There is no economic mystery in induced demand: it’s Economics 101 once you realize that drivers pay for their usage of roads in time, not money. So when a roadway is expanded, all else equal, the cost of driving on it falls—you’re no longer stuck in traffic; but the lower price (that is, travel time) increases demand for use of the road, just as low prices tend to increase demand in other markets. The result is that many more drivers use the expanded roadway, to the point that they bid the price, paid in time waiting in traffic, back up to previous equilibrium level.
(That said, there are ways in which building more roadways can increase congestion itself. Since land use is closely tied to access, building new highways can spur development in far-off areas, causing people to drive more than they otherwise would. There’s also a theoretical possibility that adding new roads can change the roadway topology in ways that cause people to change their routes in ways that create congestion.)
Or if the tolls are indexed to wealth and/or income.
But, beyond the topical similarity, the cases fall apart. The situation with traffic involves only a small incremental cost being born by each driver, whilst in housing, the internalised costs are large. The drivers of housing inflation -- the interest of numerous parties in increasing asset values (free money for no work!) most especially -- are also highly different.
Actually, that last factor tends to make me suspect that the housing arguments are at least in a significant number of cases insincere.