Falsifying causalities through generalisation
Causation in social science is hard. Beliefs about causality are based on experience* or, in modern social science, experiments. Experiments can construct causality via revealing counterfactual states (e.g. twin studies). Nevertheless, counterfactuals are difficult to experiment when hypotheses are about the whole of society. Rather, large questions in social science and economics require trend extrapolation, typically regression analysis, to prove causality. Regression analysis cannot prove that a causality A->B is not caused by a third event, C, which is causing both A and B. Due to this, regression analysis is a poor tool for proving causalities in the real world: many important phenomena in the real world consist of feedback loops (such as C->B->A->C). Thus, it almost always feels correct that
B->A and more broadly C->B->A->C.
In this short post, I propose an alternative tool to falsify causalities. I use one of the main debates in social sciences as an example: the debate between so called “classical” (aggregate supply) and “Keynesian” (demand-driven) economics.
The core promise of the industrial age to the people was:
Growth leads to work. Work leads to wellbeing. (source)
In the preindustrial era people wanted wellbeing: food, shelter, washing machines. The means to take part in the society was a job. Economic growth was the means to provide jobs to all.
The purpose of growth is economic development. Growth was a meaningful goal because it’s easier to measure and incentivize. But it’s possible to generalize growth to development: growth was a constituent of development.
The purpose of work is participation — for subsistence and for meaning. The meaning of the word “work” in the industrial promise meant jobs. We can generalize the word work and replace it with a broader concept of participation: a job is one way of participating.
Inequality of wellbeing was the main challenge of the preindustrial society. Many lacked shelter and food. In the industrial societies, the main source of inequality and unfairness is inequality of capabilities. This means: many lack education, skills and tools they could possess in the right circumstances. Both inequality in wellbeing and inequality in capabilities, when fixed, provide freedom. Freedom from need; freedom to pursue one’s full potential. We can thus replace the word wellbeing in the industrial era promise with a more abstract concept: freedom.
Thus, the industrial era promise is now replaced with a broader promise. Looking at the concepts and their interlinkages, it’s now possible to realise that the causalities in the abstract promise operate the other way around.
Through freedom to pursue one’s full potential to all, more broad and fruitful opportunities for meaningful participation arise. Through this participation, new development happens in the form of new tools and less problems hindering the freedom of man. Furthermore, we notice that there is a self-reinforcing loop, as there should be.
The promise of the post-industrial society follows from this more abstract notion, but this is not the interest of our current investigation.
Instead, we can now return to the original hypothesis but with the increased suspicion that the causalities might operate differently than what was previously considered, based on the generalized supermodel. And voilà, we can indeed see that the causalities can be turned.
Wellbeing in society increases demand for jobs, and this leads to growth. Further, there is a direct feedback from increased growth to increased wellbeing.
The initial model for industrial era replicates the assumptions of the classical economics, while the latter model reconstructs the Keynesian economics as a social promise. Is it possible to conclude based on this word play that the classical economics take on the debate is a cum hoc ergo propter hoc? Maybe. For a conclusive proof at least better resolution is required.
However, it’s possible to claim based on the thought process presented here that the Keynesian take on the debate seems to be more sustained also after a specific time in the history. In the industrial era the connection between jobs and wellbeing was indeed rather close but this might not be true in the years to come. Thus, understanding the causalities correctly is crucial to make a sustained promise for the postindustrial age.
* and experience is based on the experience based assumption that the future follows the logic of the past, a circular logic according to David Hume