Rodney Dickens doubts landlords' costs drive rents, suggesting it's more likely income growth does. But he finds there is confusion in the data with the biggest threat from inflation

By Rodney Dickens*

With the government targeting housing investors and landlords, rental housing has become a bit of a hot topic and it shouldn't be a surprise there have been claims the result will be tenants facing larger rent increases.

However, the analysis of how rents behave in this Raving suggests the government's policies aimed at boosting wages are likely to be more important in offering potential for landlords to put up rents at a faster pace.

By contrast, there is limited if any link between changes in mortgage interest rates and changes in rents. The implication is that the inflationary path the Reserve Bank is currently set on will ultimately be much more painful to landlords and housing investors than the government's policies.

This Raving reveals major discrepancies in rents reported by MBIE and Statistics NZ that can't be explained by the former only covering private rents and the latter including rents for private dwellings and social housing. Despite trying to unravel the large discrepancies I couldn't, which left me in the somewhat uncomfortable position of not knowing how much the private rents reported by MBIE versus the market-wide rents reported by Statistics NZ are providing misleading indications.

The rental bond data reported by MBIE suggests the rate of turnover by tenants has more than halved over the last 25 years. This seems more than is likely which may further add to questions over the rental data reported by MBIE that is based on lodgements with Tenancy Services. But maybe this is what has occurred?

Finally I look at whether rents have increased more or less than incomes. The answer depends hugely on what data are used in the assessment making this topic a goldmine for those wanting to mislead because they are pursuing a political agenda.

Are tenants about to face larger rent increases as a result of Labour's policies?

With the government introducing policies favouring tenants over landlords and making life tougher for housing investors, rental housing has become a hot topic. There have been some suggestions the pace of upside in rents has increased or is about to increase.

Similar claims that the pace of upside in rents is about to increase have been made on a number of occasions over the last few years especially about Auckland. But based on two national measures the pace of upside in rents has changed little since the 2008/09 recession had a temporary negative impact (chart below). The disparity between how much rents have been increasing based on the mean rents reported by MBIE that are based on bonds lodged with Tenancy Services versus the increases reported by Statistics NZ in the CPI rent component are discussed later.

By putting some people off being landlords and increasing costs for landlords Labour's policies may mean the pace of upside in rents increases a bit but there could be a larger impact from the policies aimed at boosting wages. This is because of the change in the behaviour of rents that has resulted from the new, low inflation environment that started in the early-1990s.

The evolving behaviour of rent inflation

Prior to 2001 rent inflation based on the CPI rent component was quite closely linked to house price inflation if the temporary impact the 1982-84 wageprice freeze on rental inflation is ignored (next chart). However, after 2001 house price inflation continued to experience major cycles while inflation in rents became quite stable.

The introduction of large subsidies for Housing NZ tenants in the 2001 March quarter was why annual rent inflation turned temporarily negative in 2001. The CPI rent component includes private sector tenants and tenants of Housing NZ and other social housing providers.

In the 1980s inflation in general was high including for house prices, rents, incomes and prices in general as measured by the CPI as shown in the chart above and the next two charts.

In the 1990s a low CPI inflation error emerged but house prices and rents continued to be largely linked and be much more cyclical than CPI inflation. Rents were still largely linked to house prices as a legacy of the earlier high inflation environment. However, by the time the 2000s boom in house prices arrived, low CPI and income inflation had existed for long enough they had taken over as the major drivers of rent increases; aided by how rents were set for social housing.

The chart below shows the reasonable link between rent and average hourly earnings inflation in the 1990s with both impacted by the wage-price freeze, the temporary delinking in the 1990s when general and wage price inflation had slowed but rental inflation still seemed to be linked to house price inflation and the subsequent period in which rent increases have largely returned to being linked to growth in average hourly earnings.

It makes sense that there is some link between rent increases and increases in prices in general incomes in part because of how rents on social housing are adjusted.

Based on the 2013 Census there were 453,132 rental properties including private rentals and social housing while based on the MBIE data there were 339,534 active rental bonds in June 2013 that covers just private rentals. At face value in 2013 around 75% of total rental properties measured by the Census rental number were private sector rentals. Qualifications apply to both numbers:

• In the 2013 Censes there were 97,053 households that didn't specify whether their properties were rented or owned, some of which will have been rented.

• It is an offence not to register a bond with Tenancy Services for private rentals but it is possible bonds aren't lodged for some rental properties (e.g. where people are renting from family or friends).

While it makes some sense that rents on social housing are linked to the likes of the CPI and increases in base pay measured by average hourly earnings the rental bond data reported by MBIE suggest that private sector rents have increased well above the level of inflation in the CPI and average hourly earnings. However, the chart below shows that the annual increase in the mean weekly rent reported by MBIE is reasonably closely linked to inflation in average total weekly gross earnings.

The lack of acceleration in private sector rental inflation in recent years most likely reflects a constraint from income growth that hasn't increased based on total gross weekly earnings inflation from the Quarterly Employment Survey. Total earnings include base pay measured by average hourly earnings and bonuses and other worked-related benefits, better reflecting total worker income.

The chart above suggests that the more Labour's policies boost wage and salary incomes the greater scope there will be for private sector landlords to put up rents. Labour's policies to boost especially wages include the 27% increase in the minimum wage over four years, plans for more pay equity deals and policies aimed at giving unions' greater scope to represent workers etc.

