In the fall of 2011, I joined the Department of Economics at the NYU Stern School of Business. There I am starting a new program on the wave of urbanization that is now bringing billions of people into cities.

In this century, the world’s urban population will gain more residents than in all of history to date. Because the world population will stabilize by the end of the century, humans now have a chance that we will never have again: to start dozens, perhaps even hundreds of new cities. These startups could foster a truly historic burst of economic and social progress.Prior to joining Stern, I taught at Stanford’s Graduate School of Business, which led to a teaching award and an entrepreneurial detour into educational software. Before moving to Stanford, I taught at the University of California at Berkeley, the University of Chicago, and the University of Rochester. I am a fellow of the American Academy of Arts and Sciences and received the Recktenwald Prize in Economics.

Growing up, my father, Roy Romer, was a rising star in Colorado politics. Like many young people raised in the shadow of a successful parent, I tried to strike out in a different direction, studying mathematical physics and cosmology in college.

Graduate work in economics (first at MIT, then Queens University, and the University of Chicago) was a compromise that brought me back toward the policy concerns I had been exposed to as a youth but still left plenty of room for abstract theorizing of the type that attracted me to physics. My Ph.D. thesis asked why growth rates had been increasing over time. Fresh from cosmology, I was not particularly motivated by any practical concerns. It just seemed like an important puzzle.

Existing economic theory suggested that scarcity combined with population growth should be making things worse, but in fact life kept getting better at ever faster rates. New technologies had to be the answer. Everyone “knew” that. But why do new technologies keep arriving at faster rates? One key insight is related to the special property of ideas. Because ideas are nonrival, or sharable, interacting with more people makes us all better off. These benefits show up in part as faster discovery and growth. This potential for sharing ideas not only explains why growth rates are speeding up, but as a side effect, also helps explain why we cluster in cities with so many other people. (See my recent paper with Chad Jones for more on these points.) More recently, I have shifted back toward a question motivated by a pressing policy concern: If we can share technologies, why is it that we often don’t? In a TED talk, I illustrated this painful point with a striking picture of students doing homework at night under street lights. The technology for producing low-cost light in homes is more than 100 years old. Why is it that so many people still don’t have access to it?

In thinking abstractly about this question, I found it helpful to avoid the use of a monolithic concept like “institutions” and to look instead at the atoms of predictable social interaction – rules. Bad rules can keep valuable technologies out of a country. But why do bad rules persist? Why don’t countries with bad rules simply copy better rules, which then make it possible to copy existing technologies?

I find that abstract theoretical work benefits from a careful look at something specific. Consider for example this familiar rule: “Stop on red.” In a modern urban city center, there is a better guide for drivers: “Don’t enter the intersection unless you can leave before the light turns red.” However, despite its value in preventing gridlock, ”Don’t block the box” has rarely become the rule. That is, it is not a predictable regularity in our social interaction. Laws mandate this behavior, but rules depend to a very large extent on individual norms about right and wrong. Most people feel that they are in the right so long as they enter the intersection when the light is green.

I am convinced that the most important fluctuations in the rate of human progress depend on the dynamics of rules, which in turn depend on the dynamics of norms. Because our norms are often determined through a process of social interaction in which common norms are reinforced, they tend to be stable. Because this process of social transmission operates through our preferences, outside of conscious awareness, there is little pressure for inefficient norms to change. An inefficient norm (such as “it is right to enter the intersection when the light is green, regardless of the traffic ahead”) can seemingly persist indefinitely.

Much of the texture and complexity of social progress is the result of a mismatch between stable social norms on the one hand, and the combination of continuously evolving technologies and continuously increasing scales of human interaction on the other. When technology changes and the scale of our interaction grows, our rules should change in parallel, but stable social norms can get in the way.

Nevertheless, a stable equilibrium in rules can sometimes be punctuated by a burst of change. Startups seem to be a critical mechanism in this dynamic. For example, Clayton Christensen has shown that startup firms take up new technologies much more effectively than existing firms. They do this in part by developing new norms that become part of a new corporate culture. For example, Target was an internal startup that Dayton-Hudson used to develop the new norms that supported discount retailing. A parallel, but much more important internal startup was Pennsylvania. Initially part of the British realm, Pennsylvania was used by King Charles II and William Penn to develop new norms of religious tolerance and individual freedom.

My attempts at starting a charter city come from a conviction that humans can use this startup dynamic to let progress unfold much more quickly during this century of rapid urbanization in the developing world. The progress that could result would involve not just improvements in material well being, but the deepening of such fundamental norms as freedom and inclusion.