2012 saw the passing of a great development economist, Albert O. Hirschman, at the age of 97.

Development economists spend their time these days performing randomized controlled experiments, in which a particular intervention like co-payments for mosquito bed nets are introduced into one group of villages and not into another matched set. This approach establishes causality with a level of certainty approaching that of the randomized trials used in pharmaceutical testing. But while such experiments are useful for evaluating the effectiveness of certain types of public policies, they all operate at a very micro level and don’t aggregate upwards into an understanding of the broader phenomenon of development. It is hard to imagine that all the work being done under this approach will leave anything behind of a conceptual nature that people will remember fifty years from now.

Albert Hirschman operated at the opposite end of the spectrum. He did very little quantitative work, and will be remembered for a series of slender books written in an accessible English that non-economists have no trouble understanding. He did not observe the methodological straightjacket his discipline imposed, but wandered off instead into other fields like politics and philosophy in an attempt to recover some of the unified social theory of the 18th and 19th centuries–hoping to avoid, as he put it, the “specialization-induced intellectual poverty in this field.” His legacy is not data collection or micro results, but rather some very big concepts that continue to shape the way we think about not just development but public policy more generally.

Hirschman’s best known books were Exit, Voice, and Loyalty (1970) and The Passions and the Interests (1977). In the former, he took on an issue central to public administration, namely, the problem of disciplining poor or incompetent managements like those running many American public schools. Milton Friedman had recently introduced the idea of vouchers and competition. He argued that in the private sector, bad management was disciplined by the possibility of exit, either on the part of customers who didn’t want to buy the company’s products, or by shareholders who lost confidence in the company’s management. This discipline didn’t exist in the public sector because it was often a monopoly supplier of the good in question, such as education. If parents were allowed to use a mechanism like vouchers to take their tax dollars away from failing schools and put them into better ones, there would be market-like incentives for both the competitive and failing schools to improve their performance. Since then, an exit option from state-provided public services as been a staple of public sector reform, something that spread widely after the rise of Reagan, Thatcher, and orthodox market economics in the 1980s.

Hirschman outlined the logic of the exit option and how noted how increased competition could improve government performance. But competition didn’t solve all problems, and the exit option had several important drawbacks. The freedom to exit was often used by the most ambitious, educated, or well-to-do users of a particular service, and once they exited, those remaining were even poorer, less educated, and less demanding. Moreover, Hirschman pointed out, the possibility of exit weakened the effectiveness of voice, that is, the ability to directly change the management’s behavior through feedback, discussion, and criticism. Sometimes loyalty was necessary to build the trust necessary to persuade people to change.

There are many examples of these insights playing themselves out: the end of legal segregation in the 1960s led more ambitious African-Americans to leave inner cities, condemning those who remained to even greater poverty and social breakdown. School vouchers have never quite worked as promised, because public schools were mandated to serve those residual students who couldn’t take advantage of exit. And so on.

The Passions and the Interests: Political Arguments for Capitalism before Its Triumph was basically a work of political philosophy, an interpretation of several writers using a humanistic approach that was and is utterly foreign to contemporary rational choice theorists. Hirschman argued that the rise of capitalism could not have occurred simply as a result of changes in underlying material conditions, as both Marxists and contemporary neo-classical economists believe. The very idea that it was morally legitimate to rationally maximize one’s income, far from being a universal postulate of human behavior, was something that took hold only during the 17th and 18th centuries. Earlier aristocratic societies had moral systems grounded in honor rather than gain, that were contemptuous of money-making and the calculating bourgeois way of life. Virtue lay rather in risk and glory in battle. The theorists that Hirschman covered, like Montesquieu, James Steuart, John Millar, and Adam Smith made political rather than economic arguments in favor of capitalism. They maintained that a commercial society would soften manners and morals, and in contrast to warrior societies would lead to greater international peace. Hirschman pointed out that these arguments have triumphed so completely in the modern world that we do not even perceive their historical contingency.

Hirschman was not just a theorist but a practical economist who spent a great deal of time working in the field, particularly in Latin America, advising countries like Colombia and Brazil on economic policy on behalf of various international institutions. Fernando Henrique Cardoso, the former president of Brazil, recounts in his memoirs how as a young academic he accidentally encountered Hirschman in a remote Brazilian village. I believe that some of Hirschman’s greatest insights came from his practical experience and were contained in his lesser-known books, centering around what he called “reform mongering.”

