Impacts of airport traffic volumes on house prices of New Zealand’s major regions: A panel data approach

Abstract

Chinese

The direction and mechanisms of the relationship between airport traffic volumes and property prices are somewhat unclear in the literature. This study adds to that body of knowledge by empirically investigating the role of airports as an essential driver of economic activity by creating employment and facilitating air travel between destinations. The two-stage least-squares (2SLS) approach is employed to investigate the link between house prices and the airport traffic volumes of New Zealand’s three key regions and airports (Auckland, Canterbury/Christchurch and Wellington) from July 2004 to December 2014. The empirical findings of the study suggest that airport traffic volumes positively and significantly influence the urban house prices of New Zealand’s three major regions.

Keywords

airport traffic volumes New Zealand regional housing prices two-stage least-squares approach

Introduction

To date, a substantial body of literature has investigated various aspects of the impact of airport activity on local housing markets. For example, recent studies relating to hedonic analysis of the impact of the externalities of airport activity (e.g. Chalermpong, 2010; Cohen and Coughlin, 2008, 2009; Gualandi and Mantecchin, 2008; Matos et al., 2013; Nelson, 2004, 2008; Pope, 2008; Tomkins et al., 1998) have suggested that airport noise levels and close proximity to an airport have a significant negative influence on residential housing prices in the surrounding area. In addition, other studies (Beron and Murdoch, 2001; Suksmith and Nitivattananon, 2014) have claimed that air quality (i.e. air pollution and visibility) resulting from airport operations will also have a significant negative effect on surrounding house prices. These empirical studies are in almost unanimous agreement that an increase in air traffic volumes at an airport acts as a significant negative externality that reduces property prices. Similarly, the negative effect of airport activity on property prices is reinforced by the findings of Jud and Winkler (2006) who found that the announcement of a new airport hub will reduce property prices near the airport in the post-announcement period. Similar analyses (see, for example, Debrezion et al., 2007; and Carey and Semmens, 2003 ) have been conducted in the context of major roads and railway and have reported mixed results. This is largely due to the location/distance to the road/railway of interest and types of properties in the sample (apartments, houses, commercial/residential). In the current study, the focus is on the time series dynamics of airport traffic to the regional house price as a whole.

It should be noted that a well-developed and connected key commercial airport within a country/city acts as a central node that facilitates air transport movements (air passenger and air cargo traffic), tourist flows (domestic and international) and trade volumes (Appold and Kasarda, 2013; Tsui and Fung, 2016a, 2016b), as well as creating jobs and employment (e.g. Appold, 2015; Ding and Knapp, 2002; Mukkala and Tervo, 2013; Nelson, 1980). Importantly, an airport has long been considered to be a crucial generator of economic activity, and increases the attractiveness of a country/city for business development as air services are better able to connect territories by transporting travellers and cargo (particularly perishable and/or valuable goods) between destinations (e.g. Bilotkach, 2015; Cidell, 2015; Green, 2007; Percoco, 2010; Tsui and Fung, 2016b; Zak and Getzner, 2014). In addition, Graham (2008) and Karlsson et al. (2008) point out that an airport contributes to the economic development of a country or a city via four avenues: direct impacts, indirect impacts, induced impacts and catalytic impacts. However, the impact of an airport on economic development is mainly dependent upon the areas served by a given airport (Percoco, 2010). Furthermore, Button and Taylor (2000) suggest that international air transportation has four types of effects (i.e. primary effects, secondary effects, tertiary effects and perpetuity effects) on the local economy.

A growing body of literature also examines the impact of airport activity on metropolitan employment. For example, Hakfoort et al. (2001) found that the ratio of direct employment to indirect and induced employment at the Amsterdam Schiphol Airport is approximately 2 to 1. Brueckner (2003) provided evidence of the link between airline traffic and employment in metropolitan areas in the USA. In addition, Richard (2008) also found that 1 million passengers per annum generated 500 on-site jobs at German airports. These studies have important implications for metropolitan economic development emerging from airport activities, and suggest that airports go beyond merely flight operations (take-offs and landings) but also evolve into major business enterprises with spatial impacts and functional implications (Freestone, 2009). Importantly, airport activity (particularly passenger activity) is claimed to be a strong predictor of employment and population growth (Green, 2007). In addition, McMillen (2004) studied the impact of the building of a new runway at the Chicago’s O’Hare Airport and suggested that airport expansion may increase property values in the densely populated area surrounding the airport, as well as raising service-related employment and job creation in the Chicago area. Also, the increased economic activity from airport runway expansions led to an increased in employment (Tittle et al., 2013). However, Tomkins et al. (1998: 244) emphasised that ‘airport expansion and associated growth in income and employment place significant pressures on the local environment, not only from additional passenger throughput and air traffic movements, but also from expanding freight transit and maintenance operations’.

The significance of the link between airport activity and the price of houses surrounding an airport and the link between airport activity and the economic development of a country/city have been studied in previous research. In urban economic theory, a well-functioning regional airport being served as a central point will increase the city’s land values because of agglomeration economies and positive locational externalities, although property prices in the immediate residential areas may be depressed as a result of airport traffic congestion and noise. However, the research regarding the relationship between an airport’s traffic volume and regional house prices surrounding the airport has been scant in the extant literature, particularly regarding the direction of influence. One of the possible reasons for this is the issue of simultaneity during econometric analysis (Green, 2007); furthermore, its direct impact on the housing market is also ambiguous – unlike to the negative effects of the externalities identified by hedonic analysis (e.g. airport noise levels and air pollution). In the housing literature, the growth of residential property prices is dependent on many factors including the growth of population, income, employment, rents, building costs as well as land supply. National variables such as Gross Domestic Production (GDP), interest rates and exchange rates are believed to affect the local housing market as well. More recently, Leung et al. (2013) found that the New Zealand housing market is also influenced by changes in international commodity prices.

