There are simply too many moving parts to make some sort of grand quantitative science out of all of it. That's why imo there will always be a subjective/interpretative component to analyzing economics and financial markets.
very well put
@Raoul
Is one problem with economic research that the world is simply not big enough? For example one can look at how an economic policy might have affected a country's performance, but then there will be far more factors influencing that performance than there are countries with a similar policy. If there were 10,000 nations on the planet it would be easier to get stuck into individual factors. Forgive me if this is just stating the bleeding obvious, but I haven't studied economics at any level.
It's more to do with the quality of data and the complexity of issues. While headlines like "shark attack decides election" are obviously bonkers, but the amount of difficult-to-track-variables one can control for are endless. Most of the easy-to-do stuff is done. We are talking about small effects in extremely noisy data.
Some examples:
Do early childhood interventions for (stunted) children have a lasting impact on the success of children later in life? My prior intuition is that intervention can help. My own life experience, psychology, what I learned about the biology of growing up and other different theories/”hunches” seem to all point in this direction. At least in my very subjective view.
Yet how do you track and evaluate that?
Prestigious scientists did that. Some conclusions were that these children earn 25% more compared to those without intervention. They also found, that specifically the nutritional part of the intervention has been pointless, because the families shared the supplements. So they message is, that the psychological stimulus resulted in 25% higher incomes (and many other positive things) over 20 years later.
Nobody can possibly believe their number(s). I doubt that a single of their researchers is taking them seriously. Why? Because thats just way too high of a number. Nothing is going to have such a massive effect. That’s completely out of the cards. Its not how the world works.
So I spare you the technical details, but obviously people looked into it and found many many flaws of the research. The data is so vague, that even a negative effect is entirely possible (would they have published a negative result? nope. Would they and could they have published a small positive result? Well, considering the methods and research design that would have been surprisingly difficult (which might seem counter-intuitive, but thats true) and who would have cared if they would have found that their intervention had 2% positive impact on wages?).
Other findings you could have also extracted out of their data: Childhood intervention increases likelihood of leaving Jamaica or getting fewer babies.
My point is, that this research didn't created any new knowledge. It just affirmed what most liberal minded people (including the researchers) already thought anyway due to their prior experience.
The public certainly didn't care about the very few cautious voices, who questioned this research. This study was (is) extremely influential in building political support for these programs in other countries. Thats probably a good thing, but makes an honest discussion how anyone came to this conclusion impossible.
The same problems are happening everywhere in almost all policy relevant research areas from labor market regulation, to education, to healthcare, to development policy to everything. There are good studies out there that actually create new knowledge, but its hard to find them when there are so many bad studies as well. On the first, second and third look good and bad research looks the same.
I posted initially an article from Noah Smith, who I like a lot despite me disagreeing with most of his views. His argument is:
Economists study
gender relations in the workplace,
racial gaps, changes in
labor contracts, early childhood
education,
minimum-wage policy, regional
opportunity gaps,
automation and the future of jobs, and a vast array of other highly important, immediately relevant topics. Sometimes they use theory to help them understand these phenomena, but usually the core work is empirical analysis of hard data.
Well. “empirical analysis of hard data” can make economics less scientific (and transparent), when people don't understand the current limitations of this approach. It gives the whole endevour a very shiny and objective image, but that has nothing to do with the hard reality of what actually happens in many cases.