Albert Hirschman was a progressive. He believed in the importance of economic development, social change, just distribution of resources, and the welfare state. But he also had a realistic understanding of how difficult social change was to accomplish, and spent a great deal of time dissecting the modalities of bringing it about. A book I have used frequently in teaching was his 1963 work Journeys toward Progress, which chronicled reform efforts in Chile, Brazil, and Colombia. The Colombian story was about the slow efforts of democratic governments there to bring about land reform, beginning with legislation in response to land invasions in the 1930s and culminating in passage of a landmark agrarian law in the 1960s. He notes all of the misperceptions and outright mistakes of the reformers and their international advisors, and the unintended consequences of their well-intentioned actions. But he also shows how slow reform mongering over the years eventually brought about real progress. Colombia even now has not solved this problem. After Hirschman’s book was written the narco-traffickers took over and the current government of Juan Manuel Santos is seeking once again to redistribute their ill-gotten holdings to poor peasants. But the point remains that reform by democratic governments is both possible and necessary.

Hirschman formalized these arguments in one of his last books, The Rhetoric of Reaction (1991). In it he reviews the strategies that conservatives–well, reactionaries is the term he actually uses–have used to criticize progressive reformers who attempt to bring about social change. One is the perversity argument–the case that well-intentioned social engineering always entails unforeseen consequences that ultimately undermine the reformers’ goals and often leave society worse off than before. He notes that a recent example of this was Charles Murray’s 1984 book Losing Ground, in which Murray argued that the Depression-era Families with Dependent Children (AFDC) program was creating new generations of welfare-dependent single parent families and contributing to the collapse of American inner cities. Hirschman doesn’t debate the merits of this particular case, but rather notes that Murray’s argument was nothing new. The French Revolution engendered a wave of arguments–most notably those of Edmund Burke–who warned that revolutionary change would bring about terrible consequences. Similar arguments were made throughout the 19th century in opposition to expanding the franchise and poor laws. Indeed, he notes that the moral hazard argument that is central to the contemporary case against welfare was raised by British critics of the Speenhamland poor relief measure from the 1790s. (Those of you who have read Karl Polanyi’s The Great Transformation will know all about Speenhamland and the unintended moral hazard it created.) The welfare-to-work principle embodied in the 1996 Personal Responsibility and Work Opportunity Act that abolished AFDC during the Clinton administration was actually anticipated in Britain’s 1834 Poor Law Amendment Act. This piece of legislation was so harsh in its effort to stigmatize dependency on public assistance that it engendered its own reaction, and explains, according to Hirschman, why moral hazard and perverse consequences arguments were absent from debates on welfare in Britain for the next several generations.

Hirschman did not try to argue that conservatives were always wrong in calling attention to unanticipated consequences. He simply said that to turn the possibility of unintended negative consequences into a universal principle, and a reason for opposing all deliberate efforts at reform, was wrong. Opening up the vote to all adult citizens did not undermine Western civilization, as Gustave Flaubert and Gaetano Mosca argued. He was the polar opposite of Friedrich Hayek and the latter’s theories of spontaneous order.

Hirschman did not approve of revolutionary change. His preferred course of reform mongering was one of slow but steady gradualism under democratic governments. He advocated this approach to Latin Americans in the 1960s when many were dreaming of further Cuban-style revolutions. His message to revolutionaries was that democratic change and reformism were slow and often disappointing, but that they worked much better in the end. This is a lesson many still need to take to heart.

One of my favorite Hirschmanian concepts was that of the Hiding Hand, a play on Adam Smith’s Hidden Hand, which he laid out in his 1967 book Development Projects Observed. The book analyzed a number of World Bank projects on which he consulted, and noted how a number of them failed to achieve their objectives or else produced unexpected results. But he argued that the failure to anticipate unintended consequences was actually a good thing. If we could foresee all the possible negative consequences of our actions, we would become completely paralyzed–not just as governments seeking social change, but as individuals wanting to try new things in work, love, or life in general. The Hiding Hand that blinded us in this fashion was thus Providential.

They don’t unfortunately make development economists like Albert Hirschman any more.

