I think those journals charge high prices for the same reason that niche software vendors charge high prices. Namely, the product takes a goodly number of highly educated (and thus highly compensated) people to produce, and they don't have that many customers. So if the business is going to work at all, the prices have to be high. Furthermore, they are taking steps to verify the integrity of what they sell. And of course, the more steps they take, the more the journals are going to cost...
I don't know about that industry, but generally prices don't depend on the sellers' costs. Prices are set to maximize profit, that is to maximize (volume * net profit). Sometimes that's a huge markup (e.g., the $5 soda at the movie theater), sometimes that works out to a loss (e.g., the 2-year-old smartphone on clearance). There are many complexities of course, such as loss-leaders, regulated markets, public goods, exploitative pricing, etc.
I've read that the journals realized that they had a captive audience, i.e., that college libraries had no choice but to subscribe in order to support their faculty, and the publishers raised prices dramatically. There was a backlash from some schools at the time.
These days, scientific publishers are bundling everything together into one package, commonly known as the "Big Deal". Institutions pay a single fee for access to hundreds or thousands of journals.
"I don't know about that industry, but generally prices don't depend on the sellers' costs." What you're saying is true only if the revenue at the optimal price is higher than the seller's costs. If the revenue at optimal cost is lower than the seller's costs, then generally the product won't be produced. Therefore if the product is still being produced, the seller's costs put a lower bound on the revenue. If the volume is fixed, this means a lower bound on the price. If the volume is small, this lower bound can be high.
You are aware that 98% of the highly educated people (i.e. the authors and the reviewers of papers) are not compensated at all? Sure, they're paid to do the research by their employer, but they never see a dime from the journal.
I am aware of that---I've written and reviewed several peer-reviewed papers. Maybe many of the journals are making money hand over fist. But I've become suspicious of arguments of the form "that seems like a high price, therefore they must have huge margins". Sometimes the price is high because the market is small.
