Bondi to Byron: divergent pathways in NSW’s Metropolitan and regional rental markets during the COVID-19 pandemic

ABSTRACT

In this paper, we look to compare the differential responses of metropolitan and regional rental markets in the first 12 months post-pandemic in New South Wales, Australia’s most populous state. We employ descriptive statistics to analyse broad trends in both rental prices and transfers in and out of long-term rental supply through the short-term letting and property buyer markets. Moreover, we examine how these dynamics have played out at the local level using the contrasting case studies of Bondi in the eastern suburbs of Sydney and Byron Bay in the northern rivers of NSW. The results demonstrate that there have been distinctly separate responses to the pandemic across different geographical profiles, displaying cohort-level behaviours ranging from rental declines in inner-city areas to significant increases in high-price regional areas. Furthermore, a comparison of metro and regional markets highlights a strong and symmetrical association between short-term lettings activity and changes in long-term rental prices over the initial COVID-19 period. Finally, analysis of investor activity exhibits a highly reactive and geographically differentiated response to changes in rental yields. Ultimately, the findings of this study further the understanding of the contrasting responses at the local scale to the COVID-19 pandemic, the complex interplay between long- and short-term rental markets and help shed light on necessary future policy directions for NSW’s, and Australia’s, housing market.

KEYWORDS:

• COVID-19
• Rental housing markets
• Airbnb
• property investing
• New South Wales
• housing policy

Introduction

The worldwide COVID-19 pandemic from early 2020 generated many new challenges for health services and economic activity around the world, but it also had a substantial impact on housing markets. At the outset of the pandemic, the real fear that the housing market might collapse as jobs were lost or furloughed and construction paralysed, led to widespread government action geared towards protecting housing markets, including subsidy packages to support construction, rent and mortgage moratoria, suspension of evictions, and initiatives to support the homeless (Pawson, Martin, and Aminpour 2022).

In addition, a central component of these impacts has been the major disruption to both inter- and intra-national mobility patterns. In Australia, the impact of the pandemic on global mobility and multiple prolonged lockdowns restricting internal travel after March 2020 had immediate repercussions on the housing market in key locations most affected by these shocks, particularly tourist hot spots and those frequented by international travellers, temporary migrants and students (Pawson et al. 2021; Verdouw et al. 2021). Overseas travel into Australia was put on hold and net overseas migration was predicted to be negative in both 2020–21 and 2021–22 with outflows totalling 140,000 (Centre for Population 2021). This change alone would have had major repercussion for housing demand, especially in those places where such migration is concentrated, such as Sydney and Melbourne.

However, within months, new travel patterns and changing lifestyle choices within Australia began to emerge leading to very different outcomes for property and rental markets across Australia, as they did elsewhere (Pawson et al. 2021). In Australia, changing internal migration patterns, whether by choice or necessity, stimulated a noticeable surge in moves to more ‘liveable’ ex-metropolitan locations mainly in favoured and amenity rich ‘sea-‘ and ‘tree-change’ locations, with attendant impacts on these local housing markets (Buckle and Drosdzewski 2018; Burnley and Murphy 2004). While these trends may reflect longer standing dynamics (Davis 2022) they were certainly amplified, prompting the ‘discovery’ of new demographic sub-groups such as ‘e-changers’ – households taking advantage of new work-from-home directives and IT connectivity to work away from their metropolitan-based employment (Glover and Lewis 2019). There was a related concern, albeit contested, that the cities themselves and particularly central city areas might not fully recover from the COVID-19 generated shock as working from home becomes embedded in work practices permitting greater choice, for some, in their work-home locations (Florida, Rodriguez-Pose, and Storper 2021; Wahba 2022).

