Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

A First Lesson in Econometrics (1970) [pdf]

uibk.ac.at

Post reply on HN
Print view View original

31–40 of 64 posts

Re: A First Lesson in Econometrics (1970) [pdf]

#31

Post by xapata »

boppo1 wrote: post #25
bachmeier wrote: post #18

Earlier quoted context omitted.

My monetary economics professor in grad school was teaching a paper and told us that when the authors claim it's obvious, that means it's not obvious. So he wrote out the derivation over the weekend and gave us a four-page, single-spaced handout with all the equations behind that single "obvious" result.

This is way off topic but hopefully it will get allowed because I think you have the expertise to help: It seems to me that the widely accepted practice of market stimulation by interest rate intervention has the cost of destroying price discovery. Also, that it is a primary cause of wealth inequality. These relationships seem to me actually obvious: push down DCF denominators and valuations go up, inefficient busine…

> It seems to me that the widely accepted practice of market stimulation by interest rate intervention has the cost of destroying price discovery. Also, that it is a primary cause of wealth inequality.

Those claims deserve an explanation and empirical evidence.

Measuring price discovery itself is a bit awkward. You could say that poor quality price discovery would result in more price volatility, and that increased costs of price discovery would result in lower liquidity. Unfortunately, both of those things have many other causal factors. How would you untangle the causes to isolate the effects of interest rate intervention?

Making the leap to saying it's the primary cause of wealth inequality is absurd. So long as society uses a market economy, or allows any form of individual wealth aggregation, wealth is likely to follow a log-normal distribution.

Re: A First Lesson in Econometrics (1970) [pdf]

#32

Post by chermi »

photochemsyn wrote: post #11

Econometrics is one of the most discredited branches of academics that exists. These were the people who were waving their models around in the early 1990s, trying to tell people that NAFTA would have no effect on manufacturing jobs in the United States. They also promoted elimination of Glass-Steagall and the deregulation of the housing market, resulting in the subprime fraud-based economic disaster of 2008. There's…

I suspect you may have a point hidden within here, a point that I might even agree with...

But it's hard to understand that point and believe that you are discussing in good faith when you conflate things that you must know not to, given your implied knowledge. Econometrics (methodology) and neoclassical (school of thought) are on different "axes" if you will, if you were to try to decompose economics into into orthogonal components. I know that's a simplification

Could you perhaps critique econometrics on its own?

Edit 1 because I can't type

Edit 2 to add this comment -- aww_dangs comment like 2 below here and 2 deeper is the closest to the critique this guy wanted to give

Re: A First Lesson in Econometrics (1970) [pdf]

#33

Post by huitzitziltzin »

Kranar wrote: post #30
huitzitziltzin wrote: post #27

Earlier quoted context omitted.

> 'Econometric models' have so many levers and dials on them that they be tuned to forecast whatever their operator wants them to forecast. I am going to guess you have never written down an economic model, much less presented one at an economics seminar. If your model can be "tuned to forecast whatever", you are going to be annihilated. You will never be able to get away with presenting a model like that. There is n…

To be fair, the purpose of referencing the Google Scholar citation was not to support the claim that econometricians predicted no decline in manufacturing, it was to refute your claim that: "Econometricians are unlikely to have made predictions about NAFTA." I don't know why you would have made such a claim since it should be fairly obvious that econometricians made numerous predictions of all sorts. I will respect y…

I didn’t address that point because it would seem extremely pedantic.

To an economist, “an econometrician” is almost always someone who spends their time developing new estimators and proving things about their properties.

The people who wrote the CBO report are very unlikely to have been econometricians according to nearly any economist. The CBO mostly hires applied people of various kinds, whether in micro or macro.

They are users of econometrics, as we all are, but to say they are “econometricians” as economists use that term would be wrong.

But that seemed like an extremely pedantic objection to the OPs point.

Re: A First Lesson in Econometrics (1970) [pdf]

#34

Post by aww_dang »

SantalBlush wrote: post #21
lvl100 wrote: post #17

Earlier quoted context omitted.

