To me it's more a numbers game. If we kept up the same number of ships from the Cold War and continued to upgrade them, we'd have more capable ships today. Instead we have far fewer ships and that means we're less able to absorb losses if we were to be involved in a war.
Ah but I'm not talking about old vs new. I'm talking about numbers of ships. (And to your argument, "A modern DDG could wipe the floor with a squadron of older ships" but a squadron of modern DDGs will wipe the floor with a squadron of older ships.)
On both of these, the problem you encounter is one of scarcity.
While the Federation is a post personal scarcity society (at least in some sense, given the marginal cost of energy is near 0 with cheap fusion), it's clear that there's bottlenecks in everything from ship power sources (dilithium) to physical locations to build ships to trained personnel.
So for any given amount of resources you can focus on building more capable ships or a greater number of them.
There's a reason why naval treaties focused on tonnage rather than other measures like budgets or unit numbers. The displacement of individual ships was perhaps the greatest single determinant of their capabilities. The Washington and London treaties helped keep a tight rein on battleship displacement and capabilities until they were abrogated, and once the treaties were done away with, the size and capabilities of battleships ballooned to about half again the size as the older ones.
The numbers game can be effective, but only if you have a nigh infinite naval budget and a substantial commitment that can't be covered any other way. Either way it's a great way to commit yourself to budgetary overstretch. In constant dollar terms, it's easy to forget that the DoD budget hasn't changed all that much from the Cold War era (and since the War on Terror has increased). A modern destroyer displaces as much a Cold War era cruiser with substantially better capabilities. It also has about half the crew. Maintaining numbers would always have some sort of penalty, whether in capabilities or budgets. Now given that there's plenty of resistance of a trillion dollar defense budget, I highly doubt you'd get enough votes if you decided the DOD budget needed to be closer to 30% of GDP than 10%.
Same would go for Starfleet. They have limitations, though mostly in terms of ship building/refitting capabilities (number of shipyards, slips, etc) and crew. So once the need for keeping sheer numbers around to cover every cubic lightyear of space available became lower, they'd probably shift toward getting more bang per buck.
For example, the Hermes
classes had crew complements of about 195-200. By contrast a Miranda
's complement was about 350. So for every 3 Miranda
s you could crew 5 of the scouts/destroyers. From a war time point of view, having 5 ships to send to 5 different sectors is attractive. From a more peace oriented view, you'd rather have the 3 much more capable and flexible cruisers wandering around. I'm guessing what happened in the post-Khitomer era was that Starfleet decided on ships that were most efficient from a capability for crew complement perspective. Hence they could build larger numbers of Mirandas
, replaced Constitutions
and the like. (You could crew 2 Excelsiors
for every 3 Constitutions
, and from a tonnage point of view, you get a lot more ship from 2 Excelsiors
than 3 Constitutions