4
$\begingroup$

Specifically I was wondering about this question.

It is definitely economics related, but once one writes down some equations it seems more like a computer science problem.

$\endgroup$
1
  • 1
    $\begingroup$ I think the standard website to go to is StackOverflow, e.g. with the Stata tag or R tag. $\endgroup$
    – ahorn
    Aug 12 '19 at 15:52
3
$\begingroup$

Economics continues to wed itself to statistical programming and fluency with R, Python, or some other language to do macro based problems. A question about programming a neoclassical growth model will definitely have relevance for future users of this site, and though it doesn't ask any explicit questions about the model itself and is, as you say, mostly a computer science question, I think a good answer to this question can highlight how to ensure the code for the model make economic sense. That is, it can actually solve it in a reasonable manner.

As you brought up in the comments, the form of utility and production matters. Particularly, we want to ensure the model has a solution and justify why we put certain constraints on some functional forms, and that is something that is relevant to economics. Code for this finite horizon problem might also use Euler equations, or a dynamic programming method, and that is relevant enough math for an economics problem. So that's my rough reasoning for why I'd say the question is on-topic.

$\endgroup$
2
  • 1
    $\begingroup$ I like the points you raise but I am a bit confused. From "ensure [...] the code [...] can actually solve it in a reasonable manner" it seems to me that most of your answer would be an algorithm, not actual code? You could of course also include a code snippet, but wouldn't you mostly be explaining the algorithm itself? $\endgroup$
    – Giskard
    Dec 24 '18 at 19:21
  • 1
    $\begingroup$ Sloppy wording on my part. Whether parts of actual code, or an algorithm, or a bit of both is presented as part of an answer, does not particularly matter. You are right that practically one would probably present an algorithm that can solve the problem. $\endgroup$
    – Kitsune Cavalry Mod
    Dec 24 '18 at 19:31
2
$\begingroup$

Mathematics, numerical programming, statistics and software implementations of techniques from these fields of knowledge are all an integral part of economic theory and method. I think anyone who wants to spend their time here on earth contemplating these matters deserves to be welcomed here.

For the particular problem I tried to argue that the economics of the problem mattered. Offcourse once Tara has gone from economic model to choice of algorithm to solve for equilibrium one could argue that there is nothing particularly 'economic' about the algorithm. Which is true - but it is also reductionism from a scientific point of view. When she comes back with some R code and faces a problem we could brush her off saying: oh this is just code not really anything to do with economics.

I think that would be unfortunate.

It is not about the algorithm or the code in itself but about establishing that this algorithm and software implementation can be used to solve this problem arising in economics and then going on to actually doing it. It is the why, the how and for what the algorithm is used that makes its a brilliant example of economic scientific method. Set up model, formulate equilibrium conditions, choose solution method and finally implement it succesfully.

Why would we rob ourselves of having an example of such a proces on this site even if at an introductory level? To me it seems like a good example of why mathematics, numerical algorithms and computer programming are important for economics.

"To harness the full power of computer technology, economists need to use a broad range of mathematical techniques. In this book, Kenneth Judd presents techniques from the numerical analysis and applied mathematics literatures and shows how to use them in economic analyses" Showing how to use math and programming to solve economic problems is contributing to economic scientific method.

Mathematics has not always enjoyed the place it holds today in economics but hey with general equilibrium theory who is going to argue against the importance of math for economics. So we accept question solving for Marshall demand function although offcourse it is "only math".

Statistics has also become integral part of economics much through the development of the field of econometrics. So we accept questions on choice of statistical methods we just call it 'econometrics' because then it is ok to do it as an economist.

What is the word for programs and code when used for economic problems? Do we need one in order to accept the relevance of 'code' within economic method instead of interpreting 'code' in the reductionist sense of 'just code'.

We all know that software today is important. There are several R packages or stata packages that are written by economist for economists using theory and methods developed and published in journals within the field of economics.

So in conclusion I hope that the community of moderators will choose to display a tolerant policy when it comes to questions involving programming. I would hate for EconomicsBeta to end up having just as bad reviews as StackOverflow have on Trustpilot [https://www.trustpilot.com/review/stackoverflow.com]. I realize it is small sample but I recognize picture it paints.

$\endgroup$
8
  • $\begingroup$ This meta question is from 2018, I did not write it today :) I advised you to come to meta because it seemed like you want to achieve consensus on coding, and meta is the place for that. $\endgroup$
    – Giskard
    May 17 at 16:22
  • $\begingroup$ Also, I could not care less about the Trustpilot reviews; I do not own company stock. $\endgroup$
    – Giskard
    May 17 at 16:25
  • $\begingroup$ @Giskard arhh ok I thought it was about the current question - what a waste of energy. $\endgroup$ May 17 at 16:45
  • $\begingroup$ @Giskard and with the new edits of the question it looks more like a really low effort question I would not even have bothered to answer ... $\endgroup$ May 17 at 16:52
  • $\begingroup$ I did not want to disillusion you :) I usually try to avoid spoonfeeding answers. If they understand the algorithm, great! They can easily write the code themselves. If I give code, I doubt they would learn much from it. A specific question about coding may also be good, but I am skeptical about generally welcoming coding questions. $\endgroup$
    – Giskard
    May 17 at 16:56
  • $\begingroup$ @Giskard "I did not want to disillusion you :)" - I did not answer bro. Code, math its all the same. $\endgroup$ May 17 at 17:01
  • $\begingroup$ The world is calling for a hero! $\endgroup$
    – Giskard
    May 18 at 13:55
  • $\begingroup$ Just going to keep referring these to you. $\endgroup$
    – Giskard
    May 19 at 5:50

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .