Economist-Speak

I try to keep up with the current economics literature, which means reading quite a few papers and a whole lot of abstracts. Most of the literature isn’t very interesting or meaningful to me (this is simply a matter of preference); and some of it might be interesting or meaningful but I am unable to tell. Why? Because the language of economists is often – not always, certainly, but often – deeply obtuse. Now, again, this is my problem, having to do with my preferences and my skills. Research economists, like most academics, are writing for their peers, not the laity. And as much as I might like their research to be written in plainer English – especially since it is subsidized in part by my tax dollars and yours – I understand that technical language is sometimes vital and that all guilds have their own dialects that reside beneath the very big tent that is American English.

That said, sometimes I just have to laugh. Here, from the National Bureau of Economic Research, is the abstract of a recent paper by James H. Stock and Mark W. Watson, who I assume are very smart people and good economists. The paper is called “Heteroskedasticity-Robust Standard Errors for Fixed Effects Panel Data Regression,” and is summarized thusly:

The conventional heteroskedasticity-robust (HR) variance matrix estimator for cross-sectional regression (with or without a degrees of freedom adjustment), applied to the fixed effects estimator for panel data with serially uncorrelated errors, is inconsistent if the number of time periods T is fixed (and greater than two) as the number of entities n increases.

Like I said, I don’t think this paper is for me. (But if it’s for you, please check it out here.)

So imagine how delighted I was to see another recent paper, “On the General Relativity of Fiscal Language,” by Laurence J. Kotlikoff and Jerry Green, which admits that the language of economists (well, the ones who write about finance at least) is essentially meaningless. From the abstract (emphasis added):

A century ago, everyone thought time and distance were well defined physical concepts. But neither proved absolute. Instead, measures/ reports of time and distance were found to depend on one’s reference point, specifically one’s direction and speed of travel, making our apparent physical reality, in Einstein’s words, “merely an illusion.” Like time and distance, standard fiscal measures, including deficits, taxes, and transfer payments, depend on one’s reference point/reporting procedure/language/labels. As such, they too represent numbers in search of concepts that provide the illusion of meaning where none exists. This paper, dedicated to our dear friend, David Bradford, provides a general proof that standard and routinely used fiscal measures, including the deficit, taxes, and transfer payments, are economically ill-defined. Instead these measures reflect the arbitrary labeling of underlying fiscal conditions. Analyses based on these and derivative measures, such as disposable income, private assets, and personal saving, represent exercises in linguistics, not economics.

I’d like to thank Kotlikoff and Green for daring to tackle, even if in a narrow measure, the language of economics. While I don’t have a pressing need to be able to understand heteroskedasticity-robust standard errors, it might be fun to try.


Jun Okumura

Aren't Drs. Kotlikoff and Green speaking to a very different problem, where these words and terms are familiar to laymen (laypersons?) like me through the mass media and general political discourse, but are poorly defined and therefore devoid of operative meaning?

With "deficits", et al, I think I know what they're talking about, but I actually don't know. With "hetero", etc., I have no idea what they're talking about, but at least I know that I don't know.

(Seriously, anyone who hasn't buried his head in academic journals for the last two decades knows that that the word "hetereoskedasticity" first appeared in "Mary Poppins: the Hip-Hop Version (1994).)

Justin Ross

Dubner and Levitt should be thanking Economists for our "techno-speak," that is where they have provided their value-added in Freakonomics. Fun fact: heteroskedasticity can also be spelled with a "c" in stead of a "k".

saharvetes

You should change the blog's format to have the "Posted by…" line at the top, not at the end. Some posts can be quite disorienting otherwise ;-)

Don Robertson

Philosophers know, that from the initially wrong assumption that we can speak about something, or anything, and have others know when we speak, what we are talking about, arises all the problems in misunderstanding, confusion, oftentimes offense, and certainly the most effective persuasion.

This is the problem of univeral forms and ideas, which at one time was limited in scope but has been broadened to include everything and anything we can know.

