Beating the competition: Part I - The prisoners are running the asylum

I think it’s pretty much a given that everyone understands how competition can create value. Without competition you get things like monopolies. Competition reduces prices. Competition improves service. Competition improves efficiency. It drives innovation. So to improve products, services, cost, efficiency, productivity, and innovation we should create more competition. De-regulate, open up markets, let the markets sort it all out. Right?

Well, not so fast. Yes, there is little question that competition is a useful tool. But I inserted a subtle, yet vital logical fallacy in the reasoning above. Did you see it? Hint: it’s a form of affirming the consequent, or converse error. That’s where you say “Colds cause me to sneeze. I am sneezing, therefore I must have a cold.” The error should be obvious. It doesn’t say that only colds cause sneezing, nor does it say that colds only cause sneezing.

Indeed, competition can create benefits. That doesn’t mean that competition is the only thing that creates these benefits, nor does it mean that the only thing that competition creates is benefits. Over the next few posts, I want to look at these less discussed cases. In this post, I’ll look at other things that competition can cause that aren’t so beneficial.

Reward Systems That Hurt

The problems with competitiveness are perhaps best described by W. Alan Randolph and Barry Z. Posner in their book Getting the Job Done! Managing Project Teams and Task Forces for Success. In a section entitled “Reward Systems That Hurt” (p. 23-24), the authours provide a case study in broken reward systems and contain the following excerpt:

A good example of this breakdown occurred at a naval air station that had four squadrons responsible for the maintenance of planes. The stated goal was to have 95 percent of all planes ready to go at any time. Actual readiness was running around 85 percent. The new commanding officer decided he would re-assert this goal of 95 percent and backed it up with a reward system. Each month, he would give the particular squadron with the highest percentage of planes a reward, such as a 24-hour pass or public recognition of a job well done. What happened?

The members of one squadron decided that the secret was to have a well-stocked parts inventory. Then if a plane came in with, say, a broken radar part, they could simply take that part out and put in a replacement and the plane would be ready to go again. They also realized that if they could increase their parts inventory while decreasing the parts inventory for the other squadrons, they would be well on the way to winning this competition. The obvious solution? Steal parts! This may sound ridiculous, but it is a true story.

One of the squadrons began to engage in midnight requisitions to steal parts from the other squadrons. The other squadrons figured out what was going on and decided to retaliate. Soon four squadrons were stealing from one another. Every squadron put guards on duty 24 hours a day, seven days a week to protect its inventories. Meanwhile, the percentage of planes ready to go was going downhill fast.

The problem was the reward system. People knew the overall goal but felt that their own squadron’s objectives were more important, especially since these objectives were what was being rewarded. A simple change in the reward system solved the problem. In a given month, any and all squadrons that achieved the goal of 95 percent readiness would get the reward. Now, every squadron could get the reward. What happened? First, the stealing stopped. The squadrons started sharing parts to help one another out. The readiness percentage quickly rose to 90 percent and began pushing toward 95 percent – all because the reward system was changed to support the accomplishment of the objectives and the project goal.

This case study is a prime example of a well-known problem of competitive pressure: an escalating arms race of inefficiency, or rather inefficiency towards the ultimate objective. Competition did indeed drive innovation, but it was aimed at increasingly sophisticated efforts at stealing, sabotage, and security and it subtracted from the ultimate goal of improved flight readiness. Even at the squadron level, the proximate goal of individual superiority in flight readiness over the competition did not improve flight readiness for any individual squadron. All squadrons got worse. Incentives were designed to increase competitiveness, not flight readiness.

This situation is known in economic game theory as a Prisoner’s Dilemma which is characterized by having individual incentives act against both collective and ultimate individual results. The solution to a Prisoner`s Dilemma is to set the incentives to reward efficient value. (I’ll discuss this concept in a later post and update the reference here.) In the flight readiness case, the solution had two key features: it aligned individual incentives with ultimate goals, and it created a collaborative market. Not only did all squadrons increase in flight readiness and enjoy more rewards by focusing personal attention to the task, they also benefited by collaborating. Any excess resources, whether parts, time, or ideas, could be traded with other squadrons, thereby creating a market that increases both individual and collective efficiencies through collaboration rather than competition.

This is an important lesson towards both designing incentives and measuring effectiveness. With the original naval base competitive incentives, a measure of proximate goals would show increased innovation, increased effort, increased work output, and increased competitiveness, but significantly decreased flight readiness. When these inefficiencies cost more than the benefits of the competition, a Prisoner’s Dilemma results. The effectiveness of government incentives, be they regulations or funding programs, must therefore be evaluated in the context of aligning proximate goals with ultimate goals. Competitiveness is often neither an appropriate driver nor appropriate proximate goal.

The Inefficiencies of Smoking

I’ll play devil’s advocate here and predict that someone might object to the naval air base example being too removed from normal business markets. Squadrons are not companies seeking profits, one might suggest. A base CO is not the same a government or regulator. We can’t just change market rules to “reward” all companies who produce better products and services with more money.

