Friday, September 27, 2019

Why use physics to describe economics?

Do you mistrust the predictions of mainstream macroeconomic growth models and reject the policy prescriptions of their practitioners? Many do. 

Is this fair? And what would we do instead?

How about using physics ? Certainly as a field it has a pretty good track-record for describing nature, at least as an alternative to religion and magic.  The big thing in physics as a field or any other science is that it demands falsifiable hypotheses rather than the opinion or Ivy league pedigree of its practitioners. Results should enable useful predictions, those that offer the potential for robust long-range forecasts subject to physical constraints. 

On the other hand, the scientific method is certainly not at the core of modern macroeconomics, probably because economist's are lured by influencing public policy, or they don't believe that social systems are physical systems. Making mathematical models to describe reality is common in macroeconomics. No problem there. Math is a useful tool. And quantifying things is good. But making mathematical models that can't be falsified is terrible! Neoclassical economic models employ equations for the GDP, or “production functions”, that are dimensionally inconsistent formulae that can be “tuned” to match observations of labor and capital. And they always are. It is not possible to falsify these moving theoretical targets because they can always be made “right” by adding layers of social complexity or by tweaking the production function exponents. If conditions change and the formula no longer works, economists just tune again and call it a “structural break”! This is strange,  at least if the goal is to understand how things work rather than show off one's mathematical aptitude. It would be abhorrent to imagine a basic physics equation being adjusted as time progresses or for the situation at hand. The speed of light in a vacuum doesn’t get to be different for you than for me or for last year versus this year.

Perhaps economics and science can be reconciled. It would be nice to think so. Unfortunately, I don’t think this is possible without some important adjustments. Mainstream economic models take the approach that human labor is distinct from physical capital. Labor uses capital for production. Some portion of production is short-term “consumption” of things like food and entertainment, that contributes nothing to the future. The other portion is a long-term “investment” in physical capital that enables labor to produce more in the future depending on labor productivity. The feedback loop of this "virtuous cycle" forms the basis of unconstrained long-run economic growth. The division between long-term investment and short-term consumption is set at the time of one year. In this model if you store a bottle of wine in your cellar for years it adds to capital. If you drink it next week it counts as consumption and does nothing for growth.

But from the perspective of physics this all seems a bit ad hoc. Surely, in a finite world nothing can grow forever. And why prescribe the division between consumption and investment at one year and not some other time? Other than paying annual taxes to the IRS on April 15, there's nothing inherently special about the frequency with which the high density rock we call Earth revolves about the larger accumulation of hydrogen and helium we call the Sun, especially in a non-agricultural economy. And what makes labor so distinctive from physical capital? People are not all that special in the universe, at least there’s nothing in the fundamental equations of physics that says “people” or "labor".

Thinking about the economy more generally, it might make most sense to make the following adjustments:

  1. Subsume people into a very general physical representation capital that includes all components of civilization. 
  2. Remove the one-year separation between consumption and investment
  3. Link consumption to physical resources like energy and raw materials
Admittedly, not treating people as special might seem strange at first, but let’s go with the possibility that our egocentrism doesn’t really matter.  Our personal feelings aside, we are just sacks of matter that enable electrical and fluid flows down potential gradients. It sure has been hard for neuroscientists to find any evidence for free will; so perhaps people are really no different than any other physical system in civilization, acting as conduits for energetic and material flows just like communications networks or roads. Then, the consumption/investment dichotomy of traditional models disappears. Everything that lasts, including us, is an investment in the future. Equally, everything to last must consume resources to be maintained.

The model I’ve introduced is based on the very simple premise that accumulated economic production of everything in civilization must be sustained by a proportionate amount of global primary energy consumption. Turn off all the power and civilization is worth nothing; and the more we accumulate the more power is required for sustenance. 

This is a hypothesis that might seem crazy to a traditional economist. But crucially it is an assumption that is falsifiable. A test can be set up that could potentially show it to be wrong. Making this test however, it turns out that it is a premise that is supported by available statistics: 7.1 ± 0.1 milliwatts of continuous power consumption has been required to sustain the historically accumulated global production associated with every inflation-adjusted 2005 dollar in every year statistics have been available since 1970.

