Paul Alper writes:
You often write about the psychology of color choice and this NYT article is full of (confusing) numbers/opinions and ought to resonate with your contributors. Take your pick of these forcefully stated views:
Because many other factors influence car value, color is easy to overlook. Yet both paint and car manufacturers maintain international departments of stylists and colorists who not only monitor what consumers are buying but — drawing from the fields of art, architecture, fashion, popular culture and consumer research — predict what people will want up to five years in the future.
According to iSeeCars, a search engine catering to car buyers, the worst color for S.U.V.s was beige, which lost 46 percent of its value over three years. For pickup trucks the best color was … beige. Beige pickups lost only 18 percent in value in the same time period.
In a routinely quoted poll from 2000, 39 percent of car buyers said color was more important than brand.
Cars came in more than a dozen hues by the mid-50s — the better to attract the female drivers of the family’s second car, the thinking went. Those colors became more vivid in the psychedelic ’60s.
“Rental car companies love white,” said Karl Brauer, executive analyst for iSeeCars.
But, as the iSeeCars data shows, there is a big gap between what is popular and what retains value.The 2020 Color Report from the paint provider Axalta (formerly DuPont) said fewer than 1 percent of new cars on lots in America were yellow. Yet iSeeCars data shows yellow retained the most value over all. An overwhelming 30 percent of cars on dealers’ lots are white, followed by 19 percent for both black and gray and 10 percent for silver.
“You can’t go wrong buying the popular colors — black, white or silver — but you can’t go right, either,” he added. The most popular colors generally fall in the middle of the value chart.
Rarity alone doesn’t guarantee value. Purple, brown and gold are about as rare as yellow yet retain the least value over all.S.U.V.s did best in flashy colors, possibly because the drivers didn’t want to feel like drudges.
“You are buying the S.U.V. to avoid the minivan,” said Jonah Berger, a professor at the University of Pennsylvania’s Wharton School with expertise in marketing psychology. A lively color, he said, “makes us feel like: ‘I am driving a fun car. I am a fun, exciting person.’”People who buy cars for utility, like minivan and fleet buyers, seem to value subtle colors that are easy to care for. People who buy a car as a personal statement — sports- and muscle-car owners — value glitzy colors.
The Jeep Wrangler retained the most value in Xtreme Purple, a color usually at the bottom of the overall chart. Purple Wranglers kept $2,398 more value than the same model in utilitarian silver.
Ultimately, many buyers may choose paint color disregarding both value and popularity to achieve a third goal, Mr. Berger said: “Maybe having a color that’s different than white makes you happy.”
Alper’s right. I love stories about colors! For example here. Regarding the above-quoted NYT article, I believe approximately none of the above claims. It’s just too easy for people to make up statistics, make inappropriate comparisons, etc. Even when researchers are being really careful and preparing their findings for publication, they can overstate. You can only imagine how things go when the numbers don’t have to be checked and the newspaper is motivated to print the most extravagant claims.
So, nah, I don’t trust any of those stats. But they’re fun. Colors!
Hmmm.
Seems like the relationship between the color of a car and its value would be complicated by many, many other factors.
“approximately none” =
[chef kiss]
The most extreme results (yellow, purple jeeps) come from the categories with the smallest number of observations. Sounds like normal variation, not some strange counter-intuitive behavior.
It’s not surprising that popular “color colors” retain more value. You have to pay more or wait longer for them in the first place – which is already an indicator of higher value. People who pay more or wait longer are probably more likely to buy more expensive cars that better retain their value.
It would be interesting to find out if women value red cars more when they are ovulating, and if a power pose with one’s most desired car prior to a job interview improves one’s job prospects (social program!). But here’s the big question: if we give people a car with a secretly constantly refilling gas tank, will they drive more? We’ll just drive around behind them with a tank truck, they’ll never notice.
In 1980 I wrote about John T. Malloy’s book, “Dress for Success.” I pointed out that he referred to himself as a “wardrobe engineer” who, using empirical evidence and statistical testing, could tell you what works when it comes to dress and associated colors. For example, an instructor should know
1. “One no-no is the the solid blue suit; it’s death on campus, as are pinstripe suits, Ivy League and club ties.”
2. “The prejudices of college students are violent and strong. If you turn off a college student with your clothing, you turn him off completely.”
3. “The best suit to wear in the Midwest…is a very tiny shadow plaid.”
4. But anywhere in the U.S., “Never wear green.”
5. Irrespective of color, “Never take off your suit jacket unless you have to.”
6. With or without a jacket, “The tie is probably the single most important denominator of social status for a man in the United States today. Show me a man’s tie and I will tell you who he is, or who he is trying to be.”
7. “your T-shirt should never show under an open-collared shirt.”
Of course, there was much more in the book including, “you are not what you eat; you are your product. And you must dress accordingly.”
Unfortunately, my attempt at satirizing Malloy’s dicta got my into trouble with the Administration because when it came to colors, I quoted him that as far as interacting with Jews, “do not wear brown” and when it comes to older Jews, “do not wear beige.” The moral of the story is that emulating Jonathan Swift should be left to professionals.
So is the idea that there exist confounds that stymie straightforward causal interpretation of “value lost” summary statistics? Like — it’s not that you’d have counterfactually gotten less money for your $COLOR car, it’s that people who buy $COLOR cars predominately live in the poorer parts of the country with depressed auto markets, or live in wetter, colder parts of the country where salt and water conspire to make rust, or just happen to be the sorts of owners who are lax about routine maintenance, or whatever other story we’d care to tell? I’d expect samples to be large enough that sampling variances would unlikely to be a problem, even absent regularization and given however many researcher DoF there exist in all Brand x Body Style x Color x Time comparisons.
I think there is an economics component. If 5% of people really want purple, the price is high if only 3% of the cars are purple. And the price is low if 8% of cars are purple.
Yes but we can assume that the car makers are optimizing profit and they all cost same to make. Even if they are pretty bad at optimizing each year they get to try again with last year’s sales data so the long term average number sold in a given color should match the avg percent of new car buyers (for that make of car) who want that color.
So all we really need to work out is how ppl’s color prefs change as they shift to the used cars. Now all you need to believe is that there isn’t a systematic reason used car buyers care more about getting a traditionally popular color and fact that it’s easier to increase demand for the less popular cars (if 99% of cars sold are red and 1% purple u can’t double demand for red but 1% shift doubles purple demand).
I speak to claim that max value retaining color will be some unpopular color.
Data may not be convincing on their own but it seems like what we should have predicted a priori. Even if we knew nothing else I’d think it’s easier to affect price of car that has less interest (assuming car makers optimize first sales) than more (99/1 vs 98/2..doubled later group)
Also I’d think people should become relatively less picky about color and you may get more participation by unassimilated immigrant communities. So we should start with high prior that best value retention will be in a unpopular color.