Thesis Thursday: Ayesha Ali

On the third Thursday of every month, we speak to a recent graduate about their thesis and their studies. This month’s guest is Dr Ayesha Ali who graduated with a PhD from Lancaster University. If you would like to suggest a candidate for an upcoming Thesis Thursday, get in touch.

Essays on health economics: trans fat policies, commuting, physical activity, and body mass index in the US
Colin Green, Bruce Hollingsworth
Repository link

What drew you to this particular topic and made you want to dedicate your PhD to it?

I’ve always been very fascinated about how people make decisions when it comes to health-related behaviors. My Mom was a doctor, so health has always been a main driver in what our family ate and what sort of values my parents emphasized. Growing up, I think we were always kind of “the weird family” in the neighborhood and in our extended family, because we often put health before cultural norms (i.e. changing traditional recipes to be healthier, avoiding all of those colorful and fun children’s cereals and candies, etc.), so I’ve been very aware and very curious about what drives health-related choices for different people. I’ve also lived in a number of very different communities, where norms about behaviors related to health varied significantly from place to place, and that variation has always been something that I’ve sort of observed and wondered about as well. So, I think these observations are what drew me to economics, as a way of understanding choices. My thesis work was somewhat of an attempt to crack the surface.

I don’t think I got the chance in my thesis to dig as deeply into the issue where health behaviors and culture meet as I had wanted, but I was lucky to get to use some of the types of data and some of the econometric tools that will come in handy as I continue to explore this area.

Your study focussed on US data, but you studied at a UK university. Was this a help or a hindrance?

I think it was definitely a challenge to work with non-UK data while studying at a UK University; for example, I didn’t find certain types of support (in terms of local workshops or having other colleagues using the same data) or familiarity that was often easily available to someone using UK data. However, my supervisors were very supportive of me and there were students in my cohort using other non-UK data so I was not alone in that regard.

One challenge that I had was that I couldn’t just ask someone who uses the data, but instead had to spend a bit of time trying to understand how other researchers in the literature used the data and then try to figure out whether or not their assumptions and ways of using the data were applicable to my work. I spent a lot of time reading data documentation files — both of my datasets (NHANES and ATUS) have really good online resources that I’m now very familiar with! I think that a lot of this being “on my own” with the data helped me to develop a feeling of confidence that I may not have had otherwise. I’m somewhat prone to second-guessing myself, so being able to learn to have confidence in my work was really valuable.

So, although challenging, overall I would say it was definitely more of a help than a hindrance.

Methodologically, what was the most challenging aspect of the research?

In my third chapter I use time use data, a two-part model, and a recursive bivariate probit model to estimate physical activity participation and duration decisions given an individual’s commuting time, with an instrument for commuting. There are some conflicting ideas on how best to use time-use data in the literature (see: Franzis and Stewart, 2010; Gershuny, 2012; Stewart, 2013) and few examples of instruments for commuting (see: Baum-Snow, 2007; Gimenez-Nadal and Molina, 2011).

I received a lot of different feedback on the best estimation approach to use. I wanted to estimate participation and intensity elasticities for individuals who do physical activity on a given day. There are a number of ways to deal with these two questions separately or simultaneously and also to deal with the large number of zeros in the data (i.e. individuals who did not participate in physical activity). At one point, I remember that I thought I had everything figured out with this chapter; and then following a conference presentation of this work, one commenter was dead set that I should be using a switching model. I hadn’t considered that approach and wasn’t familiar with it at all, so I had to go look it up, figure out what it was and whether or not it worked with what I was doing. So, just figuring out the best way to deal with my data and with the questions I was asking in this third chapter were probably the most challenging part of the thesis for me.

If you could have a decision maker implement one policy change supported by your work, what would it be?

If I could ideally have policy makers do one thing that is supported by my work, I think it would actually be a rather general thing, not related to one specific policy, but sort of related to the entire approach of policy-making. I would want policy makers to consider the groups they are targeting with more care. For example, my third chapter looks at time use among obese individuals and healthier-weight individuals and finds that many decisions, such as the decision of where to live and work, are often driven by different factors in these different groups. In general, my work suggests that different groups may be driven by different motivating factors and if we don’t understand what these are, policies might not successfully reach those who could most benefit. That being said, this probably isn’t as easy as it sounds, as there are a lot of political influences on how policy decisions are made.

