Credentials, yes. There is some signal there, no doubt. There's no question that the average Yale or Harvard student is going to be better than the average X student for almost all values of X. Most people agree with this. The problem is more in society's attempts to enforce a total ordering. How would compare 75th percentile Penn State vs. 50th percentile Harvard? I actually have access to numbers on how strong students are across schools, and the 75th percentile Penn State student is a better bet. (Student quality-- work ethic plus capability-- in the Ivy League seems to be a N(1.70, 0.65) distribution by most measures; at the state flagships it's about N(1.35, 0.75), so you're comparing 1.85 vs. 1.70.)

Or, here's another question. Which has more positive signal? That Tom went to an Ivy League school? Or that Bob majored in math? The latter, actually. This is borne out, for one example, in LSAT performance. The average Ivy Leaguer scores about 160-163. The average math major scores 165-166.

Connections? No, not really meritocratic at all, because there are social connections and then there's Connections. The kind you work for aren't that powerful. The kind you're born with open doors that are closed to almost everyone else (e.g. getting VC funding without having to take risks with personal finances that only 25-year-olds can afford). Connections are powerful in inverse proportion to the amount of work involved in scoring them. That's what you're missing in that analysis.