ERES Conference Edinburgh 2012 The Rise and Rise

ERES Conference Edinburgh 2012 The Rise and Rise of the Private Rented Sector: insights from supply and demand system modelling in New Zealand Prof Glen Bramley (Heriot-Watt University, Edinburgh, UK Contact: g. bramley@hw. ac. uk; +44 (0)131 451 4605) with Prof Chris Leishman (Glasgow University) and David Watkins (Heriot-Watt University) School of the Built Environment

Overview of Paper • Striking increase in PRS in UK, NZ and other countries - need to understand model change in wider market context • Selective literature review • Theory and expected determinants on demand supply sides • Formal model structure • New Zealand background & context • Local pooled c-s models for rent level and stock share • Regional annual panel model for new let rents (level & change) and lettings flow (+ turnover) • Simulations for NZ using HNZC model - baseline predicted trajectory - impacts of variant scenarios • Conclusions School of the Built Environment

Literature • Relatively little recent lit on fundamental relationships between rental & owner markets – classic US material from 1980 s-90 s • Quite a lot on deregulation impacts, on specific aspects of market behaviour (search, time on market, size of landlord) • Somewhat puzzling differences between major countries e. g. Germany vs UK/US/Australasia in significance of security & regulation • Focus on changing roles of sector in housing different groups including those in potential need School of the Built Environment

Theory & Drivers of Demand • Demand decomposed into: household formation, tenure choice, quantity/quality of housing services - PR demand is +ve rel to hhd form but –ve rel to choice/ability to buy • Affordability & ‘user cost’ of OO imp – prices, int rates, cap apprec? • Potential mobility & mobile groups imp for PRS • Increase in PRS demand reflects (some) demog change (e. g. singles), int migrn, HE, fiscal factors (marg), decl social lets, but unaffordability of OO a key factor • How to model demand: stocks vs flows; quantity or rent (inverted demand); geographical scale (sub-region) & spillovers • Demand may have increased, but so may supply, so rents might not rise so much; rents may also be slow to adjust School of the Built Environment

Theory & drivers of supply • Regulatory regime (UK regulated to 1989, NZ not) • Given deregulation, rents should impact +ve on supply • Opportunity/user cost based on price, interest rates, running costs should be –ve • Potential capital gain, probably based on recent past, expected +ve • Cyclical effects, esp ‘involuntary landlords’ (owners unable to sell) • Overall new build supply (but most supply is from tenure switching) • Particular house types e. g. flats, terraces • Tax regime (not big changes in periods studied) • Wealth distribution (potential investors with spare capital) • Alternative investment returns, eg. stock market, pensions School of the Built Environment

Formal Model Structure School of the Built Environment

Formal model contd. School of the Built Environment

New Zealand – Context & rationale • NZ PRS always larger than UK, despite OO dominance, but reflecting small SRS as well as l. t. deregulation • PRS grew from 20% to 29% 1991 -2006 • PRS main option for low income households in potential need; affordability & security issues • Relatively good data, including 5 yr censuses with rent & income data, and quarterly/annual data on rents & numbers of new lets (Rental Bond system) • Descriptive analyses show large recent and expected rise in ‘intermediate’ renters and better off renters • NZ saw big economic changes in 1980 s/90 s, with much widened inequalities • 2 distinct pairs of models (a) pooled census C-S of LA’s (4 periods); (b) annual time series panel of 15 regions x 19 years School of the Built Environment

NZ Rent Level Model (1) School of the Built Environment

Demand (Rent) findings NZ (1) • Supply (PRS %) has moderate –ve effect (low partly due to substituteability between areas); • Unaffordability of OO +ve • Household growth rate +ve; • Vacancies –ve not signif • Flats strong +ve • Working hhds –ve; • Higher (degree) qualifs +ve • Unexpected results for younger & small hhds (-ve) (reverse causation? ) • Turnover +ve; cities dummy –ve (quality? ) School of the Built Environment

Stock share supply NZ (1) School of the Built Environment

NZ Supply findings (1) • • Strong persistence in stock share (0. 83 after 5 years) Rent level sig +ve (elas 0. 31 stock) House prices sig –ve (elas -0. 15); also mort in rates (-0. 12) Recent price growth weak +ve (stock, 5 yr) New build (dwg growth) –ve (stock mod) but sig 2 nd homes & 2 -bed dwg marginal +ve Ethnic minorities sig +ve (demand effect? ) • Some concerns that specification of this model is not ideal • Simulations based on this show surprisingly large growth in sector School of the Built Environment

within-sample predictions (1) School of the Built Environment

Rent Level Model 2 a School of the Built Environment

NZ Rent Change Model (2 b) Pred & Actual Rent Change Auckland City (below) School of the Built Environment

Results of Model 2 Rents • Modified model based on subregional annual panel data – new let rents, with flows approach to supply (lettings, turnover) • Current active market, better estimate of economic & financial effects, spatial spillovers & lags • Strong effect from lagged rent • Lettings supply –ve • Mortgage cost +ve, unemployment –ve, qualifs +ve • Demographic growth +ve • Dwelling size effects different • State housing +ve – unexpected, but Auckland proxy • Partial adjustment formulation may be better • Time series drivers predominate School of the Built Environment

Private Lettings Supply NZ (2) School of the Built Environment

Supply Results Model 2 • • Model is for flow of lettings, with endogenous rents and turnover Rents +ve but relatively low elasticity Stock share and turnover +ve as expected Spatial effect (-) implies substitutability Vacancies +ve New build strongly +ve (unlike model 1) Recent house price growth –ve not signif Turnover function reflects mortgage int rate (-ve), 2 -bed (+) vacancy rate (+) high qualifications (+), younger households (+), Maori/Pacific (-) • Stock change can be inferred from lettings & turnover predictions School of the Built Environment

Baseline scenarios NZ • Using Model 2 • Rise in PRS share continues • Corresponding falls in OO and State housing • Prices more volatile than rents • Rents pretty flat in future School of the Built Environment

NZ Sensitivity Tests School of the Built Environment

Comments on NZ sensitivities • Poorer economic performance would reduce PRS % in long run, while reducing rents a bit (altho’ affordy worse), & vice versa • Worse recession would have bigger short run effect but smaller persistent long run effects, in same directions • Adverse financial conditions would increase PRS % & raise rents slightly; vice versa for favourable financials • Low supply of new build would have modest effects, reducing PRS %, & vice versa, with minor effect on rents - however current model suspect for low impact of supply on prices. . School of the Built Environment

Conclusions • • Revival & growth of PRS a big story in many countries, but not adequately understood or modelled yet Possible to develop reasonable demand (rent) & supply models at subregional scale, with mainly plausible effects Explaining rise is easy, but relative importance of factors not clear a priori In NZ (like England) both demand supply shifted out, meaning bigger sector accompanied by only modestly higher rents; timing different Demand side factors of affordability (inc credit rationing), mobility, income levels, declining SRS (UK) are supported; also demographic growth (but household type effects less clear) Supply side factors include rents as expected (moderate elasticities), but role of house price level & change less clear, with limited support for capital appreciation motive; new supply and house types are generally significant; difficult to capture wealth effects but alternative investment opportunities probably important Future prospects are probably for more growth in sector with only moderate rent rises Composition of sector broadening as scale increases School of the Built Environment