I think you’re mainly focusing on macro and it’s not completely unwarranted. That said, micro stuff is fascinating to say the least, and very pertinent to data science. In fact, much of “data science” projects until 2012-13ish were handled by economists.

They're not just focusing on macro, they are confusing econometrics with economics. Econometrics has nothing to do with any specific economic theory; it is about statistical modelling.

There are theories which dispute the underlying premises of econometrics. Economics is sometimes referred to as "the dismal science". Experiments cannot be constructed in the same way as the natural sciences. Dressing economics up in the pretenses of the natural sciences has arguably led us down a path of shenanigans. Consider the Federal reserve's machinations, deliberately opaque language and absurd narratives such as "transient inflation" or "symmetric inflation".

>The main characteristic or nature of human beings is that they are rational animals. They use their minds to sustain their lives and well-being. The usage of the mind, however, is not set to follow some kind of automatic procedure, but rather every individual employs his mind in accordance with his own circumstances. This makes it impossible to capture human nature by means of mathematical formulae, as is done in the natural sciences.

>In short, people have the freedom of choice to change their minds and pursue actions that are contrary to what was observed in the past. As a result of the unique nature of human beings, analyses in economics can only be qualitative.

>Furthermore, to pursue quantitative analysis implies the possibility of the assignment of numbers, which can be subjected to all of the operations of arithmetic. To accomplish this, it is necessary to define an objective fixed unit. Such an objective unit, however, doesn’t exist in the realm of human valuations. On this Mises wrote, "There are, in the field of economics, no constant relations, and consequently no measurement is possible."

Re: A First Lesson in Econometrics (1970) [pdf]

#35

Post by aww_dang »

barrenko wrote: post #13
photochemsyn wrote: post #11

Econometrics is one of the most discredited branches of academics that exists. These were the people who were waving their models around in the early 1990s, trying to tell people that NAFTA would have no effect on manufacturing jobs in the United States. They also promoted elimination of Glass-Steagall and the deregulation of the housing market, resulting in the subprime fraud-based economic disaster of 2008. There's…

Not sure why you're downvoted.

This is a politicized issue.

Re: A First Lesson in Econometrics (1970) [pdf]

#36

Post by smallnamespace »

boppo1 wrote: post #25
bachmeier wrote: post #18

Earlier quoted context omitted.

My monetary economics professor in grad school was teaching a paper and told us that when the authors claim it's obvious, that means it's not obvious. So he wrote out the derivation over the weekend and gave us a four-page, single-spaced handout with all the equations behind that single "obvious" result.

This is way off topic but hopefully it will get allowed because I think you have the expertise to help: It seems to me that the widely accepted practice of market stimulation by interest rate intervention has the cost of destroying price discovery. Also, that it is a primary cause of wealth inequality. These relationships seem to me actually obvious: push down DCF denominators and valuations go up, inefficient busine…

You’re not crazy, if you want to find lots of discussion about this topic head to https://alhambrapartners.com/tag/eurodollar-university/

But to be fair to the Fed (not to excuse plenty of incompetence), the banks in 08 became close to insolvency. The ‘proper’ way to inject money into the economy would’ve required Congress to authorize a recapitalization of banks - which would’ve been political suicide.

Re: A First Lesson in Econometrics (1970) [pdf]

#37

Post by iaw »

lowkeyokay wrote: post #5

This was made me laugh. So many times I read ‘it should then be obvious that…’ What the?! No it isn’t. Text book authors must hate students.

I have a friend who was in grad school and found an "...and so it follows" proof that was in standard textbooks dating back decades. To this day he can't find anyone to fill in the remainder of the proof to get the result.

Re: A First Lesson in Econometrics (1970) [pdf]

#38

Post by smallnamespace »

xapata wrote: post #31
boppo1 wrote: post #25

Earlier quoted context omitted.

