We rely on experts for a lot of our information. By “expert,” I mean someone who is paid for their opinion. Roger Koppl uses this definition in his 2018 book Expert Failure, and I use the same definition in my research, which is based on his book. This definition is useful because it allows us to sidestep the endless (and, frankly, arbitrary) discussion about who counts as an expert.
So, we rely on experts for their opinions in much of our lives. We rely on experts for weather, medicine, technical services like auto work, and so on. But organizations—both governments and private firms—also rely on experts. Often, they have their own internal teams of experts to help provide information and opinion. The expert embedded in a bureaucracy provides for an interesting dynamic that I explore in a recent working paper, “Experts Whispering Down the Lane.”
When we deal directly with an expert (our doctor, our mechanic, our meteorologist), we can get the information firsthand. But in a large organization, there are often several layers of communication between the expert (the one producing the opinion) and the non-expert using the opinion.
For example, let’s say that a VP of Marketing wants to know how best to market a new product. He asks his manager to get the information needed to make his decisions. That manager will similarly instruct his team. Assuming this team is the source of expertise, they will go out, gather the required information, and form an opinion. The team then summarizes their opinion for the manager, who similarly summarizes it for the VP of Marketing. The VP of Marketing then makes a choice, and their orders get transmitted throughout the company. In a 1966 RAND Corporation paper, Anthony Downs notes that there can be a breakdown in communication as orders get transmitted down the chain of command. Downs calls this “authority leakage.” I show that information also gets lost as it is relayed back up the chain of command. I call this “information leakage.” Key bits of information can get lost, messages may be garbled, and the VP may make a decision based on misunderstood information.
It is also worth noting that asking for clarification to clear up misconceptions is more costly in a bureaucracy because the non-expert is not directly communicating with the experts. Instead, both groups operate through formal organizational structures.
Several issues contribute to the problem of information leakage. First, there is the economic problem of scarcity. Everything is scarce, including time. When an expert is summarizing a report, they have to make choices about what information to include and what not to. There is also a tradeoff between technical precision and comprehension; jargon can be misunderstood by those not trained in a subject. But jargon is used for a reason: translating jargon into colloquial terms can strip away precision and lead to misunderstanding (I write about this problem here).
Furthermore, humans naturally communicate in both a literal and metaphorical manner, rather than always 100% literally. Take, for example, this satellite image of a record snowstorm that dumped 40 inches of snow on Massachusetts this past February. Note the well-defined eye and the spiral motion of the clouds. The storm produced sustained winds of 85 miles per hour. In describing this storm to my Louisiana students (who have only seen real snow once, last year in a freak blizzard), I described it as a “hurricane with snow.” As a literal statement, that’s not true. Hurricanes have certain characteristics that define them as a hurricane, not the least of which is that a hurricane is a tropical storm. But, as a metaphor, it served its purpose: it conveyed to the students the seriousness of the storm.
While it’s meant to economize in communication, the hurricane metaphor could be lost in translation. Someone taking things too literally could be misdirected and think a blizzard is just a hurricane with snow. Without context (and a little bit of luck), it can be difficult to know whether someone is speaking literally or metaphorically.
So, the main lesson here is that the more layers there are between the expert and non-expert, the greater the chance there is for the expert message to be garbled. It’s the telephone game. (And that’s before we account for any personal motivations of people involved in those layers.)
Why tolerate these problems that bureaucracy introduces? There is a tradeoff: a bureaucracy can gather more information than an individual. As a result of this advantage, non-experts in bureaucracies still have access to more information than they otherwise would have, despite much of it being lost in the transmission process.
As a bureaucracy of experts increases in size, the total amount of information gathered increases, but the reliability of information transmitted decreases. Thus, we have a theoretical optimal level of expert bureaucracy. (What that precise level is is a question for another time.)
A secondary lesson I hope readers take away is about the nature of Public Choice. Public Choice scholars (myself included) often make the mistake of focusing heavily on government/judicial/voter behavior. But Public Choice was originally called “non-market decision-making.” The same problems that infect governmental decision-making affect decision-making within firms and other non-government organizations. Individual firms are not necessarily more efficient than governments, but when they do make mistakes, the mistakes are less likely to cascade because the market punishes wasteful behavior with financial loss. Consequently, profit-seeking organizations face an incentive to take corrective measures. An equivalent mechanism does not precisely exist in government contexts. Nevertheless, the problems of non-market decision-making exist within the firm.
