Kitchen sink econometrics
from Lars Syll
When I present this argument … one or more scholars say, “But shouldn’t I control for everything I can in my regressions? If not, aren’t my coefficients biased due to excluded variables?” This argument is not as persuasive as it may seem initially. First of all, if what you are doing is misspecified already, then adding or excluding other variables has no tendency to make things consistently better or worse … The excluded variable argument only works if you are sure your specification is precisely correct with all variables included. But no one can know that with more than a handful of explanatory variables.
Still more importantly, big, mushy linear regression and probit equations seem to need a great many control variables precisely because they are jamming together all sorts of observations that do not belong together. Countries, wars, racial categories, religious preferences, education levels, and other variables that change people’s coefficients are “controlled” with dummy variables that are completely inadequate to modeling their effects. The result is a long list of independent variables, a jumbled bag of nearly unrelated observations, and often a hopelessly bad specification with meaningless (but statistically significant with several asterisks!) results.
A preferable approach is to separate the observations into meaningful subsets—internally compatible statistical regimes … If this can’t be done, then statistical analysis can’t be done. A researcher claiming that nothing else but the big, messy regression is possible because, after all, some results have to be produced, is like a jury that says, “Well, the evidence was weak, but somebody had to be convicted.”
The empirical and theoretical evidence is clear. Predictions and forecasts are inherently difficult to make in a socio-economic domain where genuine uncertainty and unknown unknowns often rule the roost. The real processes that underly the time series that economists use to make their predictions and forecasts do not conform with the assumptions made in the applied statistical and econometric models. Much less is a fortiori predictable than standardly — and uncritically — assumed. The forecasting models fail to a large extent because the kind of uncertainty that faces humans and societies actually makes the models strictly seen inapplicable. The future is inherently unknowable — and using statistics, econometrics, decision theory or game theory, does not in the least overcome this ontological fact. The economic future is not something that we normally can predict in advance. Better then to accept that as a rule ‘we simply do not know.’
We could, of course, just assume that the world is ergodic and hence convince ourselves that we can predict the future by looking at the past. Unfortunately, economic systems do not display that property. So we simply have to accept that all our forecasts are fragile.
Leave a comment
—– look inside —– $5.94 / $20.00
—– look inside —– $4.90 / $8.00
—– look inside —– $15.99
—– look inside —– $5.99 / 12.99
—– look inside —– $5.93 / $12.99
—– look inside —– $4.97 / $9.90
—— Ugarteche, Puyana and Madi ——
Gerson Lima / Maria Alejandra Madi
Edward Fullbrook and Jamie Morgan
————— Michael Hudson ————–
Maria Alejandra Madi / Jack Reardon
————- Edward Fullbrook ————-
—————— Steve Keen —————–
————— Richard Smith —————
————– Gustavo Marques————
– Victor Beker and Beniamino Moro –
————– Lars Pålsson Syll ————-
—————– Stuart Birks —————-
Edward Fullbrook and Jamie Morgan
———— Armando Ochangco ———-
Shimshon Bichler / Jonathan Nitzan
————— Mauro Gallegati ————–
————— Herman Daly —————-
————— Asad Zaman —————
—————– C. T. Kurien —————
————— Robert Locke —————-
Real-World Economics Review Blog 
Perhaps if we got a cleaner (most likely to be cheap labour imported from a poorer economy) she could clean up the sink a bit?
Despite the problems, econometrics has some power to falsify propositions. A correlation may not prove anything but if a model or theory implies one should exist and econometrics fails to find it, using whatever control variables may be relevant, the theory is rejected for the domain or application in question. Unfortunately academic journals are less keen on such “negative” results than on ones that appear compatible with fashionable “theory”. The representative agent with rational expectations theorising that underpins DSGE has been shown to be empirically inapplicable and no one uses it for forecasting, .though it continues to be taught in the schools.
Practical forecasting in economics, like weather forecasting, relies on data collection and the extrapolation of patterns observed in the past. Forecasts are not derived from some general theory either of the economy or of the weather. Presumably such regularities as we observe are generated by individual inertia in the face of uncertainty and the law of large numbers. Of course such forecasts are fragile. That is not news and no forecaster disputes it.
Unfortunately economists do not falsify propositions. Empirical evidence appears to be irrelevant. The quantity calculus proves that all conventional analysis is invalid. Even when presented with analysis which accords with the empirical facts and the quantity calculus they fail to understand the significance of this. What hope for valid theory to be understood?
