Doing Business with the Untrustworthy
Anticipating a Collapse of Trust Can Improve Decision Making
Anticipating and perhaps forestalling a collapse of trust in others can improve decision making when the trustworthy do business with the untrustworthy.
Doing business among Quakers is simple. Their trusting and trustworthy character facilitates trade. In the eighteenth- and early-nineteenth centuries, a seller in London could ship goods across the ocean and trust that he would be paid by a trustworthy buyer when the goods arrived in Philadelphia. As Quakers grew prosperous, others sought them as trading partners. A common contention is some of those new partners adopted Quakers as role models as a means to prosperity, becoming trustworthy in their own right. But that may be an oversimplification.
It may be that the prosperity of a Quaker has less to do with his own trustworthiness, and more to do with the trustworthiness of his Quaker partners. That is, while a Quaker can indeed prosper trading with other Quakers, extra profit may be made by an untrustworthy opportunist trading among Quakers.
Empirical observations, laboratory experiments, and game theory suggest business trades that seem honest at a particular moment can be part of a strategic plan by an opportunist. The opportunist can initially make honest trades to build a reputation for trustworthiness as a means to earn repeat business or personal referrals. Maintaining that reputation through honest trade might last most of the life of the opportunist, and beyond if he incorporates. But there may be extra profit in eventually breaking trust (milking the reputation). Anticipating and perhaps forestalling such a collapse of trust in others can improve decision making when the trustworthy do business with the untrustworthy.
Fool me once, …
“Fool me once, shame on you. Fool me twice, shame on me,” is an ancient folk proverb that can help the trustworthy do repeat business with the untrustworthy. Notice it is not “Fool me once, shame on me” because it may pay to gamble that your partner is trustworthy, or is at least motivated to act that way. “Fool me twice, shame on me” reminds one that being fooled once is sufficient to reveal your partner as untrustworthy. Thus if you cannot profit by doing business with someone known to be untrustworthy and if you don’t think they can somehow grow to be or to act trustworthy, then cut them off after their first offense.
Your threat of cutting off or punishing those known to be untrustworthy is an incentive for the untrustworthy to delay exploiting your trust and so initially allowing the mutual benefits of cooperation. There are many contexts where initial cooperation is followed by eventual exploitation. In the lab, it is found in various experiments of game theory. In a long-distance race, bicyclists or runners may cooperate for most of the race by taking turns leading and letting others follow in their slipstreams, with cooperation collapsing near the end of the race. And in college towns, “no checks allowed” signs are more common near the end of an academic year. Considering one context in detail may suffice to offer guidance for managing the rest.
Suppose you can hire a worker for each of five years. Should you employ that prospect? And if so, what terms should you offer? To make those decisions, suppose you two agree that the prospect must work hard during any year that he is on salary, but may work as he pleases if he is on profit sharing (a lower salary plus a piece rate). Before work begins, you two gather and verify the following data:
- The prospect prefers shirking (giving zero effort) while on salary to working hard while on salary; and prefers working hard on salary to working while on profit sharing; and prefers working on profit sharing to working for someone else.
- Each year profits you 75 thousand if the prospect works hard while on salary, and 25 thousand if the prospect works while on profit sharing.
- Each year loses you 125 thousand if the prospect shirks while on salary.
(For the academic reader, the next section shows how that data arises in a textbook principal-agent model of game theory. In that model, there is less profit from a worker on profit sharing because the employer must compensate the worker for the greater risk he faces when uncertain factors beyond his control affect the profits being shared.)
Option A: One simple employment option is offering the salary each year. The best possible case is if the prospect turns out to be trustworthy: he honors his agreement to work hard, and you profit 5×75 thousand over 5 years. The worst case is if the prospect turns out to be opportunistic: he breaks his agreement to work hard, and you lose 5×125 thousand.
Option B: Another simple employment option is offering profit sharing each year. Regardless of whether the prospect is trustworthy or opportunistic, you profit 5×25 thousand over 5 years. That lies in between the best-case and worst-case scenarios of Option A.