Consumption versus production
From an accounting point of view it makes a lot of sense for economists to selectively subtract short-term household and government consumption from economic output (or GDP) to obtain a long-term capital investment that adds to previously accumulated capital. Capital investments are then independent and additive; it is assumed that the whole is the sum of its parts. If saving an ounce of gold - an item that lasts - adds $1000 then it seems obvious that saving two ounces adds $2000. 

But a little added thought suggests it’s not quite so straight-forward. Neither labor nor physical capital means anything without the other. “No man is an island, Entire of itself. Each is a piece of the continent, A part of the main...” An ounce of gold has no intrinsic worth as it’s just a rock. But it has a great track record of providing value as a collectively appreciated part of society. Someone with lots of gold must be a very important person because they got the gold and others didn't. Value appears to lie in access not the thing itself.

Viewed physically, gold as a tool in society’s banking system networks, which in turn are maintained by the accumulated knowledge capital of bankers. But if the ounce of gold was left abandoned and forgotten in the middle of the desert it would currently be worthless. It only has value as part of a larger society.

And if everyone else tried to sell their gold for $1000, the value per ounce would fall, including the ounce you kept . Value, therefore, does not lie in individual “things” or people by themselves. True value lies in a larger global network and the role we and our structures play in it. Physiological, social, computer, communication, and transportation networks are all part of the living organism we call civilization. Capital value is not strictly additive because no element is completely independent of any other.

And this "super-organism" must continuously consume energy and raw materials to survive. Without energy consumption, there is no investment that will be worth anything. We would all be dead if nothing else. So there is no really legitimate separation of short-term consumption and long-term investments in capital. because each requires the other. And individually the value of any element must compete with all others as constrained by the amount of energy and raw material consumption that is possible.

For contrast with our current capitalist state, it helps to think of a subsistence society at near steady-state where nothing can be stored for the future: food rots quickly; the society maintains a more-or-less fixed population; and in its purest form there is no currency and no GDP. Even though the society has to consume food, the consumption is not part of any measurable economic output that contributes to growth. 

I have personally experienced something like a subsistence society working as a Science, Math, and Physics teacher for a couple years in the beautiful, remote tropical South Pacific island group of Ha’apai, Tonga. Even though there was a little money to go around for luxury items, it was almost totally impossible to buy traditional foods like coconuts, taro, and octopus that anyone could access. Only revolting imported “specialties” like canned beef and mutton flaps were readily found for sale in small shops. Any given root crop was more or less available from whomever had it; everyone except a handful of foreigners had direct or indirect access to a fixed, finite quantity of fertile land where they could grow throughout the year -- if you didn’t have a yam, get your own or ask (or even take). The local mantra was “Ha’apai is good; food is free”. 

In an expanding civilization, matters are rather different, as essential to expansion is the idea that products can be acquired and stored for the future. Food is treated quite differently as it is a real commodity -- in most homes we have a fridge, freezer, and larder. Owning money gives us a right to buy access to something, whether a house or a tasty sandwich. We tend to like sandwiches, and buying access to a sandwich is an investment in our well-being.  If we're really really hungry, in fact it's probably the only investment we can possibly think of. The sandwich offers the future potential - no matter how near in time - to be content, better able to interact with others, and more productive in our jobs. The wonderfully consumptive process of actually eating the energy and raw materials in the bread, mayo, lettuce, ham and cheese provides fuel for our bodies and the lingering memory of sublime satisfaction that spurs future purchase of even nicer sandwiches. Nothing about the purchase of food is “consumed” in a way that becomes totally lost to the past as expressed in standard economic models.

