This issue came up in the marine park thread:
http://www.ozpolitic.com/forum/YaBB.pl?num=1192441509/210#210PJ insisted that current catch rates could not be improved upon. He based this assumption on the fact that people had calculated a maximum sustainable yield (MSY) based on the stock levels as a percentage of unfished (virgin) stock levels. This obviously assumes a certain management regime, but PJ confused that assumption for endorsement of current management regimes. PJ went on to insist that it is therefor not possible to increase both stock levels and catch rates. That is, it is not possible to go above the catch rate corresponding to the current estimate of MSY without causing stocks to collapse - even if you change management regime. This also carries an implicit assumption that the only way to permanently increase stock numbers is to permanently reduce catch rates. Basically it is the old argument - "this is how it has always been done, therefor there is not a better way" - even though current management regimes have only been in place for a few decades.
Take cattle farming as an analogy. Compared to ranching or hunting wild stock, cattle farmers micromanage the land so that it has a far higher stocking rate. They also manage to get a far higher harvest rate than you could get from harvesting wild stock.
Fisheries have only recently come out of the management regime of unregulated harvest of wild stock. The obvious first step was regulated harvest of wild stock. Obviously there are practical issues preventing the sort of micromanagement you can achieve on land. However, the same principle applies - it is not difficult to increase both stocking rates and MSY. The current estimates of MSY assume the management regime is fixed. They are not an endorsement of the current management regime.
Back to the cattle analogy. The current management regime is similar to harvesting wild, unmanaged stock, but limiting the amount that can be taken in various ways, one of which encourages the harvest of the best (fattest, fastest growing etc) cattle and leaving the runts to breed for next year. Marine parks are analogous to introducing limited protection of the key breeding stocks and providing some areas where stocking rates can reach higher levels. It is similar to having a stocking rate of far less than the natural levels - that is, stocking rate is not limited by the risk of overstocking (as we are familiar with the issue on land), but by the inability of achieving both high stock levels and high harvest rates under the current management regime.
Marine aprks also make the management regime more stable. Imagine that for some reason it was not possible to count the number of cows you had left. You would have to take an extremely conservative approach to harvesting them, as overharvest is difficult to predict and could have serious consequences for future harvests. If you set aside a few paddocks where cattle are not harvested, and from which cattle could leave as their numbers got too high for the paddock, then you could be a lot more secure in your harvests as you would know that there is always a significant number of cattle in the unharvested paddock.
PJ's argument is equivalent to people switching from unregulated harvest to limited regulation, then deciding a few decades later (and in the face of contradictory evidence) that the limited regulation they initially implimented cannot be improved upon. It is an argument based on ignorance, and an assumption that a certain management regime is best, just because it is traditional, even though it's not actually traditional - it has just been accepted already.
Anyway, the analogy obviously breaks down as you get into too much detail. The main point is that maximum sustainable yield (as it is currently estimated) is not a fixed number. It depends on management regime. If you manage the stocks better, you can get higher stock levels, higher yields and a more sustaianble (secure) harvest - one that is less sensitive to your lack of knowledge of the remaining stock levels, and less likely to crash because you kept harvesting while you were unaware that stock levels had plummeted for some reason.
The current estimates of MSY represent the upper limit on what can be taken. In practice, fisheries must harvest well below the MSY due to the risks associated with not knowing with reasoanble accuracy what the remaining stock levels are. Marine parks increase the MSY and allow you to get closer to it in practice, because the management regime introduces an inherent stability. Current management tools make it easy to overharvest in those years where for some reason (eg natural variation) stock levels are unusually low. Marine parks make it much more difficult to overharvest because the breeding stock is far better protected. Under current regimes, only paperwork stands between fishermen and the breeding stock for next year. Fisheries managers are forced to try to limit catches in a scenario where they don't know what the remaining stock levels are and they don't know how many fish a large group of fishermen is taking. They have managed to get it to work, with only a few dramatic failures. While that is commendable, it is no reason to stick with clearly inferior managemen tools.