the random oracle

The Atlantic has been running an interesting series of articles about our modern monetary system. One thing that has struck me since the financial crisis happened is the way that we talk about money needs to change, because it is more and more out of sync with what money actually is, especially at large scales. As the Atlantic articles point out, money is trust.  It is purely a matter of how much you (and the rest of society) think it is worth; no one can eat a dollar bill, much less the electronic bits that represent the vast majority of money today. Money is also power: the flow of money governs how we allocate resources, especially that most important resource, people’s time.

One nuance I’d like to understand better is how the nature of money changes depending on the scale at which one observes it.  At a local scale, the scale at which most of us operate in our day-to-day lives, the amount of money we come into contact with is so small that it behaves “linearly”.  One more dollar buys one more dollar’s worth of goods, and we can pretty confidently understand what our paycheck means in terms of what it buys, at least over a reasonably short time frame.

Once you get to the scales that banks, governments, and multinational corporations work with, the billions and trillions of dollars, the nature of money changes completely.  Money at this scale creates feedback that money at small scales does not, and so one needs to be careful how exactly one talks about it.  In particular, money at this scale is non-linear if we view it in terms of how much stuff it buys. This I think is a mistake that we (at least we non-economists) often implicitly commit because we assume that money behaves the same way at macro scales as it does at micro scales.

Here’s what I mean by “non-linear”: suppose that the price of gasoline today is $3 / gallon.  Then we can linearly scale this relation and we know that $300 would buy 300 gallons today.  Of course, the price of oil could change from day to day (or even second to second), but fundamentally if at some point in time the price of gasoline is $x / gallon, then C * $x would buy C gallons at that time. Similarly, if I wanted to sell gas, I know that if I sold C gallons, I could get C * $x for it.

But now suppose we want to buy $10 billion worth of gasoline.  I would argue that that makes about as much sense as $10 billion worth of unicorns.  (OK, maybe a little more sense than buying unicorns.)  ”$10 billion worth of gasoline” doesn’t make sense, because the simple act of putting out an offer to buy $10 billion worth of gasoline would immediately and dramatically increase the price of gasoline. If we had $10 billion worth of gasoline to sell, then the simple act of offering to sell $10 billion worth of gasoline would decrease the price of gasoline. Thus, it would seem that $10 billion worth of gasoline means different things depending on whether we are buying or selling it, and so it’s not well-defined.

One can extrapolate and say that $10 billion worth of *anything* doesn’t make sense, and so “$10 billion” itself doesn’t really make sense, except that we know it’s bigger than $9,999,999,999.99 and it’s smaller than $10 billion + $0.01.

What $10 billion really represents, I think, is not so much a certain quantity of stuff (where stuff can be goods or services), but rather a certain amount of ability to change how people produce stuff, because if someone buys or sells $10 billion worth of anything, the most important effect of that action is not so much the transfer of ownership of that thing, but the fact that such a transaction will incite people to produce more or less of that thing. Maybe from this point of view, the effect of money can be quantified in a linear way, and so be more easily grasped by our mortal minds?

Is there a way of quantifying things and looking at things this way, or some other way of viewing money at a macroscopic scale that is less misleading than just translating it linearly into stuff?  Any pointers to relevant reading would be very much appreciated.