14 ms·

Statistical mechanics of money (2000)

86 points by hairytrog 8y ago

ferros 8y ago
ELI5?
- drjesusphd 8y ago
  Physicists have gotten very good at the statistics of interacting particles. This paper asks, what if instead of a bunch of particles exchanging energy, one interprets it as the theory of many agents randonly exchanging money. Interesting things happen even in this over-simplified model.
- drjesusphd 8y ago
  To elaborate on "interesting things", at least as I understood it from a seminar by one of the authors... - The distribution of wealth naturally forms an exponential distribution. It takes effort/energy/negentropy to depart from this situation where there will always be many more poor people than rich. - Unlike particle energy, wealth can be negative. When this is accounted for (allow for debt), the rich inherently get richer. This is not even accounting for interest being paid by the debtors to the (presumably rich) debtees. - At a certain point, money begets more money. This results in a power law tail and identifies the "upper class".
febeling 8y ago
To me reads like the first premise is already wrong: "In a closed economic system, money is conserved." The opposite is true. Debt and the loan asset (if that's correct English usage) are created from nothing, for the debtor and the creditor respectively; the loan is converted to money intermittently by the central bank. On maturity of the loan contract both liability and asset side disappear when the loan is being repaid.* The central bank creates the money in the instant it interacts with the creditor, and when the contract ends, the money is erased. Do I missread this in some way? * If you disregard the interest, which is extremely simplifying obviously; still taking it into account wouldn't make the first statement more true
- oarabbus_ 8y ago
  I recently watched Ray Dalio's video on the Economic Machine: https://www.youtube.com/watch?v=PHe0bXAIuk0 https://www.youtube.com/watch?v=PHe0bXAIuk0 he describes the monetary/credit/debit system in the same way, and I do agree the first premise is suspect depending on how you define "closed system". However, I assume the authors' rebuttal would be that real-world economic systems are not closed system. Not that it makes their claims any more applicable to the real world.
  - febeling 8y ago
    I think you're right. The "closed system" is supposed to mean "without regard for the monetary aspects." Those are where all the interesting stuff happens (like growth, crisis), though ;) But that way the statement makes sense.
- phreeza 8y ago
  I think the trick is that banks can issue loans for higher amounts than they have assets, but if you sum up all the assets and liabilities, the sum should stay constant. So the conservation works more like an electrical potential, where you can always go lower locally, rather than a number of atoms, which has a fixed minimum at 0.
- WalterBright 8y ago
  Finally, someone gets it! I.e. the creator creates an asset. The asset is then used as collateral for a loan. The loan is simply a notation in a ledger. Money is created when the loan is spent. Money is destroyed when the loan is paid off.
  - RobertoG 8y ago
    The grandparent gets it but not totally. He says: "All money is in the books of the central bank,[..]" In the case of lending, it's the commercial banks where the money is created and destroyed. If enough reserves exist already in the system the central bank don't need to act.
zwaps 8y ago
While interesting, this sort of approach starts with wrong premises, such as the simple existence of a singular quantity of "money" at any given time, or that people spend money sufficiently randomly. If any of these things - and these things being what distinguishes social interactions and human behavior from particles - were the case, we'd have "solved" our economic modeling issues long ago. From my perspective (economist with prior math background), what makes the economic interaction of humans interesting and hard, is precisely that these interactions can not be aggregated (in a well-defined manner) to simple statistical or behavioral laws - be it distributions or representations, even if certain outcome distributions are observable. To my knowledge, econophysics - despite the strong marketing effort - has yet to produce new insights into the actual economy (and it has been going on since the at least twenty years). I think this is not surprising. The economics research effort does not lack in mathematical sophistication, even if physicists like to believe that. In fact, in terms of strategic interaction, behavioral patterns and epistemology, I think the economic models are well ahead of physics models for describing behavior. But - no one is really all that close if the level of aggregation is high. That means, macroeconomics and matters of money supply.
- zwaps 8y ago
  And what further irks me about this article, and others like it, is that it does neither - produce a result that can be observed and therefore tested. Here, the money supply distribution is not observed in the real world and is not supposed to represent observables such as wealth. - Nor does the article start from principles of human interaction, trying then to aggregate. Whether talking about exchange or debt, the article assumes "particles" randomly gamble against each other. In fact, social norms, externalities, TIME and timing, herd behavior and boundedly rational strategizing is precisely why human behavior does not aggregate to a gambling distribution. I know this article is from 2000, but my outlook would not have been different back then.
- qubex 8y ago
  As an economist with a background in mathematics and physics I very much appreciated ‘econophysics’ and associated efforts, but a short while into my foray I began to realize that economics is afflicted with a dearth of conserved quantities. Physics has plenty of these (sometimes absolute, sometimes approximate) laws of conservation (conservation of energy, conservation of mass-energy, conservation of linear/angular momentum, et cetera) and these laws of conservation are critically important in constraining the dynamics a system can adhere to and thus constitute a kind of ‘shortcut’ that allows apparently intractable problems to be deftly solved by putting these constraints to masterful use. Economics has none of this. Even the apparently solid concept of ‘wealth’ (even in its weaker relative form) upon closer inspection has no fundamental basis: transactions between two parties can affect the market value of an asset and therefore increase or decrease the wealth of all of those who own (or owe) such assets. The field therefore rapidly decayed into a form of “computational economics”. This might appear similar to “computational physics” and its use of finite element models and massively parallel simulations of particle systems to solve fluid dynamics problems. Trouble is, those systems are built up from elements that respond to the net of extant forces acting upon them, whereas in economics agents can and do act on the basis of what they expect those forces might be in the future. Self-fulfilling prophecies are “very much a thing” in economic and financial phenomenology: bubbles, crashes, recessions, booms, the whole business cycle itself... Add to that that some actors, even very significant ones, in those markets might have really weird utility functions — for example, central banks seek not to maximise their utility — and that some agents are not rational at all (noise-traders, entrepreneurs misreading trends, & cetera) and you have another big flaw in the framework.
aszantu 8y ago
Is it possible that the poorer factions in a society learn skillsets that are keeping them poor? Their skillsets help them through poor life but are prolly useless when getting more money?
Confusion 8y ago
The Black-Scholes equation is literally the basic heat diffusion equation [1]. [1] https://quant.stackexchange.com/questions/84/transformation-from-the-black-scholes-differential-equation-to-the-diffusion-equ https://quant.stackexchange.com/questions/84/transformation-...

