IMO any such rule-of-thumb guidelines that may be defined are not likely to be useful, and might even contribute to poor decisions being made. There are simply too many variables and dependencies involved. A few examples:
-- The listener’s preference for how much deep bass extension should be provided by the main speakers, and also the maximum volume capability of the speakers (especially the ability to cleanly reproduce brief high volume dynamic peaks, such as frequently occur in many classical symphonic recordings that have been engineered with minimal or no dynamic compression), can dramatically influence the cost (and the size) of the speakers, for a given level of quality.
-- Low efficiency speakers will of course require more power than high efficiency speakers, and for a given level of quality and within a given class of operation (i.e., A, AB, D) more power usually means greater amplification cost.
-- Conversely use of high efficiency speakers will usually result in amplification comprising a lower percentage of system cost, for a given level of quality and within a given class of operation, while often resulting in increased speaker cost (and size) for a given level of quality.
-- Sensitivity to analog interconnect cable differences depends on the output impedance of the component providing the signal and on whether the interconnection is balanced or unbalanced, among other variables. Sensitivity to digital cable differences depends on too many variables to list here :-)
-- Sensitivity to speaker cable differences depends in part on the impedance characteristics of the speaker.
-- And of course the listener’s preference for the type of source (vinyl, CD, streamer, computer, etc.) can make a major difference in the percentage of system cost that should be allotted to it.