If you had to do it all again (perish the thought), is there anything you would have done differently?

I think overall, I’m really happy with my experience; I had some good resources and Lancaster was a great environment for me. My PhD cohort was close-knit and faculty in my department were very approachable and supportive. If anyone is interested, I also found McCloskey’s Economical Writing and Thomson’s Guide for the Young Economist to be really helpful at various stages of the PhD.

If I could have changed one thing though, it was how I dealt with insecurity. Even in such a supportive environment, the competitive nature of academia can contribute to feelings of insecurity; I worked hard to recognize my own insecurities and fight through them, but I didn’t always succeed. So, if I could do something differently, I would like to have been less afraid to own up to not knowing something and to just keep asking questions until I understood.

I really enjoyed having the chance to talk with you about my experiences and motivation. And I hope that if anyone can find my experiences useful to them at their stage of the process, they do. Thanks again for inviting me.

Sam Watson’s journal round-up for 9th January 2017

Every Monday our authors provide a round-up of some of the most recently published peer reviewed articles from the field. We don’t cover everything, or even what’s most important – just a few papers that have interested the author. Visit our Resources page for links to more journals or follow the HealthEconBot. If you’d like to write one of our weekly journal round-ups, get in touch.

Non-separable time preferences, novelty consumption and body weight: Theory and evidence from the East German transition to capitalism. Journal of Health Economics [PubMed] [RePEc] Published January 2017

Obesity is an ever growing (excuse the pun) problem associated with numerous health risks including diabetes and hypertension. It was recently reported that eight in ten middle-aged Britons are overweight or exercise too little. A strong correlation between economic development and obesity rates has been widely observed both over time within the same countries and between countries across the world. One potential explanation for this correlation is innovation of novel food products that are often energy dense and of little nutritional benefit. However, exploring this hypothesis is difficult as over the long time horizons associated with changing consumer habits and economic development, a multitude of confounding factors also change. This paper attempts to delve into this question making use of the natural experiment of German reunification in 1989. After the fall of the Berlin Wall a wave of products previously available in West Germany became available to East Germans, almost overnight. The paper provides a nice in depth theoretical model, which is then linked to data and an empirical analysis to provide a comprehensive study of the effect of novel food products in both the short and medium terms. At first glance the effect of reunification on diet habits and weight gain appear fairly substantial both in absolute and relative terms, and these results appear robust and well-founded, theoretically speaking. A question that remains in my mind are whether preferences in this case are endogenous or state dependent, a question that has important implications for policy. Similarly, did reunification reveal East German preferences for fast food and the like, or were those preferences changed as a result of the significant cultural shift? Sadly, this last question is unanswerable, but affects whether we can interpret these results as causal – a thought I shall expand upon in an upcoming blog post.

Ontology, methodological individualism, and the foundations of the social sciences. Journal of Economic Literature [RePEc] Published December 2016

It is not often that we feature philosophically themed papers. But, I am a keen proponent of keeping abreast of advances in our understanding of what exactly it is we are doing day to day. Are we actually producing knowledge of the real world? This review essay discusses the book The Ant Trap by Brian Epstein. Epstein argues that social scientists must get the social ontology right in order to generate knowledge of the social world. A view I think it would be hard to disagree with. But, he argues, economists have not got the social ontology right. In particular, economists are of the belief that social facts are built out of individual people, much like an ant colony is built of ants (hence the title), when in fact a less anthropocentric view should be adopted. In this essay, Robert Sugden argues that Epstein’s arguments against ontological individualism – that social facts are reducible to the actions of individuals – are unconvincing, particularly given Epstein’s apparent lack of insight into what social scientists actually do. Epstein also developed an ontological model for social facts on the basis of work by John Searle, a model which Sugden finds to be overly ambitious and ultimately unsuccessful. There is not enough space here to flesh out any of the arguments, needless to say it is an interesting debate, and one which may or may not make a difference to the methods we use, depending on who you agree with.

