More the reason that I said it varies is because it depends on the equipment. In the US it is nominally derived from 24x lead-acid cells yielding 48v, but in older switches the maintenance charger is connected when the switch has external power and increases battery voltage to around 50v, sometimes a bit higher, like 54v. Basically due to the charger, other countries often specify 50v as the nominal voltage (this is of course similar to how we call automotive systems 12v when in practice they're around 13.7v most of the time).
All this results in 40-50v being considered a normal on-hook line voltage, but it can vary more in the real world. Of course modern SAIs or RLCs or what have you tend to use solid-state regulators that keep a very tight 48v, so I'm sure the variation is much lower in like modern suburban neighborhoods than it is in cities or with older exchanges.
When off-hook, current starts flowing and the line relays and local loop come into play. The line relays are not tightly standardized and range from say 400-700 ohm, but unless you're pretty close to the exchange the line resistance of the local loop is greater, which can be 1kohm or more. Then the actual telephone instrument has a resistance due to the current it uses for operation, 200 ohm is perhaps average but it varies plenty, I think the WECo phones were usually 180 ohm nominal. Both line resistance and telephone resistance vary widely. A total loop resistance of 2400 ohm could be called a maximum because it allows the phone the 20 mA that's considered a minimum for reliable telephone operation, but lots of equipment will work out of that range, especially since so many newer phones have an independent power supply and digital voice circuit. On older switches, where "older" includes plenty that are still in wide service like 5ESS, the line cards have a couple of jumper options to adjust the relay resistance in order to bring loop current up or down depending on length of the loop. That's mostly because high loop currents due to a short loop can shorten the service life of equipment.
Nothing is really regulated (on older equipment and per specs, newer equipment tends to have voltage regulation as a result of using more advanced transistorized power supplies), neither voltage nor current, and so it can all vary within a fairly wide range. This is surprising from the modern perspective but not so much when you consider that the "standards" here are a hundred years old. Newer switches, RLCs, etc often measure the current on lines and raise trouble alarms when it's too high or low, which does impose certain tighter bounds.