A second important heuristic is anchoring. This heuristic uses some initial estimate as an anchor to bias or subsequent judgments.
We use anchoring, we are uncertain, we don’t have all the information needed to make an accurate judgment so we latch onto anything given to us. Think back to when you first saw the Milky Way, you are outside in the dark night away from civilization and you looked up and saw that fuzzy band of stars running across the night sky.
You only saw a few hundred or maybe a few thousand of the brightest nearby stars but our galaxy contains many more. Let me ask a question: does our galaxy contain more or fewer than 100,000 stars? Now let me ask you a second question: how many stars are in the Milky Way Galaxy?
Let’s think about that second question: when I ask you how many stars are in the Milky Way galaxy, that’s not something for which you have personal experience. Even if you’ve seen the Milky Way you’ve only seen a small fraction of its stars and unless you’re an astronomy buff, you probably have an encounter that answer recently so you can’t give an estimate based on explicit knowledge.
Defective anchoring leads to bad estimations
Without personal experience or explicit knowledge you need to find something on which you can base your estimate. The first question asked, does our galaxy contain more or fewer than 100,000 stars? That number 100,000 provides an anchor, it seems like a big number if I’m asking you about it then you might infer that’s close to the right answer.
So when faced with a difficult judgment like how many stars are in the Milky Way people often begin with any number given to them and then adjust their estimate upward or downward from that anchor. Someone who thinks that the number is larger than 100,000 might adjust upward, say to 500,000. The key idea is that the anchor like it’s nautical namesake constrains how far a subsequent estimate can move.
So what’s the right answer here? No one knows exactly, but the current best estimates are that there are several hundred billion stars in the Milky Way galaxy. In this case the anchor I gave you was about 1 million times too low so it typically leads to estimates that are likewise far too low. If I’d used a much larger anchor by asking something like does our galaxy contained more or fewer than a trillion stars the estimates would have been much much higher.
Anchoring and marketing
Marketers use anchoring when pricing goods are rare infrequently purchased or difficult to value consider expensive furniture or jewelry at a specialty store often those items will be displayed with a very high suggested retail price. That price might have no connection to reality whatsoever, it doesn’t represent the cost of manufacturing, consumer demand or anything else meaningful, it’s just an anchor.
The store want you to start with that suggested retail price and then negotiate. Downward, they know that providing a very high anchor increases customers estimates of value and leads to higher eventual selling prices. It is important to emphasize that anchoring works even if there is no meaningful connection between the anchor and the subsequent estimates.
In some laboratory experiments researchers have generated anchors by asking people to spin a roulette wheel or to write down the last two digits of their Social Security number. Even when the anchor is generated in a completely random fashion, it still can influence behavior. For example the people who write down large numbers are also willing to pay more for bottles of wine.
But anchoring is far more than a party trick for behavioral economists, it influences real-world economic transactions, even those with extraordinarily large stakes. Consider the rarefied world of fine art auctions. What is the value of said Cezane’s The card players, or Moon’s The scream or van Gogh’s Vase with 15 sunflowers? Each of these sold for tens or hundreds of millions of dollars, they are clearly valuable each represents a one-of-a-kind masterwork by great artist, but what sets their price?
Why would one van Gogh sell for $40 million and another for $80 million? Economists analyze the sale prices of more than 15,000 paintings all auctioned at major institutions in London and New York. They found the price of a given painting reflected many factors specific to the painting: who painted it, its size, its medium, whether it was signed, among many others.
All those factors makes sense as contributors to the price of painting, but price also reflects the current market. The current market for fine art is hot then paint will sell for more money the market is cold perhaps because of a down economy then paintings will sell for less. The art market is notorious for boom and bust cycles. The average price for paintings within a category might double or half within a year.
Influencing the market
So whether the current market is hot or cold should also influence the selling price of obtaining. Now let’s suppose that a painting first comes up for auction during a hot market, which inflates it selling price. So what happens when it comes in the market again in a few years? If that inflated selling price serves as an anchor for buyers something that guides her subsequent willingness to pay then the next sale price should be higher than predicted.
The reverse would happen for paintings that were first sold in a cold market, they would command less than auction was sold again in a few years. The data show that the anchor effects purveyed art auctions. Once the price was established for a painting regardless of whether that price was set during boom times or bust times, then that price influences how much people are willing to pay years later.
So if you are in the market for expensive fine art look for pieces that were last sold years ago, in a down market. Those pieces are likely to be systematically undervalued. Which brings me to the representativeness heuristic, the one that helps us estimate probabilities through the following simple rule: prototyping events are likely events. But, this is a topic for another article.