This paper overcomes the simultaneity issue by employing a two-stage least-squares (2SLS) analysis of New Zealand’s three major regional and international airports (i.e. Auckland, Canterbury/Christchurch and Wellington) whilst controlling for all necessary house price determinants at both the local and national levels. This approach allows us to correctly model the causal link between house prices and airport traffic volumes. The findings suggest that airport traffic volumes are indicators of local housing market growth. The contribution to the literature is twofold: first, it is the only study to date to study this unique linkage in detail; and second, we provide a thorough empirical analysis that uncovers causality in the airport traffic volumes and regional housing prices relationship.

Overview of the property indexes of New Zealand’s major regions and airport’s workload units, and descriptive statistics

The property index is obtained from the Real Estate Institute of New Zealand and reports the stratified median house price in the corresponding region. Developed in conjunction with the Reserve Bank of New Zealand, the stratified median adjusts for anomalies in sales across suburbs and is an ideal measure of the market trend in house prices.¹Figure 1 displays the time plots of the property indices (stratified median house prices) of three key regions of New Zealand (Auckland, Canterbury/Christchurch and Wellington) and their respective international airports’ workload units (airport activity) between July 2004 and December 2014. Figure 2 depicts the boundaries of three key regions in this study. The descriptive statistics for the three key New Zealand regions and their respective international airports are presented in Table 1. As observed from Figure 1, the monthly house price index of the three major New Zealand regions experienced a relatively steady growth rate during the study period. House prices for these three regions exhibited noticeable growth during the period of July 2004 to June 2007, followed by a fall between July 2007 and July 2008 due to the global financial crisis (GFC).

Figure 1.

Time plots of property prices and airport workload units for New Zealand’s three major regions and airports (July 2004–December 2014).

Figure 2.

Boundary of three key regions in New Zealand (Auckland, Canterbury/Christchurch, Wellington).

Table 1.

Descriptive statistics of the property price index and airport workload units for New Zealand’s three major regions and airports (July 2004–December 2014).

Region or airport	Variables	Mean	Standard deviation	Maximum	Minimum
Auckland region/Auckland	Property price index	3561	610.98	5191	2659
Auckland region/Auckland	Workload units	1,199,795	128,259.92	1,606,700	951,459
Canterbury region/Christchurch	Property price index	2969	403.84	3984	2132
Canterbury region/Christchurch	Workload units	496,001	40,254.19	584,930	387,478^a
Wellington region/Wellington	Property price index	3219	322.44	3676	2313
Wellington region/Wellington	Workload units	424,264	35,013.50	500,594	339,385

Notes: ^a appeared in June 2011. The median property price indexes were obtained from the Real Estate Institute of New Zealand (REINZ), for which the index reference point was set as at January 1992 starting from 1000.

However, house prices maintained robust rates of growth after July 2008. In particular, Auckland region’s property market (New Zealand’s largest city-region) presented the largest magnitude of growth when compared with the markets of Canterbury and Wellington during the sample period. This can be ascribed to the geographic location of Auckland with limited land supply for housing, size of its fast-growing local population,² a larger proportion of international migrants, and its role as New Zealand’s key trade and commerce centre, which facilitates local–global links (e.g. Baxendine et al., 2005; McCann, 2009; Shirely and Neil, 2013; Wetzstein, 2007; Xue et al., 2012). This may suggest that the link between property values and economic development is a function of the local economic landscape, which varies in terms of trade patterns and key industries of the city, etc. (e.g. Appold, 2015; Appold and Kasarda, 2013; Ding and Knapp, 2002; Mukkala and Tervo, 2013; Nelson, 1980).

Christchurch is the country’s second largest city and located in the South Island. The Christchurch housing market was characterised by the 2011 Christchurch earthquakes, where about 5% of the city’s housing stocks were demolished through liquefaction. Christchurch house prices dropped immediately after the earthquakes, and gradually rebounded because of a shortfall of liveable housing in the city. Wellington is the capital city and the growth of its housing market has been slow and stable since the GFC in 2007. This is partially due to the shrinking government sector since the National government took office in 2008. Meanwhile, the 2011 Christchurch earthquakes have since increased seismic awareness for residents in Wellington.³ No doubt, regional house price cycles are seen to be driven contemporaneously by regional activity (Hall et al., 2006). On average, Auckland’s monthly property index (3561) was higher than Canterbury’s (2969) and Wellington’s (3219).

There are a total of 16 commercial airports offering scheduled flight services in New Zealand. Looking at the workload units⁴ (airport activities) of New Zealand’s three key international airports, Auckland Airport is New Zealand’s largest and busiest airport as it handled the largest volume of air passengers and air cargoes within the sample. Auckland Airport’s influence and role within New Zealand’s aviation industry is mainly attributed to its hub and gateway status in transporting air passengers and air cargoes between a large number of international and domestic destinations through its extensive connectivity networks (De Wit et al., 2009; Lee, 2009; Tsui et al., 2014). Perl (1998) claims that access to international markets is seen to hinge upon an effective hub that can facilitate competition for freight and passenger traffic. Christchurch Airport is the second busiest international airport in New Zealand, located on the South Island of New Zealand, and its close proximity to one of the prime tourist destinations of New Zealand, Queenstown, attracts many international airlines (e.g. Air Pacific, Emirates, Fiji Airlines, Jetstar, Qantas, Virgin Australia, Singapore Airlines, AirAsia X⁵). Wellington Airport is located in the capital city of New Zealand (Wellington). Its flight connectivity network primarily covers domestic destinations within New Zealand and the key trans-Tasman routes and destinations in the Oceania region (e.g. Brisbane, Gold Coast, Melbourne, Nadi and Sydney).