So all financial transactions, whether for a sandwich or a bar of gold, that count in the GDP are really just monetary expressions of small, instantaneous increments in the growth of civilization’s networks of connection and access. Certainly some are larger than others, but we might expect that the accumulated GDP, adjusting for inflation, is a representation of the growth of these networks, and that the total networks must be sustained by a corresponding consumption rate of energy. Total global capital can be tallied as historically accumulated GDP  (units currency), or more strictly, as the time integral of every little differential increment in productivity (units currency per time).  Even the most ancient inflation-adjusted economic production has to some degree sustained us through to our activities today. Subsistence cavemen did nothing for our wealth today -- we should expect the GDP was zero. But growing cavemen societies, if they persisted, did, by creating fire and building language and social structures.

This approach, leads to the rather wonderful result that 7.1 ± 0.1 milliwatts of continuous power consumption is required to sustain the worth associated with every inflation-adjusted 2005 dollar of civilization, year after year after year. Simply put, consumption of energy and raw materials sustains all of civilization’s previously accumulated value as calculated by the summation of all prior economic production adjusted for inflation. This value or wealth must be sustained by a proportionate amount of energy consumption.

And with this we're off to the races. Economics can now be moved from the nebulous world of mathematical confabulation and Ivy league pedigree to a problem in thermodynamics. It may remain difficult, with much to be understood about such key problems as wealth distributions. But things can be readily said about e.g. population growth, long-run economic growth, and the fallacy of appealing to energy efficiency to solve climate change. There lies a tested physical foundation with which to attach such problems, one that rests on physical constraints and can be tested with observations like any other true science.


  1. As a school-boy, in mid-1960's London, my knowledge of Physics pointed to economic collapse within 40 years without fusion or some other type of very cheap energy. And now here we all are!

    None of this stuff is at all difficult but a lot of people of course are trapped by religion/culture/class/education and in a sense are as if already dead. IMO the Scottish Enlightment penetrated the conciosness of say 0.1% of the world pop. and what is coming now is a new dark-age and as an observer of the Human Condition it looks very dark indeed.

  2. Interesting article. So in practical terms, what does this mean for the here and now and the near-term (10yr) future? What is the global rate of power production needed to maintain our existing economy, or grow at the 2-3% rate that some economists seem to think is necessary?

    1. It's slightly more complicated, but about 2-3%, maybe slightly less.

      So assuming we do this, will double our daily energy demands in less than 30 years consuming about as much in the interim as we have in the past 100 years or so.

      Or if the feeling is that we can't accomplish this energy feat, then yeah, economic growth will be something rather different...

  3. Usually the 'economists' who use 'macroeconomic models' 'believe' that the solution to the current macroeconomic problem is the implementation of the 'correct' type of demand side policies. That is, how to increase income/output {'growth'}.
    Increase M0, M2 or M3, cut r, or make it negative. Increase G and finance it by 'borrowing' from the central bank or by borrowing from the private sector. Or ensure that private credit is extended only for GDP transactions.
    All that is lacking is a 'sufficient' increase in effective demand!
    The neoclassical/Austrian economists, who believe that income/output is 'supply determined', will argue that all that is required to generate a large increase the growth of the underlying productive potential of an economy is for taxes to be cut and more 'competition', etc be introduced!

    Aside from the negative externalities of 'growth', what they ignore is the 'energy supply side'?

    It's 'too late'?
    'We' have ten years?

    “ . . . our best estimate is that the net energy
    33:33 per barrel available for the global
    33:36 economy was about eight percent
    33:38 and that in over the next few years it
    33:42 will go down to zero percent
    33:44 uh best estimate at the moment is that
    33:46 actually the
    33:47 per average barrel of sweet crude
    33:51 uh we had the zero percent around 2022
    33:56 but there are ways and means of
    33:58 extending that so to be on the safe side
    34:00 here on our diagram
    34:02 we say that zero percent is definitely
    34:05 around 2030 . . .
    34:43 need net energy from oil and [if] it goes
    34:46 down to zero
    34:48 uh well we have collapsed not just
    34:50 collapse of the oil industry
    34:52 we have collapsed globally of the global
    34:54 industrial civilization this is what we
    34:56 are looking at at the moment . . . “