I would like to see a graph of the cost of 90th percentile weddings, private college tuition, homes and family assets over the past 25 years.
My hunch is that the graph would show that they were all strongly correlated, and that the cost of 90th percentile weddings, private college tuition (at Williams , for example) and homes is determined primarily by available assets, not by the costs of delivering the essential functions of social union, teaching, and shelter.
I'm sure this has been done, but not being an economist I don't know how to search for it. I hope Brad DeLong will pick up the topic sometime, because it has some interesting implications. Imagine you're running a private college competing for the wealthy (who fund tuition for your talented poor). Your tuition will be cost + "profit". Your cost must cover your fixed costs (faculty, etc), but the true driver is your cost of recruitment, which manifests as recreational facilities, aesthetic experiences, etc. Since your target market has essentially unlimited resources but a limited membership, your costs of recruitment will rise until the tuition costs hits a stratospheric limit.
In evolutionary terms, this is similar to the "costs" flowers pay to attract bees.
So Harvard's full price tuition will be over $1 million within 10 years ...
 Highly recommended if you are poor or wealthy. I very much appreciate the wealthy students who paid for my visiting-student time there.