The average number of monthly workload units handled by Auckland Airport was 1,199,795, followed by Christchurch Airport (496,001) and Wellington Airport (424,264). Christchurch and Wellington Airports are important in their respective regions but only generated a fraction of the air traffic volumes of Auckland Airport (Vowles and Tierney, 2007). In addition, only Auckland Airport maintained an upward trend during the study period and experienced the largest average monthly growth rate (0.43%). The air traffic activities of Christchurch and Wellington Airports exhibited a more stable trend after 2008. Importantly, the plots of the monthly airport activities (which includes tourist flows) for these three key New Zealand international airports displays different growth patterns, alongside high seasonal patterns in traffic flows (Commons and Page, 2001; Gitto and Mancuso, 2012; Tsui and Balli, 2015).

Regarding the growth trends of international air travel of the three key airports, only Auckland Airport experienced steady growth throughout the study period owing to its extensive flight connectivity and coverage,⁶ whereas the growth trends of Christchurch and Wellington Airports remained relatively stable. For domestic air travel, the three sampled airports exhibited differing rates of growth for the study period; Auckland experienced strongest growth in its domestic passenger traffic, followed by Wellington Airport and Christchurch Airport. Note that substantial fluctuations (volatility) in monthly airport traffic volumes are observed for these three key New Zealand international airports. Surprisingly, the global financial crisis 2008/2009 did not significantly impact airport traffic volumes. However, the Christchurch earthquakes of 2011 adversely affected the airport activity demand of Christchurch Airport for the period of February 2011 to January 2012, with the lowest level of workload units (387,478) occurring in June 2011.

Empirical models

Our empirical strategy is to use the region/airport-level panel data to investigate the relationship between property prices and airport traffic volumes in New Zealand. The data set covers the period of July 2007 to December 2014 for the three regions and key international airports of New Zealand (Auckland region/Auckland Airport, Canterbury Region/Christchurch Airport and Wellington Region/Wellington Airport). To control and mitigate any potential omitted variable bias, this study adopts the panel data regression model to analyse the property prices of New Zealand’s three key regions, as this allows us to control for observable region (city)/airport-level covariates and other unobservable factors in the panel aspects of the dataset (Hong et al., 2011; Wooldridge, 2002). However, one of the challenges in modelling and interpreting regression estimates of the property price data is the potential problem of endogeneity confounding any underlying causal relationships that may exist. For example, the presence of an omitted latent variable that influences both the property prices and the regressor of interest (airport’s workload units) will bias the regression estimates and our interpretation of the causal relationship, if any. As such, in the absence of experimental data, it is imperative for us to use instrumental variable(s) to overcome this estimation issue (Wooldridge, 2002). Finding a strong and valid instrument can be a difficult task, as two important conditions must be upheld (Villas-Boas and Winer, 1999):

The instrumental variable is correlated with the endogenous regressor of interest, and

The instrumental variable is uncorrelated with the error term of the regression of interest.

In this study, consistent with Stern (1996) and Mumbower et al. (2014), a measure of airport competition (an airport’s activity concentration or market power) is used as the instrument for an airport’s workload units. Airport competition, theoretically, satisfies the conditions for a valid instrument, as it should have an influence on an airport’s workload units. For example, Forsyth (2006) discussed and found significant links between airport competition and traffic volumes, and highlighted its welfare implications. Cui et al. (2013a, 2013b) claimed that the improvement of an airport’s competitiveness will decrease the market share of other airports. Similarly, the competitiveness of an airport will largely depend on ‘five core factors’: spatial, facility, demand, services and managerial (Park, 2003). Wang et al. (2016) also claim that airline market competition and airline market concentration have had strong and robust relationships with airline traffic volumes in China.

In this study, we analyse the workload of three major airports in New Zealand. When there is little competition, characterised as high concentration in national aircraft movement market share, the major airports are likely to experience relatively higher levels of workload. Similarly, in periods of high competition and low concentration in market share, major airports are likely to experience relatively lower levels of workload. Thus this satisfies the first condition for the adoption of a concentration/competition-type instrumental variable. However, there is no convincing argument that factors influencing property prices are related to the level of competition faced by local airports beyond the channel of airport activities/throughput. As such, it is likely that the second condition of a valid instrumental variable is also upheld. These two conditions will be empirically tested in this study using the F-statistic of the first-stage regression and the Hansen–Sagan J-statistic for overidentifying restrictions.⁷

The fixed-effect panel regression of interest is written as equation (1):

Y = β_{0} + β_{1} A + β_{2} X_{1} + β_{3} X_{2} + \dots + β_{k + 1} X_{k} + ε_{i t}

(1)

where $Y$ is a measure of the property prices of the region where an airport is located; $A$ is the measure of an airport’s workload units; $X_{i}$ is the control variable i; $β_{0}$ is the constant term; $β_{1}$ is the slope coefficient of $A$ ; $β_{i}$ is the slope coefficient of the control variable $X_{i}$ ; and $ε_{it}$ is the error term.

Note that the control variables included in this study are those typically employed when modelling property prices: (1) national economic variables, namely national gross domestic product (GDP) per capita, migration, and number of unemployed persons; (2) variables that are likely to influence property investment: the interest rate (the cost of borrowing in a house purchase) and the exchange rate; (3) rent;⁸ and (4) exogenous shocks: the global financial crisis of 2009–2012 and the Christchurch earthquakes of 2011.

As $A$ (an airport’s workload units) is likely to be an endogenous variable in equation (1), the following first-stage regression in equation (2) is estimated using the conventional airport concentration and competition measure (the Herfindahl–Hirschman Index (HHI)) as an instrumental variable:⁹

A = α_{0} + α_{1} Z + μ

(2)

where $A$ is the measure of an airport’s workload units; $α_{0}$ is the constant term; $Z$ is the measure of airport concentration; $α_{1}$ is the slope coefficient of $Z$ ; and $μ$ is the error term.

