by ARGeezer
Wed Sep 23rd, 2009 at 01:18:09 AM EST
I infamously, if unintentionally, hijacked a Chris Cook diary back in April and turned it into a diary on economic modeling. Mig subsequently turned the hijacked portion into a separate diary. Then I turned a comment by Marco on Phillips' Moniac Machine into a diary and there have been several previous modeling diaries.
There has been lots of discussion about the uses and benefits of a computer simulation of the economy that could run on a P.C. Meanwhile, Steve Keen has actually created a dynamic model for an economy and has used it to critique Krugman and the NCE mainstream for remaining tied to equalibrium analysis and to critique Obama's approach to the GFC and the TBTF banks:
I've recently developed a genuinely monetary, credit-driven model of the economy, and one of its first insights is that Obama has been sold a pup on the right way to stimulate the economy: he would have got far more bang for his buck by giving the stimulus to the debtors rather than the creditors.
The following figure shows three simulations of this model in which a change in the willingness of lenders to lend and borrowers to borrow causes a "credit crunch" in year 25. In year 26, the government injects $100 billion into the economy--which at that stage has output of about $1,000 billion, so it's a pretty huge injection, in two different ways: it injects $100 billion into bank reserves, or it puts $100 billion into the bank accounts of firms, who are the debtors in this model.
The model shows that you get far more "bang for your buck" by giving the money to firms, rather than banks. Unemployment falls in both case below the level that would have applied in the absence of the stimulus, but the reduction in unemployment is far greater when the firms get the stimulus, not the banks: unemployment peaks at over 18 percent without the stimulus, just over 13 percent with the stimulus going to the banks, but under 11 percent with the stimulus being given to the firms.
The time path of the recession is also greatly altered. The recession is shorter with the stimulus, but there's actually a mini-boom in the middle of it with the firm-directed stimulus, versus a simply lower peak to unemployment with the bank-directed stimulus.
Steve has previously discussed his plans to use computer based dynamic models and noted that he had such models working. I believe he said they were developed in Math Cad, but I can't find the reference.
In a later article Steve Keen comments on Krugman's recent column critiquing mainstream economics for falling in love with mathematical beauty. Keen notes that economics has isolated itself from the mainstream of scientific development by eschewing the use of differential equations, which are essential to analyzing dynamic systems. Krugman had suggested that meteorology might be a better anology to economics than Newtonian physics. Keen responds:
Well Paul, to understand meteorology, you're going to have to give up on equilibrium-or rather on the fantasy that a complex system can be modelled as if it is in equilibrium. There are, for example, 3 equilibria in the Lorenz Attractor-all three of which are unstable. Equilibrium is a state in which the Attractor will never be.
I also find it remarkable that intelligent people like Krugman can be so ignorant of developments outside their own field of endeavour. Notice the comment he made that not applying the standard neoclassical economics simplifications like equilibrium "will, inevitably, lead to a much messier, less pretty view".
Prettiness itself should not be an objective of science. But nonetheless, some of the prettiest objects in science are the result of non-equilibrium dynamics. The Lorenz Attractor is clearly one-take a look at it and you'll see why people often talk of "the Butterfly Effect" when describing the instability of the weather. But there are many others.
I think my own non-equilibrium models of the economy have a similar beauty-here for example are two from my model of Minsky's Financial Instability Hypothesis. The first is of an economy without a government sector which undergoes a debt-driven Depression:
The fate is not pretty-the economy collapses as debt rises-but the image is "pretty".
Both the image and the fate of the next model are pretty-this is an economy starting from the same initial conditions, but which has a government sector that practices counter-cyclical fiscal policy and therefore prevents a debt-induced crisis:
The above is the time path of employment and wages in this model, which is actually a 2D slice through the 4-dimensions of the model: (1) the wages share of output; (2) the employment rate; (3) the debt to output ratio; and (4) government spending as a percentage of output). This next view shows 3 of those dimensions (excluding government spending), and the 2 dimensional view of the previous simulation is shown as a shadow below the 3D shape):
So the output of non-equilibrium models can be "pretty"-it's just the picture they craft of capitalism that isn't pretty. It isn't a system that automatically reaches equilibrium and ensures the best outcome for the largest number of people. It may not be "pretty" in that way, but it is the world in which we live.