Whatever the actual outcomes of this shock to the economic life of Australia’s metropolitan areas will be, there have been unquestioned, and potentially longer-term, impacts on Australian housing markets. This paper explores one aspect of this, namely the response of local rental markets to the pandemic within New South Wales (NSW), Australia’s most populated state. While research has started to unpack the response of housing markets to the pandemic in Australia (see Pawson et al. 2021; 2022), apart from media and market commentary, how the shock of the pandemic on rental markets across Australia has played out at the local scale has received little scholarly analysis to date. Moreover, while globally there have been ample studies evaluating COVID’s housing market impacts at a city- or region-wide level (see Gallent and Madeddu 2021; Verdouw et al. 2021; Witherspoon and Chang 2021), there is a scarcity of research directly comparing the sub-regional responses.

In this paper, as an exemplar of national trends, we examine the diverging stories of the COVID-19 pandemic inspired impact on NSW’s metropolitan (Greater Sydney, hereafter GSyd) and regional (Rest of NSW, hereafter RNSW) rental markets through a multi-scale data analysis approach that triangulates long-term rental prices, trends in short-term lettings and investor activity. By constructing an analysis of the sub-regional and local scale impacts of the pandemic on housing markets across NSW over the broad period from pre-COVID late-2019 to mid-2021 when the country was beginning to emerge from the initial pandemic conditions, the paper makes several key contributions.

Foremost, it is the first study to directly contrast the responses of comparable metropolitan and regional housing markets at a local scale for the long-term rental (LTR) market. Second, the paper utilises the divergent pathways of metropolitan and regional rental prices to explore relationships and outcomes on LTR lettings and rent levels resulting from the differential transfer in and out of the LTR market via the short-term rental (STR) markets. Here, the study utilises the contrasting Airbnb trends (as a proxy for the STL market) at a variety of scales to examine the symmetricity of Airbnb’s impact on rental prices and supply. While previous studies have examined the effect of an increase in Airbnb lettings, this study uses the pandemic to study the rental market response to reduced STR activity. Third, as an exemplar of the rental market trends discussed, the case studies of Bondi in Sydney and Byron Bay on the north coast of NSW provide local evidence of the distinctly varying responses of metropolitan and regional markets to the COVID-19 pandemic and the highly diverse local interplay between the LTR and STR markets. Finally, we review evidence of the way the pandemic resulted in changes to landlord investment behaviour, in particular, in the relationship between rental yields and investment behaviour in the LTR sector, again drilling down to explore the sub-regional and local variations.

In the next section, a brief review of a key driver of housing markets, namely migration and temporary visitors (including tourists), sets the scene for the following section reviewing the literature and trends in rental market responses to the COVID-19 pandemic in Australia and overseas. After the data and methods are described, the analysis then turns to a detailed analysis of local impacts within NSW of trends in the LTR market and transfers with the STR and property buyer markets.

Literature review

The impact on Australia’s migration and population movement

Perhaps the most immediate impact of the pandemic on the housing market was on the movement of people, both into and within the country. With overseas travel banned from 20 March 2020, international movement virtually ground to a standstill (Figure 1). Total international arrivals fell from 2.257 m in January 2020 to just 21,170 in April 2020, recovering to 265,000 by January 2022 following the easing of restrictions during the latter half of 2021 (ABS 2022a). Total short-term (less than a year) visitor numbers – basically tourists – suffered a 93% contraction from 766,000 in January 2020–2,250 in April 2020, recovering to 54,000 by January 2022. Movements out of Australia were similarly impacted. The disruptions to international movement were particularly significant for the rental market, with an estimated 70% of incoming international migrants to Australia starting out in the private rental market (Hulse et al. 2012). International student arrivals, even more dependent on rental accommodation, were equally decimated, suffering a 99% decline between pre-COVID January 2019 and January 2020, from 91,600 to just 360, albeit recovering to 28,011 in January 2022 (ABS 2022a).

Figure 1. Total monthly overseas arrivals and short-term visitor arrivals, Australia Jan 2018 to Jan 2022. Source: Australian Bureau of Statistics, Overseas Arrivals and Departures, Australia December 2022.