Here's why: Virtually all of our sensory data, which is virtually everything we get from the thing in itself, a philosophic phrase describing reality outside our minds, must be converted into universal forms and ideas before our minds can grapple with that data.

The anology of trying to pour hot maple syrup into a laptop computer for analysis is not un-akin to trying to see the thinking mind any other way.

From this known conundrum, all collective knowledge progress comes from 1)re-defining universal forms and/or ideas, or, 2)inventing or discarding universal forms and/or ideas.

This however is wholly more complicated than one might assume because no two people have ever had the exact same notion of what any universal form or idea means, and in fact, it is easy to dissect any universal form or idea to the point of it being meaningless and illusory.

So, when I read a term like "heteroskedasticity", I am reminded of a story about the King of Siam, Thailand, who when in the eighteenth century was told by Englishmen about ice, decided all Englishmen were liars through and through.

In this sense, we are all liars. Nothing is believable unless we are enchanted enough by the teller of the fib to make the attempt to construct what he or she might mean by their utterance or scribe.

So then, these two economics authors have apparently unconsciously decided heteroskedasticity is an enchanting term, a term that will inspire others to try to construct for themselves in their own minds what they are talking about.

Philosophers know, it is of course impossible to figure out what anyone else is talking about, and the best anyone can do is to build it all up from scratch with the few hints we have from the last time we attempted to exercise the universal forms and ideas of that part of our brain.

The risk is of course, as I stated in my first paragraph, all the problems in misunderstanding, confusion, oftentimes offense, and certainly the most effective persuasion.

Don Robertson, The American Philosopher
Limestone, Maine

An Illustrated Philosophy Primer for Young Readers
Precious Life - Empirical Knowledge
The Grand Unifying Theory & The Theory of Time
http://www.geocities.com/donaldwrobertson/index.html
Art Auctions:
http://www.artbyus.com/auctions.php?a=6&b=4807

Read more...

zbicyclist

Oddly enough, the paper “Heteroskedasticity-Robust Standard Errors for Fixed Effects Panel Data Regression,” is actually relevant to me.

Perhaps due to your plug, when I tried to access the paper, I get the message "Server load is too high. Please try again later".

GamblingEconomist

I actually enjoy predicting who made the post while I'm reading it. At the beginning of the blog it was usually obivious but lately it has been a little harder to tell as Stephen has become more of an economist (and Steven has become more of a 'blogger').

FYI, Stephen, this is a technical paper that would probably interest Steven since he does a lot of panel data analysis. There are many papers who are saying something that the 'laity' but are filled with Math and econospeak that make them unreadable. This is not one of them.

T

http://pdos.csail.mit.edu/scigen/
Although this site is specifically for computer science papers, I think this "tool" has relevence on academic papers in general.

pyesetz

It is not clear that NBER actually intends to make any money from their $5 charge for this paper.  Anyone whose IP address is for a "developing country" gets free access.  The article summary includes a direct link to author Laurence Kotlikoff's homepage, which in turn includes a link to a free copy of the PDF.

While the paper's abstract seems to portend a linguistic approach, the paper immediately goes heavily into Greek letters with multiple italicized subscripts.

"If m>1, the economy's debt and deficit will head to infinity with no affect whatsoever on the economy or any agent in the economy."

Mack

Just FYI:
obtuse is not the word you're looking for.

Jun Okumura

"arcane"?

Ragout

The Stock-Watson paper sounds very important. Thanks for calling my attention to it.

I'll try to rewrite the abstract in English, to demonstrate how pointless this would be.

A certain "heteroskedasticity-robust" technique for calculating statistical significance has become widely used in the past decade. We show that this technique gives the wrong answer in certain common situations, and propose a technique that gives the right answer.

My point is that the less-technical version gives at best the same information to a non-statistician ("this article is not for me"). But to someone in the field, it gives a lot less information.

John Fembup

"not un-akin to"

Got it. Thanks . . .

