If we did have $30,000 more on average, as the article states, how would it be used? The case is obvious for the poor, but as you go to higher income brackets more and more of the possibilities fall into "luxury" and "status symbol".
Take, for example, the rise in the cost of college education. It's absolutely bizarre that in an era of low fiction information, tuition has actually gone up. That is, until you consider the credentialing as a form of signalling for limited jobs at the high end. The colleges aren't getting funded to expand the core mission of education, they're chasing each other to provide a premium experience that will attract people with a pedigree, who subsequently raise the status of that institution.
From that angle, automation's effect is to shuffle around the job landscape, not to directly increase productivity. More minds on higher value jobs - but eventually we start cutting into the highest value stuff we can think of. What automation doesn't do for us is expand our ability to be creative about what work is and what jobs could be done that aren't. That capability is directed through our social structures and "what people will pay for." Anyone good at conversation knows that you can have the same room and the same people and achieve wildly varying outcomes in discussion. You can have an economy that "shrinks" because less is measurably produced, yet people feel wealthier on average. Indeed, that's central to discussion of open source software and its commodifying effect.
In conclusion, no, we don't know. Social science is a fragile thing and the things that seemed obvious to one generation have a habit of being discarded by the next. Maybe the $9000 of income inequality is, in fact, the important number, even if we can fabricate another bigger number with linear extrapolation.
Your single example is a university education, which is at least arguably a positional good. But tons of luxury items aren't positional.