ONE baker in Lahore guarantees the quality of her product by saying that even if you have eaten half a cake and think it is not of the quality it should be or is not fresh enough, she will replace the product for you.

This shows the confidence that the manufacturer has in her product. She is essentially saying that she will supply only quality products and if there is a problem, she will offer a full replacement.

Warranties are quite common in the goods market. They are harder for services: if a dentist has extracted your tooth, it is hard to put it back. But for most consumer products, especially durables, warranties are quite ubiquitous.

If honoured, they are an excellent means of reducing the risks of malfunction and poor service for the customer. Durable goods last over time and the customer wants a stream of services over this period. Warranties ensure that if things go wrong — durables are usually big-ticket items — her purchase will not go waste. This confidence is important for a customer deciding to purchase a durable as it is almost impossible for most people to judge how a car or a television will perform six months down the road.

Why do all manufacturers of durables not offer warranties? If it is a simple matter of giving the customer confidence about service and guarantees of performance, what stops them? Simply put: the quality of the product. For producers of high-quality products, offering warranties is easy and is a cost-effective means of saying that they stand behind their product. Producers of poor-quality goods cannot do that.

Suppose there are two manufacturers. A produces excellent televisions while B produces televisions of poor quality, where quality is defined in terms of probability of breakdown. A’s televisions have a low probability of breakdown over the first three years while B’s televisions break down very often. If both producers price their product at the same level, where A breaks even because of his higher cost of production and B makes a lot of profit because his costs are lower, then clearly B is doing very well in this market. A now offers a warranty that he will replace any broken televisions with new ones. If B follows suit, she will have a lot of televisions that she will have to replace while A will have very few. B’s costs will escalate if she lives up to the warranty promise. If B’s quality is poor enough, the cost will rise so much that she will not be able to honour the warranties that she offered and will either be forced to withdraw the warranty or go out of business.

A high-quality producer offering warranties will also get repeat customers while a low-quality one without warranties will have one-time customers. Under certain conditions the market could become divided along these lines, and this could be a stable state. Those who have the money and want the assurance of quality would go for the high-quality product. For others, taking the risk of buying the cheaper product without warranty might still be better than not having the product.

But if quality differences are not large enough and the low-quality producer can mimic the high-quality one successfully, the equilibrium above might not obtain. In some cases the low-quality producer, mimicking equal quality but having the ability to lower prices compared to the high-quality producer given her lower cost, can drive the high-quality producer out of the market. This is definitely a poorer outcome for customers, but it can happen.

Warranty can thus be a signal of quality. But it is not the only signal of quality and it does not work in all situations. For services and goods that are not durable but get used in consumption, such as food and medicines, warranties are not easy to devise. The cake example was special as it is possible to return half a cake. But it is hard to take a medicine and know if it was good enough.

Brand names and advertising expenditures work in the same way warranties work for durables in many such situations. Companies that produce high-quality products and services will work to establish brand names and will advertise because they can be confident that their product quality will get them repeat customers. For low-quality producers, advertising or establishing a brand name is not worth it as they will not get repeat customers. Advertising expenditures and spending on brand names also offer the kind of separating equilibrium that was mentioned in the case of warranties. But again, if the space between the quality of products is not large enough, it will be hard to separate the companies and markets could lose high-quality producers.

Local manufacturers and retailers often opine that customers are not quality conscious. Could this be a reflection of a lack of separating equilibria in many markets? It would be worth studying this issue and seeing if any intervention is needed to ensure warranties and other promises are kept and that high- and low-quality products can be separated and that both can exist.

I would rather have cake from a baker who guarantees that if the quality is not good she will give my money back. It gives me confidence about the quality she is selling. I would prefer to buy durables that have money-back or similar warranties. For services I would prefer companies that are looking for repeat customers and are willing to spend money to cultivate them. Fly-by-night operations are inevitably low-quality producers. These issues have not yet made it into the public-policy space in Pakistan. It is time we start thinking about them and any public-policy implications that might go with them.

The writer is senior adviser, Pakistan, at Open Society Foundations, associate professor of economics, LUMS, and a visiting fellow at IDEAS, Lahore.

fbari@osipak.org