By contrast, there is no clear link between rent inflation and mortgage interest rates (chart below). As covered in our economic and housing reports, the Reserve Bank is set on a path that will cause an inflation problem; ultimately requiring lots of OCR hikes to fix, similar to what occurred last decade.

Interest rates aren't likely to increase as much as over the next several years as they did between 2002 and 2008. But sizeable/painful increases are eventually likely as a result of the Reserve Bank seeming to be set on repeating the mistake it made last decade of allowing labour cost and domestic price inflation problems to develop. This time around it will be encouraged by Labour's policies aimed at boosting wages but also residential building and infrastructure spending.

Why the major discrepancies between the MBIE mean rent and CPI rental component?

As shown in the chart below, since 2002 private sector rent increases as measured by the mean weekly rents reported by MBIE have generally significantly exceeded overall rent increases as measured by the CPI rent component that includes rents on social housing.

Based on the mean rents reported by MBIE for the three months ended March 2018 the average private sector rent increased 4.6% compared to the three months ended March 2017 while the CPI rental component increased only 2.1% over the same period. At face value this suggests that private sector rents must be increasing dramatically more than rents on social housing.

The difference is massive when the cumulative increase in the mean weekly rent is compared with the cumulative increase in the CPI rental component (chart, adjacent column that shows when Housing NZ subsidies impacted).

Excluding the temporary fall in the CPI rent component in the 2001 March quarter caused by large subsidies for Housing NZ tenants, since 2001 the mean rent reported by MBIE has increased on average by 4.5% per annum versus a 2.5% for the CPI rental component (arithmetic not geometric averages). On the basis that social housing makes up 25% of rental housing the implication is that rents on social housing have fallen by an average of 2.2% per annum since 2001. This could occur in one year if Housing NZ increased rental subsidies but there is no way that social housing rents have fallen 2.2% per annum on average since 2001.

As Statistics NZ clarifies:

The housing rentals component of the CPI includes private and local authority rentals, State rentals, and student accommodation. The CPI housing rentals index measures changes in the rents of all private, local authority, and State houses – not only those for which new tenants have recently been found.

By contrast, the MBIE rental bond data only reflect rents for newly lodged bonds each month for private rentals.

Rent increases for social housing will have been less than those for private sector rentals but it takes much more than this to explain the discrepancy between annual inflation in the mean rents reported by MBIE and in the CPI rental component compiled by Statistics NZ. The difference has meant the mean rent reported by MBIE has increased 46% since 2001 versus a 112% increase in the CPI rent component (chart above).

From year-to-year the mean rents reported by MBIE may overstate how much private sector rents on average increase. Based on the number of new rental bonds lodged with Tenancy Services each year compared to the active number of bonds at the start of the year, in the last year around 42% of tenants changed rental properties. Mean rents only reflect new rental bonds. They should give some reflection of what is happening to private rents at the margin but if sitting tenants on average face smaller rent increases than tenants who change rental properties then the overall increase in private rents will be less than the increase in mean rents for new bond lodgements. Statistics NZ alludes to this potential shortcoming of the mean rents reported by MBIE when it states that the CPI rental component is "not only those [properties] for which new tenants have recently been found."

This can explain why private rents on average most likely increase less in any one year than suggested by the mean rents reported by MBIE. But it is questionable whether it explains all of the large difference in the increase in the mean rents and CPI rent component since 2001. Over time mean rents reported by MBIE should reflect what private tenants in general have experienced in terms of rent increases. It is likely that tenants who don't change property for one or more years end up facing sizeable increases to catch up with current market rents when they do change properties. Over periods of several years and more the mean rents reported by MBIE most likely provide a reasonably accurate indication of how much private sector rents have increased on average.

I am left with a sense of disquiet because I haven't been able to explain to my satisfaction why there is such a large discrepancy between the mean rents reported by MBIE and the CPI rent component.Setting aside my disquiet, to the extent government policies boost the pace of growth in employees' incomes the greater scope there should be for landlords to increase rents. I suspect that this more than landlords responding to government policies aimed at landlords and housing investors will justify some upside in the pace of rental inflation.

The MBIE data may reveal an interesting evolution in the behaviour by tenants

The percentage of tenants changing properties each year has fallen dramatically since 1994 based on the rental bond data (the next chart). While in the 12 months ended March 2018 there were 161,871 new bonds lodged compared to a total of 380,725 active bonds at the start of the year (i.e. approximately 42% of tenants changed properties) in the 12 months ended March 1994 99,551 bonds were lodged compared to 95,082 active bonds at the start of the year (i.e. there was slightly more than 100% turnover of tenants during the year).

The chart below suggests tenants have become progressively less inclined to change properties over the last 25 years. Having only become a landlord relatively recently I'm not sure if this is what has occurred or whether it reveals something fishy about the rental bond data?

If I had the time I would check the Census data to see if turnover rates by tenants had increased significantly since 1994, assuming it would reveal this. I'm not going to in part because Statistics NZ doesn't make it easy to extract data that covers numerous censuses.

You can tell whatever story you want to depending on the numbers you use

Have rents increased more or less than incomes? It turns out that the answer depends on the data you use (chart below). To have a bit of fun the chart compares how much house prices, incomes, rents and prices in general have increased since 1990 including two measures of incomes and rents. Probably of most important, mean rents reported by MBIE have increased significantly less than average total gross weekly earnings.