Improving citizen happiness is an important goal for many, if not all, governments. But what is happiness really? Can it be objectively measured? Can we discover the key factors that best correlate with happiness? And ultimately, can governments implement policies and programs that maximize happiness?
Is maximum happiness nothing more than a non-linear conjugant gradient optimization?
In the late summer last year, I had the pleasure of spending about a week in the United Arab Emirates, participating as a speaker in the National Security Middle East 2016 event in Abu Dhabi. It was the second time I travelled to the UAE last year, and I found the Emiratis to be warm, friendly and welcoming without exception.
It turns out that the Emiratis’ warmth is something they're now attempting to measure, and in some sense optimize across society. During one of the breaks at the event, I was speaking with my SAS Middle East colleagues, and they shared with me that the UAE government had recently created a Cabinet-level agency, the Ministry of Happiness, led by Her Excellency Ohoud bint Khalfan Al Roumi.
On International Happiness Day (March 20th – come on, you know you have it marked on your calendar…), Minister Al Roumi provided her definition of happiness in her article, A Quantum of Happiness:
Happiness is knowing that you and your family are safe; that there is opportunity open to you and your children; and that you can depend on a high degree of care, dignity, and fairness in your society.
She goes on to say that happiness is “…a flourishing and ambient joy” – a definition I will return to at the conclusion of this blog.
After my colleagues finished telling me about this new agency, I wasn't quite sure what to think. To American ears, the idea of a US Federal Department of Happiness with a big Corinthian-column ordained building in downtown Washington DC, sounds like some kind of dystopian alternate reality (despite the fact that the "pursuit of happiness" follows right behind those unalienable rights of life and liberty in our Declaration of Independence).
Part of the UAE happiness plan includes changing the names of government services centres to “customer happiness centers.” When April 14th rolls around each year here in the US, I am not quite sure what most Americans would think about sending off their 1040 tax returns to the “internal revenue happiness center.”
But this discussion did get me thinking. Having spent twelve years in the US Government, building analytics into government decision making processes, and as an analytics and data science guy at heart, how would one go about measuring something like this in society? An even more challenging question is how we would go about understanding the criteria that most correlate and predict the societal outcome we are driving towards (in this case, happiness). Then, how would we seek to optimize those criteria?
Minister El Roumi has recently outlined more of her agency’s strategy for achieving their goal, releasing the customer happiness formula, “unveiled by the Ministry of Happiness to boost happiness in the UAE through government services.” Just last month the agency released the first-of-its kind national survey aimed at collecting happiness data from 14,000 participants around the country. Specifically, the survey aims to collect feedback around happiness, positivity, quality of life, education, learning, healthcare, society, culture, environment, infrastructure, government services, living standards and work.
This data will present a large opportunity and many challenges at the same time. The ministerial agencies mandated with evaluating happiness measures will likely need sophisticated approaches to determine which objective measures most correlate to the subjective “happiness” responses.
There is a large body of global data that compares happiness around the world. Of course, simply knowing which objective measures correlate to happiness isn’t enough to improve anything. The next steps, from an analytics perspective, are to:
- Identify what policies and programs (both existing, and non-existent but possible) impact the objective measures.
- Quantitatively assess the degree to which those policies or programs will impact the objective measures alone or in combination.
- Optimize the portfolio mixture of policies and procedures to maximize the desired subjective outcome (here, happiness).
Even when this effort succeeds, I wonder about the long term ability of any government, no matter how sophisticated, to sustain programs that maximize happiness. Many of these objective measures are not in our control. What if there is a global recession? A disease pandemic? A war? Or any number of other possible objective measures that are known to negatively affect happiness? More broadly, what keeps cultures and societies from slipping into despair when things outside of human control occur?
I would offer that resilience – the ability to maintain one’s position and attitude even in the face of things around us that might make us unhappy – is just as important to society as happiness.
Many cultures around the world have been assessed for their outlook and well-being. A number of them maintain quite a positive outlook, despite circumstances like disease, high infant mortality rates, ravages of war, etc.
Maybe there's a difference between happiness and joy, that has to do with resilience in the face of our circumstances. While joy has a number of definitions, one that I like is right in line with this idea: joy is “a state of mind and an orientation of the heart… a settled state of contentment, confidence and hope.”
As we kick off 2017 around the world, thinking about all of the ways that analytics can help improve our world and help its people, may the New Year be both happy … and joyful!
14 Comments
Unless your optimization builds in constraints to exclude policies which benefit some people's happiness at the expense of others, such a program will become a vehicle to implement communism, and as we have seen historically, this results in nobody being truly happy.
