Inequality is the result of many different phenomena. Some of them should be a source of policy concern while others should not. My main problem is the inequality that arises from differences in productivity—namely, differences in productivity across regions, across cities, within cities and across social groups. We know that there are huge differences in income across countries of the world: the richest countries are 200 to 300 times richer than the poorest countries in per capita terms. That’s inequality at the global scale.
That is mostly caused by differences in productivity. It’s not because there’s a global pie and it is shared unequally between the rich countries and the poor countries. These are just independent pies of radically different size. At the global level, the bulk of inequality across countries is inequality in productivity.
Our research has uncovered that in the developing world, there are enormous differences in productivity within countries, across their different regions. For example, in the US, the richest state, which is probably Connecticut, is about twice as rich as the poorest state, which is either Mississippi or West Virginia. The difference is a factor of two. In Mexico, the difference between Chiapas and Nuevo León is a factor of nine. Similar differences exist between the Indian states of Bihar and Goa or between the cities of Patna and Bangalore. These differences in income are mainly differences in productivity. It’s not the result of what share of the pie goes to capital and what size of the pie goes to labor. It is differences in the sizes of the pie.
So there are these enormous differences in productivity that make the productive places rich and the unproductive places poor. The poor people are not being exploited. They’re being excluded from the higher productivity activities. It’s not that the capitalists are taking a very large share of what they produce. It’s just that they produce very little in the first place.
Many of those that worry about inequality blame capitalism for it. Even Pope Francis has been framing the issue in this way. Now, let’s define capitalism the way Karl Marx did. It is a mode of production where some people own the means of production and others work as wage laborers for them. But if this is the case, capitalism hires 8 out of each 9 workers in the USA, 2 out of 3 in Nuevo Leon, 1 out of 7 in Chiapas and 1 out of 19 in India. Places where more of the labor force works for capitalist firms are richer, because capitalist firms allow for much higher productivity.
Poor places are characterized by the absence of capitalist firms and by self-employment, employment: these are small peasants and farmers or owners of small shop. In these settings, there are no wages, there’s no employment relationship. There are no pensions. There is no unemployment insurance. The trappings of a capitalist labor market do not exist.
While Marx thought that capitalism, as a form of organizing production, would take over the world, poor countries and regions are characterized by the absence of capitalism, of capitalist forms of production.
So the question we should ask ourselves is why did capitalism not succeed in these regions, leaving huge differences in productivity between the places where it succeeded and the places where it did not? The answer we have found is that modern capitalist production requires the simultaneous access to many different inputs.
For example, let’s look at Harvard Kennedy School: to operate, it needs electricity and access to the Internet. It needs an urban transportation system for its diverse staff to be able to go to work. It needs the ability to hire a faculty with very different talents, so as to produce what we produce. The lack of any one of these inputs has disastrous consequences. The day the lights go out, the school cannot operate. The day the Internet goes out, our productivity suffers: students will not know what to read for what courses, which events to attend, and we wouldn’t be able to do any research.
So access to all these inputs is necessary for productivity to happen. Absence of any one of these inputs has devastating effects. So this characteristic of modern production means that for places to be productive, they have to have everything.
The conditions for high productivity are very hard to achieve everywhere, but much easier to achieve in a few places. So governments are faced with the dilemma between concentrating all the inputs in a few places, and then getting the benefits of that concentration but also the inequality between those areas and the rest of the country, or trying to be very democratic in assigning inputs—say, electricity in a few places and roads in other places and internet access in some other places. Then no place has everything and if no place has everything, modern production becomes impossible everywhere.
I think that the deep underlying reason for this dilemma is the presence of increasing returns to the inputs. What do we mean by that? Simply that the cost structure of the input involves some fixed cost and then some variable costs. Consider each time we connect a house to the water network, the electricity network, the urban transport network, the road network, the educational system network, the labor market, or the banking system. All of these require someone to pay a fixed cost of connection. The fixed cost may be the copper wire or the pipes or the road that hooks up your home to these networks, the bus line that goes by your home, the accessibility to a labor market that you can go to work in and get back home in the evenings. These fixed costs are independent of whether a household is going to consume 100 kilowatts, 1000 kilowatts, or 5000 kilowatts [of electricity], or whether it’s going to consume 10 liters of water, 50 liters of water, or 1000 liters of water.
Then there’s a variable cost. That depends on how many kilowatts you consumed or how much water you consumed. But first you have to put in the wire or you have to put in the pipes or you have to build the road.
These fixed costs create increasing returns because the more you consume, the cheaper is the total cost per unit. Paying for these fixed costs becomes a headache because if somebody is expected to be poor, you don’t want to open a bank account for him because the fixed cost of opening a bank account is not going to be recouped through the little money or the few transactions that a poor person is going to make. So banks decide not to include the poor. The same thing happens with other services: if you are going to consume very few kilowatts or kilobytes, it doesn’t pay to connect you and if your expected wage is low relative to a bus ride, it does not pay to commute to work. As a consequence, this generates a trap in which you don’t connect people because they’re poor and because they’re not connected, they’re unproductive and hence poor.
This is a fundamental dilemma that needs to be addressed if we are to tackle the inequality problem and I think it is an issue that hasn’t been sufficiently emphasized.
There are two classes of solutions to this problem. The first one is that some technological innovations might reduce those fixed costs and if the fixed cost is reduced, more people can be included. For example, in India today, cell phone penetration is upwards of 80 percent. Landline penetration is, on the contrary, two percent. Why would this be the case? It’s not because landlines are a more recent technology that has not had the time to diffuse. It is because the fixed cost of connecting a home to the landline network is much higher than the fixed cost of buying a cell phone. As a consequence, cell phone technology diffused at light speed while landlines have not diffused. In fact, cell phones have a larger penetration than piped water, which covers less than 50 percent of the population. So technologies diffuse when the fixed costs are low and if technologies can be invented to lower these fixed costs, it facilitates diffusion.
Lowering the fixed cost was also the idea behind micro lending. Traditional banks don’t give small loans because the fixed cost of processing them is too high and would require unaffordable interest rates. So they exclude the customers that would have required a small loan, mainly the poor. Innovations in micro lending are all about reducing that fixed cost of lending, through ideas such as group lending. Mobile banking might allow us to further reduce these fixed costs.
The alternative to a technological solution is to have a policy that shares the fixed cost. A good example comes from the U.S. The Continental Congress in 1775, a year before the Declaration of Independence, decided to create a U.S. post office. They decided to put a post office in every incorporated city in the US. The postal system was the Internet of its time. They decided to pay collectively for that system and to have a flat rate so that any place within the country could communicate with any other place within the country. That is an example of sharing the fixed cost. Had it not been designed that way, small or poor towns would have been excluded and everybody else would have lost the opportunity to communicate with them. We can paraphrase the policy as saying: “We want a network where everybody is connected and we will use policy to make sure it happens.”
So policies can be very important in determining the universality of access to some inputs. I think it’s very important to have a serious discussion of what are these inputs that need to be accessed universally and what is a reasonable strategy to get there.