I want to make one correction to an assertion I made in my last blog post.  In it, I said that delivery of services like education and health care is something that “states accomplish, and not the institutions that check them.”

This is a big overstatement.  The checking institutions actually play a big role in improving the delivery of services and controlling corruption.  Courts are often used to force executive branches to carry out mandated tasks, and of course to prosecute corrupt officials.  Democratic accountability, free media, and open information are  critical in disciplining corruption and keeping on pressure to improve performance.  Much of the international donor community has been promoting mechanisms to increase transparency and accountability in governments as the primary route towards good governance.  The Open Government Initiative developed by my colleague Jeremy Weinstein is another worthy effort in this direction.  Clearly, the more information that’s out there about corruption and bad governance, the more people are likely to mobilize around pressuring executives to fix things.

I was trying to make a different and more complex point.  There is no question that greater transparency and accountability, as well as strict application of the law, are critical to improving the performance of governments.  However, without basic capacity, no amount of transparency and accountability will produce good services.  If you look around the world at all of the great bureaucratic traditions—Germany, Sweden, Japan, Singapore, etc.—not one of them became great because of democratic accountability.  In fact, many great bureaucracies were created by authoritarian regimes that needed efficient services, primarily for the sake of national survival.  This was true of American state-building as well in the Progressive Era–something that I will have to address in a later post.

Moreover, democratic accountability is a double-edged sword.  We of course want bureaucracies to be broadly responsive to public needs, rather than serving their own interests, or those of the elites that appoint them.  On the other hand, there is a permanent tension between accountability and bureaucratic autonomy based on expertise and merit.  Democratic accountability produces its own forms of corruption, as when politicians mobilize voters clientelistically.  In clientelistic systems politicians don’t just set policy and appoint heads of agencies, they choose personnel throughout the bureaucracy.  Politicians representing, in theory, the public will, often make populistic choices at odds with long-term public interest.  This is why executive branches have to be shielded from day-to-day legislative oversight.  The need for bureaucratic autonomy is why we don’t turn monetary policy or military strategy over to our elected representatives for management.  Most Americans seem to recognize (at least implicitly) the importance of expertise and autonomy; surveys like General Social Survey tend to show that the most respected parts of the US government are actually the least democratic:  the Supreme Court, the military, the Centers for Disease Control, etc. By contrast, the part of the government most immediately accountable to public opinion, the House of Representatives, is the least respected of all.

So the checking institutions (law and accountability) have to do their job to force executives to serve the public will.  But checks by themselves do not produce the expertise and enforcement power needed to govern effectively.  I argued last year in the Financial Times that the American system has too many checks, when compared to other democratic systems, and has been steadily accumulating more as time goes on.  Vetocracies can stop bad things from happening, but they also don’t provide for much by way of effective collective action.

It is a curious fact that in contemporary American political science, very few people want to study the state, that is, the functioning of executive branches and their bureaucracies. Since the onset of the Third Wave of democratizations now more than a generation ago, the overwhelming emphasis in comparative politics has been on democracy, transitions to democracy, human rights, ethnic conflict, violence, transitional justice, and the like. There is of course interest in stability, but primarily as the absence of violence and conflict. Studies of non-democratic countries focus on issues like authoritarian persistence, meaning that the focus still remains the question of democracy in the long run or democratic transition. In other words, most people are interested in studying political institutions that limit or check power—democratic accountability and rule of law—but very few people pay attention to the institution that accumulates and uses power, the state.

The relative emphasis on checking institutions rather than power-deploying institutions is evident in the governance measures that have been developed in recent years. There are numerous measures of the quality of democracy like the Freedom House Freedom in the World and Polity IV measures, as well as newer  ones like the Varieties of Democracy project led by Michael Coppedge, John Gerring et al. What we do not have is a good measure of Weberian bureaucracy—that is, the degree to which bureaucratic recruitment and promotion is merit-based, functionally organized, based on technical qualifications, etc. One of the only studies to attempt to do this was by Peter Evans and James Rauch back in 2000, but their sample was limited to 30-odd countries and produced no time series data. There is also a proprietary cross-country measure, the Political Risk Service’s Group (PRSG) International Country Risk Guide, but because it is proprietary we don’t really know what goes into it. Several of the World Bank Institute’s Worldwide Governance Indicators purport to measure state aspects of state capacity (government effectiveness, regulatory quality, and stability and absence of violence, control of corruption), but these are aggregates of other existing measures and it is not clear how they map onto the Weberian categories. For example, does a good absence of violence score mean that there is effective policing? I suspect that there isn’t much street crime in North Korea. (There are similar problems with the Bank’s internal CPIA scores.)

