How can prices be compared over time?

Such estimates have actually been made.

Roger Bagnall in a recent book writes that the price of a sheet of papyrus was "something like a quarter to a third of the value of the food for an active adult for a day". Or, to put it more succintly "a sheet of paper cost you as much as a hamburger". (p. 134).

A similar estimate (probably derived from the same sources) is quoted by QuintusCinna Cocceius from a book by Harris here:

Papyrus was extensively used by the elite, and all well-to-do Romans were familiar with it. But in spite of some assertions to the contrary it must have been quite expensive for most people's purses, currently outside Egypt, which remained the main source of supply. The price at Tebtunis in the period 45-49 seems normally to have been four drachmas a roll, and a single sheet might cost two obols- this at a time when skilled labor earned about six obols a day, unskilled three. The price is analogous to one of, say, thirty to thirty five dollars for a sheet of paper today. The real price of papyrus would have been much higher in Greece or Italy, not to mention Spain or Britain, than in Egypt.

Also, in another book by Bagnall the sixth-century cost of a single gospel book is estimated to be a third of a solidus which he renders as about 1,000 dollars in today's money. (See here)

To declare I don't have a definite answer, but with my answer I might get you somewhat closer to the ultimate method.

To estimate a value back in time is especially hard because of the following factors:

• There were various monetary systems (not just currencies) over time. For example the silver/gold backed 1 Krone in Austro-Hungarian empire worth 0,3 grams of gold back in 1892 which is today roughly 9,5 EUR. But since then the Krone collapsed, in Hungary the Pengő collapsed, in Austria the Schilling inflated then switched to EUR. You can use any inflation calculator which will be more or less accurate, but I am sure they don't work back into middle ages.

• Gold and silver were used for money for more time than they weren't. You can measure things in gold and silver, these two commodities were money for very long time, and they were rare. The measurement won't be precise since as other people wrote too: the historical exchange from gold to silver was usually close to 1:15 to 1:16. Mike Maloney conducted fine research on gold and silver, you can watch his videos on youtube and read his books. That surely helps understanding the mechanics of gold/silver based monetary system and its relationship to the modern monetary system. Be careful with the comparasion, in 1900 there were less than 2 billion of people, and not mentionable use of industrial gold, today, the extraction is higher, but more costy, we now have industrial use of gold and soon we will have 8 billion people.

• As you said: you can measure things in labour. The only common ground value is the uneducated AND two handed hard labour's value, since in that case you don't have to count with the expertise extra fee of education, cost of machines etc... Take note that in 19th century people worked harder in sweatshops in Europe than today's Europe, so you might need to compare first the economic value of a normal 8 hours workday with a 12 hours workday on the field, which didn't happen in winter since they had no work with the frozen ground.
  The productivity increased by the new scientific discoveries, more advanced machinery etc...

• property, land etc... these are also problematic to compare since the population increased a lot, but the available territory was pretty much the same.

If we take these measurements to account, you can get a certain precision comparasion, but never very accurate.
If I would need to make an estimate, I would take some sort of an average of the following:

1. Gold value back in time compared to 1900's USD (backed by Gold) and then use inflation adjustment.
2. Labour value between the two eras, first I would need to find out for example how many hours a roman peasant worked on the field in an average year and use it as base in comparasion to modern days work.
3. property, land, etc...
4. other methods (?)

I would say, this kind of comparasion requires tons of time to research for each eras.

If I have to take a method, I would choose a cost efficient version. It would be the first method since history of Gold usage for money. Take the comparasion between a fixed Gold standard USD from the very beginning of the 20th century (let's say 1900) and then compare that USD to modern USD inflation adjusted. For silver, be careful, since the gold-silver ratio was fixed for long time, but was released in the middle of the 19th century.

And an additional chart for Gold price history

Understanding historical prices is a difficult problem, and even historians of economics use methods that I would consider inadequate. Many published estimates of ancient prices and value are often wildly wrong. Regular historians (not economists) are even worse at value estimation and rarely suggest modern equivalents with any meaningful accuracy.

There is actually a very simple way to relate ancient values to modern values and that is through the medium of gold. It is an approximate measure, obviously, but much better than the typical methods you will find economists using.

For example, in Pompeii which was covered in ash in 79 A.D. a loaf of bread was listed at 2 asses (an as is a copper piece). If you work out the gold equivalent (the aureus), it comes to about $1 today, which interestingly enough is about what a loaf of bread costs today (on sale). In the Koran it lists the price of a sheep in gold dirhams which comes out to $400 which is about what it costs for a sheep today.

In general, everything is cheaper today than it was in the past due to mechanization, and the effects are not equal. For example, in the old days silver was about 1/20th the price of gold, but nowadays due to the Comstock lode and other factors (such as the silver recovered from copper mining), silver is only about 1/60th the value of gold.

You can also relate money via work or wages. In other words, you can say: making a Roman tunic took an ancient tailor 3 hours. If we convert this to a modern skilled worker wage of $20 per hour, the tunic is "worth" $60. So, value is determined and synchronized as a unit of a worker's time. Of course you have to remember that different civilizations have different wage scales. For example, workers in ancient Rome would make a lot more than the same worker out in the boonies in Britain, just like today a worker in Manhatten will make more than some guy in Nigeria.

Economic history's great attempt to produce a great price/wage series for Southern England's monasteries, agricultural labourers, and modern workers failed. It leaves us with the series used by Measuring Worth, but the series is untrustworthy:

1) Wage labour was not generalised until the 19th century, this makes "wages" and "prices" meaningless.

2) Series construction raises great problems

3) The reliance upon southern English monestaries' as a record source means a poor generalisation even to people in receipt of wages of account; and cannot really be generalised out to the surrounding local markets.

So there's no way to reconstruct a wage price series beyond 1700 with any confidence.

There are certainly goods that still exist in comparable quality (I could think of gold, sheep, cows, stone or ground), the ratio might be different.

No, these weren't "goods" in the past, their social meaning was entirely different. We can try to evaluate their social importance by understanding the past economic and social system as it was, but a comparison can only be made in terms to evaluating and comparing concrete past societies. We can't run a sheep/cow index any better than wage/price.

As soon as we have coins, we can probably compare what people have for a living (per month).

Wages were not socially determinate before 1790, and then only in England. Wage labour, and the "wage" as we understand it as the only basis of subsistence, is a very very new thing.

Finally, what are we comparing? Paper for personal consumption? Paper as a production object in the mode of production? Paper as a dedication of social resources? These are all very different time meanings of social worth.

I relate to economic history as a modern social historian, but reading Measuring Worth's discussion of the problems with post 1950 wage/price series ought to be illustrative, as should the history of the Australian C series index which is available on Australian Bureau of Statistics' website for the Australian CPI index.