This is way off topic but hopefully it will get allowed because I think you have the expertise to help: It seems to me that the widely accepted practice of market stimulation by interest rate intervention has the cost of destroying price discovery. Also, that it is a primary cause of wealth inequality. These relationships seem to me actually obvious: push down DCF denominators and valuations go up, inefficient busine…

> It seems to me that the widely accepted practice of market stimulation by interest rate intervention has the cost of destroying price discovery. Also, that it is a primary cause of wealth inequality. Those claims deserve an explanation and empirical evidence. Measuring price discovery itself is a bit awkward. You could say that poor quality price discovery would result in more price volatility, and that increased c…

The proper operation of monetary policy (which you can verify by cracking open any macroeconomics 101 textbook) calls for the central bank to control interest rates via the creation and destruction of money, such that the interest rate in the market for money (via fixed income instruments) always remains at an equilibrium rate.

The Fed failed to effectively create money in ‘08 and afterwards, and instead settled for more or less directly controlling the price of fixed income instruments. This is bypassing the textbook cause and effect mechanism and is indeed the equivalent to a price control in the fixed income market.

You can see negative effects by realizing that price controls always widen the gap between the marginal borrower vs. non-borrower — in a ‘true low rates’ scenario, loans would’ve been freely available at the low rate, but in post-Recession US, loan rates were low, but only strong borrowers could get liquidity. That directly widens the gap between the haves and have-nots, and was a direct result of Fed policy.

Re: A First Lesson in Econometrics (1970) [pdf]

#39

Post by xapata »

smallnamespace wrote: post #38
xapata wrote: post #31

Earlier quoted context omitted.

> It seems to me that the widely accepted practice of market stimulation by interest rate intervention has the cost of destroying price discovery. Also, that it is a primary cause of wealth inequality. Those claims deserve an explanation and empirical evidence. Measuring price discovery itself is a bit awkward. You could say that poor quality price discovery would result in more price volatility, and that increased c…

The proper operation of monetary policy (which you can verify by cracking open any macroeconomics 101 textbook) calls for the central bank to control interest rates via the creation and destruction of money , such that the interest rate in the market for money (via fixed income instruments) always remains at an equilibrium rate. The Fed failed to effectively create money in ‘08 and afterwards, and instead settled for…

Asset inflation is exacerbating inequality, but to say it's the "primary cause" isn't supported.

Re: A First Lesson in Econometrics (1970) [pdf]

#40

Post by yourqiwi »

huitzitziltzin wrote: post #27
photochemsyn wrote: post #23

Earlier quoted context omitted.

'Econometric models' have so many levers and dials on them that they be tuned to forecast whatever their operator wants them to forecast. Hence they mainly serve as a marketing tool to defend policies that their sponsors desire to get through Congress or the next board meeting or whatever. As far as things like model projections for NAFTA, any relatively simple search turns up dozens of references. Google Scholar isn…

> 'Econometric models' have so many levers and dials on them that they be tuned to forecast whatever their operator wants them to forecast. I am going to guess you have never written down an economic model, much less presented one at an economics seminar. If your model can be "tuned to forecast whatever", you are going to be annihilated. You will never be able to get away with presenting a model like that. There is n…

> I am going to guess you have never written down an economic model, much less presented one at an economics seminar. If your model can be "tuned to forecast whatever", you are going to be annihilated. You will never be able to get away with presenting a model like that. There is no place for it in the field.

A poorly specified model will get ripped apart when presented in an academic context. I've seen it happen in seminars. Questions about the standard errors (robust? clustered?), omitted variable bias, whether things have been detrended, fixed effects, quality of the instrumental variable, etc. When econometrics is done properly and by the book, it can be a useful tool for analyzing observational data.

But these tools get regularly applied outside of academia and the results are presented to people who can't critically evaluate them. Nobody knows that you ran a million other regressions in STATA, and kept the significant one. Or that your result isn't robust when you excluded/included such-and-such variables. When research is motivated in a particular direction (as it often is in politics and consulting), then you can often pull the levers and turn the dials until you get the result you want.

Post reply on HN