One important measure that would be great to have but which no one has ever attempted to create, to my knowledge, is a measure of bureaucratic autonomy, that is, the degree to which bureaucrats are under day-to-day control by their nominal political masters, both with regard to policy and with regard to control over cadres. This is utterly critical in understanding bureaucratic quality, and yet is totally unavailable for any kind of quantitative analysis.

The bias against thinking about state capacity is particularly strong among rational choice institutionalists, who dominate many of the best political science departments these days. Most in this school begin with Mancur Olson’s assumption that states are predatory, and that the chief aim of political development is the creation of institutions like rule of law and accountability that limit the state’s ability to act arbitrarily and allow it to commit itself credibly to the protection of property rights.  This school assumes that all states have the power to be predatory, and seldom raise the question of where state capacity comes from in the first place, or how it increases or decreases over time. Frankly, it would be very hard to develop a rational choice theory of state capacity, since capacity in any organization is so heavily influenced by norms, organizational culture, leadership, and other factors that don’t easily fit into a model based on economic incentives.

It should be public administration departments that deal with these sorts of issues, but they have been falling down on the job. With all due respect, when is the last time you heard about an important, policy-relevant theory coming out of a public administration school? When was the last time that this field has had anything useful to say from a policy perspective about controlling corruption, either at home or in a developing country?

I of course am a big fan of democracy and rule of law and applaud all of those studying these institutions of constraint. But successful government is the product of a balance between a strong state that can deliver services and enforce laws, and checking institutions that insure that the state acts predictably and in accord with general interests. Many of the problems we see with democratic consolidation around the world has to do with absent state capacity on the part of countries that have recently made a transition to democracy. The moment the crowds bringing down the dictator have dispersed, the next question is always, Where am I going to get a job? or Why isn’t my child getting a decent education or health care? Unless democracies have the ability to deliver on these kinds of issues, unless they can deal with pervasive corruption, they will lose legitimacy fast. This is something that states accomplish, and not the institutions that check them.

There’s some evidence of change in assessing the state, such as the interesting work that’s being done by Bo Rothstein in Gothenberg with his Quality of Governance Institute. But we need a lot more work in this area.

A lot has happened since I last reported on my surveillance drone. My fleet has grown to three drones: in addition to the DJI quadcopter, I have a Bixler Sky Surfer equipped with a GoPro camera that can send a live video feed back to a base station. This is what’s called FPV, or “first person view” that allows you to see the world as if you were in the drone. The video feed can be viewed either on a laptop or through a pair of video goggles, which are necessary if you really want to pilot the plane via FPV. In addition, there is an Eagle Tree telemetry system in the loop that sends real-time telemetry data back that includes GPS data on the drone’s location, bearing,  battery power, and the like.

The following is a video taken by the Sky Surfer flying over the oval at Stanford:

This video was taken at Bayfront Park at the edge of San Francisco Bay:

The next video is a flight over Lagunita at Stanford; at around second 43 of the video the telemetry kicks in and you can see the data being transmitted back to the ground station. So I’ve now fulfilled my goal of having a drone like the Army’s Raven that you can throw in the air and use to see the back side of a building in front of you.

The second addition to the fleet is a 3D Robotics six-rotor Arducopter, pictured below. There’s a GoPro camera mounted underneath it which is on a gyroscopically-stabilized two-axis gimbal. This one does not send back a live video feed (though that would be very easy to rig up); it does however transmit telemetry data back to a laptop that will display the drone’s location on Google Earth via the Mission Planner software.