The predicted prices of $A$ in equation (2), $\hat{A}$ , will be exogenous to $ε_{it}$ in equation (1). If we substitute $A$ into equation (1) with $\hat{A}$ , this can produce an unbiased slope coefficient from which the causality between the regressors and regressand can be inferred. The 2SLS estimators will be compared to the fixed-effect regression estimators to determine if the problem of endogeneity is present. Clustered (region) bootstrapped standard errors are estimated for inference testing in the second-stage regression.

In the 2SLS model, all the control variables in equation (1) are treated as covariate variables that may influence the relationship between property prices and an airport’s workload units. This is due to the complex and often confounding nature of macroeconomic linkages, such as unemployment, property prices and unobservable demand shocks in the local economy (Gerlach and Peng, 2005). As such, $X_{i}$ were not included in equation (2) as instrumental variables. Consequently, it is difficult to infer causality from their coefficient estimates, $β_{i}$ , in the 2SLS estimate of equation (1).

The data used in this study were collected from various sources. Airports’ workload units (airport activities) were obtained from the airports’ annual reports and management. National GDP per capita, regional unemployment (the number of unemployed persons), migration, the supply of new houses, the material costs for the house construction and the median salary were sourced from Statistics New Zealand. New Zealand Airways provided the flight movement data for all New Zealand’s commercial airports (used to compute the HHI Index). The mortgage interest rate and exchange rate data were sourced from the Reserve Bank of New Zealand. The regional house price indices and rental data were retrieved from the Real Estate Institute of New Zealand and the Ministry of Business, Innovation and Employment, respectively.

Empirical findings and discussion

Unit root tests

In order to estimate the panel regression model, all variables must be stationary to avoid the modelling of spurious correlations (Aizenman and Jinjarak, 2009; Alba and Papell, 2007; Leung et al., 2013). We conduct panel unit root tests (i.e. Levin, Lin & Chu; Im, Pesaran & Shin; Augmented Dicky–Fuller (ADF), and Phillips–Perron (PP) tests) to ensure the stationary properties of all the variables included in this study. Table 2 presents the results of the panel unit root tests, and confirms that all the variables are non-stationary except for ln(rent), ln(exchange rate), and ln(HHI Index). We apply first-order differencing to all non-stationary variables in order to arrive at stationary transformations.

Table 2.

Panel unit root tests of the variables of interest (July 2004–December 2014).

Time series and variables		Level				First-order differencing
Time series and variables		Levin, Lin & Chu	Im, Pesaran & Shin	ADF	PP	Levin, Lin & Chu	Im, Pesaran & Shin	ADF	PP
ln(property price index)	AIC	0.788	0.877	0.240	0.163	0.452	0.000**	0.000**	0.000**
ln(property price index)	SIC	0.393	0.681	0.206	0.163	0.000**	0.000**	0.000**	0.000**
ln(airport’s workload units)	AIC	1.000	0.963	0.937	0.000**	1.000	0.000**	0.000**	0.000**
ln(airport’s workload units)	SIC	1.000	0.963	0.938	0.000**	1.000	0.000**	0.000**	0.000**
ln(rent)	AIC	0.392	0.748	0.030**	0.969	1.000	0.000**	0.000**	0.000**
ln(rent)	SIC	0.048**	0.362	0.026**	0.969	0.000**	0.000**	0.000**	0.000**
ln(interest rate)	AIC	0.293	0.652	0.836	0.904	0.001**	0.000**	0.000**	0.000**
ln(interest rate)	SIC	0.267	0.647	0.811	0.904	0.000**	0.000**	0.000**	0.000**
Migration (%)	AIC	1.000	0.861	0.953	0.000**	1.000	0.001**	0.001**	0.000**
Migration (%)	SIC	0.996	0.198	0.029**	0.000**	1.000	0.001**	0.001**	0.000**
ln(national GDP per capita)	AIC	0.809	0.993	0.999	0.9999	0.812	0.048**	0.093	0.000**
ln(national GDP per capita)	SIC	0.591	0.986	0.996	0.999	0.000**	0.000**	0.000**	0.000**
ln(exchange rate) (NZD vs USD)	AIC	0.086	0.005**	0.01**	0.118	0.000**	0.000**	0.000**	0.000**
ln(exchange rate) (NZD vs USD)	SIC	0.372	0.064	0.109	0.118	0.000**	0.000**	0.000**	0.000**
ln(unemployed persons)	AIC	0.320	0.474	0.703	0.621	0.006**	0.000**	0.000**	0.000**
ln(unemployed persons)	SIC	0.320	0.474	0.703	0.621	0.006**	0.000**	0.000**	0.000**
ln(HHI Index)	AIC	1.000	0.999	1.000	0.000**	1.000	0.000**	0.000**	0.000**
ln(HHI Index)	SIC	0.000**	0.000**	0.000**	0.000**	0.000**	0.000**	0.000**	0.000**
ln(new houses)	AIC	0.733	0.519	0.602	0.000**	0.000**	0.000**	0.000**	0.000**
ln(new houses)	SIC	0.519	0.651	0.405	0.000**	0.000**	0.000**	0.000**	0.000**
ln(material costs) ^a	AIC	0.936	1.000	1.000	0.2053	0.936	1.000	1.000	0.205
ln(material costs) ^a	SIC	0.002**	0.055	0.000**	0.699	0.664	0.436	0.655	0.016
ln(median salary)	AIC	0.135	0.706	0.612	0.996	1.000	0.020**	0.038**	0.000**
ln(median salary)	SIC	0.341	0.869	0.806	0.996	1.000	0.000**	0.000**	0.000**

Notes: The values indicate the p-values. The test is shown for the specification with the constant term only (no time trend). ** indicates that the rejection of the null hypothesis ( $H_{0}$ ) that the time series has a panel unit root. ^a Two-order differencing was applied to make this time series become stationary. AIC denotes Akaike Information Criterion; SIC denotes Schwarz Information Criteria; ADF, Augmented Dicky–Fuller; PP, Phillips–Perron.

