The inventory of homes for sale continues to shrink. There were 2.02 million homes listed for sale at the end of October, representing only 4.3 months’ supply. A balanced market would be closer to 6-to-7 months’ supply. From a year ago, the raw inventory count was down 4%, which marked nearly two straight years of decline.

This shortage of housing inventory is the principal reason why home prices have been outpacing people’s income growth for the past five consecutive years. From 2011 to 2016, the median home price will have risen by 42% compared to the median household income gain of only 17%. Such disparity hurts affordability and is unsustainable over the long haul. The only way to lessen home price growth is to bring in more supply. It cannot be a simple case of existing homeowners listing their home. Keep in mind that nearly all home sellers are also home buyers, and thereby not truly providing a net increase to the inventory. The same logic applies to underwater homeowners who come above water after home price gains. What is needed is for homebuilders to boost construction and/or for investors who bought for the purpose of renting to unload those rental properties onto the market soon. There is no indication of the second occurring because of nice rental income flows. The only way to bring additional supply, therefore, is for homebuilders to get really busy.

Economic logic says that about 1.1 to 1.2 million net new households are formed each year. So that is the number of new homes needed to be built just to accommodate this rise in housing demand. In addition, 300,000 to 400,000 old, uninhabitable homes are demolished. Therefore, additional new homes of the same amount are needed just to replace the demolished ones. That puts the logical need for new home construction at right around 1.5 million per year.

In fact, the 50-year annual average for housing starts, up to the year 2000, was 1.51 million units. Quite comforting to know logic and the long-term statistical average matches up. But from 2001 to 2006, which covers the early years of the housing bubble to the peak, housing starts averaged 1.8 million per year – an oversupply. From 2007 to 2016, from the crash and subsequent recovery, the average new home construction clocked in at only 870,000 per year. Over the total big cycle from 2001 to 2016, the average is 1.25 million, and not the prior historical average of 1.5 million.

But as said above, an average of 870,000 new units per year over the past decade imply 8.7 million cumulative new units, when 15 million units would have been needed. Taking the difference between the two figures, the country is short by 8.3 million housing units. Part of this shortage has been absorbed from people moving in to what had been empty buildings and hence falling vacancy rates. But as evidenced by fast-rising rents and fast-rising home prices, we cannot expect a further fall in vacancy rates to handle the ongoing and growing housing shortage gaps.

The bottom line is that we need a few years of above-normal construction activity, say 1.7 million housing starts per year. Only then will we see a slight rise in vacancy rates to help lessen the rent growth pressure and bring the inventory of homes for sale to a more balanced market. However, based on various economists’ consensus projections of housing starts of 1.3 million in 2017 and at best 1.4 million in 2018, if proven true, then we are in for a housing shortage for at least four more years.