I can not really give you an intuitive explanation, but economists usually refer to bubbles (or disequilibrium), when actual asset prices deviate from the prices estimated based on the identified long term equilibrium properties of often co-integrated variables.

To identify such models, VAR and SVAR (Structural Vector Autoregressive) analysis are often used, where up to 15-20 time series fed into. Such approaches of course come with their sets of necessary assumptions, but wich are in the case of VAR in levels quite reasonable...

Basically the assumption behind such kind of analysis is, that in the long run, there is some kind of equilibrium path were forces of nature/system are pushing towards to, but to identify disequilibria, you need to look at the whole system and not only 1 or two (as a ratio) variables.

An interesting application about identifying "bubbles" in housing prices with VAR you can find here [1].

[1] https://docs.google.com/viewer?a=v&pid=sites&srcid=ZGVmYXVsd...