Measuring “Gross National Happiness”: My initial reflections

A few days ago, in honor of World Metrics Day, I spoke on a panel to the Investment team at the Acumen Fund.

I focused my short discussion on GNH and the measurement of happiness. Coming into this internship I was very skeptical about the practical and technical feasibility of measuring something like “Happiness”. Unlike the many tourists and researchers who flock to Bhutan mesmerized by its philosophy and vision, I’ll admit to being a bit of a hard-nosed cynic. I choose Bhutan not for its idealized image, but for the other reasons that make Bhutan and the GNHC a fascinating place with exciting opportunities.

I focused my comments to Acumen on just the facts of how Bhutan measures GNH and how it integrates it into policymaking. Thinking through what I was going to say, though, I had a few opinions of my own that I wanted to get off my chest. You guys will have to put up with some of these initial reactions from my first few weeks on the job (I choose not to share these with Acumen).

So, in my humble, naïve opinion: How well do I think Bhutan is doing at measuring National Happiness? (click below to read more)

First the pros:

• The GNH Index is multidimensional and rigorously constructed
  The challenge for policymakers is boiling a number of different objectives (deemed of intrinsic importance) down to a single indicator of the “progress of the nation”. In part, this is why GDP has grown with such prominence as the key indicator of the development of nations—it is relatively easy to measure, consistent across countries and extremely salient.
  
  The Capability Approach argues that income is generally of only instrumental purpose in achieving substantive human development. Money is useful only in as much as it gives us the ability to pursue a life we consider valuable—a long, healthy life, pursuit of an education, spending time with our family and friends, travelling, taking part in political decisions, etc. The instrumental role of money I think this is something that a lot of people identify with. The Capability Approach provides the technical and philosophical grounding for it.
  
  The GNH Index is an operationalization of the Capability Approach for Bhtuan. It boils nine dimensions of well-being (like Education, Living Standards and Health) down to a single indicator. Based on cutting-edge work, the GNH Index uses the same methodology as the recent MPI index from Oxford’s OPHI. It is a bit of a technical process but results in a single indicator that satisfies a number of convenient and useful properties. While multidimensional indicators have existed for a while, GNH is the first case (I know of) of a country developing one for use in its own policy screening and welfare measurement.
  
  
• The GNH index represents[1] the unique values that define Bhutan and therefore help frame and direct national debate 
   Perhaps since it is a small country, Bhutan has had the policy flexibility to pursue welfare measurement on its own terms and rooted on its unique value-system. This is fantastic. Beyond some of the more traditional dimensions mentioned above, the GNH index also captures things like environment, cultural preservation and community vitality that Bhutanese consider important for their own development. These are normative, values-laden decisions that each country must make for themselves.
  
  Indeed, despite occasional pronouncements to the contrary, the history of development has generally been characterized by international agencies coming to tell developing countries what to do and how to do it (e.g. improve national income by liberalizing markets). Bhutan has made explicit its own values and, in doing so, has been able to resist policies that may erode (for example) the strong emphasis they place on cultural and environmental preservation. Recently, for example, Bhutan chose to postpone WTO accession. Postponing WTO accession may—in theory at least—have come at a cost to GDP, but Bhutan was able to make a specific policy trade-off by having an explicit, well-articulated set of government objectives.

Now my concerns:

• It does not follow good practice in sampling design.
  The sampling methodology is not articulated, it is not nationally-representative (for such an endeavor of supposed national importance, some districts weren’t even included in the survey), the analysis is not reproducible, and it has a small sample size which makes it extremely sensitive to the weights chosen. It also doesn’t yet capture the fact that the indicator is measured with error (that is to say, it doesn’t have the confidence intervals you see on most surveys). Since it’s such a small sample, the district-level decomposition is questionable at best.
  
  
• It seems to have difficulty at actually measuring happiness. 
  I’ll go into more detail on this in another post, but my short answer is that I’m still mostly skeptical of the technical and practical feasibility of measuring happiness and that I think this has important implications for policy. The GNH Index does not use current good-practice for measuring happiness (which I’ll admit is evolving at a rapid pace).
  
  
• It’s a long survey.
  I generally refuse to sit through an online survey longer than 15 minutes. The first draft of the GNH survey took 7-8 hours per respondent. It was then pared down to a mere half-day-long survey. In my (albeit limited) experience, after an hour of questions, the quality of answers from respondents drops dramatically. You’ll also have selective attrition in the types of people who choose to answer a 5 hour survey versus those who give up half-way through.
  
  
• It’s way to large—a cacophony of 72 indicators across 9 dimensions.
  The sheer scale of the index is a bit boggling. In its complexity it loses clarity and coherence. In my opinion, a multi-dimensional index does not have to capture everything that a society values, rather just a few key attributes with signaling value (that is to say, they correlate well with other variables of interest deemed valuable by society).
  
  
• It asks questions which may not elicit the underlying responses.
  Questions which seek to elicit truthful responses on what are deemed “socially-desirable behavior” (or alternatively socially undesirable behavior) are particularly difficult since there is a natural inclination for the respondent to want to please the surveyor (or hide their ignorance/misdeeds).  GNH’s indicators like “Do you attend festivals?” “Do you know how AIDS is transmitted?”, and “Do you donate to charity?”, all are likely biased estimates of the true underlying indicator since there is a natural inclination to simply answer “yes” to each.

In short, what I like about the GNH index is the way it articulates a national vision. You hear people in the streets talking about GNH and it provides a common mission that runs through everything the Ministries do.

I’m afraid that where GNH currently fails is in the technical aspects of operationalizing this vision. These shortcomings are not insurmountable, though, and a lot of interesting work is going on in this area. Indeed, the Center for Bhutan Studies is supposedly already in the process of conducting a second round of the GNH index.

---

[1] Or at least purports to represent—that’s another issue