Mitchell’s laws: The more budgets are cut and taxes inceased, the weaker an economy becomes. To survive long term, a monetarily non-sovereign government must have a positive balance of payments. Austerity = poverty and leads to civil disorder. Those, who do not understand the differences between Monetary Sovereignty and monetary non-sovereignty, do not understand economics.
==========================================================================================================================================
I predict that within 20 years, the science of economics, and the economists who ply that trade, will be gone, and the world will be much better for it.
Wikipedia and Websters New Collegiate Dictionary both define “economics” as: “The social science that analyzes the production, distribution, and consumption of goods and services.” The key word is “analyzes.”
How do economists “analyze” production, distribution and consumption? They begin by assembling information. They ask, “What and how many?” “Who?” “Where?” “When and how often?”
From these data, they come to conclusions (“Why?”), and from these conclusions, they make predictions and recommendations for actions. It is the conclusions and recommendations that give economics value. Without them, economics would be useless.
In a previous post, How IBM can change the world, I described how the IBM super computer named “Watson” played Jeopardy and beat the best, two, human Jeopardy players in history.
Those who know Jeopardy understand what an amazing accomplishment this was. The game is question and answer, with the twist that the answer is given, and the contestant must come up with the question. That twist essentially is meaningless; the contestant merely precedes his answer with the words “Who is” or “What is” and voila, an answer becomes the required question.
The difficulty lies in the nature of the questions. Not only do they ask for a broad knowledge of obscure trivia, but the wording of the questions can be ambiguous and non-specific, involving puns, analogies, metaphors and axioms.
Even a large group of programmers could not possibly input Watson with the infinite question and answer variations, so they used machine learning — a method by which a computer repeatedly answers questions, then is given the correct answers. Over time, the computer detects patterns that allow it to answer future questions more accurately.
The computers review data and the proven-wrong or proven-right answers, and doing this often enough, allows them to calculate degrees of “correctness” — the odds of various answers. Computers, of course, can handle massive data tirelessly. Humans cannot.
Every hypothesis and theory in economics evolves exactly the same way. An economist looks at data and determines its meaning. He does that by seeing repetitions and calculating correlations. At its essence, economics is mathematics.
As an economist who fancies himself somewhat creative, I believe there is no creativity or genius in economics. What passes for creativity and genius is just discovery. The economist compares two sets of data and discovers they seem to correlate — or not. So he adds more data, and soon he comes to the conclusion, or rather, discovers, that one factor seems to precede the other — most of the time.
Then he adds other data to see how they affect the results. Economics is a trial and error game.
Upon seeing what Watson can do, and understanding the nature of economics, I have come to this conclusion: There is not one thing human economists do that a computer cannot do faster, more accurately and without the hubris that affects human economists’ judgement.
Actually, what Watson did with Jeopardy is much more difficult than what a Watson clone could do with economics. Jeopardy involved interpretation of the nuances in the English language. Economics is much more straightforward — made for a computer.
Watson was primed with information, statistical and nonstatistical. An economics Watson, let’s call it “Econoputer,” would receive data from: Encyclopedias, the complete text of every book ever written, census tables, phone books, statistical abstracts, voting data, Macroeconomic and Regional Data, public finance, all federal laws, transportation, health, education, economic history, weather history, agricultural, astronomy, tax tables, physics, chemistry, crime and punishment — trillions of words and pieces of data that continuously are collected from existing sources.
Then our “Econoputer” would begin to correlate its information, trillions upon trillions of calculations, to establish probability tables, showing the likelihoods of cause and effect. No problem for Watson’s future “children.”
The President of the United states might ask a clerk in the Treasury Department, “What will happen to the economy, if I cut FICA to 5%?” Or he might go further: “What level of FICA will yield the greatest GDP and employment growth?
Econoputer would present its results: “Given all historical data, and the current situation, cutting FICA to 1.3% has a 92% probability of raising GDP 4.6% and a 94% probability of increasing employment 5.1%”
Or the President might go even broader: He might ask, “What can I do with all taxes for maximum economic growth?” Econoputer already has correlated all past tax rates and tax laws with all past tax collections, and correlated that with every other economic factor affecting GDP growth. It might tell the President to cut certain tax rates and increase others, while increasing the IRS staff by 7%.
This is what human economists attempt to do now, but no economist can do it. We are forced by human limitations, to use much less data, which is corrupted by personal biases. Can you imagine a human economist trying to answer the question: What will happen to farm income in Idaho, and total U.S. GDP, if we bring 25,000 soldiers back from Afghanistan?”
Wild-ass guess is what you’d get from a human economist. Econoputer would provide a far more precise calculation. It would include civilian clothing sales, new births and deaths, baby food sales, imports and exports, hospital visits, vacation travel — millions of changes no human could take into account.
