Really? The Fed doesn’t make predictions either.
Hedge fund choices don’t conclusively prove anything, but it’s highly amusing to see the very adivsory firms the market selects take the opinions they do. Which brings us to the next question: Hot libertarians?
I like this picture better:
what auto is doing at his desk
Pretty accurate I’d say.
Really? The Fed doesn’t make predictions either.
Correct, it is an impossible action.
Read Aristophane’s signature when you get the chance.
Ah yes, the old “I know you are, but what am I?” trick. Nice try. Actually no, scratch that. You get a yawn instead.
Here we go again with the misused obscure references. As I said, I realize you’re trying to look hip and in-the-know when you do this kind of thing, but it actually does the opposite and just shines a light on your utter ignorance. If you want to come off knowledgable, it might help if you actually became knowledgable.
What? I actually took an economics class there. Don’t blame me, it was called “Economics of Development” and I had no idea!
Next you’ll be saying that your recent stay at a Holiday Inn qualifies you to knock down the strawmen you’ve set up.
+1 Clayton. Good post.
Once again you completely misunderstood what I said. The point was that you cannot know the exact future level of savings from past data. Do you really deny this? The irony is that your point about the identity still existing shows that you actually draw upon praxeology, which all economists must ultimately draw upon even if they disparage it.
Also, I was talking about physical constants.
Once again your entire argument in this thread is based on the post hoc ergo propter hoc fallacy. Do you believe that the ‘economists’ that use past prices in order to forecast future prices yet go bankrupt prove that empirical evidence has no place in economic theory? Didn’t think so. Your argument is simply ‘heads I win, tails you lose’: if the entrepreneurs do well, they prove that empirical evidence can predict the future, while if they do poorly, they simply didn’t understand the evidence correctly. Forecasting is not predicting. It is ultimately based on what we can call a ‘hunch’, and not on anything that we can call a science - even if the forecast turns out to be correct.
Damn… 70 posts, 1 thread, 12 hours, 5 participants.
This thread is on fire for this site…
Anyway, carry on.
@ Aristippus
+1, Great post. I suppose it will be said as long as it needs to be…
Mustang19, I think you should read this article that was recently recommended to me. It’s a bit long, I know, but here’s the enlightening passages.
http://mises.org/daily/5158/Mises-on-Mind-and-Method
Experience is necessarily of past events. It can be resorted to for the prediction of future events only with the aid of the assumption that an invariable uniformity prevails in the concatenation and succession of natural phenomena.
[…]
The “category of causality” is our inbuilt conception that past regularity predicts future regularity. That conception is a prerequisite to all reasoning regarding cause and effect in the natural world. As Mises wrote, agreeing with Hume, there is no deductive proof that past regularity predicts future regularity.
[…]
First let us reflect on the methodology of geometry. In preclassical times, geometry was largely empirical. For example, ancient Egyptian surveyors called “rope stretchers” used to find right angles by stretching out knotted ropes to make a 3-4-5 triangle (a triangle with sides in the ratio of 3 to 4 to 5). The geometric principle that all 3-4-5 triangles form right angles might have been discovered by the Egyptians through experience, by simply noticing that such triangles generally have perpendicular legs by eyeballing multiple instances of them.
However, the ancient Greeks pioneered using long chains of deductive inferences to discover geometric principles (fortunately, there was no ancient Greek Congressman Clay at the time to deride this deductive scientific enterprise). And so the Greeks could discover the same principle (that all 3-4-5 triangles form right angles) by reasoning from the Pythagorean theorem. And the Greeks were able to use deductive geometry to discover a great many geometric principles that would have been too subtle to find with empirical methods.[32]
Nowadays the deductive method for geometry is universally embraced. A good geometry teacher will not demonstrate the Pythagorean theorem with measuring tape and a pile of plastic right triangles, except perhaps as a preliminary exercise. She will teach her students to derive the theorem from prior premises.
What if, when measuring plastic triangles, one of the geometry teacher’s students discovers something that does not seem to fit the Pythagorean theorem as it was taught to him by the teacher? There are a few possible explanations for such an occurrence.
Perhaps the teacher, bizarrely, doesn’t know the correct Pythagorean theorem herself, or how to derive it. An economic analogy to this situation is when a theoretical economist introduces vicious formulations (as the classical economists did with their value theory) and invalid reasoning (as Keynes did with much of his work) to build a false economic theorem.
Perhaps the student measured the right triangles incorrectly. This would be analogous to an economic statistician gathering faulty data.
Perhaps the student is not even dealing with right triangles in the first place, and so the Pythagorean theorem is not applicable. This would be analogous to an economist trying to explain a rise in a good’s price using the quantity theory of money when the quantity of money didn’t actually change, and instead the price rose because of an increase in demand.
