Economics, Literature and Scepticism

Powered by Blogger.

About Me

My photo
I am a PhD student in Economics. I am originally from South Africa and plan to return there after my PhD. I completed my M. Comm in Economics and my MA In Creative Writing (Poetry) at the University of Cape Town, where I worked as a lecturer before starting my PhD.

Saturday, October 22, 2005

The Philosophy of Economic Policy Making (or 'Work Avoidance on a Saturday Afternoon)

Posted by Simon Halliday | Saturday, October 22, 2005 | Category: |

Subsequent to having studied a course in Policy Analysis in my Honours year of Economics and now having completed a development style course in Masters, something has become horribly apparent to me. This problem is that people pre-judge the discipline of Economics because of its idealism. Not only is this 'sin' committed by policy-makers external to Economics and by politicians attempting to derogate the discipline, but also members within the discipline itself.

What makes this doubly frustrating is that Economists have to sell their models and their methodologies to government in order to create some kind of actionable change in policy or in the economy itself. However, if Economists themselves cannot agree on what is necessary in order to have a decent Economic model, then this is more than problematic. When I talk about this dispute I am not referring to the mundane nature of the New Keynesian-New Classicist debate, but rather an approach to statistics and economics combined.

Now, every economist who reaches a high enough level of competence and study inevitably encounters the problems with assessing whether a program or intervention works or not. In order for us to know whether an economic or policy intervention works, we need to be able to answer a counterfactual question: How would those who were in the presence of said intervention have fared were they not treated or in the presence of the intervention? Or equally, how would those who were not in the treated group have fared were they treated? If we cannot answer these counterfactual questions then the whole point of applying policy in the first place is irrelevant. The reason for this is that policy is meant to enact positive change, change that will help the economy and that will provide for people in an optimal way. Moreover, it is meant to do so in the most cost effective way. If a policy is rolled out to the entire population, but it is ineffective then spending the money in the first place was inane and a poor choice when there are so many opportunity costs to poorly spent resources.

Now, the methodology of Randomised Evaluation1 is easily understood by Econometricians and by Statisticians. However, policy makers in attempting to address specific population groups and constituents often ignore both statistics and econometrics in favour of attempting to look as though they have done something when in reality nothing has changed. This is all to reminiscent of the Peron style policies of giving people pots and pans to make them like you, but these pots and pans DO NOTHING to help these people in reality.

Where the problem lies for Economists is that people often don't like 'The Truth'. When I talk about 'The Truth' in this context, I refer to statistics or qualitative statements which provide us with some reflection of what occurs in reality, rather than some peoples rhetoric about 'poverty alleviation' or 'employing the unemployed'. These themselves are worthy goals, but claiming that a project or a policy will do one of these things is inaccurate when we have no way of measuring if a change occurred.2

This is where the problem arises, both within Economics and with those who attempt to deride it. Two main problems that people have with such statistical practices are as follows: 1) it's unfair, and 2) it (randomised evaluation) is ivory tower intellectualism that doesn't work in the real world. I will deal with both statements.


As far as the first is concerned, this is a drastic misrepresentation of what Randomised Evaluation is and does. A randomised evaluation attempts to find two very similar groups, intervene in one with some treatment (say an extra teacher in a classroom, food provision or some such) and compare this treated group against the similar group which remained untreated. This is often performed on a fairly small (depending on your interpretations) scale, for example looking at 20 schools: 10 of which get the intervention and 10 of which do not.

Now, if we need to roll out a program to enhance school quality, BUT we are unsure of how to do this just reading what other people have done and claiming that 'This is what is good for 'our country' does not actually tell us if it is. In order for it 'to be good for our country' it should fulfill some criteria – test scores should improve, attendance should increase, numeracy and literacy levels should improve. However, the only way we can know if they have improved at all is if there is some basis for comparison. If we expend money without knowing whether a program will do anything, then how is that at all fair? As far as a utilitarian argument progresses, welfare has not been improved by any measurable outcome, but large amounts of money have been spent, resulting in an overall decrease in welfare (same people, less money).