Pierre Simon Laplace said:
We may regard the present state of the universe as the effect of its past and the cause of its future. An intellect which at a certain moment would know all forces that set nature in motion, and all positions of all items of which nature is composed, if this intellect were also vast enough to submit these data to analysis, it would embrace in a single formula the movements of the greatest bodies of the universe and those of the tiniest atom; for such an intellect nothing would be uncertain and the future just like the past would be present before its eyes.
The quantum mechanics’ Heisenberg Uncertainty Principle proved Laplace wrong at the atomic level, but on a macroeconomic level, the more you know about the past and present, the more you can know about the future.
Because of Watson and its future progeny, economics will be reduced to information collection, and economists will be clerks. I’m an economist, but I hardly can wait for this to happen. It will mark the end of uninformed argument and decision-making, and humans will benefit from a better world.
Rodger Malcolm Mitchell
http://www.rodgermitchell.com
==========================================================================================================================================
No nation can tax itself into prosperity, nor grow without money growth. Monetary Sovereignty: Cutting federal deficits to grow the economy is like applying leeches to cure anemia. Two key equations in economics:
Federal Deficits – Net Imports = Net Private Savings
Gross Domestic Product = Federal Spending + Private Investment and Consumption + Net exports
#MONETARY SOVEREIGNTY
#Monetary Sovereignty - Mitchell 
Very valid observations.
But to be truly useful the programming must be truly empirical and not have any “modeling assumptions.” The problem with economics today is that many economists (and most of the very influential) do all their analysis starting with neoclassicla assumptions which have been little changed in more that a century.
Whe their models give a result that proves contrary to observations they seldom do the obvious for most scientists and test their assumptions. Instead they add another level of tweeking to make the old model adjust to the new history. This goes on ad nauseum.
Watson should be assumption independent and make everything for all relevant history fit an empirically derived performance function, free of any assumptions.
When new data contradicts the old empirical function, Watson should produce a new function.
This process should be implemented provided real scientists (those of the physical and mathematical science fields) dominate the developement of the algorithms rather than having them the sole purview of economists.
Make the system one of “artificial intelligence” rather than one that is “intelligent artificially.”
Another way of saying it: there is a difference between “thinking robots” and “robotic thinking.”
John Lounsbury
LikeLike
Good observations, John.
You’ve described the hubris I mentioned. It’s possibly the single greatest impediment in economics. Thinking tends to stop with the development of a hypothesis. A computer has no pride.
LikeLike
They’re trying to model the weather with computers, and only having partial success, because one little unanticipated blip throws the whole model off. “Chaos Theory.”
Economics should be less messy, but still, you have the psychological aspect.
And then there’s the values questions.
When I worked in the semiconductor industry, we used a design-of-experiments method to model multiple variables. I’d like to see that applied more in economics, instead of relying on the 2 variable studies that are commonly used. For example, if you try to establish a correlation between the top tax rate and GDP growth, as several people have tried to do recently, you may be misled, because there are so many other variables — private savings, trade balance, deficit spending, state and local taxes, etc..
LikeLike
@Dan Lynch – – –
There are a few economists who have done / are doing multivariant analysis, such as Steve Keen, Richard Goodwin, Marc Lavoie and Wynne Godley, We posted a good review of that area: http://econintersect.com/wordpress/?p=9318
As Steve Keen has told me personally, the surface has just been scratched. I will ask him about the Watson learning model approach the next time we make contact.
John Lounsbury
LikeLike
Dan,
All the useless arguments, about hypotheses, theories and laws, would disappear. A super computer has no beliefs. There would be no “Watsonians.”
To my knowledge, multivariant analysis has been limited to a handful of variables. A Watsonesque approach would handle billions.
One disadvantage: We would not know exact cause and effect. We would not be able to point at one or two variables and say, “This will grow the economy.”
We merely will know what groups of variables are most likely to have that effect.
LikeLike
Dan, you’re right.
Meteorology is an apt comparison. And think of what meteorology has become. More and more it’s a computer science, with less and less human analysis. Can you imagine trying to predict the weather based on a couple of variables, like an economist does?
The partial success actually is pretty good success. Today, a week’s prediction is about as good as a two-day prediction of 20 years ago. And of course, meteorology is on a different time scale — hour by hour — as opposed to economics’ month-by-month.
Interestingly, I gave a speech titled, “The Meteorology of Economics.” That was seven years ago. It’s reprinted in this blog.
Rodger Malcolm Mitchell
LikeLike
That’s pretty hilarious, to think that the Very Serious People will allow anything resembling empiricism or objectivity to influence the distribution of resources. Economics aren’t like weaksauce human concerns such as medicine or agriculture where science has had pretty much the last word. People will be sniveling to the gates of hell that their intuition and prejudices are better than a machine.
LikeLike