Within Australia, as noted above, the pandemic also instigated a well-documented uptick in outflows of population from the major cities during 2020 (Figure 2) reaching a record high in early-2021 despite lockdown restrictions that varied significantly by state and locality over the period (ABS 2021; Borsellino et al. 2022). The losses were largely accounted for by the two largest cities, with Melbourne recording a net loss of 8,300 people in the March 2021 quarter and Sydney losing 8,200, with the other capital cities recording either net gains (especially Brisbane) or only marginal changes (ABS 2021). While the emergence of sea- and tree-change migration patterns was a key contributor (Li et al. 2022), evidence on counter-urbanisation suggests that regional residents not moving to metropolitan areas was perhaps more significant in explaining the net outflows from metropolitan to regional areas (Buckle and Osbaldiston 2022; McManus 2022). Indeed, total arrivals to Australia’s greater capital cities from within Australia decreased by 28.2% between December 2020 to September 2021, while internal departures also fell 14.5% over this same period (ABS 2021).

Figure 2. Quarterly net internal migration greater capital cities combined, Australia Q1 2016 to Q1 2021. Source: Australian Bureau of Statistics, Regional internal migration estimates, provisional March 2021.

The impact on rental housing markets within Australia

In this context, the housing market went through several phases as the economy and the population adjusted to the new ‘normal’ of the pandemic and its associated outcomes. In Australia, as observed in several other comparable housing markets (e.g. Ramani and Bloom 2021), a drop in property prices in the immediate aftermath of the pandemic shock into the second half of 2020 (Hu, Lee, and Zou 2021) was followed by a period of unprecedented, and largely unpredicted, price rises through 2021, with annual increases in the year to February 2022 of 19.2% in the combined Australian capital cities and 25.5% in regional Australia (CoreLogic 2022a). However, property price changes, while playing an important context within which rental market changes have played out, are not central to this paper and will not be considered further here.

• i) The long-term rental market¹

In common with the experience in other comparable countries, rents in Australian metropolitan areas initially dropped globally during the onset of the pandemic. First, demand for inner-city areas from several different demographics fell dramatically. International students, traditionally accommodated in inner-city share houses and apartments, but more lately housed in new-build high-density student lodgings, were among the first to leave major cities in the face of travel restrictions and lockdowns (Evans, Rosewall, and Wong 2020; Halliday 2021). While the initial student exodus had an immediate effect on inner-city rental markets, the inability of potential new students to acquire student visas and the greater uptake of full-time digital delivery courses are projected to have a more enduring impact on international student volumes in Australia (TEQSA 2021). Total international student visas in Australia decreased by 70% between 2020 and 2021 (Hurley, Hildebrandt, and Brisbane 2021), while international student enrolments by 2025 are projected to be 25% less than pre-pandemic numbers (TEQSA 2021). These figures suggest a longer-term impact on inner-city rental markets from the initial and sustained reduction in international students coming to Australia. The impact on employment also hit inner city rental markets as jobs in the accommodation, tourism, food, and entertainment sectors, concentrated in inner suburbs and heavily reliant on younger workers who disproportionally rent, were hit hardest (Evans, Rosewall, and Wong 2020).

But as the pandemic continued, people’s housing choices began to change as domestic, work and leisure spheres were rolled into one under lockdown conditions. This prompted those households who can to increasingly make choices based on space and lifestyle preferences rather than convenience or pricing (Carmona et al. 2020), opening up a marked difference in rental demand between metropolitan and regional locations. For inner city renters, the immediate outcome was that some tenants were able to relocate to take advantage of falling rent levels. For example, Reserve Bank of Australia analysis showed rental bond lodgements in inner Sydney rapidly escalated in mid-2020 as such moves took effect (Evans, Rosewall, and Wong 2020). These trends paralleled broader impacts. In America, Liu and Su (2020) noted the desire for greater housing space, finding a decreasing relationship between housing demand and population density. Gallent and Madeddu (2021) identified the inner-city ‘London Doughnut’ where vacancy rates increased and rental prices fell, while Verdouw et al. (2021) documented greater falls in unit rental prices than houses for all major cities in Australia. Outside the cities, lifestyle attractions and revival of domestic tourism significantly increased local demand for rental in key regional areas, as mentioned earlier. In the UK, rental prices outside of London were up 3% year on year in Q1 2021 (Zoopla 2021), while prices for smaller regional cities have increased across the board in the US (Witherspoon and Chang 2021), a trend also documented in Australia (Pawson et al. 2021).

