This piece from Sierra Bullets summarises the main reasons:
http://bulletin.accurateshooter.com/?s= ... mit=Search
Case capacity has a major effect and 223 Rem sees a considerable variance across makes (and even periods - eg Lapua went to thinner wall / higher capacity manufacture some years ago). There is a table in here of 223 case weight / internal capacity which although rather old is still useful.
https://www.accurateshooter.com/cartrid ... es/223rem/
Powder and bullet manufacturers may make certain assumptions when compiling testing loads and compiling data. Because of the large variety of 223 brass around - commercial and arsenal - and differences between makes, the manufacturer may build a 'safety discount' in, reducing charges across the board. Printed manuals usually say somewhere in the 'How to reload' section and in Notes to reloading data, that those data apply only to the exact components listed. However, many users either don't understand the issues, or never read the 'small print' anyway, or just say 'B*gg*r it! - an xxgn bullet is the same whoever made it, likewise there's no difference between makes / models of primer' etc, etc. So the manual compilers add a little safety in. Conversely, some use the lowest possible capacity case available and say so in the cartridge notes. For instance, Speer deliberately chose Israeli IMI 223 brass for its tests loads for many years as this is a very heavy / low capacity design. (Using thin Winchester brass could add as much as 1.5gn to allowable max loads.)
However, with a few exceptions such as Speer, we don't know whether the manufacturer has or hasn't built in any such adjustments and it's extremely unwise to assume they have (as some handloaders do as a matter of course!) How often do you hear the comment on the range that Vihtavuori loads are always lower than they need be so you can play fast and loose with them. Some Viht data are 'modest', indeed inexplicably so, but not all!
Then Hornady is American and Vihtavuori is European working with different standards bodies - SAAMI and CIP. This may or may not affect the specs for the industry-standard test barrels used in their load development. Certainly, the two use different pressure measurement techniques even though both are Piezzo crystal based. In the case of 223, SAAMI MAP is 55,000 psi whilst CIP is 43,000 bar / 62,366 psi. Even with modest differences in measuring methods, this standards difference in itself would make CIP-compliant Viht maxima higher than Hornady's SAAMI equivalent. That's why there are two QuickLOAD 223 entries.
There is a growing gap too between US loads for some cartridges largely used in AR type semi-auto gas-guns and the same numbers when used in bolt-actions. Sierra produces two sets of 223 load tables for this reason. Those cartridges based on the 7.62X39mm M43 case design - Grendel, new 6mm Hornady ARC etc have modest SAAMI MAPs - 50,000 and 52,000 psi - despite people knowing the parent PPC case is a strong design. It's because three-quarters or more users in the US will shoot these numbers in AR-15s and the larger diameter case-head (over 223) makes them marginal for the AR-15 bolt size / strength. Put the same cartridges in a bolt-action, even the small Cz527 and Howa Mini 1500 and they'll take another 5-7,000 psi easily and safely.
At the end of the day, loads tables aren't tablets written on stone. There are so many variables involved with both the user's firearm and ammunition components, even the effects of the environment they're used in - ie ambient temperatures - that they are advisory rather than 100% prescriptive. It is up to the user to start low and work up in small steps, look for pressure signs, and stop / move back down a bit when things start to look too 'hot'. Given the improvements and prfice reductions in chronographs, use of one really is mandatory too nowadays. If you're 'getting higher MVs than you should', then it's a pretty fair bet you're also producing higher pressures.