If economic data do not seem to demonstrate the playing out of a certain market process described by economic theory (assuming the theory is sound and the data are correct), that would indicate that circumstances must have been dominated by another market process (also described by pure economic theory), another set of factors, or the interplay of several market processes/sets of factors.
The economic historian uses data to determine which economic laws are most relevant in any given episode. If, for example, the economic historian discovers trustworthy data that show that after an increase in the supply of a certain good the price for that good increased, instead of falling, that would not testify against the law of supply. That would instead be an indication that other relevant factors are at work, like perhaps a precipitous drop in the supply of another good for which the first good can serve as a substitute.
Whatever the case ultimately proves to be, what should the geometry student do when faced with data that do not harmonize with the theory he is using? Should he proudly announce that he has refuted the “orthodoxy of Pythagoras” and publish a new theorem, based on his measurements? Should he gather a larger sample set of plastic triangles and perform statistical analysis on the data gathered?
Of course not. He should check the soundness of the reasoning used in deriving the theorem, check his measurements, and check the applicability of the theorem to the data.
Insofar as the student derived the Pythagorean theorem using discursive reasoning, he was a geometer. But insofar as he was using geometry along with measurement to study plastic triangles, he was a topographer, not a geometer. While geometry is indispensable for topography, it is completely invalid to try to derive geometric laws from topography.
This is analogous to the crucial distinction between economics and economic history. Insofar as a scholar derives economic theorems using discursive reasoning, he is an economist. But insofar as he uses economics along with data-gathering to study actual events, he is an economic historian, not an economist. While economics is indispensable for economic history, it is completely invalid to try to derive economic laws from history.
Engineer / Data Analyst here
Empirical evidence is composed of factual data collected through controlled experiments. Observational evidence is composed of data collected through uncontrolled experiments. In the field of economics, very little evidence is empirical. The field is dominated by observation evidence. The methods employed in generating knowledge in the field of economics are deduction or induction.
The method of induction by its nature requires evidence. This evidence in economics is strictly observational. Generally observational evidence suffers from multicollinearity. Data which you collect for your specific inquiry may be influenced by events/processes/dynamics not related to your inquiry. Any conclusions that you may draw will be highly uncertain. Further complications arise when judging the predictive power of models produced from data collected from uncontrolled experiments. The data used to compare predictions against is collected from uncontrolled experiments. Uncertainty regarding conclusions increases.
Now quantitative models in economics are complete rubbish. They are statistical models produced by data collected by uncontrolled experiments. Ergo, whatever dynamic the model is supposedly trying to predict, it is not. It is predicting the data with high multicollinearity. Causality is simply ignored in time series and stochastic models. Over-determination is rife. Most papers do not even attempt to engage in validation processes because, more often than not, measures of accuracy will display that model is less accurate than the measures of accuracy involved in the construction process. Quantitative models in economics amount to nothing more than voodoo intertwined with mathematical symbols.
The deductive method of inquiry can produce qualitative models which the predictive power of the model can be judged. An example would be the principle of supply and demand in determining being indirectly validated by interventions in the economy such as rapid expansion of the money supply. It is observed that the prices of goods and services increase.
Solid post!
I know what you mean. I checked the forum at about 9 this morning and see 7 responses. I checked again and there were two pages. We may have to make a thread to talk about this thread.
Just as a bit of support proxyamenra’s excellent post, I thought I might link here, to where a most recent Nobel Prize winning economist just got through admitting how “[economic models] are abstractions that make very simple assumptions that aren’t realistic.”
+1 proxyamenra… you speak my language!
You claimed there are no constant quantities in economics. I noted that the definition of a “constant” is arbitrary and any such number could be input in place of a variable in order to apply a theory. Need a theory for the case when “savings is $1”? Set savings constant to 1. This is very semantic.
This is all true. However, apart from your use of praxeology (which mainstream economists do not necessarily draw on), the same could be said for any number of predictive models in biology, physics, or just about any other science. Predictive models are used often in other scientific fields to forecast (not “know”) future values of variables.
Not all values in physics, etc, are derived from physical constants. If you were to split off the parts of engineering, like strength of materials, which rely on empirical data, would that suddenly invalidate them? No.
Forecasts are not predictions! There we go.
I agree it doesn’t prove anything. It’s just an amusing phenomenon to see capitalist market agents have their own opinions.
And yet, banks still use them a great deal in economic forecasts. That might just make them worth something. Microeconomic evidence in particular does allow for controlled behavioral experiments.
You hear social sciences declaring “laws” from time to time, but these aren’t really meant to be so. An empirical/observational economist is an “historical economist” in the same way that many biologists are “historical biologists” or the friction coefficients were determined by a “historical physicist”. It’s just a matter of different scales.
There’s nothing to “admit”. He acknowledges that models aren’t perfect- big whoop. He then goes on to note that models are potentially very powerful. And that doesn’t do much to answer why models aren’t useful in economics alone among other sciences.