Estimation results

Table 3 presents the 2SLS regression results of the relationship between ln(property prices index) and ln(airport’s workload units), using ln(HHI Index) (airport concentration) as the instrumental variable. It should be noted that the panel model specifications are also based on the within-airport variation (rather than between-airport variation) in an attempt to estimate the unbiased model coefficients. The advantage of within-airport variation is that it permits stronger causal inferences and is more resistant to the endogeneity bias than in the cross-sectional regression analysis (Bilotkach et al., 2012).

Table 3.

Fixed-effect 2SLS models of the property prices of New Zealand’s three major regions and airports (July 2004–December 2014).

First-stage regression
Dependent variable = ln(airport’s workload units)
Explanatory variables	Coefficients
Constant	−1.368** (0.0826)
ln(HHI index)	0.199** (0.569)
Second-stage regression
Dependent variable = ln(property prices index)
Explanatory variables	Model (a) Coefficients	Model (b) Coefficients	Model (c) Coefficients
Constant	0.004** (0.002)	−0.001 (0.004)	−0.001 (0.099)
ln(airport’s workload units) ^a	0.268* (0.144)	0.335* (0.182)	0.427** (0.213)
ln(interest rate)	–	–	0.007 (0.039)
ln(rent)	–	–	−0.002 (0.016)
ln(national GDP per capita)	–	–	1.343* (0.767)
ln(exchange rate) (NZD vs USD)	–	–	−0.031 (0.210)
Migration (%)	–	–	2.923 (7.373)
ln(unemployed persons)	–	–	0.006 (0.006)
Global financial crisis 2009–2012	–	–	−0.003 (0.004)
Christchurch Earthquakes 2011	–	–	0.007 (0.006)
ln(new houses)			−0.009 (0.008)
ln(material costs)			−0.888 (1.233)
ln(median salary)			0.020 (0.020)
Seasonal dummy (1)	–	0.011** (0.005)	0.012* (0.006)
Seasonal dummy (2)	–	−0.000 (0.006)	−0.002 (0.006)
Seasonal dummy (3)	–	0.013** (0.005)	0.014** (0.006)
Seasonal dummy (4)	–	0.013** (0.005)	0.015*** (0.006)
Seasonal dummy (5)	–	−0.008 (0.007)	−0.006 (0.008)
Seasonal dummy (6)	–	−0.016*** (0.005)	−0.018** (0.007)
Seasonal dummy (7)	–	0.023*** (0.006)	0.025*** (0.008)
Seasonal dummy (8)	–	0.017*** (0.006)	0.020** (0.009)
Seasonal dummy (9)	–	−0.002 (0.006)	−0.002 (0.008)
Seasonal dummy (10)		−0.002 (0.006)	0.0005 (0.008)
Seasonal dummy (11)	–	0.002 (0.006)	0.0001 (0.006)
Adjusted R²	0.01	0.17	0.19
Observations	375	375	372

Remarks: *, **, and *** indicate that the explanatory variable is significant at the 0.10, 0.05, and 0.01 significance levels, respectively. Bootstrapped standard errors are contained in parentheses. ^a indicates the predicted value of ln(airport’s workload units) for the second-stage regression.

To better understand the inherent relationship between property prices and an airport’s workload units, we empirically estimate the relationship in Models (a) and (b) by excluding the control variables. Note that Model (b) considers the effect of seasonality on the housing markets in the three regions under consideration. The statistical significance of ln(HHI Index) on airports’ workload units in New Zealand in the first-state regression is reported in the top panel of Table 3. Significance in the first-stage regression facilitates the computation of the predicted values of ln(airport’s workload units) required in the second-stage regression analysis.

A statistically significant and positive coefficient estimate for ln(airport’s workload units) is reported in Model (c) at the 5% level. Together with Models (a) and (b), the empirical findings of this study support the literature’s general perception that the level of concentration at an airport has a significant impact on its air traffic volumes (Bilotkatch, 2015; Bilotkatch and Lakew, 2014). Overall, the coefficient estimate of ln(airport’s workload units) reported in all three model specifications are consistent (albeit with small differences in their significance levels).

Several key observations from the estimation results of Model (c) can be made. First, the statistically significant positive impact of ln(airport’s workload units) on the property prices of New Zealand’s three key regions suggests that air transport activities (air passenger and air cargo movements) passing through New Zealand’s three key international airports can generate more trade/business/related logistical activities and employment opportunities for the respective regions and cities. These transmission mechanisms are expected to have a positive influence on regional housing prices, ceteris paribus. Recent studies such as Bilotkach (2015) and Green (2007) have demonstrated the impact of airport activity on economic development.

