When the federal government made it effectively illegal to originate mortgages to many families after the financial crisis, the effect of that development was clear in the prices of homes. Low tier home prices collapsed and sales of new homes at low prices collapsed. The combination of those effects meant that the median price of new homes moved much higher than the price of existing homes. The median new home has long had a price about 30% or 40% higher than the median existing home, and that ratio was slowly declining during the housing "bubble" because the "bubble" was mostly facilitating the construction of more affordable homes and the mass migration of Americans out of the expensive, housing-deprived cities. New homes were being built, mostly, where they could be built, and that means they were built where they were cheaper.

When the feds quashed that process with draconian new lending regulations, the median price of new homes shot up to about 70% more than the price of existing homes. For the past several years, that has been moderating. Some of that moderation has been because low tier home prices have done some catching up over the last few years. Some of it might have been due to some recovery in building, but sales of homes under $200,000 is still sitting near cycle lows. So, I don't think the decline in the median price of new homes is due to a compositional shift back toward entry-level units. It must be due to a pullback in buying among the existing, "qualified" buyers with high incomes. The same basic group of buyers are buying the same number of units that they were a couple of years ago but at slightly lower prices. Maybe the intrinsic value of homes was knocked down a bit by the 2017 tax bill. I would say that is a bearish development, even though homebuilding has such pent up demand that it's tough to be too bearish. On the other hand, we just had this blowout number in housing starts for December. Housing will be interesting to watch this year.