"when the government got involved in the student loan business."
Its more general than that. System wide. Same scenario during the same timeframe in cost of housing, cost of day care, cost of automobiles... Anything involving loaning money from a bank and possibly paying it back, or anything where the .gov hands out money to certain favored groups.
"the rise of the ridiculous requirement to have a college degree to get anything close to a decent job"
At least in the USA there are fewer jobs every year. So its not ridiculous so much as market forces. Also note that society wide (ignoring the very narrow IT code monkey situation solely in NY and SF, I mean country-wide) we produce more college grads than we have job openings.
This is kind of an important lesson for HN startups etc. Yes is it true there is a shortage of IT folks in NY/SF. But in the rest of the country, we're rapidly moving from merely needing any bachelors degree to be a receptionist or call center script reader, to perhaps needing a STEM degree for those kind of jobs. I know experienced and productive STEM people lucky to be working in call centers. The lesson is if you intend to scale your startup outside SF/NY you need to face jobforce realities outside SF/NY. For example say you create a startup to sell to education. Don't do your marketing plan aimed at Ed degree holders (maybe a .edu discount or some kind of viral marketing via sponsorship of alumni activities or something) because most of those grads are either working in the hospitality industry as waiters / bartenders unable to get a job, or got a teaching job and burned out in a decade and are already on their next career. So most Ed degree holders are going to be unable/uninterested WRT your startup. You could aim at people actually employed in the field, which will be a pretty small subset of those educated in the field.
Its tempting, if you're in silicon valley, to try to scale, based on what works in silicon valley where you live. The rest of the country that you're trying to scale into is basically a different, failing economic model. Sometimes this is important when scaling, sometimes it isn't.