Sunk Costs and Driving Decisions?
Do people who buy more expensive vehicles drive more? At first, the answer to that question seems obvious. Within a certain income group, people who are going to spend a lot of time in their cars are willing to invest more – get those leather seats if you’re going to be sitting in them several hours a day.
So, yes, we should see a positive relationship between the amount people pay for their cars and the miles they drive per month, say.
But, in a recent working paper, my colleague Teck Ho and co-authors show that the relationship between vehicle purchase price and miles driven extends further. They study driving behavior among people in Singapore, and because roads are heavily congested there, the government imposes a hefty combination of taxes and fees on new-car purchases. One result of this is that cars are extremely expensive. The most popular model in their data set (they don’t identify the brand or model, but I’m imagining something like a Toyota Camry) sold for US$168,000 in 2011.
Changes in these fees meant purchase prices varied from $142,000 in 2001 to a low of $130,000 in 2009 before rising sharply to $168,000 in 2011. Given restrictions on resale and scrapping, a good chunk of the fees are not recoverable if the car is sold or junked.
Ho and co-authors find that people who happened to pay more than their luckier neighbors drive the car more. Given that the researchers are looking at drivers using exactly the same car model and people paid more for it for reasons due to a policy change, Ho and co-authors argue that the effect I identified in the first paragraph – people who know they’re going to drive more self-select into the cars when prices are higher – is not at work.
Instead, they say, it’s an example of what social scientists call the sunk-cost bias. With cars, the sunk-cost bias reflects mindset where people think, “No way am I walking or taking the bus. I paid $168,000 for my car, so dammit, I’m going to use it.”
Most of us can imagine this psychology, but it is irrational. The $168,000 cannot be recovered by driving more and rational decision makers should not drive more based on money they have already spent. Nonetheless, the impacts they find are sizeable. For every 20% increase in sunk costs (and the whole vehicle purchase price is not sunk), people are driving 10% more. The close relationship between miles driven and purchase price is depicted in the above figure for two models in their data set.
There are many classic examples of the sunk-cost bias, which I relay to my students as I try to convince them to avoid falling prey to it in business settings. For example, Ho and his co-authors describe a study where researchers randomly gave discounts to theatre subscribers. The people who didn’t get the discounts sat through more of the plays. (“I hate musicals, but I dammit I’m going given what I paid for my subscription.”) There’s a great article in the New Yorker asking whether the New York Jets were falling prey to the sunk-cost bias as they continued to start flailing quarterback Mark Sanchez after they’d signed him to an expensive contract.
Implications for Energy Policy?
Although this study is based on data from Singapore, there is good reason to think that similar forces are at play in the rest of the world. Singapore simply provides a convenient setting for the study given the widely varying prices for the same car model.
So, would curing the sunk-cost bias improve energy efficiency and encourage people to drive less? Ho and co-authors argue that it would in Singapore if the government tried to discourage driving through higher per mile charges rather than unrecoverable taxes and fees on vehicles.
In the US, a chunk of the purchase price of a new vehicle is sunk since cars reportedly lose around 30% of their value the second they are driven off the dealer’s lot. It’s possible that drivers of more fuel-efficient vehicles, which are generally smaller and cheaper, are driving relatively less than their brethren in large SUVs because they believe they have fewer sunk costs to recover.
Ho’s findings remind me that there are a lot of factors influencing people’s decisions about how to use energy, including how much to drive. As policymakers put increasing emphasis on improving energy efficiency as a means to combat climate change and ensure energy security, we need to draw on all the quivers in our social science cap to understand these factors.
Catherine Wolfram View All
Catherine Wolfram is Associate Dean for Academic Affairs and the Cora Jane Flood Professor of Business Administration at the Haas School of Business, University of California, Berkeley. She is the Program Director of the National Bureau of Economic Research's Environment and Energy Economics Program, Faculty Director of The E2e Project, a research organization focused on energy efficiency and a research affiliate at the Energy Institute at Haas. She is also an affiliated faculty member of in the Agriculture and Resource Economics department and the Energy and Resources Group at Berkeley.
