Some ideas areintuitive. Others sound so obvious after they are expressed that it is hard todeny their truth. They are powerful, because they have many nonobviousimplications. They put one in a different frame of mind when looking at theworld and deciding how to act on it.

One such idea is the notion that cities,regions, and countries should specialize. Because they cannot be good ateverything, they must concentrate on what they are best at – that is, on theircomparative advantage. They should make a few things very well and exchangethem for other goods that are made better elsewhere, thus exploiting the gainsfrom trade.

But, while some ideas are intuitive or obvious,they can also be wrong and dangerous. As is often the case, it is not what you don’tknow, but what you mistakenly think you know, that hurts you. And theidea that cities and countries actually do specialize, and thattherefore they should specialize, is one of those very wrong anddangerous ideas.

When an idea is both intuitively true and actuallyfalse, it is often because it is true on one level but not on the level atwhich it is being applied. Yes, people do specialize, and they shouldspecialize, too. Everyone benefits from each of us becoming good at differentthings and exchanging our knowhow with others. It is not efficient for adentist and a lawyer, for example, to be the same person.

But specialization at the individual levelactually leads to diversification at a higher level. It is precisely becauseindividuals and firms specialize that cities and countries diversify.

Consider a rural medical facility and a majorcity hospital. The former probably has a single general practitioner who isable to provide a limited suite of services. In the latter, doctors specializein different areas (oncology, cardiology, neurology, and so on), which enablesthe hospital to offer a more diverse set of interventions. Specialization ofdoctors leads to diversification of hospital services.

The scale at which specialization of individualsleads to diversification is the city. Larger cities are more diversified thansmaller cities. Among cities with similar populations – say, Salvador andCuritiba in Brazil, or Guadalajara and Monterrey in Mexico – more diversifiedcities are richer than less diversified cities. They tend to grow faster andbecome even more diversified, not only because they have a larger internalmarket, but also because they are more diversified in terms of what they cansell to other cities and countries.

What is true at the level of cities is even moreapplicable at the level of states and countries. The Netherlands, Chile, andCameroon have a similar population size, but the Netherlands is twice as richas Chile, which is 10 times richer than Cameroon. Lookingat their exports shows that the Netherlands is three times more diversifiedthan Chile, which is three times more diversified than Cameroon.

As my colleagues and I recently argued, one way to understand this is to think ofindustries as stitchingtogether complementary bits of knowhow, just as words are made by puttingtogether letters. With a greater diversity of letters, the variety of wordsthat can be made increases, as does their length. Likewise, the more bits ofknowhow that are available, the more industries can be supported and thegreater their complexity can be.

Cities are the places where people that havespecialized in different areas congregate, allowing industries to combine their knowhow. Richcities are characterized by a more diverse set of skills that support a morediverse and complex set of industries – and thus provide more job opportunitiesto the different specialists.

In the process of development, cities, states,and countries do not specialize; they diversify. They evolve from supporting afew simple industries to sustaining an increasingly diverse set of more complexindustries. Achieving this implies solving important coordination problems,because an industry that is new to a city will not find workers with industry experience or specializedsuppliers. But policymakers can do a lot to solve these coordinationproblems.

This is why the idea that cities, states, orcountries should specialize in their current areas of comparativeadvantage is so dangerous. Focusing on the limited activities at which theycurrently excel would merely reduce the variety of capabilities – or “letters”– that they have. The challenge is not to pick a few winners among the existingindustries, but rather to facilitate the emergence of more winners bybroadening the business ecosystem and enabling it to nurture newactivities.

This is all the more important today, becausethe globalization of value chains is delocalizing supplier-customer relations.Cities and countries would be ill-advised to focus on a few “clusters” andconsolidate the value chains in their location, as is so often recommended.Instead, they should worry about being a node in many different value chains,which requires finding other industries that can use their existingcapabilities if they were somehow expanded and adjusted to new needs.

Competition inevitably tends to winnow out the lessefficient firms and industries. It is not the policymakers’ role to hastentheir death. Their task is to identify productivity-enhancing interventionsthat can harness economies of agglomeration by adding new activities andproductive capabilities, making the whole bigger than the sum of the parts.