Household Income or Rents
Which drives which?
A friend challenged me with an assumption I had not previously thought about, that as rents rise, people will increase HH income by taking in sub-tenants. Fair enough idea, so I decided to take a close look, a bigger task than I expected, so this post is a bit later than expected;
My assumption; that tenants look for a better home when they can afford it driving prices higher. ie HH Income rises lift rents. I also accept that landlords raise prices when the market rises, ie prices rise to what the market will bear.
The challenge; as rents rise tenants take in more people who can help to pay. ie Increasing rents drive higher dwelling density, ie more people per household so higher income per household to meet rent demands.
I have already assumed that increasing household income is a good forecaster of rent rises based on the evidence I can see, but I could be simply forecasting Household Income based on rents.
I decided I needed more granular income details, since the only data I have found that is current is Household Income by region, published for every June by the Stats department.
I also needed some detail on household composition, ie does the number of people per tenancy rise with rents? That has a fundamental problem that if it were so, then we would have extraordinary numbers of people per tenancy after 20 years of rent as a proportion of income being static. However take a look at the data.
There has been an increase in people per household over these 12 years, but it is a very small increase. Three or more families per household rose in 2011 but fell in 2018, to 0.65% or 22,317 households. One-family households dominate, sitting at 76% for all census, in fact increasing from 2.3m to 2.6m households.
Rents as a proportion of household incomes have been remarkably stable since 1998;
2006 it was 23.5%,
2013, 24.8%,
2018, 24%
2023, 24.1%
ie very little change, but is it enough to have driven increased household sizes? The household sizes suggest that again very little change has occurred. Yet rents have risen 61% and household incomes have risen 58% from 2006 to 2018.
Note that not all things are equal, so comparing one year with another brings serious issues associated with supply and demand. Were any of the census years impacted by shortage of supply creating a temporary higher rent or vice versa? The impact of Covid can easily be seen in the rent/HH Income chart when rents went up (or incomes down maybe with service industry staff shortages?)
Rent bands compared to HH Income bands in 2018 Census
How many households are impacted by high rents and low incomes? Fortunately Stats department publish some stats that covers this question, but unfortunately the stats raise more questions than answers.
The charts that follow are derived from the Dept of Stats raw data;
Rents divided by Income
Rents shown as a proportion of Household Income grouped by rental bands. The Legend shows the Annual Rent bands, eg of those earning less than $20,000pa, there are a group (2,619 households) paying $36,400pa rent, ie 180% of their Household Income.
Households paying more rent than income total 26,415 households. That seems impossible but the data is from a survey - the 2018 Census, where cross-checks on supplied incomes are not easy. I expect most if not all of these households receive the Accommodation/Income Supplement (AS) plus Working for Families(WfF) tax discount amounting to over $36kpa, yet some may not report AS as income, thinking of it is a payment, not an income. WfF is most likely not recorded anywhere, it’s applied as a tax deduction so many would forget they even get it.
Families that pay more than 40% of income on rent (the rent that is impossible to live on, most likely receive AS and WfF, maybe at a lower level for some. However there are 120,765 households paying more than 40% of income on rent. Note Stats own quote on the quality of this data; “The total household income variable is rated as moderate quality”.
Number of Rentals (in Bands) for each Income Band
The shape of each chart makes sense, ie decreasing numbers as income increases and vice versa. However the outliers do not make sense - 3000 households earning more than $150,000 yet renting at less than $10,000pa ($200pw) and 3000 households earning less than $20,000pa but paying greater than $36,000pa ($600pw).
Note that again the Legend is Rents pa in this count of households in each rental band. Total renting households surveyed was 468,216, ie 100,000 more than the number of bonds which was 367,275 in September 2018, the date of the census.
As you can see the data is barely reliable, yet I have seen it quoted. I doubt we will get reliable data until the lack of quality is exposed. Decisions on affordability and “accommodation” or “income” supplements, will potentially be made based on this data from the 2023 Census.
Conclusion
Data quality fails to provide full proof, but leads to the most likely conclusion that the number of people per household does not increase with rents. A simple test of this theory is that if it were true then people per household in rentals would continuously rise and we know they do not.
Benefits such as the Accommodation Supplement and Working for Families provide support for those on lower wages. It is likely from these charts that many prioritise higher rents over other expenditure, so the benefits may support rents, but may equally support lower wages, we cannot tell.
Rents (MBIE) rise at almost the exact same rate as Household Incomes (Stats) I have reported this relationship in the second chart above from high quality data for over 10 years. Stats: “Weekly household income is the sum of weekly income of all people in the household from wages and salaries, self-employment and government transfers. Average weekly household income is total weekly household income, divided by the number of households.” This history and quality gives me confidence that rents rise in response to Household Income, ie ability to pay.
My rental property management experience is that we monitor the market (hence the data I have) and raise rents to what the market would bear. If we got prices wrong we either lost tenants to a cheaper home (rare because we always kept prices lower for existing tenants) or failed to get tenants on the open market. We regularly tested the market when we advertised by using a price that was at the top end and reducing the asking price quickly if not rented.
I have reached the end of this analysis partly because I want to move on, after a lot of checking, but if you see something odd, please comment below.