Wednesday, December 14, 2016

The shortcomings of Economics II (interlude)

In my last post I argued basically that Economics (the discipline shaped and practiced by university professors, students, uncountable employees and executives distributed in tens of thousands of private and public enterprises along the world, as well as in government positions and as consultants and accountants and occasional entrepreneurs) is a sham. A con job. An ideological construct (in the old Marxist sense) without an atom of empirically verifiable truth within it.  

At the core of such accusation was the idea that economics offers only the skeleton of a theory in the form of equations that are not really equations. So in any Econ101 text you will find that demand (number of units demanded) is a function of the price: D = f(p), and is a monotonically decreasing function at that (so its derivate must be negative): df/dp < 0. Similarly we are told that supply (amount of units supplied) is a function of price, and labor (total number of hours offered by employees) is a function of wages (which is but another name for the price of hours worked in exchange for a salary), etc.

Which sounds “mathematicky” and “scienty”, but is really either a tautology (a statement that, being always true, tells us nothing new about the world) or empty verbiage with very little explanatory power. Putting seemingly common sense ideas (“the law of diminishing returns”, “the equalization of marginal costs and marginal utilities”) in a mathematical formalism apparently allows to understand them better and explore better their implications. It’s called “building a model”, it is understood that the “equations” are not really tools for providing testable predictions of how the real world works, but simplified representations that can indeed help us, in practice… predict how the world works.

According to this line of defense, the shadow of an equation is the second best thing (I guess after a “real” equation) to understand the infinitely complex realities of economic interactions, and not that different from what we encounter in physics or chemistry. When we read that the gravitational attraction between two bodies is directly proportional to the product of their masses and inversely proportional to the square of their distance (F = G x m1 x m2 / d2) that is also an approximation (we can only know G, the gravitational constant, up to a certain level of exactitude, which could mask that the formula is not as precise as it sounds, but an approximation that masks the contribution of additional factors, not to mention that it is valid, even approximately valid, for non-relativistic conditions -when both bodies are relatively stationary or move at speeds much below that of light). Although the engineering part of me feels strongly tempted to dismiss such argument as ludicrous, wooly-headed and downright idiotic, I’m going to give it some unmerited credit and expose its tomfoolery by devoting the remainder of today’s post to highlight the difference between what the knowledge of “economic laws” allow us to predict of how the world behaves and what the knowledge of good ol’ plain physical laws allow by imagining a world where the later are as well known as the former.

Thus, let’s imagine a world where all we know about how the natural world behaves is that it follows “pseudo-equations” of whose real content we know nothing at all. We don’t know what variables underlie the exhibited properties of any system, and what the variation of those variables may cause in the observed output we may choose to measure. All we know is there are “relationships”, and we may have some (not very reliable) inklings about the monotonicity of such relationships. To make it more understandable, I’ll illustrate such a world by the interaction between two characters, a Young Engineer called YE that has just started to work in an engineering company, and an old hand (unimaginatively called OH) that has to teach him the tricks of the trade:

YE: Good morning, Mr. Oh, I was told to ask you for some help in my new assignment

OH: Well, good morning, kid. Pleased to meet you. You’ve come to the right place, yes you have, as I am well known in our company for being not only very knowledgeable, but also kind and patient with young guns wanting to prove their mettle and find their way around. What is it that you have to do?

YE: Well, I have to calculate the diameter and the thickness of the pipes that go into the boiler of a new combined cycle power plant that we are designing.

OH: Now that’s a honorable and exciting task, isn’t it? And a serious one to boot! Make the pipes too big and thick and they will be too expensive, and difficult to support for the metallic structure that is being designed at the same time. Make ‘em too small or thin and they may crack or, even worse, blow up when their valves are closed, and seriously damage the plant. What input data have they given you?

YE: I know the height of the boiler intakes, and the water flow it needs (in gallons per minute) under different operational modes. I also have the performance curve for the pumps that have been selected to be at the base of the boiler and that will push the water

OH: Good, good, so the first thing you have to know is that the pressure loss the pumps will need to overcome is a function of the diameter of the pipes, and of the total flow that goes through them: TDH = f(D,F)

YE (smugly): Yep, I already know that, they teach that to us in engineering school

OH: Oh, that’s great, isn’t it? You’ve got a lot of  the bases covered, then. The other thing you need to know is that the loss of pressure grows with the flow of water, and decreases with the diameter of the pipe: df/dD < 0 and df/dF > 0

YE (a whiff of impatience starts showing in his voice): yep, I also knew that already

OH: well, my boy, then this is the final thing you have to know: WE DON’T HAVE THE DARNEDEST IDEA OF ANYTHING ELSE ABOUT WHAT THE RELATIONSHIP BETWEEN TDH, D AND F IS.