86 points by hairytrog 8y ago

ferros 8y ago

ELI5?

drjesusphd 8y ago
Physicists have gotten very good at the statistics of interacting particles. This paper asks, what if instead of a bunch of particles exchanging energy, one interprets it as the theory of many agents randonly exchanging money. Interesting things happen even in this over-simplified model.
drjesusphd 8y ago
To elaborate on "interesting things", at least as I understood it from a seminar by one of the authors... - The distribution of wealth naturally forms an exponential distribution. It takes effort/energy/negentropy to depart from this situation where there will always be many more poor people than rich. - Unlike particle energy, wealth can be negative. When this is accounted for (allow for debt), the rich inherently get richer. This is not even accounting for interest being paid by the debtors to the (presumably rich) debtees. - At a certain point, money begets more money. This results in a power law tail and identifies the "upper class".

febeling 8y ago

To me reads like the first premise is already wrong: "In a closed economic system, money is conserved." The opposite is true. Debt and the loan asset (if that's correct English usage) are created from nothing, for the debtor and the creditor respectively; the loan is converted to money intermittently by the central bank. On maturity of the loan contract both liability and asset side disappear when the loan is being repaid.* The central bank creates the money in the instant it interacts with the creditor, and when the contract ends, the money is erased. Do I missread this in some way? * If you disregard the interest, which is extremely simplifying obviously; still taking it into account wouldn't make the first statement more true