After my first drone post I was introduced to Chris Anderson, editor-in-chief of Wired Magazine and founder of DIY Drones, the organization that developed the Arducopter, Arduplane, Ardurover, and other Arduino-based autonomous drones and cars. Arduino is an open-source board that’s very popular in the robotics community that serves as a digital interface between various real-world processes. The Arducopters have a controller that includes a full autopilot that allows you to plan missions for your drone by simply clicking on a series of waypoints on Google Earth. If you’ve got it set up correctly, you can send the drone off to any set of GPS coordinates you specify and then have it fly back to you, without any pilot input.

In building the hexacopter, I initially couldn’t get the controller board to work properly. I emailed Chris, who told me to bring the board to the 3D Robotics booth at the Maker Faire in San Mateo. The latter is a huge event (more than 100,000 attendees) dedicated to hi-tech do-it-yourselfers who want to build their own drones, robots, computers, make objects with 3D printers, etc. One of Chris’ engineers discovered that I had forgotten to solder a couple of spots on the board, and got the thing working. While we were talking different programmers would drop by and start chatting about their own modifications to the Arducopter software, what features it might have in the future, and the like. A perfect illustration of how social capital operates in Silicon Valley, and why it is such an innovative place.

I’ve now gotten diverted from the drones into Arduino itself. You can hook up a wide variety of “shields” to the basic Arduino board that can be a platform for sensors of various sorts (barometric pressure, accelerometers and gyros, temperature, distance, light, gas, smoke, radiation, etc.), keyboards, joysticks, RC receivers, etc., which can then be used to manipulate any number of real-world processes (motors, servos, displays, and so forth). The programming language is pretty straightforward if you already know C or Java, and it is a lot of fun to actually get these boards to control things like rovers or report on conditions around the house. There are a number of (mostly free) development tools like Fritzing that allow you to prototype a circuit, turn it into a digital printed circuit board design, and have the board custom manufactured to your specifications.  I’ve installed an Arduino in an RC car chassis and programmed it to avoid walls and send back telemetry.

It’s pretty clear to me that we have not begun to think through the political consequences of living in a world where sophisticated drones can be operated by lots of other countries besides the United States, and by private individuals within the US. After two large and unpopular land wars, we are shifting to a military posture that is heavily reliant on drones and special forces to carry out what amount to targeted killings of our political opponents. This is a very seductive path to follow because it is relatively cheap, lacking the huge logistics trains that accompany conventional force deployments, and seems for the time being to be a monopoly of the United States. The residual forces we are going to leave in Afghanistan will largely be of this nature.

We’ve already eroded to some extent the prohibitions coming out of the Church Committee hearings in the 1970s on assassinations of our political opponents with the targeted killings in Pakistan and Yemen.  The lawyers will explain why these are not legally assassinations, but I suspect that the American public is getting used to the idea that we can simply eliminate individuals we deem to be threats with drones.  The problem will come home to roost when other countries develop the same capabilities and start using them against us.

Of the technologies available to hobbyists now, the most troubling are the autopilots.  You can defend yourself against remotely-controlled drone by detecting and disrupting the radio signal from whoever’s controlling it on the ground.  But the autopilots emit no RF signature and depend only on their ability to pick up a GPS signal to get to their destinations.  I leave it to your imagination to think of some of the bad uses to which this can be put.

For more than 2000 years, the Chinese political system has been built around a highly sophisticated centralized bureaucracy, which has run what has always been a vast society through top-down methods.  What China never developed was a rule of law, that is, an independent legal institution that would limit the discretion of the government, or democratic accountability.  What the Chinese substituted for formal checks on power was a bureaucracy bound by rules and customs which made its behavior reasonably predictable, and a Confucian moral system that educated leaders to look to public interests rather than their own aggrandizement.  This system is, in essence, the same one that is operating today, with the Chinese Communist Party taking the role of Emperor.

A high-quality centralized government with few checks on its power can do wonders when the leadership is good:  it can take large decisions quickly because it doesn’t have to form coalitions or wait for consensus; it is not subject to second guessing or legal challenges; and it can ignore populist pressures to undertake questionable policies.

The issue that Chinese governments have never been able to solve is what was historically known as the “Bad Emperor” problem:  while unchecked power in the hands of a benevolent and wise ruler has many advantages, how do you guarantee a continuing supply of good Emperors?  The Confucian educational system and Mandarinate was supposed to indoctrinate leaders, but every now and then terrible ones would emerge and plunge the country into chaos, like the Evil Empress Wu who killed off much of the Tang Dynasty’s aristocracy, or the Ming Dynasty’s Wanli Emperor who in a fit of pique refused to come out of his palace or sign documents for nearly a decade.