I love this comment Aaron; you're exactly right. "Optimization" type approaches without the sorts of constraints you're referring to can end with least-common-denominator policies or "tyranny of the majority" sorts of policies. Western liberalism (and here I mean liberalism as in the form of liberal democracy we have in the U.S. and the west, not as in the left kind of ideology liberalism) is built on John Rawls' "veil of ignorance" thought experiment in which we consider the needs of minority groups, not just the rule of the majority. You're right that optimization of government policy is a tricky endeavor involving more than just analytics.
Thanks, Steve! I looked at your linked "veil of ignorance" concept, and it seems to me that it does not go far enough in protecting individual rights, and gives governments too much redistributative room in their policies. I would be fine with governments using analytics to choose the "optimal" policy, as long as their optimization includes constraints preventing the taking from any individual to benefit any other individual or group.
Yes - I think you're fundamentally right that the harder part of applying analytics to policy development is not so much the technical aspects of the optimization, but rather the choice of objective functions to maximize and against what constraints. It will be interesting to see how the UAE tackles these kinds of challenges as their data collection continues...
Aaron: The first thing you have to get past is the idea that 'redistribution' in any form is wrong. All tax policy is redistributive, in that all tax policy takes money from the people and uses it in ways that is never going to be equal to the amount taken. The person who pays $1,000 in income tax is not going to receive $1,000 in benefit. Thus, the idea of government action being redistributive is a given - money will be taken from those who have it, and some of it will likely be used to benefit those who do not. However, as with all things, it can be taken too far. The goal, then, would be to enact policies that take the least (since paying taxes is not happiness-creating) while providing the greatest level of happiness-inducing benefit.
Hey Jim, thanks for the thoughtful comment - I am glad that the UAE's foray into this, and the nexus with analytics is drawing such great discussion! I think there are likely a full spectrum of beliefs as to what the right balance is between how much to take from citizens, verses how many benefits to provide. But wherever readers come down on that question personally, I nevertheless found it fascinating that a government would step into this realm to try and make something like happiness a formal policy objective in the first place. This will be an interesting project to watch over the next year to see 1) what they do with the survey data they collect, and 2) what policy implementation takes place as a result.
Start with this film: http://www.imdb.com/title/tt1613092/?ref_=login
I should have added more context. This film is a documentary - a qualitative study on communities throughout the world that have a high level of happiness. As i recall, it also contrasts them with communities that have a very low level. Anyway, it presents Mihaly Csikszentmihalyi's notion of Flow. I suspect from studies like these, you could identify variables that could be used for measurements and analysis.
Hey Ted, thanks for the comment - I think this topic of analytics and data science applied to broader societal objectives has gotten a lot of folks talking - whether or not governments can successfully optimize these kinds of things, we will have to wait and see - I will definitely check out the movie however.
Yeah, I agree. While having a Minister of Happiness seems like a good idea on the surface, this would be an area where governments should probably follow the research and the markets rather than lead.
Quoted from our founding fathers, the pursuit of happiness is one of natural human right. We need a society/platform that provides opportunities to individuals to pursue their happiness.
But it doesn't mean we need a government agency, who thought they knew better than ourselves, to micromanage fabric of our lives and figure out how to maximize everyone’s happiness by mandating government policy. Such ideology sounds great in theory but had killed many people in socialist countries.
Hey Ricky - your point is well taken. I think what fascinated me most about the travel to the UAE and hearing about this project, was the fact that they were trying to address happiness as a policy goal as head-on as they were. Your thoughts are right in line with the challenges they will likely face, and which Aaron has previously touched on as well. Here in the west, liberal democracy follows this Rawlsian idea (see the 'veil of ignorance' reply to Aaron's comment) that we should make policy as if we were in this position of not knowing whether or not we will be a potentially negatively-affected minority. This 'thought experiment' (theoretically) keeps western democracies from falling into the tyranny of the majority, in which some supposed overarching objective function gets maximized, but in which a great number of people can be on the losing end. Once the UAE collects its 14,000 data points, it will be interesting to see how the analytics would proceed. Will the constraints that Aaron discussed in his comment be incorporated to guarantee policy protections for all? How will the results of the survey be operationalized in their government? All that said, for me personally based on my faith tradition, the part of this exercise that was most interesting was whether or not we should focus on happiness alone (which is based on external events and circumstances), or in addition, look to cultivate resilience and joy which can endure regardless of external circumstances.
Here is a TED talk that covers the problems that infect a society that has wide income inequality. And happiness is one of the casualties.
https://www.ted.com/talks/richard_wilkinson
And here is another on happiness:
https://www.ted.com/talks/nic_marks_the_happy_planet_index