Applying Behavioral Economics to Supply Chain Decisions

B2B relationships aren't the rational arena classic theories would suggest

Say you’re the maker of widgets, with a big new customer that builds and sells machines, including a model that will now use your product. Before the customer places an order, you will have to add plant capacity, so there’s an up-front investment to factor into your pricing decision.

But you can’t know how much business you’ll get, so you’re guessing at how big of an expansion to fund. Your sales volume to this new customer depends on eventual demand for his equipment, and he knows that market better than you do.

This puts you at a disadvantage in negotiating the most profitable contract.

One way to discover the buyer’s knowledge of demand is to present him with a menu of contract options, each one for a different quantity of widgets. The buyer will choose the contract that, given the expected demand, earns him the highest profit. With that knowledge, you can expand appropriately and maximize your profit and reduce your risk.

Anyway, that’s what classical game theory predicts, based on the expectation that the buyer will make a rational, fully informed choice and pick the option that’s most profitable for him. But studies have found that it doesn’t always work that way. Buyers frequently will choose a contract that is less profitable for themselves, lowering earnings for the supplier, too.

As we’ve learned from years of research in behavioral economics, people aren’t the rational actors that traditional theory says they are, whether they’re at home or at work.

A paper by University of Hamburg’s Lennart C. Johnsen and Guido Voigt and UCLA Anderson’s Charles J. Corbett digs into how a B2B buyer’s behavior can cause a menu of contracts to perform more poorly than expected. In a set of experiments, they find that buyers are likelier to pick contracts they think are “fairer” — wherein the spread between a supplier’s and buyer’s profits is narrower — even if a less equitable option delivers higher earnings for the buyer.

Other factors, such as information overload, also can lead buyers to select less-than-optimal contracts, the researchers find. To design a menu of contracts that deliver the highest profits to both suppliers and buyers, the authors propose new models that take into account the behavioral dynamics causing less-than-rational choices in buyers.

The recognition that people in real life behave far differently than economic models predict goes back at least to the work of Nobel Prize-winning economist Herbert Simon, who described how limits to our abilities, information or time can lead to a “bounded rationality” in our decision making. The field of behavioral economics took off in the 1970s and ’80s, but only more recently have its insights been applied to the problems of operations management and supply chain contracts.

Problems like that of the widget maker: How to maximize profits when she knows less about market demand than her buyer? A menu of contracts might be one way to deal with this information asymmetry, but not if the contracts don’t actually deliver the results standard models promise.

To test the role of behavioral factors, Johnson, Voigt and Corbett ran a series of lab experiments with a group of German undergraduate and graduate students. Specifically, they looked at how a buyer’s preference for fairness, along with his bounded rationality (ability to handle only so much information), can lead him to accept lower-than-ideal profits.

In the experiments, students playing the role of buyers were matched with either a computer program or other students, each acting as supplier. (The idea is that buyers do not care about the profits of a computer.) Buyers received a packet of information about customer demand for their product, their potential profits, the contract terms and, notably, the suppliers’ profits under each contract option.

In the experiments, the supplier always earns more than the buyer, but the gap in profits is different in each contract. For example, in one contract, the buyer receives a profit of, say, $835, while the supplier’s profit is $7,317. In the second best option for the buyer, the buyer’s profit is $834 and the supplier’s is $1,766.

In the experiment, buyers working with a human supplier choose a contract with the highest profits for themselves only 53% of the time, while those who knew they were dealing with a computer chose the maximum-profit option 77% of the time. While in standard theory it shouldn’t matter whether the buyer was dealing with a human or a computer, the difference suggests that buyers believed the human seller was unfairly taking a larger share of the transaction’s profits.

As a result, the authors write, “subjects are willing to sacrifice their own profits to reach a more equitable split.” No one wants to be a chump.

A preference for fairness didn’t account for all the difference between experimental results and the classical model, which predicts buyers should always choose the option that maximizes their own profits. Something else contributed to the buyers’ choices — in this case, too much information.

Running the experiment again and giving buyers fewer details about the contracts (and with all suppliers identified as computers), researchers found that buyers chose the contract with the highest profits for themselves 91% of the time.

The authors then devised and tested new models for a menu of contracts that incorporated buyers’ preference for fairness and less information to deal with. These offered a wider spread between the profits for the buyer (the optimal choice might be $2,669, compared to $1,307 for the next best option) and narrowed the payoff differences for the supplier ($5,531 and $2,084). As a result, buyers chose the option that maximized their profits 92% of the time. Profits for both buyer and seller rose in this approach.

The result is “a more equitable payoff distribution,” the authors write. “We conclude that if we wish to encourage managers to adopt more complex contracts, we should recommend that they use the behaviorally adapted versions.” This work sheds new light on why more complex contracts have often performed poorly in experiments in the past, and suggests that practitioners should be aware that classical theory overstates the gains that can be achieved from such complex schemes.