Engine of what?

Mainstream economics fails to help us with good policy decisions on our critical problems because, as a discipline, it does not practice anything close to the radical skepticism that is necessary for real science, social or not. Microfoundations, which is about as scientifically rigorous as eugenics or astrology, is a good example.

By microfounded, I mean that every macroeconomic relationship has to be formally derived from optimisation by individual agents –Simon Wren-Lewis

Microfoundations is premised on the theory that things like unemployment, wage increases or decreases, gross national product, and whether manufacturing expands or contracts are completely determined by the market interactions of individuals or families or individual firms all looking out for themselves (optimizing). This is a theory that relies on there being no society – no economic effects of racism, or advertising, or patriotism, or religion or fashions or culture or geopolitics or technology.  The utility and desirability of analyzing the economy via microfoundations is widely assumed in the Economics literature despite contrary evidence and a complete failure to even attempt to ground it in scientific research. And despite all the references to  “rigorous microfoundations” one sees constantly referenced in the Economics literature, the work itself is  appallingly shoddy- a mishmash of mathematical-ish nonsense that attempts to disguise a morally repellent political program as science.

Here’s an example from Nobel Prize winner, Professor Robert Lucas.

Start with that “subjective rate of discount”. If you suppose utility grows 5%/year then the value of this formula at net discount 5% is 50% less than the value when net discount is 3%. Which is right? Everything depends on a subjective rate, pulled out of the air or somewhere else. The formula is apparently adapted from finance where something similar is used to get an intuition about the value of investments that pay over time. Given a stream of payments p₁,p₂ …. p_n over time and a safe interest alternative rate r we can estimate a present value P as the sum of the payments discounted by the safe interest level. Pretending that we have continuous compounding allows us to approximate with the exponential function. The result is an approximation  of how much we’d have to be paid now to end up with an equivalent to  the sum of all the payments over time. If treasury bonds over 10 years pay 4% then

P= sum[t=1…10] e^{-(0.04 * t)} * p_t 

If payments are continuous or close enough, the summation gets replaced by integration. This formula can only give an approximation of the present value for a number of reasons such as the risk of future payments, the effects of taxes, and the need for money at the moment.

No sane investor would automatically accept or reject a deal based on simple present value calculation – it is, at most, a data point.  Lucas however, claims this computation provides a rigorous measure of success for the economy. He  proposes to treat “utility” given by the function U as if it were cash and to discount it by the rate ρ – λ where λ is the rate of population growth and ρ is “the subjective discount rate”. Why the discount rate should be the difference between a subjective rate and an estimate of population increase is never explained. And we certainly don’t get an explanation of why utility which is another woozy function, a monotonically increasing function of consumption and leisure, is like money. What is the counterpart of safe interest, continuously compounding, that one could earn on some hours of sitting in the backyard, consuming beers now rather than 3 years from now?  What is represented by the time periods before and after the discounted value goes to zero? The life and death of the household? Supposedly population increases, but at some point in the future, depending on the subjective discount rate, the economy loses all utility. What does that mean? This is not even an analogy, it’s a formula pattern match, like that of a desperate student hoping a wild guess might get partial credit on the exam.

Why is it even plausible to assume that individuals in a market care only about consuming and leisure? What about job security or the costs of moving from place to place, or working conditions? Why is it remotely rational to assume that these individuals have no limits in how much leisure or stuff to consume they want? Isn’t it more than a little creepy to assume that a person/family who never works but spends all day eating caviar and snorting coke (or otherwise consuming a lot of expensive goods)  is maximizing utility from the economy? Real scientists actually do research on what motivates people – why is that research not even mentioned? And notice that because it is really hard to make a mathematical description of millions of individuals interacting in a market, Lucas is doing all this for a single “representative” household. A zoologist claiming to be able to analyze a herd of goats from a single “representative” would be laughed out of the house, and goat herds are less complex than human economic systems. This is a household that, unlike 80% of Americans, starts off with significant investment income. And it never makes bad investments. But somehow this infallible investor represents and determines the whole economy. And it’s an economy where it turns out that government or other public economic operation cannot play a positive role.

How do we know things like manufacturing policy, public education, social welfare programs, public childcare, and public investment in research and development don’t work? Why it’s in the model, by definition. Public goods like – water and sewer systems, education, research and development (like that which produced the Internet), fire and police protection, even national defense: none of that has any value in this model. Increased national income comes only from the labor and investments of the  representative agent. Taxes are purely a cost. Taxes can recycle only as welfare payments and government consumption.  The teaparty view of government is correct, mathematically rigorous, and scientific, because Professor Lucas assumes it.  Shockingly, Lucas discovers from this model that he was right all along: income from investment should not be taxed.

