The benefit isn't in the files. Once you've loaded a file into a hash table using the /hload command the data is stored in memory, which means it's several orders of magnitude faster than accessing a file directly each time.
Alternatively you could read the contents of a text file into multiple global variables in order to speed up access, but it's generally uglier in terms of code and you lose efficiency of the powerful built-in searching capabilities available to hash tables via $hfind.
