The impact of upzoning on housing construction in Auckland∗

Abstract

In 2016 the city of Auckland in New Zealand upzoned approximately three quarters of itscore suburban area to promote and facilitate construction of more intensive housing. We use aquasi-experimental approach to analyze the impact of upzoning on house construction over thefour years subsequent to the policy change. Our analysis permits potential shifts in constructionfrom non-upzoned to upzoned areas (negative spillovers) that would, if unaccounted for, leadto an overestimation of treatment effects. These spillovers are accommodated through a partialidentification approach to the estimation of treatment effects that extrapolates pre-treatmenttrends in the control group of non-upzoned areas to define a set of counterfactual outcomes. Thecounterfactual set permits local departures from linearity to permit nonlinear trends. We findthat treatment effects remain statistically significant even under implausibly large counterfactualsets that include seven times the consents implied by the extrapolated pre-treatment trend. Toproduce a spillover-robust estimate of the additional construction enabled by the policy, wecollapse the counterfactual set to the extrapolated trend, finding that the policy generated anadditional 26,903 consents by 2021. This corresponds to 5.07% of the city’s dwelling stock, andeffectively doubles the rate of housing construction immediately prior to the policy change. Asof 2021, construction of attached dwellings was trending upwards in upzoned areas, indicatingthat the long-term impacts of the policy are yet to be realized.

More information

Identifiers

ORCID for Peter Charles Bonest Phillips: orcid.org/0000-0003-2341-0451

Catalogue record

Export record