Second, the statistically significant coefficient of ln(national GDP per capita) is consistent with standard economic theory, as it is an indicator of the purchasing power of a country (Eckes, 2011; Le Gallo and Ertur, 2003; Yueh, 2008). That is, higher growth in personal income and national economic output is associated with a greater purchasing power within the population’s ability to afford housing (Agnello and Schuknecht, 2011), ceteris paribus. It should be noted that New Zealand’s GDP per capita grew from approximately NZ$35,631 in 2004 to NZ$ 51,319 in 2014, equalling a 44.03% increase (Statistics New Zealand, 2015b). Arguably, the regional per capita GDP may be a better suited indicator of the level of economic development and purchasing power of a region. In unreported results, we re-estimate Model (c) using regional GDP per capita in place of national GDP per capita. The results remain consistent with those contained in Table 3. That is, there is a positive and statistically significant coefficient for ln(airport’s workload units) in the second-stage regression.¹⁰

Lastly, it is surprising to find that the empirical results of ln(interest rate), ln(rent), ln(exchange rate), migration, ln(unemployed persons), the global financial crisis 2009–2012, and the Christchurch earthquakes 2011 have had no impact on the house price indices, ceteris paribus. The empirical results show that, except for national GDP per capita, none of the remaining control variables exhibited a statistically significant marginal covariation with property prices during the study period. This may be due to correlations with airport activities or other confounding factors. Note that the control variables are not instrumented and hence their estimated coefficients do not represent causation. Importantly, the findings of this study highlight the need to incorporate airport traffic volumes/activities when forecasting local housing market price movements in New Zealand cities.

Validity of instruments

An alternate model specification was employed to facilitate formal tests of the validity of the proposed instrumental variable, ln(HHI Index). In order to create an overidentified system of equations, we used the lags of ln(airport’s workload units) as the instruments. The use of internal instruments is possible under the assumption of sequential exogeneity (e.g. Arellano and Bond, 1991; Roodman, 2009; Schultz et al., 2010; Wooldridge, 2002). The second-order lag of ln(airport’s workload units) was used as an additional instrumental variable, as the panel tests for serial correlation reported that serial correlation of order one was present in the model. This invalidates the first-order lag as the valid instrument. The residuals of the model specification employing ln(HHI Index) as an instrumental variable are uncorrelated with the second-order lag of ln(airport’s workload units). Likewise, the residuals of the model specification employing the second-order lag of ln(airport’s workload units) as an instrument are uncorrelated with ln(HHI Index).¹¹ Coupled with the significant first-stage regression results, these results suggested that ln(HHI Index) is a strong and valid instrument.

Concluding remarks

Using the 2SLS model, this study examines the relationship between regional house prices and the airport traffic volumes of New Zealand’s three key regions and international airports (i.e. Auckland, Canterbury/Christchurch and Wellington) for the period of July 2004 to December 2014. The empirical findings of this study reveal that airport traffic volumes (an airport’s workload units) significantly influenced the house prices of New Zealand’s three major regions and cities. This confirms that, whilst airports serve an important role in a city’s transport infrastructure, they are also major generators of economic activities, such as the facilitator of international and domestic traveller flows, and the creator of jobs and employment. Among all covariates incorporated in this study, only national GDP per capita has a positive significant association with the property housing markets of New Zealand’s major regions, ceteris paribus. This may reflect the relationship between higher growth rates in personal income/economic activity and housing affordability.

The main contribution of this study is the comprehensive analysis of the impact of airport traffic volumes on regional house prices using a 2SLS panel model. More importantly, this particular issue has not been extensively researched and explored by prior studies. The 2SLS approach allows this study to produce a robust regression coefficient estimate for the property prices–airport traffic volumes relationship by mitigating the estimation issues caused by endogeneity.

The findings of this study highlight opportunities for developers to use value capture techniques in order to fund the building of housing and off-site airport infrastructure to support an airport’s operation and growth. A good example is the urban development of the new town of Tung Chung near Hong Kong International Airport (HKIA) that was aggressively and successfully developed by property developers through the transit value capture strategy. The railway infrastructure was financed with the ‘Rail + Property’ development programme (Cervero and Murakami, 2009). Owing to its close proximity to HKIA and Hong Kong hub-based airline headquarters (e.g. Cathay Pacific Airways, Dragonair, Hong Kong Airlines and Hong Kong Express), this area is attractive to airline personnel who wish to purchase or rent residential properties, which subsequently increases the house prices of the Tung Chung area. In addition, aviation-related infrastructures have been established in the precinct of HKIA, such as air freight terminals (Asia Airfreight Terminal Company, Cathay Pacific Cargo Terminal), aircraft engineering and maintenance providers (Hong Kong Aircraft Engineering Company Limited – HEACO, Hong Kong Aero Engine Services Limited – HAESL). Similar development of off-site airport infrastructure would likely support the future growth of air traffic volumes of New Zealand’s three key international airports.

In terms of future research, similar studies along this avenue can be conducted on other secondary/provincial cities and regional airports in New Zealand, and to explore whether airport traffic volumes and the growth of the aviation sector in those secondary cities may influence their respective property prices. Moreover, the methodology and the results of this study are likely to be of interest to other countries and regions (e.g. Australia and North America). This is particularly the case for those destinations that have been experiencing rapid growth of airport activities (increasing international/domestic air passenger and air cargo numbers) and rising property prices, such as in Sydney, Australia. Gaining a more accurate understanding of the link between airport workload and property prices will allow policymakers to implement value capture techniques to improve the welfare of the community and make better strategic decisions regarding airport construction and local development.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Notes

References

Agnello

Schuknecht

(2011) Booms and busts in housing markets: Determinants and implications. Journal of Housing Economics 20(3): 171–190.

Airways New Zealand (2015) Aviation movement statistics. Available at: https://www.airways.co.nz (accessed 23 December 2015).

Aizenman

Jinjarak

(2009) Current account patterns and national real estate markets. Journal of Urban Economics 66(2): 75–89.

Alba

Papell

(2007) Purchasing power parity and country characteristics: Evidence from panel data tests. Journal of Development Economics 83(1): 240–251.

Appold

(2015) The impact of airports on US urban employment distribution. Environment and Planning A 47(2): 412–429.

Appold

Kasarda

(2013) The airport city phenomenon: Evidence from large US airports. Urban Studies 50(6): 1239–1259.

