This note was stimulated by Tatsuo Tanaka’s paper: Possible Economic Consequences of Digital Cash. The original DSR paper mentions in passing that packet values would be multiplied as packets crossed currency boundaries. We explore this issue further here.
Some who have commented on DSR imagine that it necessarily entails occasional heavier weight digital monetary instruments to handle settlements between node operators. That may indeed be the dominant link settlement scheme but I think that it is important not to weld such a plan into DSR. Some of the flexibility of DSR stems from the fact that some settlements may be done in terms of monthly exchange of chicken thighs.
A node operator may have links to nodes that use different currencies. More concretely the agreements with other node operators is that payments are to be settled in currencies that vary with the node. To simplify matters we imagine a node with two links to two different currencies.
We will call this node a cambio for obvious reasons. Node operators will have something like credit limits with each of their neighbors. There might be two kinds of credit limits:
Arbitrageurs also solve another problem. A systematic cyclic flow of money in the net would trigger limits. If such a flow first triggers modifications in the exchange rate, then arbitrageurs will discover this and set up countercyclic flows before the hard limits are reached.
DSR supports domino style financial failures of nodes. Small players with link limits smaller than their personal assets would survive if they chose to. The network might become disconnected, however.