Hacker News Reader

Comment by adekok | Hacker News Reader

A personal anecdote about Economics...

A friend was doing a Masters in Nuclear Physics at the same time as myself. She decided she didn't like it, and went to do a Masters in Economics. 4 months later, she was back in Physics.

Over a beer, she told me that the comment which did it for her was a professor who claimed "The math you're learning now is more complex than nuclear physics." Uh.... right. She showed me her notes (math, no explanation), and I correctly identified it as Bayesian probability calculations.

It's hard to have respect for a field when they're don't know what they're doing, and still think they're better than everyone else.

I'm sure there's lots of things in economics which are real and useful, but there's a lot of woo in it, too.

View on HN

Your comment reinforces my cynical opinion of economics.

A friend of mine who is a noted economist with a Ph.D. in statistics, commented to me with something along the lines of:

> The only reliable correlation in economic papers is the probability that an author will be published.

I take her comment to mean that big names from big universities get papers published, but publication is not based on the merits of the paper.

More to the point, she said that the statistics used for almost all mainstream economics is flat-out wrong.

> More to the point, she said that the statistics used for almost all mainstream economics is flat-out wrong.

As someone who has worked extensively with both PhD statisticians and economists, but practices neither beyond what I've learned via osmosis, I couldn't care less what the statisticians think anymore. They were definitely smart, but completely useless. We had several projects where the economists came up with solutions that measurably and objectively worked to the order of 9 figure changes in the bottom line and reversals of decades-long trends...and they did it using data sets that the statisticians wouldn't even touch due to some religious moral panic from observational data. Statisticians always found some way to object to everything under grounds of being proper in some way or another, and rarely came to the table with anything more useful than "First we need to find some way to create a perfect alternate universe".

Econometrics has very rightly diverged from statistics. Economists do not have the luxury of the theoretical environments that statisticians require but never have to create. And as a social science, they don't have the luxury of measuring outputs of perfectly understood processes, but rather the ever changing output of cultures and personalities guided by 3.5 billion years of evolution. Maybe their ideas offend statisticians, but it's pretty hard to find something that exists in the real world that doesn't offend them, so I'm not concerned.

I don't know about your industry but the abiding memory i have from before the financial crisis of 2009 was of economists fanning the property bubble with the mantra ' the fundamental are sound' and how the worst scanario would be a "soft landing"

The main problem with economics is that it's hopelessly corrupted by political and financial interests, in a way that hard sciences only intermittently have to deal with, except with stuff like global warming.

When people wave around the conclusions of scientific papers as proof that some political action is required which is going to impact the lives of millions of people, you need to seriously take the papers with a grain of salt.

I don't really know any way to get around it. Any time you're studying the impact of human collective activity the stakes are far too high to trust that anyone is taking a disinterested view of reality.

To support your point, the economist I can most clearly remember and who seemed to be on tv every other day, was chief economist with a bank that later needed rescuing.

> I don't know about your industry but the abiding memory i have from before the financial crisis of 2009 was of economists fanning the property bubble with the mantra ' the fundamental are sound' and how the worst scanario would be a "soft landing"

If you'd like to read about it, the best contemporaneous piece I've read on it is "All the Devils Are Here" - https://www.amazon.com/All-Devils-Are-Here-Financial/dp/1591...

I once sat in on a friend's final year economics course.

They talked about the "gravitational model" of economics, that the amount of trade between two economies is proportional to the product of the size of the economies each raised to some power, divided by some power of the distance between them. They showed a best fit of the 5 parameters to under a dozen data points and it looked rather awful. (I am not sure how well defined those variables are). Without any theoretical justification it seemed like a model stolen at random from physics that didn't fit well.

I've also read a published economic paper written in terms of measure theory, where it was clear the author did not understand measure theory at all. The ideas were misapplied to a simple problem in a way that made it sound very difficult. Admittedly this is partly the fault of mathematicans who teach measure theory as statistics; while you can use that model it makes things very complicated and you almost surely don't need to deal with infinities in most real problems, let alone non-measurable sets.

This arm of economics trying to ride the coattails of prestige worn by maths and physics is void of meaning.

You seem to at least have some personal experiences/familiarity behind your criticism, unlike almost everyone in this thread. So here's a chance to substantiate your claim. Here is a paper that was recently published in the Journal of Economic Theory https://dl.dropboxusercontent.com/u/17516137/RapidWeaverSite...

