Interview with Professor Tana Johnson
Image Source: Sanford School of Public Policy, Duke University
Several of our posts this year have covered talks by visiting academic and practitioners, but Sanford has impressive thinkers within our own community too! I recently sat down with Professor Tana Johnson to discuss her new book Organizational Progeny: Why Governments are Losing Control over the Proliferating Structures of Global Governance. Professor Johnson’s motivation for Organizational Progeny arose from an interest in the concept of delegation, which is at the center of not only politics but also what we do in life. Delegation’s role in politics is obvious – for example, individual citizens cannot possibly be informed and involved in every policy decision. In the U.S. for example, citizens send representatives to DC to make choices on their behalf. In a similar way, the U.S. government often delegates policy decisions to the UN or other international intergovernmental organizations (IGOs). We know from our own relationships with family members and co-workers that sometimes the delegation process breaks down and dividing responsibility does not work effectively. And yet we often seem to be surprised when the same problems arise with government officials and international organizations.
When Professor Johnson started researching delegation in graduate school, she thought that a principal-agent model might hold the key for solving delegation problems. That model suggests that principals (e.g., countries) would create and delegate to agents (e.g., IGOs) that have been incentivized to do what the principal wants. Unfortunately, the world is really complicated (as many of us have learned as policy students), and models don’t always capture the whole picture we need for effective policymaking. Exactly how do we design institutions so that incentives are aligned – and what happens if we don’t?
As she delved farther, Professor Johnson realized that principal-agent models aren’t very good at capturing something that’s been going on in international politics for decades. Right after World War II, governments created and delegated to the United Nations and several other big IGOs. But since then, these IGOs have been creating and delegating to offshoots, or “progeny,” of their own. For example, the Intergovernmental Panel on Climate Change (IPCC) is an offshoot of the United Nations Environment Program (UNEP), which is an offshoot of the UN. The international bureaucracy has become a complex family tree, with interactions among multiple generations. The newer generations, which aren’t created by governments alone, also are less likely to be designed with features (such as vetoes or control over funding) that governments traditionally have used to manage delegation relationships.
In other words, governments have a hard time controlling the international bureaucracy, for at least two reasons. First, the number of IGOs keeps increasing, so governments have to divide their time and oversight over many more organizations. Second, organizations that are progeny of other IGOs have been designed differently, without some of the most conventional mechanisms that governments like to use to exert control. With case studies and an original dataset, Organizational Progeny gives a sense of the prevalence of IGO offshoots and the differences in how those offshoots look and operate. Professor Johnson’s research indicates that progeny are about 65% of the total IGO population, and progeny can be found in nearly every policy area.
I asked Professor Johnson whether there was an optimal level of autonomy for IGOs. She explained that having some autonomy is useful for enabling agencies to make their own decisions, provide honest recommendations, and develop expertise. However, having a lot of autonomy can strain the relationship between countries and IGOs, sometimes even reducing the organization’s effectiveness. That’s one of the reasons why governments often want to have “ex post” controls, such as approval over organizational budgets. That way, an IGO could have some autonomy to try various things, but if governments don’t like the results they can punish and rein in the organization, for instance by withholding funding.
Professor Johnson also said the “best” level of autonomy can also vary by policy area. Very scientific areas often require deeper levels of expertise and can benefit from a more hands-off approach from governments. The IPCC, for example, is an organization that defies what some of the most powerful states would want. Its autonomy is important because it works closely with the scientific community, which does not want to be or to appear co-opted by government interests. Because the IPCC has a lot of expertise but little authority to create or implement policies, one of its most important roles is sharing information with citizens. In this way, the organization can shape policy by building consensus around the science of climate change.
At the end of our discussion, Professor Johnson offered some research tips for students. First, she encouraged us to consider using both quantitative analyses and case studies. In Organizational Progeny, these two methods work together to illustrate the full picture. Quantitative analyses were useful for demonstrating that a pattern existed across many organizations, and case studies were useful for showing the history of specific important organizations. Second, even though she put together an original dataset for the book, she warned that this is a difficult and time-consuming process. Before following a similar path, grad students should do a lot of legwork to see whether any existing datasets would suffice. It’s good to know that sometimes if you email individual researchers and ask nicely, they might let you use data that they haven’t made publicly available yet. Third, Professor Johnson advised that if you do end up collecting your own data, first focus on a sample instead of immediately trying to tackle the overall population. For example, if you use random sampling, then you can feel reassured that the sample is representative of the overall population, even if time or resource constraints prevent you from coding every piece of the overall population. And if those constraints never become a problem, random sampling is still a good idea: you just keep expanding the sample until it eventually captures the entire population.
To learn more about the ideas in Organizational Progeny or read sample chapters, visit the book’s website at http://www.organizationalprogeny.com/.