In the view of many Chinese, the last Bad Emperor to rule China was Mao Zedong, who in the Great Leap Forward and Cultural Revolution unleashed unspeakable suffering on the Chinese people, and whose power could not be checked until his death in 1976.  The current rules governing decision-making and leadership at the very top of the party reflect this experience:  responsibility is shared among the nine members of the Standing Committee of the Politburo; there are ten year term limits on the tenure of the president and prime minister; no one over the age of 67 can be considered for membership on the Standing Committee.  These rules were designed explicitly to prevent the rise of another Mao, who would use his personal authority to singlehandedly dominate the party and the country.

China’s authoritarian system is distinct because it follows rules regarding term limits and succession.  Tunisia’s Ben Ali, Egypt’s Mubarak, or Libya’s Qaddhafi, not to speak of authoritarian African leaders like Robert Mugabe or Meles Zenawi, would be much more fondly remembered by their people had they stepped down after their first ten years in office and arranged for an orderly transfer of power.

This is why, then, the recently purged Bo Xilai was such a threat to the system:  using his base in Chongqing, he used the media effectively to build his own charismatic authority, which was strong already given his status as a Princeling or son of a revolutionary hero; he was ruthless in the use of state power to go after not just criminals and corrupt officials but businessmen and rivals who had accumulated too much power and wealth; and he revived Mao-era mobilization techniques like the singing of revolutionary songs at mass rallies.  Unlike his gray compatriots, he could potentially dominate the leadership with an independent power base if he were promoted to the Standing Committee.  It therefore makes sense that Hu Jintao and the existing leadership should use the scandal of a coverup and murder to eliminate him from consideration and remove the Bad Emperor threat.

The commentary to date has noted how the Bo Xilai affair has demonstrated serious cleavages in the senior leadership of the party, corruption and turpitude among its members, and weakening control over what the Chinese public can say on vehicles like Sina Weibo, the Chinese Twitter.  All of this is true, but the incident reveals an even deeper problem, which is the lack of formal institutions and a real rule of law.

The rules that the Chinese leadership follows are neither embedded in their constitution, clearly articulated, or enforced by a judicial system.  They are simply internal rules of the Party, which actually have to be inferred from the Party’s behavior.  Had Bo Xilai succeeded in getting onto the Standing Committee and increasing his personal authority, he could easily have overturned any one of them.  Latin American presidents who want to linger in office still have to go through a process of constitutional revision, and every now and then the rule of law is strong enough to prevent them from doing so (as was the case recently when Colombia’s Alvaro Uribe was denied a third term in office by the country’s Constitutional Court).

So the apparent institutionalization of the Chinese authoritarian system is largely a mirage.  The Communist Party has not solved the Bad Emperor problem, nor will it until it develops something like a genuine rule of law with all of the transparency and formal institutionalization that entails.

I had a meeting a couple of years ago in Beijing with a mid-level official heading a Central Committee office, who told me over a long lunch that I could not possibly understand contemporary China without appreciating what a total disaster the Cultural Revolution was, and how the current system was organized to prevent that from happening again.  Looking around at the books and memorials to Mao Zedong that the Party was still promoting, I asked him how that could come about unless the Party was more forthright in telling the truth about Mao’s legacy.  His generation had personal experience of those terrible events, but people growing up since then did not, and could be seduced into viewing it with nostalgia.  It was precisely that lack of historical remembrance that Bo Xilai was exploiting.  The official, by the way, didn’t have an answer to my question.

So in the end, informal rules observed by a small clique of insiders can’t really substitute for a formal rule of law.  As we can see today, modern liberal democracies constrained by law and elections oftentimes produce mediocre or weak leaders.  Sometimes democracies elect monsters, like Adolf Hitler.  But at least the formal procedures constraining power through law and elections put big roadblocks in the path of a really Bad Emperor.  Despite having beaten back Bo’s challenge in the short run, the Chinese system has not solved this institutional problem yet.  It now has a real opportunity to do so, which we can hope the new leadership coming into power will take up.

[A version of this piece was published in the Financial Times on May 10, 2012]