rides a day to executives from a prominent entertainment company, as well as whatever else he could cadge up on Facebook.

Attempts by government to protect the consumer too often fail to consider the counterfactual, if policymakers and regulators took substitution effects seriously, many regulations wouldn't see the light of day.