Heterogeneity in smokers’ responses to tobacco control policies. Health Economics [PubMedPublished 4th January 2016

In an ideal world, public health policy with regards to drugs and alcohol would be designed to minimise harm. However, it is often the case that policy is concerned with reducing the prevalence of use, rather than harm. Prevalence reducing policies, such as a Pigouvian tax, reduce overall use but only among those with the most elastic demand, who are also likely to be those whose use leads to the least harm. In this light, this study assesses the heterogeneity of tobacco users’ responses to tobacco control policies. Using quantile regression techniques, Erik Nesson finds that the effects of tobacco taxes are most pronounced in those who consume lower numbers of cigarettes, as we might expect. This is certainly not the first study to look at this (e.g. here and here), but reproduction of research findings is an essential part of the scientific process, and this study certainly provides further robust evidence to show that taxes alone may not be the optimum harm reduction strategy.


Some comments on obesity

A recent post over at Ezra Klein’s excellent Wonkblog caught my attention, both for being a post about a health economics issue, but also due to comments users had posted below. The post is about a recent set of projections about obesity and reports that there are new estimates that the burden of obesity in the US, if current trends continued, will be $550 billion over the next two decades. But this could be cut by almost 10% if trends continued 5% slower than they currently do. While this raises awareness of the importance of targeting healthcare interventions at preventing obesity and the diseases associated with obesity such as diabetes, it also raises important questions about responsibility for health. The comments section raises some important points:

MikeSoja wrote:

It wouldn’t cost “us” a dime, if “we” weren’t forced to pay it.

How about fat people pay for their own problems?

This, perhaps uneloquent, point illustrates the quite common opinion that behaviour matters. Those individuals to whom a causal responsibility could be assigned for their condition, such as smokers, should have a smaller right to healthcare resources. The negative externality that arises from unhealthy food may be countered with a Pigouvian tax – fat people ‘paying for their own problems’. One country, Denmark, is already trying this by taxing butter. However, this tax is foolish – it is not butter consumption that leads to obesity! People are becoming ever more aware that it is refined carbohydrates that are more to blame. So, should a tax be placed on these products?

Obesity, and other diseases that may be viewed as being caused by personal behaviour, are more prevalent among the poor. A tax on the products consumed more by the poor (inferior goods) would be regressive and would contribute to inequality. You might argue that the (threat of) restriction to healthcare is enough to reduce consumption of these harmful goods, but, it is unlikely that that threat will make much difference, particularly since the negative effects occur in the future and time preference matters. This could lead you to the comment made by AnonymousBE1:

But, if obesity affects mortality, which it does on average, then there will be huge cost savings in terms of unpaid Social Security and other pensions. In addition, people who live longer do not have less expensive final years – it’s just that those final years come later. Obesity is a quality of life issue, but I am not convinced that it is a fiscal issue….

This is a salient point, and it is something which health economists know much about. But as liamdc710 points out (in not a strictly polite way):

Your analysis is as stupid as those who say smokers save people money because they die young from cancer. In fact, far fewer than half of smokers die from cancer at an age younger than the national life expectancy. For the majority of obese Americans, much like smokers, insurance companies, families, and the public health systems in place will see increasing expenditures on related diseases that require expensive life long treatment.

To which AnonymousBE1 replied:

Okay, well, there needs to be comparison studies of LIFETIME expenses, medical and pension and disability.

I think this final point is something on which most economists would agree. But, even if costs are greater for the obese, does this mean they have less of a right to quality of life?

Many commentators ascribe to a luck egalitarian point of view whereby health differences that are due to sheer luck should be ‘evened out’ after which personal responsibility should play a role – we should be responsible for the consequences of our actions. It is what John Roemer calls ‘the cost of freedom’. However, how individuals respond to the same diet or lifestyle is often a matter of sheer luck; an individual’s genetic lot plays a big role in their propensity for obesity or the damage they receive from alcohol, for example. Separating luck from behaviour is highly difficult, if not impossible. Furthermore, saying that an individual ought to behave in a certain way or else face the consequences does not necessarily mean that an individual can behave in that way.

arm3a posted:

Man, is this troubling. Looks like a Pigouvian moment to me.

Perhaps it is, but I think it comes down to what you view the function of the health system to be. Some view it as an insurance system. So those most likely to require healthcare should pay more which could be funded by such a Pigouvian tax. But, I believe a national healthcare system acts more as a system of redistribution. As I have mentioned in a previous post, health is a special good and a precondition for achieving any of the things we have reason to want in life. The socioeconomic differences in unhealthy behaviours are large. Working class men smoke more than the middle classes. But these differences in behaviour only account for some of the socioeconomic differences in health outcomes. We should see reducing inequalities as a greater social responsibility than punishing those who become ill partly as a result of their behaviour.