If a small pilot project is run with some people getting the intervention and some not, then we can know whether it works. However, people then claim that on the ground if we have two households one receiving a treatment and one not, then the one not receiving the treatment believes that it is being prejudiced against. Notwithstanding the myopia of individual consumers who could voice frustration at not receiving some patronage, the individual unfairness is overridden by social benefit. Hence, again using a utilitarian argument, the overall utility is improved because of us knowing what programs to run and what the impacts are regardless of a small number of peoples' discontent. Contemporary society is riddled with personal injustices which are for the overall benefit, to take a few examples: progressive tax rates are individually unfair on the rich – they give away money for which they have some claim because of work, but this is taken away from them in the name of a greater good, equally so with government intervention in a number of other spheres: government providing free water to the poor (or providing a market for water to the rich), government re-modeling education structures so that they are more equitable and less competitive. Individually, this irks people and people find it individually unfair. However, they are intended to provide for a greater good. People then re-frame the questions as a one of poor vs. poor. 'How can you tell one poor person that you are going to employ their equally poor neighbour, but not them?' Government does this all the time anyway when it initiates public works programs, or when there are programs by international bodies as part of disaster relief or some such. This means that there must be some other motivation – the 'fairness' argument does not stand up to scrutiny. In fact I believe that it comes down to the second point which is the attack on 'ivory tower intellectualism'.

The problem with this second point is that it is predominantly a normative or perspective-based debate. Almost inevitably it remains as a post-modern attack on Statistics and Economics which I foreshadowed earlier with my comments on 'Truth'. People (myself included), often like to believe that they are individuals, that they cannot be described by observable characteristics and that they (because they are special) do not fall into any stereotypical roles or cultural boxes. The problem with this is that it ignores a host of economic, psychological, sociological and anthropological evidence to the contrary. As much as we would like to believe that our free will can and should override any genetic or social imperatives, it often does not. People DO fit into general and describable categories. This is where statistics and econometrics are useful because they DO REFLECT REALITY. Granted this reality is not in an individual or unique sense, but as far as most people are concerned providing them a situation in which they can begin to have food, family and something more than just subsistence living is good.

This then begs the question of where this attack is directed? This is a far more controversial question. My belief is that it is pointed at the perceived elitism of the intellectuals, i.e. people who are intelligent are obviously out to get everyone who isn't. This can possibly also be seen as a misinterpretation of Marx, who wanted equal provision for people but without preferential treatment based on capability. So the problem here is that people perceive that intelligent people are rich people and that rich people abuse poor people. Therefore Academics (in their ivory tower institutions and intellectualism) are out to get the poor. This is a drastic misrepresentation of many academics' intentions. In fact the causation underlying this, possibly generally held, misperception is horribly flawed. It is actually the reverse – the appreciation of an intent to find truth from rhetoric is the intent of the academic. Hence the academic is trying to distill reality to find the constituent truthful parts. This can only be done in a situation where it is not being undermined by rhetoric or by the polemics of politicians whose own interests may be undermined by the 'truth' that is found through a rigorous statistical analysis.

In conclusion, it has been argued (possibly overly vociferously) that the attack on statistically rigorous processes such as Randomised Evaluations is actually based on a normative position, which is inevitably based on the self-interest of the politicians (and maybe other economists whose findings may be undermined) as a result of it. Sadly, it seems as though economists who are striving to find 'truth' are in the minority, or if this quest is their initial intention they become bogged down by political rhetoric and doublespeak which undermines that which they try to achieve. This leads me to my last appeal, that of a belief or at least an acceptance of the hopeful ideal situation of good econometric and statistical analysis in an ideal world, if economists begin to continuously excuse the rhetoric which undermines their work, then their work will suffer. In order for good econometric work to continue to be done in the future we cannot bow down to the intentions of too many self-interested individuals who target a polemic of elitism at those who are, in reality, trying to find new and innovative ways of solving problems of unemployment, poverty and welfare. It should be noted that this is not an attack on government as a body, but rather at individuals who may believe that government and academia are incompatible because of cursorily observed superficial differences, this is not the case. The intentions of government and a host of academics are those presented above – without by in and collaboration from both the possibility of brilliant solutions fades.


1See for example Duflo and Kremer (2003) or Duflo (2003) for explanations of RE.

2This does open me up to many of the problems with the hermeneutics and ontological arguments of what constitutes 'Truth', but in this context I believe that this conceptualisation of 'Truth' is sufficient. I am not about to get all European and post-modern on the discipline of Economics.

Currently have 0 comments: