Immigration, housing formations create uncertainty for housing trends
A major uncertainty for trend housing starts
In April, I wrote a Spotlight that addressed the question of how a major reduction in immigration would affect trend economic growth by lowering labor force growth. In this Spotlight, I will address what a reduction in immigration could do to the annual growth in net household formations. Household formations are the primary driver of housing start demand. Household formations accounted for 70% of trend housing demand between 1980 and 2010. Thus, a decline in household formations would lower trend for both housing starts and for wood products demand.
Forecasting Net Household Formations (HHF)
Before we dive into the impact of lower immigration on HHF, a brief refresher on what is involved in forecasting HHF might be helpful. The process appears fairly simple. You merely have to take the U.S. Census forecast for population and multiply that number by the propensity to form households to arrive at the annual level of households. The number used to estimate this propensity is called the headship rate. The headship rate is merely the ratio of the number of households divided by the population.
Base Case HHF Forecast for 2015-2016
Step 1: Census forecast of population growth by age group. The Census makes a forecast of population growth every few years after the actual Census is taken (last was in 2010). The most recent population forecast by the Census was made in 2014 for 2014 to 2060. We are only interested in the forecast to 2025 in this Spotlight. The Census population growth forecast made in 2014 is shown in Graph 1. Population growth was forecast to be near 2.6 million per year over the next decade with immigration share rising to above 50% by 2025. The natural increase is based on an estimate of the birth rate (number of babies born every year) and on the mortality rates for each age group.
Since our concern is the growth in HHF, we are only interested in the growth in the adult population (over age 15). In this case, there are only two risk factors. First, what is the mortality rate for each age group. Second, what is the annual level of immigration by age group. The Census estimate for annual immigration by age group is shown in Table 3. Over 80% of projected immigrants are under the age of 44. Thus immigration is particularly important for growth in the younger age groups, with almost no impact on the growth in the over 65 population.
Step 2. What headship rate to use? The headship rate fell significantly between 2006 and 2016 for some age groups. FEA has addressed this question in a variety of ways before. This decline was the reason HHF and housing demand fell so dramatically in 2006-2016. But the focus here is not on what happened or if the headship rate will change again, but rather what will happen if the population growth is lower. So for this Spotlight the headship rate is assumed to stay at the 2016 level until 2025.
An alternative scenario: what if immigration is cut significantly?
Efforts by the Trump Administration to make immigration more difficult, plus a bill introduced by Republican Sen. Tom Cotton last March to cut legal immigration by 50%, mean that immigration will fall short of the 2014 Census Forecast. In fact, the latest update of population estimates by Census for 2015 and 2016 make it clear immigration is already below the 2014 projection. The estimate for 2016 population levels were about 700,000 below the 2014 projection, primarily due to immigration, but also a higher mortality rates for the 35-44 age group. (I am not certain, but could reflect the drug overdose problem that is epidemic.)
While we wait for new population projections from the Census Bureau, we thought it would be a worthwhile exercise to construct a plausible alternative “What If” scenario.
To see the impact of lower immigration, what if immigration were to drop to 700,000 per year, rather than rising to 1.3 million per year (Graph 2). This is not the fully 50% reduction proposed in the Republican plan. (This case has also used the more recent estimates by Census for 2015 and 2016 as well).
Since the headship rate is assumed to be the same as in the base case, the only change is in the population growth by age group due to lower immigration. The shares by age group were used to allocate the reduction in immigration. The overall HHF formation growth was only 200,000 per year lower than the base case over the next decade despite a drop in the population growth of about 500-600,000 per year. This is because 35% of the immigration loss is in the under 25 age group which has a low headship rate. Even so the decline in immigration lowers household formations to near 1 million per year after 2020. This would suggest a trend housing start number of only 1.3 million per year.
It is now clear that the 2014 population forecast by the Census department was too optimistic. Immigration is clearly going to be significantly less than forecast at that time. Just as important is the implication for the age distribution of HHF growth. There is positive growth in the 25-44 age group, but that is offset by the large annual decline in the 45-64 age group. The decline reflects the Baby Boom group moving into the over 65 category.
In the What-If scenario over 100% of household growth after 2020 will be in the over 65 age group. The front end of the baby boom group (born in 1947) is just turning 70. Based on personal experience, people in their 70’s are focusing on downsizing. This means moving to ramblers, maybe smaller, but with smaller yards that are easy to maintain. People that have lost a spouse are moving to condominiums or to retirement communities that offer potential services for aging people.
In other words, these people will be vacating their current larger homes and shifting to a different lifestyle. What this means for the type of housing, the average size and the regional location will be significant strategic issues for the housing and wood products industry over the next 10 years.
Source: LBM Journal Lumber Partner Post