Steve,
"Most theories agree with findings" Much data is 'fitted' to extant theory to derive 'interpretations'. The meaning of 'findings' can be very wide wide, which is why I also specified 'data'.
As an example; you say; "The luminosity curve is the findings" That's the conventional view, allowing the LC much more import than it may contain. The LC is a 'CURVE' built from data, not the data itself. It is CREATED FROM the data by applying certain assumptions, then implying interpretations (really other assumptions). Believing the curve has some 'reality' beyond our assumptions is what gives rise to all the paradox (the SM is riddled with it).
LC's can in fact be constructed in a number of ways, including in perfect agreement with a recycling model; We have no proof about implications in either case, only consistency of other evidence with each case. By the way I forgot to attach the significant DR9 adjustment to the population distribution you showed, which I attach below.
A 'recycling' hypothesis fits wide data rather better than the SM. In fact all but one small part is present agreed science. It is this;
A spiral galaxy blends to a disc and accretes matter from the disc to the bulge with increasing OAM. Bi-polar outflows commence, forming 'lobes'. These gradually build in energy (and LF). At a certain a energy they become full 'quasar' jets, which are often in pulses matching solar accretion rates and can sometimes be intermittant, but all have the same collimated (layered Lagrangian bulk flows) structure. Now the problem. Nobody has worked out what happens then! 'Feedback' was the watchword, but NONE has been actually found at the scales required.
Now here is the only new part, solving a gamut of issues including overabundance of galaxies apparently viewed 'on edge' after the quasar era's, and then of 'young', 'open' spirals.; ALL the central disc matter is accreted and ejected, (re-ionised), the column of matter then commencing rotation at a virial axis (bar radius) on the perpendicular axis. (explaining bars, perpendicular halo rotation, stellar ages, re-ionization, red/blue z population distribution etc etc.) LF data can be fitted to such a model perfectly, but that means far less than it's power to resolve anomalous findings.
It's only a theory. It may well be wrong, but it is rather more coherent and consistent with the DATA (if not the 'findings' derived from current 'interpretation') than the long standing SM. I don't try to PROVE this model. I try to DISPROVE all possible models. With this I've so far failed. If you can help to falsify it I'd be delighted. But of course it fails if we use past alternative theory and interpretation as the 'measure', which we do by (bad) habit.