Wolfram has published extensively on the economics of energy markets. Her work has analyzed rural electrification programs in the developing world, energy efficiency programs in the US, the effects of environmental regulation on energy markets and the impact of privatization and restructuring in the US and UK. She is currently implementing several randomized controlled trials to evaluate energy programs in the U.S., Ghana, and Kenya.
She received a PhD in Economics from MIT in 1996 and an AB from Harvard in 1989. Before joining the faculty at UC Berkeley, she was an Assistant Professor of Economics at Harvard.
A long-forgotten paper investigated a related phenomenon, and may provide some insight into the source of a relationship between car purchase price and usage, particularly where transit service provides as attractive an alternative to automobile travel as it does in Singapore; see Roger Sherman, “A Private Ownership Bias in Transit Choice,” American Economic Review, Vol. 57, No. 5 (December 1967), pp. 1211-1217. As I recall — and I admit that’s only hazily — Sherman argued that the fixed costs of owning an auto caused the average cost of driving it to decline with increasing usage, thus “biasing” auto purchasers’ routine travel choices between using their cars and paying average-cost transit fares toward driving. (I always thought that a simpler but perhaps more plausible explanation might be that car owners compared the relatively low marginal costs of driving with average-cost transit fares, but that depends on the fraction of automobile depreciation that is related to usage rather than simply to the passage of time, about which we — or at least I — don’t know very much.) Maybe I have it exactly backwards, but it seems like this effect should be stronger for purchasers of more expensive cars. This could contribute to the effect Ho and his co-authors find, alhtough the widespread use of fixed-price monthly transit passes could tend to counteract some — but probably not all — of it.
Anyone that has lived/been to Singapore know that driving is a luxury, not a necessity due to incredibly efficient public transportation.
In general, people in Singapore drive very short distances, anything above 20mins is typically considered a long drive.
I feel that the author might fail to understand the reasons and how exactly prices of cars fluctuate in Singapore. Perhaps Prof Teck Ho can provide some insight on how it works.
High car prices is a very complex move by the gov to reduce the number of cars on the road. Thus, higher prices as reflected in the graph also means less number of cars on the road. Given equal amount of travel, wont the amount of usage per car increase?
I dont disagree with sunk-cost bias theory in general, but I believe using Singapore cars as a primary example without understanding/explaining a lot of what goes behind those fluctuations can leave readers very misinformed.
How about ‘I have to drive a lot so I better get a comfortable, even if expensive, car’.
I dont quite get the logic of driving a car more or less, when you dont really have a choice, as in Singapore … all cars are expensive.
In certain societies some people get expensive cars [when there is a choice] only to show to their friends and neighbors: “See how well off I am.” Somewhat like some societies where fat (obese) is considered ‘healthy’ because it used to imply wealth.
So the question should be: “Do people who have to drive a lot buy expensive cars?”
The same question: “I bought this food, and even I am not hungry I am going to eat it before it goes bad, even though it may make me sick or fat.”
Given that hybrids and EVs have a higher sunk-cost than their conventional counterparts, what lessons should we draw from this study to improve subsidy programs?
“I paid $168,000 for this car; there’s no way I’m going run up its miles and wear it out if i don’t absolutely need to!”
“I paid $168,000 for this car, not $150,000, so every mile I get to show it off on the street pays off with more status.”
Thanks for the comment.
The paper works very hard to rule out selection, and I’m ultimately convinced, but decided it was too complicated to get into in the post. Essentially, they look at changes in miles driven as the car gets older as a function of the initial purchase price. With higher sunk costs, miles driven should go down more over time, based on their mental accounting model. This could be some kind of mean reversion (people who know they’re going to have a driving SHOCK, not just drive a lot forever, are more likely to buy the expensive cars), so they also estimate a model where they let the marginal benefit of driving scale with the purchase price. This is probably identified off the functional form of their scaling but the estimates are really pretty close and don’t suggest to me that they’ve captured the wrong functional relationship.
I don’t quite understand how you could truly parse out selection effects vs sunk-cost effects unless prospective buyers were forced to purchase the car regardless of the costs (or, equivalently, if the costs were imposed on buyers after the purchase in an unpredictable manner). This would be true even if sunk costs were randomly assigned in a true experiment.
That said, I’ve long believed that cars should be cheaper on the extensive margin (purchase) and more expensive on the intensive margin (operation). This work clearly supports that view.