The paper DoDo linked uses soldier wages to get some grasp. The annual wage for a legionaire during different emperors (table 1, page 13)

Augustus (27 B.C. - A.D. 14) 150 → 225
Commodus (180-92) 300, 375
Septimius Severus (193-211) 400, 450, 500, 600
Caracalla (211-17) 600, 675, 750, 900
Maximinus Thrax (235-38) 1200, 1350, 1800
Diocletian (284-305) 7500, 12400

Using Egyptian wheat prices (table 2) they estimate annual price increases of wheat at 8-9% during the last half of the third century compared to around 1% in the two preceeding centuries.

JakeS:

On the subject of coin debasement, debasement of the metal content of coins is not necessarily indicative of inflation. It could also be indicative of price stability in a growing economy with fixed metal supplies, or a monetization of the economy under fixed metal supplies.

Which is probably what we see during a large period of the debasement (this is also argued by the authors of said paper, in contrast with the common view of a prolonged inflation), while the radical debasement in the second half of the third century was connected with inflation. That inflation presented a problem to the Roman state as its taxes were collected in fixed sums. The inflation was also a symptom of larger political problems that were probably driving down the amount of goods.

Same paper (p29):

It would appear that at the time, devastations caused by civil wars and foreign in-vasions (which led to loss of production, infrastructure, and the decline in inter-regional trade), along with a rise in barter practices, adversely affected the physical volume of transactions, T. This development would have certainly stimulated price increases, within the Equation of Exchange framework. Other factors that could have further boosted price increases are (a) an increase in the velocity of circulation, a slight or modest rise of which is conceivable; and/or (b) an influx of currency (by the Government or rival governments), which has to be explained in terms of (i) pay-rises, (ii) pay-compensations for equal reductions in provisions in kind, (iii) an expan-sion in the numbers of administrative staff, (iv) an expansion in the size of the army. Of these, item (i) is unattested, (ii) and (iii) seem unlikely during the climax of political instability/anarchy, and (iv) is plausible (even somewhat attractive) in that it allows for Diocletian to have found a larger-than- usual standing army instead of nearly doubling it (alluded in footnote 38). In short, we propose that the price increase that transpired a few decades prior to the reign of Diocletian ought to be attributed to a decline in T, and, perhaps (unmeasured, though conceivable) rises in V and M.

They also argue that Diocletian might have triggered 100% inflation in a year by doubling the value of the coins. This being the grounds for his largely failed price edict.

By and large, I think treating Roman inflation as an economic issue misses the political side. The realm was being torn asunder, this had effects on the coins of the realm.

Also to add to my theory about Christianity (which is btw not a part of the standard historical narrative) some link I followed noted that Constantine and his successors got a lot of the silver and gold they minted from the pagan temples they closed. So maybe Christianity was a good religion to embrace partly because of its relative lack of precious metals?

Sweden's finest (and perhaps only) collaborative, leftist e-newspaper Synapze.se