This was a co-authored publication of an economist and a political scientist (at the University of Rochester). It's one of their several co-authored pieces; and it was published in a good field journal, but not a top general-interest journal. It has so far received a couple of cites from papers on the same topic. My point is, this paper is fairly "standard" for contemporary economic theory.

It also happens to use a decent amount of measure theory.

Now, I will personally venmo you $200 if you are able to find a single mistake or "abuse" of measure theoretic concepts in that paper in the following sense: a measure-theoretic theorem, definition, or proof technique was unnecessary for mathematical purposes, and you can demonstrate why that's true by writing an alternative proof, with the same generality but without such a concept.

I'm not baiting you, I'm genuinely curious to see if you or anyone else in this thread can do it, because I had taken plenty of measure theory in my PhD days, and this paper's usage of it seems perfectly on point to me. I'm also using it as an example because the authors' language is very clear and meticulous, and you shouldn't have trouble accessing that paper without knowing economic jargon.

I can also link you a few other random economics papers if you want. Alternatively, you could admit your disparagement of economists' intellectual ability, and intellectual honesty, was perhaps misplaced.

Solovay showed that assuming the negation of the axiom of choice, there exists a model of real analysis where all sets are measurable. So really you only need measure theory if the paper depends on the axiom of choice in some fundamental way. But would you really trust an economics paper that depended on the axiom of choice?

For me, a good analogy is with engineers and real numbers. Engineers use real analysis all the time. But the results of an engineering paper should not depend on the definition of a real number, nor should the engineer care.

Edit: Hamming made the point better than I can:

"Does anyone believe that the difference between the Lebesgue and Riemann integrals can have physical significance, and that whether say, an airplane would or would not fly could depend on this difference? If such were claimed, I should not care to fly in that plane".

If you go on Arxiv and Journal of Finance, the quant papers there are very technical (very advanced statistical and PDE methods). Not saying it's harder than physics, but it's not a walk in the park either.

Thanks for the tip in broadening my negative experience. It makes sense that big banks do employ some brilliant people to do some VERY complex things which would blow my mind :)

It's also worth noting that a lot of quants have a pure maths/physics background rather than economics

Yes, "quants" tend to have undergrad math/physics backgrounds, but "quants" aren't the ones writing academic papers in the Journal of Finance. You, and many others in this thread, seem to have a painful lack of awareness that there are things called PhD programs, and that they select and train students to specifically do research. Here's an analogy: a random comp-sci undergrad/master's isn't going to be publishing in a conference any time soon. Most of you have no chance of getting into a PhD comp-sci program, just as most random econ students have no chance of getting into a PhD econ program. Both of these select heavily from mathematics.

Here's the editorial board of the Journal of Finance:

http://onlinelibrary.wiley.com/journal/10.1111/(ISSN)1540-62...

You can feel free to check their affiliations for yourself, but almost every single person has a PhD in Economics or a PhD in Finance/Financial-Economics from a top department. PhD Finance programs are structured like PhD econ programs (you take the usual micro/macro/metrics), and were historically rooted in economics departments. And for what it's worth, almost all of them will have a bachelor's in mathematics, or a master's in mathematics, or in the case of international students, a master's in economics.

The topic of this thread is about academic economics, yet you and everyone else are attacking it by citing the irrelevance of undergrad economics in the U.S.. Which we're all painfully familiar with. But believe it or not, the best aspiring comp-sci academics in the United States rarely take a full major of comp-sci courses at the undergrad level either, because U.S liberal arts majors are frankly geared towards students who have no interest in academia. This does not imply a single thing about graduate-level comp-sci, economics or finance research.

Another anecdote (engineer here): We were told in my econ class how difficult the math was and I always struggle with core math classes like calculus and differential equations, so I was worried. I spoke to the professor and he asked my major. When I replied with "engineering", he asked my GPA, to which I replied 3.6. He then laughed and said I'd be fine. In the next few classes I found out that your beginner Macro & Microeconomics courses basically need you to be able to add/subtract/multiply/divide and find the area of a triangle. So basically any 5th grader could do the problems although terms like marginal cost might sound scary. I know in more advanced courses things like probability & statistics come into play and that there are many professors who know what they're doing. Despite that, I feel like many graduates probably don't know the background, just the tooling to use for a particular class of problems. As much as I hate to admit it, the derivations typically done in Math/Physics/Engineering serve a real purpose (besides annoying students) in that they show you how things really work and came into being. Without knowing the why, you're basically plugging numbers into a black box and hoping you don't do something that causes it to break down. Despite all this, there are some real gems in Economics that are very useful. I'm re-learning econ at the moment and it is helpful to know how markets work.