oarabbus_ 8y ago
I recently watched Ray Dalio's video on the Economic Machine: https://www.youtube.com/watch?v=PHe0bXAIuk0 https://www.youtube.com/watch?v=PHe0bXAIuk0 he describes the monetary/credit/debit system in the same way, and I do agree the first premise is suspect depending on how you define "closed system". However, I assume the authors' rebuttal would be that real-world economic systems are not closed system. Not that it makes their claims any more applicable to the real world.
- febeling 8y ago
  I think you're right. The "closed system" is supposed to mean "without regard for the monetary aspects." Those are where all the interesting stuff happens (like growth, crisis), though ;) But that way the statement makes sense.
phreeza 8y ago
I think the trick is that banks can issue loans for higher amounts than they have assets, but if you sum up all the assets and liabilities, the sum should stay constant. So the conservation works more like an electrical potential, where you can always go lower locally, rather than a number of atoms, which has a fixed minimum at 0.
WalterBright 8y ago
Finally, someone gets it! I.e. the creator creates an asset. The asset is then used as collateral for a loan. The loan is simply a notation in a ledger. Money is created when the loan is spent. Money is destroyed when the loan is paid off.
- RobertoG 8y ago
  The grandparent gets it but not totally. He says: "All money is in the books of the central bank,[..]" In the case of lending, it's the commercial banks where the money is created and destroyed. If enough reserves exist already in the system the central bank don't need to act.

zwaps 8y ago

While interesting, this sort of approach starts with wrong premises, such as the simple existence of a singular quantity of "money" at any given time, or that people spend money sufficiently randomly. If any of these things - and these things being what distinguishes social interactions and human behavior from particles - were the case, we'd have "solved" our economic modeling issues long ago. From my perspective (economist with prior math background), what makes the economic interaction of humans interesting and hard, is precisely that these interactions can not be aggregated (in a well-defined manner) to simple statistical or behavioral laws - be it distributions or representations, even if certain outcome distributions are observable. To my knowledge, econophysics - despite the strong marketing effort - has yet to produce new insights into the actual economy (and it has been going on since the at least twenty years). I think this is not surprising. The economics research effort does not lack in mathematical sophistication, even if physicists like to believe that. In fact, in terms of strategic interaction, behavioral patterns and epistemology, I think the economic models are well ahead of physics models for describing behavior. But - no one is really all that close if the level of aggregation is high. That means, macroeconomics and matters of money supply.

zwaps 8y ago
And what further irks me about this article, and others like it, is that it does neither - produce a result that can be observed and therefore tested. Here, the money supply distribution is not observed in the real world and is not supposed to represent observables such as wealth. - Nor does the article start from principles of human interaction, trying then to aggregate. Whether talking about exchange or debt, the article assumes "particles" randomly gamble against each other. In fact, social norms, externalities, TIME and timing, herd behavior and boundedly rational strategizing is precisely why human behavior does not aggregate to a gambling distribution. I know this article is from 2000, but my outlook would not have been different back then.
qubex 8y ago
As an economist with a background in mathematics and physics I very much appreciated ‘econophysics’ and associated efforts, but a short while into my foray I began to realize that economics is afflicted with a dearth of conserved quantities. Physics has plenty of these (sometimes absolute, sometimes approximate) laws of conservation (conservation of energy, conservation of mass-energy, conservation of linear/angular momentum, et cetera) and these laws of conservation are critically important in constraining the dynamics a system can adhere to and thus constitute a kind of ‘shortcut’ that allows apparently intractable problems to be deftly solved by putting these constraints to masterful use. Economics has none of this. Even the apparently solid concept of ‘wealth’ (even in its weaker relative form) upon closer inspection has no fundamental basis: transactions between two parties can affect the market value of an asset and therefore increase or decrease the wealth of all of those who own (or owe) such assets. The field therefore rapidly decayed into a form of “computational economics”. This might appear similar to “computational physics” and its use of finite element models and massively parallel simulations of particle systems to solve fluid dynamics problems. Trouble is, those systems are built up from elements that respond to the net of extant forces acting upon them, whereas in economics agents can and do act on the basis of what they expect those forces might be in the future. Self-fulfilling prophecies are “very much a thing” in economic and financial phenomenology: bubbles, crashes, recessions, booms, the whole business cycle itself... Add to that that some actors, even very significant ones, in those markets might have really weird utility functions — for example, central banks seek not to maximise their utility — and that some agents are not rational at all (noise-traders, entrepreneurs misreading trends, & cetera) and you have another big flaw in the framework.

aszantu 8y ago

Is it possible that the poorer factions in a society learn skillsets that are keeping them poor? Their skillsets help them through poor life but are prolly useless when getting more money?

Confusion 8y ago

The Black-Scholes equation is literally the basic heat diffusion equation [1]. [1] https://quant.stackexchange.com/questions/84/transformation-from-the-black-scholes-differential-equation-to-the-diffusion-equ https://quant.stackexchange.com/questions/84/transformation-...