Enormously sweeping statements, Mr Salter. “Economists” are a diverse crowd. I don’t think we advance discussion by turning the whole trade into a uniform bogeyman. There isn’t a universally “valid theory” of the behaviour of the macroeconomy and it is doubtful if there could be. The search for “generality” has led to much empty theorising in the neo-classical tradition. Behaviour depends on institutions and memories, which are always evolving. Ad hoc theories that acknowledge a limited domain of application are more useful. As we build a library of such theories they provide insights as to how to understand a current situation through analogy with past situations. This is unsatisfactory to people who want a “theory of everything” but when studying a non-stationary system it may be the best we can do.
.
Of course Gerald from Frank’s perspective you just don’t understand the math:
.
.
A brief history of an exchange between Robert Locke and Frank Salter gives some historical context:
.
.
As Robert Locke points out, human “activities” like all human behavior take place in the mind of the human actor and are determinative and causative and override elegant but utterly useless logical axiomatic system of applied mathematics. Salter applies his deductivist thinking to economics in the same manner as neoclassical economists do, then becomes infatuated with the beauty of the mathematics and falsely believes that if his first principles axiomatic premises are true, the conclusions necessarily follow, when in fact, unfortunately for Salter, they don’t (Syll 2016). The snag is his model isn’t relevant to the real world, where human beings make mind-value-decisions based on diverse mind meanings and then change their minds, abandon the factory (despite its entire output from zero to infinity is predicted in hypersurface to nowhere generated by transient analysis) as a sunk cost and move it to Mexico or Vietnam. Or close it down due to a Trumpian trade war.
.
.
If it cannot be falsified it is not a scientific theory at all but merely a tautology that predicts the end from the beginning—a deterministic hypersurface no nowhere. It is not tested against reality, cannot be tested against reality (so Salter claims) and therefore is utterly useless in telling us anything about reality.
Don’t be fooled by his opening statement he doesn’t resort to the “you don’t know the math” argument, for he has done so on this forum multiple times (along with Shiozawa) as Lars and others know full well.
.
Consider the following. Salter argues that his hypersurface to nowhere — aka Buzz Lightyear Hyperbolics — predicts the “physical realization of decisions already made,” but clearly when plants are planned, built, and then scarped because of those pesky things in the mind of the observer that he has such a hard time modeling in his mathematics and wrapping his mind around.
.
.
My Response.
Scraped as sunk cost not “scarped” …
I graduated in chemical engineering.
With this idiotic tirade Rob proves he does not understand the mathematics at all.
… despite the multitude of words at no point has Rob ever been able to demonstrate that my analysis is invalidated by empirical evidence. But then he does not rely on evidence only his gut feelings. Of course he is unable to invalidate my analysis as it accords with the empirical facts.
I understand your reluctance to discard conventional quantitative analysis in the way I did but if you apply the strictures of the quantity calculus, there is NO valid quantitative analysis. All econometrics is merely curve fitting with what may as well be arbitrary equations. The path being taken can never lead to a valid theory. The quantity calculus is a potent tool to separate equation forms which may be valid from those which never can be.
Two passages from the posting and a few comments.
“Predictions and forecasts are inherently difficult to make in a socio-economic domain where genuine uncertainty and unknown unknowns often rule the roost. The real processes that underly the time series that economists use to make their predictions and forecasts do not conform with the assumptions made in the applied statistical and econometric models. Much less is a fortiori predictable than standardly — and uncritically —assumed.”
This is important since every aspect of human life is situated in the socio-cultural domain. That means all the sciences have their origins in this domain. As do the arts, law, religion, and so on. When physicists talk about string theory or any other physics esoterics, this originates in this domain. When mathematicians invent new forms of mathematics, this originates in this domain. Everything humans do, say, think, or believe originates in this domain. This is the ground of human existence. The only ground. In simple terms, humans create themselves and with themselves they create the context fro their lives.
“The forecasting models fail to a large extent because the kind of uncertainty that faces humans and societies actually makes the models strictly seen inapplicable. The future is
inherently unknowable — and using statistics, econometrics, decision theory or game theory,
does not in the least overcome this ontological fact. The economic future is not something that
we normally can predict in advance. Better then to accept that as a rule ‘we simply do not
know.’”
This situation arises out of first, just described. For humans, as Lars points out “the future is inherently unknowable.” Ontological is a big and scary word. Better to follow the observation
of Jean-Paul Sartre, “Man is condemned to be free; because once thrown into the world, he is
responsible for everything he does.” There’s nothing to hang on to for humans. All alone, making it up as they go.