First, you should determine how you are going to judge film speed. There are a few methods.
Many – probably most – people like to use the very lowest values as a way to judge film speed, and then tweak development to match a predetermined contrast of their (or usually of Ansel Adams') choosing. They usually do this if they routinely practice tonal placement, because to do this accurately, calibration to a low tone is needed. It is these people who definitely should rerate their films based on individualized testing in order to get the results they desire.
Others – like myself, most of the time – like to develop in a certain way, and then use the printed values of the mid tones to decide how to rate the film. In this method, shooters are generally using an incident meter, and use knowledge of where low and high tones will fall in relation to a placed mid tone (placed via incident metering) to decide how to alter exposure and development to change the way the scene is captured.
The first method forces any film and developer combination to roughly fit a set idea of what a negative should print like. The second method is a little more laid back, and you simply analyze the way different films behave when given different development, and choose a film/development combination that suits what you need for the shot. For instance, you shoot one emulsion, and develop it five different ways, and you have five tools from which to choose to capture a shot.
Because of the preference given to shadow detail and texture by many photographers, the method of testing the low tones a la Ansel Adams is usually what people use. This method is almost certain to produce results indicating that you should use an EI lower than box speed, though I do uprate one b/w film and developer combination. (With new T-Max 400 in D-23 1:3, I rate the film at 640). This is not because the film makers are lying. It is just that different criteria are being used to judge the film speed.
With the midtone method, the box speed is usually right on IME, or within one DIN number. I shoot most films at box speed with incident metering. (I shoot Kodak Portra 400NC at 500 in order to get the mid tones as dark as I think they should be, and I also often push this film 1/2 stop or one stop, even when I have plenty of light to expose it "properly.") Because of this, I would make the general statement that box speed will be acceptable to use without a lot of EI testing, if using an incident meter, though you can always benefit from testing. I would definitely recommend testing for the range of your film, however. If using incident meters, you have to have an eye for what luminance ranges your film will or will not be able to capture in order to get best results.
I think testing low tones for your EI is the way to go if you are using a reflected meter, especially a spot meter with tonal placement, and testing for midtones and learning the range of your film in a few different methods of development is the way to go if you use an incident meter.