However, the initial downward pressure on rents began reversing by the end of 2020, with rental levels escalating rapidly in a range of countries, including Australia (Pawson, Martin, and Aminpour 2022). In Australia, annual rental increases hit a high of 9.4% in November 2021, with lower turnover rates and low levels of new supply, particularly for apartments following the housing market downturn after 2018, compounding the situation (CoreLogic 2022b). Total rental listings fell to 25% of long-term averages by September 2021, well below their historic levels (Pawson et al. 2021). As a result, rents recovered strongly during 2021. Reflecting this, investors moved back into the market with finance for investor lending reaching $10.9bn by January 2022 after falling to a 20-year low of $4.2bn in May 2020 (ABS 2022b).

Again, there is a distinct sub-regional geography to these trends. Pawson et al. (2021) have shown that rents in inner and middle Sydney suburbs fell or were stagnant at best between the onset of the pandemic and mid-2021 while those in outer suburban areas and beyond increased. Rental market turnover and vacancy rates also showed a variable response to the pandemic by initially bouncing up in inner and middle Sydney and Melbourne after which they settled back, while turnover and overall supply was little changed in regional areas (Pawson et al. 2021). Subsequently, however, all areas have experienced a major inflation in rents.

• ii) Short-term lettings through the pandemic

In the two decades prior to the pandemic, STR platforms such as Airbnb experienced near uninterrupted growth, disrupting housing markets globally (Adamiak 2019). The ‘Airbnb effect’ on the housing market is well documented, referring to the displacement of long-term rentals and increase in rents associated with an increase in Airbnb activity, as well as negative impacts on the established tourism accommodation market (Barron, Kung, and Proserpio 2020; Che et al. 2020; Garcia-Lopez et al. 2020; Horn and Merante 2017; Thackway et al. 2022). The onset of the pandemic and the associated global travel restrictions was a severe shock to the short-let tourist accommodation (Baxter 2020). Moving STRs into the long-term rental pool was a logical move as investors sought to offset the disappearance of both international and local tourists, thus increasing the supply of rentals in the long-term market and putting downward pressure on rents. Evans, Rosewall, and Wong (2020) noted that listings of Airbnb properties fell by around 40% or 40,000 properties across Australia at this time, increasing national long-term vacancy rates by an estimated 1%. Boros, Dudas, and Kovalcsik (2020) documented significant declines in Airbnb listings for 15 Airbnb hotspot cities throughout 2020. In Sydney, Thackway and Pettit (2021) documented increasing LTR supply in tourist areas where Airbnb listings had declined in Q2 2020. Furthermore, they evidenced a direct relationship between Airbnb activity and reductions in rental prices, with rental prices declining by up to 7.1% in the most active neighbourhoods. Similarly, Buckle and Phibbs (2021) estimated that a transfer of STR lettings into the LTR sector accounted for two-thirds of a 9% reduction in median rents in Hobart during the first COVID-lockdown (March–June 2020).

However, this retraction did not last long in some areas as the domestic tourist market rebounded later in 2020 as restrictions eased (SQM 2021). AirDNA (2021) data showed that demand, as measured by new bookings, recovered strongly for regional Australia after the initial lockdown in 2020 compared to an ongoing slump in metropolitan Australia. In the latter, bookings were down by 58% by July 2021 (having fallen by 72% in September 2020) almost certainly due to the lack of overseas tourists, compared to just 12% in the former (having peaked at 22% growth in May 2021). The recovery of STR demand in regional locations clearly followed the easing of internal lockdown restrictions and the return of local tourism within Australia, compounded by the pent-up demand from those unable to holiday abroad.

