What I learned from my first focus group discussion

As I mentioned in my previous blog post, my first focus group discussion among SACCO members is conducted with boda boda drivers. As a novice researcher and a mzungu, I didn’t conducting the discussion perfectly. But I did learn a lot and I am confident that I will do better next time.

Only one boda boda driver showed up at 2PM, our scheduled time of discussion. My translator and I welcomed him into our discussion room, introduced ourselves and told him the purpose of the discussion, and started asking him questions about basic information (number of people in the household, number of people employed. But we soon entered a phase of awkwardness when we didn’t know if we should proceed to the other questions or wait until the others are here. The second driver came in and melted this awkwardness. Since I was not sure whether the others will be coming, I started a small discussion with the two of them. Somewhat unexpectedly, four more drivers showed up. I was happy that I reached the minimum number of participants for a focus group discussion, but I didn’t provide the newcomers with enough context of our discussion.

The first three questions went fine. The respondents were comfortable to discuss with each other. However, one driver said he was too busy and wanted to leave early. Not wanting to disappoint us, he proposed that we ask him the questions first and ask other people after he leaves. Without too much consideration, I agreed. So we spent some fifteen minutes just talking with him. In retrospect, this is a bad strategy because 1) talking with only one participant will not generate a diversity of ideas (which is the aim of the focus group discussion) and is no different from a one-on-one interview 2) the explicit time constraint places the facilitator (in this case, me) under great pressure and lowers the quality of answers, and 3) other participants will feel being left out and may be less involved when they are the ones to speak. 1) and 2) happened this afternoon, and indeed I had to cut some questions short for that one driver. I didn’t finish all the questions for him because I could sense the impatience of other members.

Several points about the skill of asking the questions:
– Not all questions should be asked separately to each respondent. Many can be asked to one respondent first, and then ask the others to share their thoughts.
– Avoid multiple answers at the same time. If more than one is speaking, the translator will have a hard time conveying all the information to the facilitator and the facilitator will be less likely to fully observe everyone’s reaction. This is hard in practice though.
– If you have a translator, ask the translator to translate every sentence that each participant says. This generates a spontaneity between the facilitator and respondents. It also allows the facilitator to probe the answers. My translator was a bit too involved in the discussion herself (as she is a bank officer and knew all the respondents), and often waited to gather all the information before translating to me.
– Have a guideline, but make changes when necessary. While a facilitator should always make sure the discussion is heading for the right direction, she does not need to stick to every single question written in her guidelines. As long as the flow of the discussion is well controlled, the facilitator should focus on probing the most important questions instead of trying to touch upon every detail.

About logistical issues:
– Expect some participants to be early and some to be late. Socialize with the early arrivers and try to keep them in a good mood. I have to admit I did a poor job today: the guy who arrived the earliest was so tired by the time we officially start the discussion that he nearly slept through the rest of it.
– Remember to turn on your recording before the real discussion starts, and ask for permission from the respondents before doing that. I was too nervous today and completely forgot to record (an unforgettable mistake).
– Have a good control of time. If a respondent is repeating what another person has already said or is going off course, politely acknowledge his sharing and steer the discussion back to the right direction.
– Sit in a circle and let everyone see each other. This fosters eye contact and creates a sense of closeness between the facilitator and the respondents.


Thoughts about Seeing Like A State (1)

Seeing Like A State by James Scott is in my reading list for the vacation. So far I have found the book insightful and provocative. As it says on its cover, the book is about “how certain schemes to improve the human condition have failed”. I’ve been jotting down my thoughts as I read along. The following are my thoughts on the first two chapters.

The author started off by comparing the state to a forest. To get an idea about the composition of a forest, one has to classify the trees into several categories with certain standards, and record the number of trees in each category. This procedure sacrifices precision for the general idea (the “big picture”). Likewise, if a states needs to manage its people for fiscal or political purposes, it has to assign unified measures to people to categorize them.