>We were told in my econ class how difficult the math was and I always struggle with core math classes like calculus and differential equations, so I was worried. I spoke to the professor and he asked my major. When I replied with "engineering", he asked my GPA, to which I replied 3.6. He then laughed and said I'd be fine

Was this in the US? Undergrad econ is very watered down compared to Europe. A common lament here is that people who get their undergrad in econ in the US are woefully unprepared for the math they'll need in grad school.

Econometrics uses as much calculus as engineering does (PDE's, etc). In other countries, they do this at the undergrad level.

Yes in the U.S. If econometrics in Europe actually goes over partial differential equations, then it is an order of magnitude more complex than schools in the U.S. If they do cover that in the U.S. it would only be done at a glance as they don't take a real calculus or differential equations or other mid-advanced math classes. They do have business calculus, but that is a joke.

Oddly, economics is one of the fields where you're probably better prepared for graduate work in economics if you major in something other than economics. Math, physics, or engineering may be better prep.

Of course, one solution is to just major in Econ and make sure you take those classes. The US undergrad path is very different, it's not as specialized as it is overseas. There's flexibility - well, use that flexibility to take math through diff eq, linear algebra, some proof based stuff, and perhaps some numerical analysis or computing.

Also - some universities essentially offer two tracks. My alma mater, UCSD, does this. When I was there, you could major in Econ or "QEDS" - quantitative economics and decision science. I think they've changed the name and modified the curriculum since then but it's a similar track - it requires that you sit with the engineering, math, and hard science majors for the first couple years of calc, linear algebra, and DE's, and offers upper division work that draws on this background. The basic Econ major required very minimal calculus from a shorter and less rigorous sequence.

This split is present in a lot of US universities, as far as I understand. You really can't know just by looking at the name of the degree, because a "BA" in economics at some universities actually does reflect a mathematically rigorous program.

It's also present in other fields. Maybe in most fields. We're kinda looking at the gem of knowledge from our own perspective thinking we found the best one.

Sadly, no. Or, not anywhere near that level of BS.

There are fields where the practitioners actively try to dis-prove themselves. Where they are aware of the limits of human psychology, and design experiments / analsyses to work around their own limitations. These fields are called "science".

There are other disciplines which don't do that, but who do fake it. See Noam Chomsky for a great explanation:

https://www.youtube.com/watch?v=1M-2_OD3oMc

It seemed to cut off before he got to the "faking" part.

Did you watch / understand the video? The entire thing was about non-science departments faking the trappings of science.

For another explanation, see "Cargo cult science"

https://en.wikipedia.org/wiki/Cargo_cult_science

His whole point (and mine) is that a field of study can follow the superficial trappings of science while at the same time not understanding the core principles of science. If that isn't "faking it", then you should explain why. Use examples.

I think he must have been too subtle.....a longer version of that video autoplayed a bit later and made it a bit more clear. People from those departments are often so deluded that if you didn't know who this man was, you may very well have thought he was advocating that they are equal.

> It's hard to have respect for a field when they're don't know what they're doing, and still think they're better than everyone else.

So you judge a field from one conversation that a friend told you happened? From one professor? Who might have even been joking.

And you think that is a reasonable proposition?

In my school, business, compsci and psychology had a GPA requirement.

All of the fratboys who washed out of the Business department either landed in sociology, MIS, or economics. They seemed to do well there, and get decent jobs as long as they had 15 credits of accounting.

I was told that they didn't really wash people out until the graduate level.

Wash out from business and psych TO econ is strange. My impression of the rigor of those disciplines:

Econ > Psych > Business

Econ/psych debatable, but business definitely isnt

You're considering the supply side, but not the demand side. If the business major required a certain GPA to stay in, the competition among students could kick someone out, especially if grades were calculated on a bell curve. Someone who can't hack it would consistently score lower than their peers in that scenario and not meet the GPA cutoff.

Reply on news.ycombinator.com