Hacker News Reader

Comment by ucaetano | Hacker News Reader

Economics is a science as much as hydrology is a science: it isn't exact, but it is based on solid science.

Nobody derived Navier-Stokes equations to reach the formulas that show the erosion of a river bank, or the discharge calculation of a weir. Doing so would be nearly impossible given the extremely high level of complexity.

So instead of trying to start with the fundamental equations and build "up", you start by analyzing the behavior and modeling it.

It isn't any different than what Kepler, Newton, Galileo, etc. did: you look at the real world and build models that describe the world with enough precision for what you need.

Sure, you might say that hydraulics isn't a science per-se, but an applied field and so on. In that case, so is economics.

View on HN

I think the truth is somewhere in between because of the relationship of economics and politics.

There are absolutely academic economists who are doing a reasonable job trying to find the objective truth underlying human behaviour in economic life, understanding market forces, understanding relevant institutions, and so on.

However, the majority of what actually makes it into the public discourse is simply politically motivated reasoning, often in the form of hacks writing for think tanks that are paid for by special interests.

It's amazing to me that, at least when I was at university, the computer science curriculum had more content concerning ethics and our responsibility towards society than the economics curriculum.

No other field of knowledge is as poisoned by politics as economics, by far. That's why even though economics can be a decent science, a healthy amount of scepticism towards anything you read by (even supposedly academic) economists in the public discourse should be the default attitude.

Politics is power and power does corrupt. I am sorta glad that CS and the pure sciences remain largely free from this sort of perverse influence, although it does make me wonder how much more progress we could make as a species to fully explore and understand economic behaviors of societies.

Economics is not based on solid science, especially not macroeconomics. There is no way to run proper experiments and have counterfactuals. So it ends up being mostly opinions wrapped into a 3 variables linear regression fitted on 4 data points.

"Many empirical controversies in economics are essentially disputes about whether or not certain variables constitute valid instruments" - Econometric Theory and Methods, Davidson & MacKinnon

There are also no ways to run experiments in cosmology.

The difference is that you don't try to act on the cosmology models to change the way galaxies behave. Most of practical economics tries to derive policies or make choices out of its models.

No, you don't. But having the complex macroeconomics models far beats relying on layman's guesswork.

You've never looked at a macroeconomics model, have you?

Economics is complicated because it is modeling the behavior of human beings.

Math is easy, computer science is easy, modeling the collective behavior of humans is hard.

That's irrelevant as far as the designation of "science" is concerned. This is a question of knowledge and certainty, not practical application. The point is that results in cosmology (and other fields) are very well confirmed despite the fact that they don't come out of strict experimentation. In that same vein, when someone denies that economics is a science I'd argue they're really talking about whether its results are robust, not whether economics uses a certain methodology or is put to certain ends.

On the contrary, cosmology is based on physics (gravitation, Maxwell equations & so on) and a good bit of chemistry. 99% of the knowledge used in cosmology is based on science science developed and tested "in vitro" on Earth.

But you still can't run experiments on cosmology. SO if your definition of "science" is "based on experimental data", cosmology isn't a science.

"Economics is not based on solid science"

Pretty broad statement. Some micro-economic theories, like the game theoretic models in auction theory, contract theory, mechanism design and behavioral economics are probably the most successful, accurate and impactful theories in all of social sciences, including psychology.

You may not realize how much your buying behavior, your salary, your interactions online and your use of technology are dependent on these models. But comparing their success and use to fields like sociology, management and even psychology, their track record is frankly amazing.

We already know that Macroeconomics amounts to modeling systems that do not aggregate meaningfully from these successful micro-models.

Saying economics is not a science based on this, is like saying physics is not a science because we still can not find a unified theory. Which is a pretty good analogy for a Macroeconomic model that forecasts GDP (aka, everything).

Economics also has a pretty wide range of subjects. I would argue that something like macroeconomics isn't based on particularly anything (and has dubious value), but certain aspects (some microeconomics, game theory, etc) are somewhat more based on either logic or empiricism.

Indeed, economics end up being the study of resource allocations by human beings, both at the individual and collective levels, and the resulting system.

But macroeconomics is indeed based on a lot of things, and you seem to be very misinformed. Macroeconomics relies very heavily on mathematical tools such as calculus and topology to model the collective behavior of humans.

If you are really interested in learning more, I recommend "Knowledge and the Wealth of Nations: A Story of Economic Discovery".

Math is a language, it doesn't make something scientific (or not scientific). The issue of math and science are unrelated. The problem with macroeconomics is not the tools they use, it is that it does not have predictive power. We cannot meaningfully (better than martingale or historical MA) predict GDP/inflation/SP500 return etc. A lot of macro assumptions are obviously wrong (people do not optimize marginal utility, people have transitive preferences, nobody can meaningfully compute externalities of basically anything, etc)..

Reply on news.ycombinator.com