The tension between local powers and the central government is present in all states. The central government is the political focal point and steers the policy direction of the whole country. But it lacks information about local situations, which may discount the effects of policy implementation. This is especially evident in ancient and contemporary China given its size and heterogeneity among regions. Qin Shihuang (the first emperor in China) unified units of measurement and currency. This accommodated for market activities between regions and fostered political unity. Without comparable units of measurements, market activities are limited to a small, local scale. Units of measurements also comes into play in tax collection.

These arguments remind me of what I learned in my development seminar. One feature of the development microeconomics is nonanonymity. Everyone in the village knows each other well. Trust can be established easily through repetitive interaction. When market transactions are limited to the local community, it seems safe, but it also greatly restricts the resources available. As the market grows, people will need to interact with the outside to consume a wider array of goods, to gain more information, and to make more diversified investment portfolios. These requires the adoption of unified measurements and currency, accepted rules about market order, and legal enforcement institutions.

The author argues that the administrative simplification of the state progresses from control over nature (e.g. forest) to control over space (e.g. land tenure), and then to inherited, permanent patronyms. In China, many patronyms were given in the Qin dynasty for better identification and easier tax collection.

The state can design cities to be more legible to outsiders and therefore increase its control over local areas. An example is the redesign of Paris. The military security of the state was the central concern of Louis Napoleon’s and Haussmann’s plans. The boulevards were designed to be wide and directly accessible from the center, which allows the army to march quickly in case of insurrection.

Democracy and its consolidation — thoughts from The Economic Origins of Dictatorship and Democracy

My thoughts are based on the book The Economic Origins of Dictatorship and Democracy by Daron Acemoglu and James Robinson. The book discusses how the economic conditions of countries affect their paths of political institutions. Here the word “democracy” is defined by Schumpeter to be

The institutional arrangements for arriving at political decisions in which individuals acquire power to decide by means of a competitive struggle for the people’s vote.

There are four paths of political development: nondemocracy to democracy (Britain), unconsolidated democracy (Argentina), nondemocracy without repression (Singapore), and nondemocracy with repression (South Africa before the collapse of the apartheid regime).

Social choices are conflicting because there are two groups: the elite and the masses. Here the elite might be the (relative) rich and the citizens be the (relative) poor. In a nondemocracy, the political institutions represent only the interest of the elites who have de jure political power. But the citizens have de facto political power in a sense that they can launch a revolution to topple the existing political institutions. This is referred to as a revolution threat. Faced with a revolution threat, the elite can repress or make a concession. When costs of repression are low (South Africa in apartheid regime), they choose this. Otherwise they make a concession. For the latter choice, they have to transfer the de jure political power to the citizens. Because de facto political power is transitory, citizens require long-lasting political institutions to ensure that majority preferences will be the yardstick of future policies.

For democracy to make it, the society should be moderately unequal, because there’s no reason to adopt democracy in an egalitarian society and too much inequality will greatly increase the burden of the elites once resources are redistributed. I find the inequality argument convincing. If the country is too unequal, on the one hand the citizens want very much to revolt, on the other hand the elites take all means to repress. The result is bigger loss of physical and human capital, and an unstable regime or even one switching between democracy and nondemocracy

The composition of wealth also matters in the process of a country’s potential move to democracy. If the major source of income is land, it can be easily taxed, so land owners are more afraid of democracy for its redistributive nature. By contrast, if the main source of income becomes human or physical capital, it is harder to tax them and the elites who own them fear democracy less.

The level of democracy cannot be easily measured, mainly because not all the countries have the institutional arrangements of democracy. There are measures like the Freedom of Houses Index and Composite Polity Index and so on. Countries with higher income and education level are generally more likely to be democratic, but correlation is not causation and once we take out many country specific factors, the relationship is not so significant. 


Daron Acemoglu and James Robinson. The Economic Origins of Dictatorship and Democracy. Cambridge University Press, 2005, chap 1-3.