Arellano

Bond

(1991) Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. The Review of Economic Studies 58(2): 277–297.

Baxendine

Cochrane

Dharmalingam

et al . (2005) The New Zealand Population: A Synopsis of Trends and Projections 1991–2016. Hamilton: The University of Waikato, Population Studies Centre.

Beron

Murdoch

(2001) The benefits of visibility improvements: New evidence from the Los Angeles metropolitan area. Journal of Real Estate Finance and Economics 22(2/3): 319–337.

10.

Bilotkach

(2015) Are airports engines of economic development? A dynamic panel data approach. Urban Studies 52(9): 1577–1593.

11.

Bilotkach

Lakew

(2014) On sources of market power in the airline industry: Panel data evidence from the US airports. Transportation Research Part A: Policy and Practice 59: 288–305.

12.

Bilotkach

Clougherty

Mueller

et al . (2012) Regulation, privatization, and airport charges: Panel data evidence from European airports. Journal of Regulatory Economics 42(1): 73–94.

13.

Brueckner

(2003) Air traffic and urban economic development. Urban Studies 40(8): 1455–1469.

14.

Button

Taylor

(2000) International air transportation and economic development. Journal of Air Transport Management 6(4): 209–222.

15.

Carey

Semmens

(2003) Impact of highways on property values: Case study of the superstition freeway corridor. Transportation Research Record, pp. 128–135.

16.

Cervero

Murakami

(2009) Rail and property development in Hong Kong: Experiences and extensions. Urban Studies 46(10): 2019–2043.

17.

Chalermpong

(2010) Impact of airport noise on property values: Case of Suvarnabhumi International Airport, Bangkok, Thailand. Journal of the Transportation Research Board 2177: 8–16.

18.

Cidell

(2015) The role of major infrastructure in subregional economic development: An empirical study of airports and cities. Journal of Economic Geography 15(6): 1125–1144.

19.

Cohen

Coughlin

(2008) Spatial hedonic models of airport noise, proximity, and housing prices. Journal of Regional Science 48(5): 859–878.

20.

Cohen

Coughlin

(2009) Changing noise levels and housing prices near the Atlanta Airport. Growth and Change 40(2): 287–313.

21.

Commons

Page

(2001) Managing seasonality in peripheral tourism regions: The case of Northland, New Zealand. In: Baum

Lundtorp

(eds) Seasonality in Tourism. New York: Taylor & Francis, pp. 153–172.

22.

Cui

Kuang

(2013a) The evaluation of dynamic airport competitiveness based on IDCQGA-BP algorithm. Mathematical Problems in Engineering 2013: 1–8.

23.

Cui

Kuang

et al . (2013b) Dynamic formation mechanism of airport competitiveness: The case of China. Transportation Research Part A: Policy and Practice 47: 10–18.

24.

De Wit

Veldhuis

Burghouwt

et al . (2009) Competitive position of primary airports in the Asia-Pacific rim. Pacific Economic Review 14(5): 639–650.

25.

Debrezion

Pels

Rietveld

(2007) The impact of rail stations on residential and commercial property value: A Meta-analysis. Journal of Real Estate Finance and Economics 35(2): 161–180.

26.

Ding

Knapp

(2002) Property values in inner-city negiborhoods: The effects of homeownership, housing investment, and economic development. Housing Policy Debate 13(4): 701–727.

27.

Eckes

(2011) The Contemporary Global Economy: A History since 1980. Oxford: John Wiley and Sons.

28.

Forsyth

(2006) Airport competition: Regulatory issues and policy implications. In: Lee

(ed.) Advances in Airline Economics: Competition Policy and Antitrust. Bingley, UK: Emerald Group Publishing Limited, pp. 347–368.

29.

Freestone

(2009) Planning, sustainability and airport-led urban development. International Planning Studies 14(2): 161–176.

30.

Gerlach

Peng

(2005) Bank lending and property prices in Hong Kong. Journal of Banking & Finance 29(2): 461–481.

31.

Gitto

Mancuso

(2012) Two faces of airport business: A non-parametric analysis of the Italian airport industry. Journal of Air Transport Management 20: 39–42.

32.

Graham

(2008) Managing Airports: An International Perspective. 3rd edition.New York: Routledge.

33.

Green

(2007) Airports and economic development. Real Estate Economics 35(1): 91–112.

34.

Gualandi

Mantecchin

(2008) Aircraft noise pollution: A model of interaction between airports and local communities. International Journal of Mechanical Systems Science and Engineering 2(2): 137–141.

35.

Hakfoort

Poot

Rietveld

(2001) The regional economic impact of an airport: The case of Amesterdam Schiphol Airport. Regional Studies 35(7): 595–604.

36.

Hall

McDermott

Tremewan

(2006) The Ups and Downs of New Zealand House Prices. Moto Working Paper 06–03. Wellington: Moto Economic and Public Policy Research.

37.

Hong

Chu

Wang

(2011) Transport infrastructure and regional economic growth: Evidence from China. Transportation 38(5): 737–752.

38.

Jud

Winkler

(2006) The announcement effect of an airport expansion on housing prices. The Journal of Real Estate Finance and Economics 33(2): 91–103.

39.

Karlsson

Judders

Wilde

et al . (2008) Airport economic impact methods and models. Transportation Research Board, Project 11–03, Topic 03–03. Washington, DC: Transportation Research Board. Available at: http://onlinepubs.trb.org/onlinepubs/acrp/acrp_syn_007.pdf.

40.

Le Gallo

Ertur

(2003) Exploratory spatial data analysis of the distribution of regional per capita GDP in Europe, 1980–1995. Papers in Regional Science 82(2): 175–201.

41.

Lee

(2009) The networkability of cities in the international air passenger flows 1992–2004. Journal of Transport Geography 17(3): 166–175.

