Millennia-Scale Inflation is About Negative 0.09%

https://www.researchgate.net/figure/Wheat-Prices-1300-1600-in-Tuscany-soldi-per-staio-and-England-pence-per-125-bushels_fig1_253633731
From 1300-1600, approximately from the earliest reliable records to the age of the Bank of England and fractional reserve, the price of a bushel of wheat was 10p. That's 0.84 shillings, or 4.7 grams of silver. We can consider that the non-buggered price.

There was also a multicentury period of very stable wheat prices in Rome.
http://web.mit.edu/14.731/www/jep05.pdf

“[Rickman] said that the price of wheat was between three and four sesterces per modius around 150 BCE“

4 sesterces = 1 denarius, denarius = 0.84 shillings worth of silver. 

A modius is about 1/4 of a bushel though, so wheat in Rome was about four times the price of wheat in England. 

The price did go up in the next few centuries, and that's reflected in the above link, but only because they were debasing the denarius. (Even then the debasement was sub-1% per year on average, though of course Rome debased in job lots rather than every day. If you run the numbers you'll find it's pretty much the same weight of silver per bushel.) You can check because they later debased like crazy and prices went crazy as expected.

If you do 400% deflation over 1600 years you get that the price of silver denominated in wheat grew by 0.09% per year. Over a normal 70-year lifespan prices would go down by about 6%. Obviously these are rough figures, but the data isn't all that accurate in the first place.

About 1/3 of the grain shipped from Egypt, so a lot of that was shipping costs.

 

In conclusion long-run prices are indeed very stable. Stuff is worth what it's worth. But not exactly stable: economic growth does indeed lead to deflation. Gregory Clark made a chart of long-run economic growth and published it in Farewell to Alms, which you can refer to if you want to check my ballpark estimate against someone who isn't entirely lazy. I'm confident it will be thematically accurate. 

 

Speaking of, let's do Egypt. http://www.touregypt.net/featurestories/prices.htm

1 sack of wheat (c.58 kg) 1600BC to 1100BC, very roughly:
1 to 2 deben
1 senyu = 7.6 grams silver = 5 deben

58 kg => 2/5 7.6 silver = 3.04 grams.
1 bushel wheat = 27 kg, so 1.42 grams silver.

Way cheaper than England. About 1/3. However, we can see this isn't a good apples-to-apples comparison because the Nile Valley is absurdly fertile, which results in very high supply and thus very low prices. That's why the Romans conquered it for cheap grain. In other words we expect Egyptian grain prices to be very low, but we have no measure of how low, exactly, so we can't factor it out. Given ancient shipping was often oar-powered, obviously Egyptian prices had to be less than half Roman prices, but at the same time we can see Roman prices are fairly representative of Mediterranean/European prices in general.

 

Fun fact at very small numbers percentages work like adding. 0.09% growth * 0.09% growth ~= 0.09% + 0.09%. As a result if you want the prices of wheat in silver it went down by 0.09% per year, 99.91% as high. Although if you multiply that through all 1600 it goes to 25% as per the correct ratio.
Given the large exponent if you want anything that's really accurate you would need many many more significant figures, and given the price reports have uncertainties of at least ±33%, good  luck with that. 

Next fun fact: seems every European and even North African society had a denarius/drachma/shilling/senyu of about the same size hence performing the same social function. Until someone diluted the silver in it or otherwise assbushed the economy.

 

P.S. With great futility, I dream of a world where I don't have to do all my error-handling in-house.

I can't help imagining a world where I face rebuttals when I'm wrong, instead of when I'm not wrong.
I suppose it comes with the territory if you decide to be a sacred cow butcher. You only get a response if the cleaver hits a real sacred cow.