There are many textbook case studies and much empirical research  where experiments in profit sharing (Option B) yield higher profits than guaranteed salary (Option A). Profit sharing may seem like the best option when most workers are untrustworthy, but the presence of even a few trustworthy workers may cause the opportunistic workers to initially mimic the trustworthy workers. That suggests experimenting with a third option.
Option C: Employ the prospect on a year-to-year contingent salary: offer the prospect a salary for a year if the prospect has worked hard in every previous year, but offer the prospect profit sharing (a lower salary plus a piece rate) if he has ever shirked. How will the smart opportunist respond to that offer? Consider two possibilities. One response is the opportunist shirks in his first year, and so gets one year of shirking while on salary followed by 4 years working under profit sharing. The other response is the opportunist works hard in his first 4 years then shirks in his last year, and so gets 4 years of working hard while on salary followed by one year of shirking while on salary. Either way, the opportunist gets one year of shirking while on salary, but he chooses the latter response because his other 4 years are working hard while on salary, which he prefers to working while on profit sharing.
Putting it all together, the contingent salary (Option C) is the superior option in this context. In the best-case scenario of a trustworthy worker, the contingent salary (C) generates the same profit (5×75 thousand) as the guaranteed salary (A) and more profit than profit sharing (B). And in the worst-case scenario of an opportunistic worker, the contingent salary generates 4×75 thousand over the first 4 years (while the opportunist mimics the trustworthy) followed by a loss of 125 thousand in the last year, which exceeds the worst-case scenarios under the other two options. The superiority of the contingent salary is even stronger if employment extends beyond 5 years: the worst-case scenario under the contingent salary gets closer to the best-case scenario under the guaranteed salary because there are more years of the opportunist working hard but still only one year of the opportunist shirking. That reasoning suggests that employers that have already run successful experiments with profit sharing over guaranteed salary should further experiment with contingent salary for their long-term employees.
A principal agent model of profit sharing
“An unthinking person believes everything, but the prudent one thinks before acting.” Proverbs 14:15 (International Standard Version)
Empirical observations, laboratory experiments, and game theory offer coherent guidance for a prudent person doing repeat business with potentially untrustworthy partners. Such a prudent person does not trust in their partners’ character, but simply in their partners’ ability to recognize their own self interest.
- Even untrustworthy partners may initially act as though trustworthy in order to earn repeat business or referrals. So, do not trust a partner with a large deal just because they have previously acted trustworthy in small deals.
- Be especially careful toward the natural end of a business relationship because that is when untrustworthy partners are especially motivated to break agreements.
- The number of periods where untrustworthy partners act as though trustworthy increases as the total number of periods of the relationship increases. So, seek partners with which you can potentially do more repeat business. And, if possible, divide a big deal into a series of smaller deals (such as paying workers monthly rather than yearly).
- Punishing those acting dishonestly is an incentive for the untrustworthy to delay exploiting your trust, and so may forestall a collapse of trust and cooperation.
 Surowiecki, James. The Wisdom of Crowds (New York: Anchor Books, 2005), 120.
 Kagel, J., and Roth, A.E., (eds.). Handbook of Experimental Economics (Princeton, N.J.: Princeton University Press, (1995), 8-10, 26-28.
 Sameulson, William F. and Marks, Stephen G. Managerial Economics (Hoboken, N.J.: Wiley, 2014), chapter 14.
 Banker, R., Lee, S., Potter, G., & Srinivasan, D. “An Empirical Analysis of Continuing Improvements Following the Implementation of a Performance-based Compensation Plan. Journal of Accounting and Economics 30 (2000): 315-350.
 Lazear, Edward P. “Performance Pay and Productivity,” (July 1996). NBER Working Paper 5672.
 Sameulson, Managerial Economics.
About the Author(s)
Jonathan L. Burke, PhD, is a professor of economics, in the Business Administration Division of Seaver College, Pepperdine University, Malibu, CA. His research specialty uses general-equilibrium theory to offer a single model of perfect competition as a benchmark to compare various economic models. He seeks to find the breadth of economic environments that are consistent with perfect competition, and establish the determinacy and efficiency of perfectly-competitive equilibrium for intertemporal economic models. His primary teaching blends intermediate microeconomics and game theory to help managers make profitable strategic decisions. He can be reached at firstname.lastname@example.org.