42.

Leung

CKY

Shi

Tang

ECH

(2013) Commodity house prices. Regional Science and Urban Economics 43(6): 875–887.

43.

McCann

(2009) Economic geography, globalisation and New Zealand’s productivity paradox. New Zealand Economic Paper 43(3): 279–314.

44.

McMillen

(2004) Airport expansions and property values: The case of Chicago O’Hare Airport. Journal of Urban Economics 55(3): 627–640.

45.

Matos

JCB

Findell

Masurier

et al . (2013) A comparison of hedonic price and state preference methods to derive monetary values for aircraft noise disturbance and annoyance. Transportation Research Part D: Transport and Environment 20: 40–47.

46.

Mukkala

Tervo

(2013) Air transportation and regional growth: Which way does the causality run?Urban Studies 45(6): 1508–1520.

47.

Mumbower

Garrow

Higgins

(2014) Estimating flight-level price elasticities using online airline data: A first step toward integrating pricing, demand, and revenue optimization. Transportation Research Part A: Policy and Practice 66: 196–212.

48.

Nelson

(1980) Airports and property values: A survey of recent evidence. Journal of Transport Economics and Policy 14(1): 37–52.

49.

Nelson

(2004) Meta-analysis of airport noise and hedonic property values. Journal of Transport and Economic Policy 38(1): 1–27.

50.

Nelson

(2008) Hedonic Property Values Studies of Transportation Noise: Aircraft and Road Traffic. New York: Springer.

51.

Park

(2003) An analysis for the competitive strength of Asian major airports. Journal of Air Transport Management 9(6): 353–360.

52.

Percoco

(2010) Airport activity and local development: Evidence from Italy. Urban Studies 47(11): 2427–2443.

53.

Perl

(1998) Redesigning an airport for international competitiveness: The politics of administrative innovation at CDG. Journal of Air Transport Management 4(4): 189–199.

54.

Pope

(2008) Buyer information and the hedonic: The impact of a seller disclosure on the implicit price for airport noise. Journal of Urban Economics 63(2): 498–516.

55.

Richard

(2008) The impact of additional passengers on airport employment: The case of German airports. Journal of Airport Management 2(3): 265–274.

56.

Roodman

(2009) How to do xtabond2: An introduction to difference and system GMM in Stata. The Stata Journal 9(1): 86–136.

57.

Schultz

Tan

Walsh

(2010) Endogeneity and the corporate governance-performance relation. Australian Journal of Management 35(2): 145–163.

58.

Shirely

Neil

(2013) New Zealand: Auckland. In: Ian

Carol

(eds) Asian and Pacific Cities – Development Patterns. New York: Routledge, pp. 184–199.

59.

Statistics New Zealand (2015a) Estimated resident population for regional council areas. Available at: http://www.stats.govt.nz (accessed 16 October 2015).

60.

Statistics New Zealand (2015b) Gross domestic product per person. Available at: http://www.stats.govt.nz (accessed 16 October 2015).

61.

Stern

(1996) Market definition and the returns to innovation: Substitution patterns in pharmaceutical markets. Working paper. Cambridge, MA: Sloan School of Management, Massachusetts Institute of Technology.

62.

Suksmith

Nitivattananon

(2014) Aviation hubs on property values and management: The case of Suvarnabhumi International Airport. IATASS Research 39(1): 58–71.

63.

Tittle

McCarthy

Xiao

(2013) Airport runway capacity and economic development: A panel data analysis of metropolitan statistical areas. Economic Development Quarterly 27(3): 230–239.

64.

Tomkins

Topham

Twomey

et al . (1998) Noise versus access: The impact of an airport in an urban property market. Urban Studies 35(2): 243–258.

65.

Tsui

WHK

Balli

(2015) International arrivals forecasting for Australian airports and the impact of tourism marketing expenditure. Tourism Economics. Epub ahead of print 30 June 2016. DOI: 10.5367/te.2015.0507

66.

Tsui

WHK

Fung

MKY

(2016a) Analysing passenger network changes: The case of Hong Kong. Journal of Air Transport Management 50: 1–11.

67.

Tsui

WHK

Fung

MKY

(2016b) Causality between business travel and trade volumes: Empirical evidence from Hong Kong. Tourism Management 52: 395–404.

68.

Tsui

WHK

Balli

Gilbey

et al . (2014) Operational efficiency of Asia–Pacific airports. Journal of Air Transport Management 40: 16–24.

69.

Villas-Boas

Winer

(1999) Endogeneity in brand choice models. Management Science 45(10): 1324–1338.

70.

Vowles

Tierney

(2007) The geographic impact of ‘open skies’ policies on Trans-Tasman air passenger service. Asia Pacific Viewpoint 48(3): 344–354.

71.

Wang

Bonilla

Banister

(2016) Air deregulation in China and its impact on airline competition 1994–2012. Journal of Transport Geography 50: 12–23.

72.

Wetzstein

(2007) Managing for a prosperous Auckland? Critical reflections on regional interventions for growing New Zealand’s economic centre. Sustaining Regions 6(1): 28–49.

73.

Wooldridge

(2002) Econometric Analysis of Cross Section and Panel Data. 2nd edition.Cambridge, MA: The MIT Press.

74.

Xue

Friesen

O’Suillivan

(2012) Diversity in Chinese Auckland: Hypothesising multiple ethnoburbs. Population, Space and Place 18(5): 579–595.

75.

Yueh

(2008) Multinationals and economic growth in East Asia: Foreign direct investment, corporate strategies and national economic development. Economica 76(301): 213–214.

76.

Zak

Getzner

(2014) Economic effects of airports in Central Europe: A critical review of empirical studies and their methodological assumptions. Advances in Economics and Business 2(2): 100–111.