RDP 2018-03: The Effect of Zoning on Housing Prices 1. Introduction

Some government policies, which we will refer to as zoning, restrict the supply of housing. Examples include minimum lot sizes, maximum building heights and planning approval processes. Although these restrictions may confer benefits, they also raise the price of housing. This paper attempts to quantify the effect of zoning on housing prices in Australia's four largest cities.

Anecdotal evidence suggests that zoning can have a huge effect on land values. For example, a 363 hectare site in Wyndam Vale (40 km west of Melbourne) increased in value from $120 million to $400 million following its rezoning from rural to residential (Schlesinger and Tan 2017). Examples like this are common – Appendix A provides more. Such large increases in values as a result of zoning changes are inconsistent with the view that a physical shortage of land itself is the main cause of high land values and housing prices – and instead point towards a high ‘shadow price’ of government permission to build dwellings as a likely explanation. It is difficult, however, to gauge how representative these anecdotes are, or to analyse how they change over time or place.

A natural approach to this issue might be to examine the effects of differences in zoning regulations. However, for reasons we discuss in Section 2, that approach is not appropriate for estimating zoning's overall effect on prices. Instead, we follow the approach of Glaeser and Gyourko (2003) and Glaeser, Gyourko and Saks (2005), which involves comparing property prices to the marginal cost of supply. For detached houses, this means decomposing property values into the cost of the structure and the value of physical land at that location, with the remainder being attributable to zoning restrictions.

To illustrate, and to preview our results, we estimate that the average Sydney house, valued at $1.16 million in 2016, represents a $395,000 structure on a $765,000 block of land.^[1] However, the reason land is expensive is not because it is physically scarce. Hedonic regressions indicate that home owners do not value it, as land, especially highly. Specifically, they value land as worth about $400 a square metre on the margin, or $277,000 for the average Sydney block. This ‘marginal’ or ‘physical’ value of land represents the opportunity cost of extra land, as judged by what people are prepared to pay for it.^[2] This is $489,000 less than its market value. This difference represents what home owners need to pay for the right to have a dwelling at that location, or the cost of ‘administrative’ scarcity. The academic literature refers to this difference as the ‘zoning tax’ because the wedge between prices and costs is induced by government-determined scarcity and to draw an analogy with a Pigouvian welfare tax. We, however, use the less-loaded term ‘zoning effect’, which avoids implying that government revenue is involved.

The difference between the average (or market) price and the marginal (or physical) value of land represents an arbitrage opportunity. In the absence of zoning, an investor could purchase properties where the marginal value of land is lower than the average value, subdivide them to create multiple smaller properties, and make a profit. More concretely, excluding the effect of zoning, the marginal Sydney house buyer could have been supplied with an average house for $671,000 – it would have cost $395,000 to build the structure and landowners (existing or potential) would have been prepared to forego the land for $277,000. Instead, buyers needed to pay $1.16 million. The extra $489,000 reflects administrative restrictions. That is, zoning restrictions raised prices 73 per cent above the cost of supply.

This effect is similar but smaller in other cities. We estimate that zoning restrictions raised the average price of detached houses, relative to supply costs, by 69 per cent in Melbourne, 42 per cent in Brisbane and 54 per cent in Perth. As a share of the total price, these contributions are 42 per cent (Sydney), 41 per cent (Melbourne), 29 per cent (Brisbane) and 35 per cent (Perth). Higher-density dwellings require a slightly different approach. As discussed in Section 8, we estimate that zoning restrictions raised average apartment prices, relative to marginal cost, by 85 per cent in Sydney, 30 per cent in Melbourne and 26 per cent in Brisbane.

Figure 1, a standard supply-demand diagram, may help to interpret our estimates. Point A represents a frictionless market, with the number of dwellings equal to Q_E supplied and demanded at price P_E. However, if restrictions limit supply to Q_Max, the price would rise to P_Zoning as Q_Max intersects the demand curve at point B. The gap between what home buyers pay (P_Zoning) and the marginal cost of supply (P_Supply) is our estimate of the zoning effect. This zoning effect or ‘shadow price’ can be thought of as the price a developer would be willing to pay for permission to build one more dwelling at a given location.

As this diagram indicates, our focus is narrow and partial equilibrium in nature: specifically, we focus on the effect of zoning on prices under current conditions. While this allows us to answer some important questions about the effect of zoning regulations, there are others that we cannot address. In particular, we do not examine effects on quantities, incomes or other measures. Nor do we examine the benefits of zoning, such as the externalities of congestion and overdevelopment. Our approach does not allow us to estimate the difference in house prices compared with a counterfactual of no zoning restrictions (P_E in Figure 1). Estimating this difference would require estimating the slope of demand and supply curves, together with general equilibrium effects. This would be an important exercise for a calculation of welfare costs, but it is beyond the scope of this paper. Kulish, Richards and Gillitzer (2011) address this issue.

Another limitation of our approach is that because we estimate the effect of zoning indirectly, as a residual, any errors in our estimates of the other components of marginal costs will flow through to our estimated zoning effect. This matters because our estimates of costs rest on some assumptions that strike us as reasonable, but are difficult to quantify.

Footnotes

To provide context for international readers, the mean annual household disposable income in Australia was $97,000 in 2015/16 (based on the ABS Survey of Income and Housing). Fox and Finlay (2012) discuss internationally consistent comparisons of housing prices and incomes. [1]

Although the value of physical land is low relative to the sale price, it is high in Sydney relative to other Australian cities. Moreover, physical land values in Australian cities are higher than in some overseas cities, a point we discuss in Section 4. [2]