What can we say about an academic discipline in which Lucas and his colleagues are lauded, rewarded and cheered for claiming that this nonsense is rigorous science? Here is how Lucas characterizes this work:

 I hope as well that my story will serve as illustration of the way in which the search for theory at a more fundamental level can revolutionalize our thinking about important practical questions, and hence of the way in which progress at the most purely technical, abstract end of economics serves as the fuel for what Alfred Marshall called our “engine for the discovery of truth.”

Take a look at how learning (education) works in this “engine of truth”. Lucas set up the representative to split time between leisure, working, and learning. He even has an equation to illustrate how mathematically rigorous he is: u is work time fraction, x is leisure time fraction, and v is skill enhancement time fraction and:

u(t) + x(t) + v(t)=1.

The representative agent has two sources of income: wages for labor and earnings from investment and two different tax rates: one on wages and one on investment income.  Oh yes, the taxes are completely flat and don’t distinguish between types of investment income – because, because that is how the model works. Learning increases the value of labor:

so a worker with skill level h receives wh per unit of time worked

How do we know skill increases wages linearly ? Yes, well,  this is science so shut-up. But Lucas’s one representative is not just a worker, it is both an investor and at the same time it is the representative of the whole workforce. When its skills (and thus the skills of the workforce as a whole) increase, investment income increases – also in proportion to change of skill. Time spent on education then directly increases both salary and investment income. Investment income takes no work time.  How surprising is it that this representative, who begins with 2.4 times more capital than salary will find a low tax on investment income pretty convenient? Especially since there are no money losing investments in this rigorous, scientific model of the economy. 

Suppose we could change the model so that increased taxes automatically increase the general skill level – on the theory that the government invests some of that tax money in providing free education – not necessarily of the representative but of other workers – so the representative does not have to get up from the couch or work less. In that case, given the right parameters, it turns out that the liberals were right after all: increasing taxes increases prosperity. But we cannot tie increased government spending to better workforce skills in Lucas’ model because that’s not how the model works.

So what?

The general defense of this nonsense is that physics 101 is full of formulas that don’t take things like friction into account, and besides, um, quantum mechanics. But, dear Lord, that’s just general purpose misdirection that could be applied to anything at all. You could try to defend numerology with exactly the same argument. Or we often hear that it’s just an approximation. But like much in Economics, microfoundations is not an approximation, it is an ideological proposition. Microfoundations is  Margret Thatcher’s proposition that there is no such thing as society, just individuals. That’s not an approximation, though: it is a political program that flies in the face of lived experience. And finally, we can be pointed to many dissents, often tremulous ones, in the Economics literature. But that’s the real problem. The aura of credibility around microfoundations and other dubious truisms of mainstream permeate the discipline.

Here’s Paul Krugman – about as far from Lucas as one can get in respectable Economics. After complaining that there has been a decades long “de facto blockade of the journals against anything without rational-actor microfoundations” (something that in itself is an indictment of the entire profession), Krugman writes:

I would agree that being willing to use models with hyperrational, forward-looking agents was a natural step even for Keynesians. The Faustian bargain, however, was the willingness to accept the proposition that only models that were microfounded in that particular sense would be considered acceptable. It’s one thing to accept that models with an Euler condition at their core can sometimes be useful; it’s quite different to restrict your discourse to models with that characteristic, while ruling out everything else.

When, exactly, can these kinds of models be useful? Shouldn’t the eagerness of “Keynesian” economists to latch on to such mythology give pause? Make one fear their judgement and standards are irreparably deficient and they never understood Keynes in the first place? Keynes even said that he had to work very hard to extract himself from the conceptual errors of Marshall’s economics, the same errors that Krugman tells us were so alluring to “Keynesians”.

Here’s Simon Wren-Lewis again:

I doubt that most New Keynesian modellers adopted the microfoundations perspective against their better judgement. Instead I suspect most saw the power of the microfoundations approach (in analysing consumption, in particular), recognised the dangers in ad hoc theorising about dynamics (as in the traditional Phillips curve), and thought there was no contest

This is a remarkable passage because he is making a favorable contrast between Microfoundations (remember that “subjective discount rate”) and something else he calls ad-hoc? Was that other stuff even less grounded in scientific research than what Lucas describes? But even “left” economists have internalized the microfoundations ideology. Here’s Mike Begg, writing in Jacobin to insist that the value of money is:

determined by countless price-setting decisions by mainly private firms, reacting strategically to the structure of costs and demand they face, in competition with other firms.

That’s a claim of neoclassical economic faith, not an actual fact. But if you believe this stuff you will perhaps unconsciously incorporate right wing ideology into your economic analysis. This is why, for example, Christine Romer dismisses industrial/manufacturing policy and Robert Reich opposed the auto-rescue. If you believe, as a matter of faith or axiom, that economics emerges from the interaction of individual optimizing agents, then you carry a whole lot of political baggage with you, knowingly or not.