When is meritocratic work structure a rational choice? It depends. Presently, meritocratic work structure is the basic nature of general employment - albeit in what is still relatively tight monetary circumstance. Meritocratic compensation draws from other resource capacity in the form of already existing wealth. When monetary policy proves insufficient, employers (in aggregate) can also wait longer before they hire workers. Hence employers logically "hold out" for the highest degree of skills capacity possible, to effectively compete with other institutions.
Meritocratic compensation is a natural component of institutions which tend to a single given spectrum of economic activity. However, asymmetric compensation processes now contribute to an incomplete marketplace. Everyone could ultimately benefit from a broader institutional framework - one capable of coordinating time and skill on more egalitarian terms. The symmetric compensation of time arbitrage, would eventually lead to greater labor force participation than is presently possible.
Even with the egalitarian (non rival) coordination of time arbitrage and knowledge sets, competitive constraints would still of course exist for time use options. However, those constraints would exist in group capacity frameworks. Given the changing nature of individual demand for services product, skills options wouldn't necessarily be abandoned due to the "all or nothing" skills demands of traditional employment. Non rival compensation can generate new wealth, but it needs carefully thought through organizational capacity in order to do so effectively.
Choice in group context also means one's skills sets will sometimes be matched for reasons of personal convenience. Those with desirable skills capacity always face real time constraints, a factor which would be taken into consideration for ongoing education in a wide variety of skills capacity. Coordination exists in an immediate and spontaneous sense, and also in terms of long term planning for ongoing services capacity.
One consideration in this regard, is system mobility. Just as some might choose meritocratic compensation for work where possible, others would likely desire to return to knowledge use systems as well. Likewise, some individuals may prefer the labor force participation of time arbitrage, who are not born into already existing systems. In some instances, some would gravitate towards non rival knowledge use, because of a lack of available work (in one's preferred disciplines) elsewhere. Occasionally, individuals might choose to work in knowledge use systems due to proximity, should local corporations set up operations where others would like to live.
Meritocracy is mostly problematic, insofar as the work conditions it creates are insufficient for full employment over long periods of time, particularly as resource capacity shifts into new frameworks. Even so, there may be factors in time arbitrage which would allow for greater knowledge use dispersal, than the rival conditions of meritocratic employment are able to allow.
Another thing to consider is that meritocratic compensation is always a greater possibility when economies are strong, and supply side conditions sufficient for hiring across a wide range of disciplines. Knowledge use systems could provide an additional support mechanism for nations, when it is otherwise difficult to maintain the work challenges which so many individuals seek.