If the sole objective of an economy is to survive and reproduce, then the currency (that gives agents permission to act using the systems resources) should be increased to match a recent or expected increase in the ability of the infrastructure survive and reproduce. Deciding how to inject the new currency is part of the intelligence of the governing of the economy and currency. Evolution indicates that "happiness per person" is not the "infrastructure" goal that is being increased. The goal appears to be "harder" memory systems. It is not only longer lasting "genes" (in addition to ability to reproduce accurately* and plentifully**) in the face of assaults that benefits from "hardness" (strength of bonds) but hardness also results in fewer possible states (lower entropy) that seems to give rise to better communication within the system (i.e. better control, cooperation, and modeling even in competitive interaction that is seeking better end results). The hard structures support highly controllable and flexible movement of the much smaller "thought" elements (molecules, ions, electron, and/or photons) that are needed for modeling possibilities or algorithms and catalysts implementing known actions working towards the hard support structures. Harder structures result from strong bonds which are more dense, not counting gaps or air pockets that might be needed for lightness, insulation, or encapsulation. Stronger bonds means lower entropy because fewer states per molecule are possible (for a given temperature, S=k*ln(1.46*px/h for a single harmonic oscillator (atoms or phonons) in constrained 1D).
Happiness in a majority of individuals is a side effect of a succeeding system, not an apparent goal of evolution. Infrastructure is a mass of matter that includes algorithms, people, and other machines. It appears the primary purpose these days is to remove oxygen from metals, metalloids, and carbon which creates an infrastructure that has lower entropy, and to discover and tap into sunlight or more ores that are low-entropy concentrations that can be used to incorporate more mass into hard structures that are within a cooperative system. It emit oxygen gas which does not cause a net entropy reduction, but the mass within system itself is lower entropy. More efficient algorithms are lower real entropy due to Landauer's limit.
Is lowest entropy/mole the most survivable and expanding system? Should a governing system make laws and expand currency where it projects the greatest entropy/mole reduction will occur? Does this follow from and/or assist the least action principle? Is lower entropy/mole = happiness/mole? Is happiness = simplicity, a single state, or the progression towards it? More likely (more traditionally) evolutionary success = dGibbs F.E. = dU - T(dS/mole)/dt? Is the efficiency sought the ability to use external Gibbs free energy to increase controllable Gibbs free energy?
*or modified if it's necessary to hide from viruses or exploit nearby solutions. **Also, the entire idea of faithful reproducing is not a "goal" of "selfish" genes but is just the best solution found so far that potential energy gradients are pushing into existence.
Still weak on my practical entropy equations, but here goes:
If each independent agent (or mole of matter) has lower kinetic energy or entropy, then less heat (random kinetic energy) is present via Q=S*T. If incoming Sun energy is not simply being reflected back to space (albedo factor has not changed), then a lower Q means more potential energy is being stored. This lower average K.E. minus average potential energy which is a lower value for least action, as I have been expecting (evolution towards lower entropy is simply an expected result of dynamics). If there are more oxygen atoms in atmosphere instead of CO2, T will be lowered which may decrease local Q without decreasing S. Solar cells increase T which increases Q without necessarily lowering S.