Fine, I won’t argue with that, though I wouldn’t entirely agree. However, they do translate into those terms. The Prisoner Dilemma is a generalized abstraction of a principle. It does not depend on the type of reward, incentives, or regulatory organization. Let’s consider the case of cigarette advertising.

Cigarette advertising has been banned in the United States, Canada, and other countries in various form for a few decades now. When these bans started, the major tobacco companies saw profits soar on the order of hundreds of millions of dollars per year, as that had been the size of their advertising budgets. With those sorts of savings, one might wonder why they didn’t individually volunteer to cease advertising prior to the ban if such profits were available. The answer is competitive advantage. Advertising was not a big driver for the size of the cigarette market, but it was a big driver for the market share of each company. Any company that volunteered to cease advertising would lose market share to the competition and eventually go out of business. This is why the major tobacco companies voluntarily assisted in the drafting of legislation that banned them from advertising. They couldn’t individually stop advertising to save costs, but they could if they knew nobody else would. Binding everyone to a policy that benefits them all is a solution to the Prisoner’s Dilemma.

The price of cigarettes are typically regulated so as to keep customers from buying them, and with high taxes to cover the associated health costs. However, supposing the prices were not regulated, competition between companies could then drive the price down using those freed-up millions from the advertising budget. Hence we get an interesting outcome. If we de-regulated prices but regulated advertising, the customer would get the best value. It is not the principle of regulation that is a problem; it is the form of regulation, just as on the naval base it was the form of reward. Regulations are essentially punishments (fines for breaking rules), so the incentives can be described in general terms of punishments and rewards.

Now I’m not suggesting that we de-regulate cigarette prices. It is unfortunate that I’ve used such an unhealthy example to make the point. It is simply a convenient one and could apply to other industries. Advertising costs come at a cost to consumers. If we banned advertising on running shoes, customers could get better value as well, both in terms of saving the expense and because sales would better match the quality of the shoes rather than the quality of the advertising. But, that sort of policy would be somewhat against the idea of a free country, allowing sellers to advertise what they wish. It is a waste cost that we citizens accept.

Transactional Inefficiency

The problems I’ve described here fall into the category of transaction costs. These are costs associated with carrying out the transactions. In order to buy shoes you don’t just pay for the shoes. You pay for the advertising, the stores, the employee salaries, the transportation, and so on. This is one reason why mail order was very popular and now online shopping has replaced it. There can be a big benefit in terms of efficiency and cost. Of course, there is greater risk in some cases. Buying clothing online might not be so great. The shipping time and inconvenience of returning ill-fitting clothes, or ones that don’t look good on you, are transactional costs.

Competition does a good job of taking care of those inefficiencies. Right? Sure, to some degree and in some cases. What about health care? With the health care debates during the last U.S. election there was much discussion of the benefits of universal (single-payer) health insurance versus competitive insurance. Can competition be the better model?

Well, what do you pay for? In a competitive model you have to pay for marketing, sales departments, contract writers, claims review boards, redundant office overhead, claims litigation, and of course, profit. You also pay for collections staff at hospitals, qualification checkups to get coverage, and any sort of kickback agreements that happen between doctors, hospitals, and insurance companies. In a universal coverage system, you pay for none of those.

Can competitive pressure drive down prices to compensate for all of these added expenses (and more)? Don’t forget we’re talking about insurance here, not running the medical care or services. The product of insurance is payment for need. Can that be made cheaper? Well, you can refuse to pay some people in need. But then that doesn’t create value to customers, so that’s not better. Perhaps you can cut down on fraud. Which is more likely to be better at that, a bunch of small, independent companies with their own fraud investigators (that you pay for in premiums), or a much larger single entity with the power of government policing behind it?

What else? Well, perhaps the insurance companies can influence the cost of care by finding the cheapest means to perform it (and study effort you pay for) and pressure hospitals and doctors to implement it. Sure, but assuming the effort costs less than its outcome, who can put more pressure to do that, small separate insurance companies or one massive agency that dictates how much it will pay for each type of care? Seems obvious. Universal coverage in theory should be cheaper, and empirically it is. Perhaps I’ve missed a means by which competition might do better, so I’m always open to hearing other ways (and empirical evidence).

When Competition Works

So am I saying competition is bad or should be done away with? No, of course not. Under some circumstances it does a much better job. These circumstances are almost all in the development of new markets and products, or where there is a private monopoly or oligopoly. Competition can bring down prices very quickly. But when competitive efficiency is exhausted and the market matures, improvement can typically be done by forcibly removing the transactional costs of the competition by changing the incentives (rewards and punishments) such as regulations.

This means there is no hard rule. Deciding what measures are best depends highly on the circumstances. The open market doesn’t always win. Sometimes it loses big time.

The rest of the story: Part I, Part II, Part III, Part IV, Overview