You're right that information doesn't perfectly filter through and starts aren't a particularly reliable metric. My intended point was very narrow though: your

You're right that information doesn't perfectly filter through and starts aren't a particularly reliable metric. My intended point was very narrow though: your default assumption should be that changes in expectations of future cashflows have an immediate impact on price, unless you can point to a particular market friction that prevents this adjustment. Basically, you should distinguish between timing of cash flows (which don't generally matter except for discounting), and the timing of the information flowing into the market (which does). If there were a credible plan to build more much housing and the market believed in it, it should affect the price more or less immediately. Discount rates are very low -- $1 a year from now is more or less worth $1 today.