Can good rules be replicated? — thoughts from the proposal of charter cities

My thoughts are from Paul Romer’s (bio) radical idea about charter cities (TED link). He talked about building cities with good rules (institutions) which drives people to move there and businesses to grow, finally accelerating urbanization process in underdeveloped areas around the world. According to Prof. Romer, we need rules which offer more choices to people and more choices to leaders in order to change bad rules. The government can be too strong (North Korea) or too weak (Haiti) to ensure the provision of public goods (in Romer’s case, electricity).

The idea seems to be borrowed from China’s Shenzhen. The Chinese government was able to carry out the Special Economic Zone (SEZ) policy and replicate successful models because 1) China was very big and Shenzhen was relatively poor, so failure is bearable, and 2) the economic and political structure of Chinese cities are highly similar to promote good practices around the country.

Prof. Romer mentioned three necessary conditions for setting up a charter city:

1) the charter, i.e. a set of rules that are beneficial to the development of economy and welfare of people;

2) a piece of uninhabited land to which the people can opt to go;

3) collaboration between countries (which offers a choice for the leader).

The first question that came to me was: “How is this different from colonialism?” Prof. Romer suggested in his speech that this idea differed from colonialism in that the latter had “an element of coercion and condescension”. But even if the theory works fine, there might be ideological barriers in some developing countries, as was the case when China first created its SEZs.

The second question is:”Once a charter city succeeds, is the experience applicable to other underdeveloped regions?” China’s SEZs could be replicated because the Chinese government at local levels are similar to each other (relics from the three departments six ministries system 三省六部制度 in ancient time and decentralization of power in modern time). China itself is like a huge laboratory. But few other countries share these characteristics with China. The basic administrative organs are pre-existing before any kind of charter is built and they can be destructive to the functioning of a charter city. How to coordinate a balance between the new rules and the existing ones will be a great challenge for charter city program.

More info on charter cities is available here.

Reflections on Collective Action by Russell Hardin

In my economics course State, Law, and the Economy, the professor has assigned us a series of readings to better understand the course materials. I will write reflection articles on them and enhance my own thinking.

In the first three chapters of the book Collective Action by Hardin, he synthesized existing literature on two questions: First, what are the preconditions of forming an interest group? Second, once an interest group is formed, how to ensure that members do contribute to promote the wellbeing of the organization?

Regarding the first question, the answer is short and long. The short answer is: we form groups to achieve something that cannot be done by individuals out of their own interests. In a classical Prisoner’s dilemma situation, the pursuit of self-interest leads to an inferior outcome for both players. The same logic applies for a larger group of players. An organization is set up to promote a certain extent of cooperation for players to improve their payoffs. Also, there might be a subgroup (privileged group) which can still benefit if they were to provide the full costs to accomplish group benefits. A good example is big companies lobbying for tariffs out of their own interests while creating benefits for smaller companies in the same industry.

Another example is China’s household responsibility system reform in the late 1970s. The farmers who initiated the reform (locally) had to endure high political risks, but they were also eager to improve their own wellbeing and abandon the inefficient commune system. The close-knit nature of the local community means that if anyone betrayed, the rest of the village would never forgive him and he would for sure live a miserable life afterwards. Therefore the bystanders didn’t create any troubles.

I have found Hardin’s explanation on the second question enlightening. Apart from the fact that organizations can help improve individuals’ welfare, other factors also contribute to the organization once it is created. Mancur Olson put forward the by-product theory and suggests that sometimes organizations succeed in realizing certain goals simply by stimulating group-oriented behaviors. For instance, women’s liberation movements benefited a lot from the women’s leagues which were originally created for social life. The existence of an organization provides the convenience to coordinate and to gather power to fight for members’ rights.

Another intriguing question is about whether group size affects the formation of an organization (or, whether a group is latent or not). There’s dispute on this issue. Some suggest that as the number of individual increases, an individual is more reluctant to pay for the common good. But if there is increasing returns to scale, a coalition can still be formed and led by some “big” players (those who will gain significantly from collective action).