YE: What?

OH: You heard me right. Just no clue.

YE: What do you mean no clue?

OH: what you just heard. There are lots of theories, of course. A bunch of people call themselves “pipe neokeynesians” and maintain that after reaching some maximum the pressure loss may start increasing with an increase in the diameter of the pipe. Another lot (“pipe monetarists”) say that when the flow diminishes beyond certain point (“zero boundary”) it doesn’t matter what you do any more because the pressure loss remains constant. Nobody has been able to prove one or the other, and of course they don’t know either what those conditions may be beyond which the behavior of the water in the pipe changes

YE: But, but… couldn’t this things be, I don’t know, like, be tested? Couldn’t some experiments be run and so those “equations” be a bit more fleshed out, so for example we could determine the coefficients and the relationships between the variables. To arrive at something of the form, I don’t know, v x v/2+ gz + p/d = K (and then determine the value of K for different fluids and different materials of the pipe)???? So I could know the pressure of the fluid, and then see how that pressure affected the section of the pipe at each point?

OH: No, no, nonononononono… that’s a rookie’s mistake, thinking that such relationships can be teased out. You have to realize that nature is really complex, and that this field of research is the most complicated you can imagine. You really would have to take many, many more things into account than the fluid and condition of the inner surface of the pipe: what about the fluid temperature? How viscous it is because of impurities? The air temperature? The possible vibrations due to wind? The noise of the turbine nearby? Lingering effects of past earthquakes? The position of the moon and the stars? Any attempt to try to define anything more precise than TDH = f(D,F) is doomed to fail once and again, to be thoroughly disproved, to throw the whole field of hydrodynamics into disrepute and to reveal more exceptions than rules and to sow more confusion than clarity

YE: But… again, even if we can’t find a detailed relationship to define the pressure within the pipe, and the diameter to limit the internal speed of the fluid, and matrix of forces at each point of the pipe (depending on the thickness), can’t we at least plot the main variables and thus go to such graphic and find the values I need so I can solve the problem I’ve been given

OH: Well, we could do some nice graphs, sure, but they could not have any units on them, so they wouldn0t be of much use

YE: Why is that?

OH: At the beginning of time that method was attempted, and essentially found wanting. Depending on who drew the graphs the units were different, but for some unknown reason the shape was always the same. Thus they lost all credibility. Furthermore, used by different people they predicted wildly different values: where one engineer understood that a 25 inch stainless steel pipe two inches thick was needed, another concluded that a 2 inch PVC pipe 2/16 inches thick would be more than enough, and there was no way on Earth to make them agree. If you asked for the opinion of a third engineer with more experience, he would proffer a third combination, and the ensuing discussion could last for months, so no projects could ever be completed. So the profession decided it was best not even to attempt to reach numerically precise values

YE: So, how should I answer my supervisor? How can a respond to her request for a numerically precise value?

OH (speaking very, very softly): well, kid, this is the dirty little secret of our trade I’m about to tell you. Do as we all end up doing: go to the archive, find a previous project we have already design, take whatever number we used there, and use those

YE: But what if she asks me to justify those numbers?

OH: Again, do as we all do. Invent any contorted justification, make it as confusing and difficult to follow as possible, and the defend it so vehemently that nobody can reasonably judge it may conceivably be worth the effort to contradict you

If nothing better could be said, I’m pretty sure Ye would leave somewhat dejected and disappointed, but would in the end follow Oh’s advice, and in the end being as self-assured and confident in his engineering abilities as you could dream of.

Of course, in such a world power plants would be either outrageously expensive (but as in our world, that would greatly depend on what you compared them with…not that they would have a labor theory of value to explain in crystal clear terms why things cost what they cost) or dramatically unreliable. Just to be clear, that’s NOT the world we live in. We are clueless about what effect a raise in the interest rate may have. We are clueless about what may happen if the minimum wage is raised (heck, it is not as clear as I thought it was that the overall level of employment will fall, as that happens ceteris paribus… but all the rest never stays the same). But we do know pretty well how big and thick the pipes going into the boilers should be. And how much fuel (or gas, or coal) those boilers should burn to send a certain amount of steam to the turbine. And how many megawatts the turbine will in turn allow the generator attached to its shaft to pour in the electrical grid.

Now for how those megawatts should be priced, that is an entirely different story… 

No comments:

Post a Comment