Tuesday, September 16, 2008

More Seatbelts = More Deaths?

Suppose for a particular location, we have the following two facts:

Fact 1: Over the past 20 years, automobile companies have introduced a number of safety devices such as seatbelts and air bags.

Fact 2: Over the past 20 years, the number of highway fatalities per mile driven has increased.

How can you reconcile Fact 1 with Fact 2?

Congratulations to Jeff Staudt for being the first to submit a satisfying answer to the above conundrum. Take a hard look at the picture above and then read his answer in the comments section. Essentially, Jeff is describing what is commonly referred to as a "moral hazard" problem. The idea of using a spike on the steering column has been attributed to Sam Peltzman.  Here's a related cartoon.


dante sherman said...

Manufactures have produced more and safer seat belts due to the increase in highway deaths per mile.

Greg Delemeester said...

Dante, your explanation is an example of what I would call the "conventional wisdom." If the conventional wisdom is correct, then what does that say about society's efforts to improve highway safety? Have 20 years worth of safety improvement simply been a game of "catch up."?

Can you (or anybody else) think of another explanation for these two "facts" that relies on economic reasoning involving incentives and individual behavior?

Cody Meglio said...

I can reconcile the two facts by showing that the number of drivers and the number of cars on the road has increased faster over the last twenty years than has the introduction and spread of safety features. More roadway traffic leads to more crashes and, subsequently, more fatalities. For example, let's say the number of vehicles on the nation's roadways have increased by 50%, safety features have improved by 30%, and the percentage of accidents (number of accidents per driver) remains the same. That means that, as a result of pure growth, more accidents will occur, 10% more, in fact.

To illustrate using the figures from above, let's say there were 100 million cars in the US in 1988 and 100,000 fatalities as a result of accidents that year. Now, in 2008, let's say there are 150 million cars on the road, (artificially low, but just an example) and the same percentage of accidents as before, but fewer fatalities per accident due to improved safety features. Without improved safety features, there should be 150,000 fatalities, which is higher than 20 years earlier. With the effect of safety features, there are only, 105,000 fatalities in 2008, for example. This number is still more than 1988, but is smaller as a percentage of drivers on the road. Thus, though the number of fatalities are greater now than before, even with more safety features, the number still does not keep up with the increase in the number of total drivers on the road.

In summary, though the number of fatalities has grown as a result of higher population, the percentage of fatalities has actually shrank.

Jeffrey Staudt said...

Because auto manufacturers have implemented more safety devices, drivers feel more at liberty to drive more aggressively and take greater risks due to their increased safety margins.

Some economists have suggested that if you want to increase automotive safety (ie, greater driver caution) the auto manufacturers should place sharpened spikes in the center of steering wheels.

Greg Delemeester said...

Cody, you are talking about total fatalities rather than what the question is concerned with, namely, fatalities per mile driven. An increase in the absolute number of cars or drivers will not necessarily have any impact on fatalities per mile driven.

Anonymous said...

My guess is that the increase is due less to moral hazard than to secular trends. Over the last 20 years the driving behavior of under-25 female drivers as become to resemble that of their male contemporaries. This might be a good thing for issues of sexual discrimination(??), but it is bad for traffic safety.

Thus, let me ask: Has the number of fatalities per mile increased for the 45-65 age cohort? If so, then I would buy the moral hazard explanation. If not, I would think one should look at the secular trend. Of course, moral hazards could be part of that trend.

Dave Prychitko said...

I've heard that the steering wheel spike was attributed to Gordon Tullock, but Peltzman (given his work in the 60's) wouldn't surprise me.