Steve

The Stock and Watson paper was important enough that STATA and other statistics software used by researchers throughout the social and natural sciences has incorporated their results as a default setting...

http://www.stata.com/stata11/pdmixed.html

Jun Okumura

Aren't Drs. Kotlikoff and Green speaking to a very different problem, where these words and terms are familiar to laymen (laypersons?) like me through the mass media and general political discourse, but are poorly defined and therefore devoid of operative meaning?

With "deficits", et al, I think I know what they're talking about, but I actually don't know. With "hetero", etc., I have no idea what they're talking about, but at least I know that I don't know.

(Seriously, anyone who hasn't buried his head in academic journals for the last two decades knows that that the word "hetereoskedasticity" first appeared in "Mary Poppins: the Hip-Hop Version (1994).)

Justin Ross

Dubner and Levitt should be thanking Economists for our "techno-speak," that is where they have provided their value-added in Freakonomics. Fun fact: heteroskedasticity can also be spelled with a "c" in stead of a "k".

saharvetes

You should change the blog's format to have the "Posted by…" line at the top, not at the end. Some posts can be quite disorienting otherwise ;-)

Don Robertson

Philosophers know, that from the initially wrong assumption that we can speak about something, or anything, and have others know when we speak, what we are talking about, arises all the problems in misunderstanding, confusion, oftentimes offense, and certainly the most effective persuasion.

This is the problem of univeral forms and ideas, which at one time was limited in scope but has been broadened to include everything and anything we can know.

Here's why: Virtually all of our sensory data, which is virtually everything we get from the thing in itself, a philosophic phrase describing reality outside our minds, must be converted into universal forms and ideas before our minds can grapple with that data.

The anology of trying to pour hot maple syrup into a laptop computer for analysis is not un-akin to trying to see the thinking mind any other way.

From this known conundrum, all collective knowledge progress comes from 1)re-defining universal forms and/or ideas, or, 2)inventing or discarding universal forms and/or ideas.

This however is wholly more complicated than one might assume because no two people have ever had the exact same notion of what any universal form or idea means, and in fact, it is easy to dissect any universal form or idea to the point of it being meaningless and illusory.

So, when I read a term like "heteroskedasticity", I am reminded of a story about the King of Siam, Thailand, who when in the eighteenth century was told by Englishmen about ice, decided all Englishmen were liars through and through.

In this sense, we are all liars. Nothing is believable unless we are enchanted enough by the teller of the fib to make the attempt to construct what he or she might mean by their utterance or scribe.

So then, these two economics authors have apparently unconsciously decided heteroskedasticity is an enchanting term, a term that will inspire others to try to construct for themselves in their own minds what they are talking about.

Philosophers know, it is of course impossible to figure out what anyone else is talking about, and the best anyone can do is to build it all up from scratch with the few hints we have from the last time we attempted to exercise the universal forms and ideas of that part of our brain.

The risk is of course, as I stated in my first paragraph, all the problems in misunderstanding, confusion, oftentimes offense, and certainly the most effective persuasion.

Don Robertson, The American Philosopher
Limestone, Maine

An Illustrated Philosophy Primer for Young Readers
Precious Life - Empirical Knowledge
The Grand Unifying Theory & The Theory of Time
http://www.geocities.com/donaldwrobertson/index.html
Art Auctions:
http://www.artbyus.com/auctions.php?a=6&b=4807

Read more...

zbicyclist

Oddly enough, the paper "Heteroskedasticity-Robust Standard Errors for Fixed Effects Panel Data Regression," is actually relevant to me.

Perhaps due to your plug, when I tried to access the paper, I get the message "Server load is too high. Please try again later".

GamblingEconomist

I actually enjoy predicting who made the post while I'm reading it. At the beginning of the blog it was usually obivious but lately it has been a little harder to tell as Stephen has become more of an economist (and Steven has become more of a 'blogger').

FYI, Stephen, this is a technical paper that would probably interest Steven since he does a lot of panel data analysis. There are many papers who are saying something that the 'laity' but are filled with Math and econospeak that make them unreadable. This is not one of them.

T

http://pdos.csail.mit.edu/scigen/
Although this site is specifically for computer science papers, I think this "tool" has relevence on academic papers in general.