Therefore, literature both at home and abroad has identified differing trajectories for metropolitan and regional rental markets during the COVID-19 pandemic. However, apart from Pawson et al.’s (2021) national study of the impacts of COVID-19 on rental markets and homelessness, there has been little detailed attention paid to a direct comparison of the pandemic’s impacts on local rental markets in Australia. We seek to address this gap by contrasting the short- and long-term rental market movements in Sydney and regional NSW in the aftermath of the pandemic.

Data and methods

Data

The data for this paper have been gathered for the whole of NSW over an 18-month period between Quarter 4 2019 to Quarter 2 2021 from several sources:

• Address-level new rents and letting data from the Australian Property Monitors (APM), who compile new rental data monthly.

• Address-level lettings data from the NSW Office of Fair Trading’s Rental Bond Board (RBB), which records all lettings in the private rental market, including the rent level and date of letting. These data are used to define the LTR sector in this analysis.

• Postcode-level Airbnb data from the online web scraping company AirDNA, who collect information on all Australian Airbnb listings. While AirDNA collects data for several short-term rental platforms including Stayz and others, this study isolated Airbnb lettings given the more extensive research on its housing market impacts (e.g. Horn and Merante 2017; Thackway et al. 2022). In this paper, we have taken Airbnb lettings rates as an indicator for the short-term tourist lettings market more generally.

• Rental vacancy rates taken from the Real Estate Institute of NSW (REINSW) self-specified areas – see Appendix.

• Address-level property sales data from the Australian Property Monitors (APM), who compile property sales data monthly.

• Dwelling counts from the NSW Valuer General All Property Dataset for July 2020.

While this misses out later trends, including those for the period after the onset of the Delta variant in October 2021, this is the latest period for which detailed data were available at the time of writing.

The address-level property price and rental data were geocoded using the GNAF geocoding software thereby allowing aggregation to local government area (LGA) level and Census geographies, notably Statistical Area 2 (SA2) and Statistical Area 3 (SA3) geographies.² AirDNA data were similarly coded to comply with SA2 and SA3 areas. This procedure also allowed for addresses to be matched for the final section assessing the transfer of properties to and from the rental market through property sales transactions.

It should be noted that the paper does not assess the impact of the pandemic on the informal rental market, such as informal leases beyond short-term rentals, or its interaction with the STR and LTR markets. While informal lettings may represent a material proportion of all rentals and potential driver of growth and/or loss to the formal rental market (Gurran, Pill, and Maalsen 2021), the absence of any formally collected data on this sector of the rental market the analysis precludes its consideration here.

Methods

This section tracks the movements of NSW’s metropolitan (GSyd) and regional (RNSW) housing markets during the first year of the COVID-19 pandemic (March 2020–March 2021). The study's methods are not intended to be proof of any causal relationships between different sectors of the housing market. Rather, the study takes a straightforward data analysis approach, using descriptive statistics to contrast the major spatial and temporal trends in metropolitan and regional NSW’s rental market movements at different stages in NSW’s COVID timeline.

To investigate these outcomes during the first year of NSW’s COVID response, three 6-month study periods were selected to reflect important lockdown and reopening dates (Table 1). The first 6-month period (P1), between the last quarter of 2019 and the first quarter of 2020, represents NSW’s housing market prior to the onset of the pandemic, and effectively acts as a control period for the study. The second 6-month period (P2) begins in Q2 of 2020, immediately following the onset of the state-wide lockdown in March 2020. This period captures the initial market response to the shutdown of economic and social activity. Finally, with NSW’s state-wide lockdown finishing in May 2020, the 3rd period (P3) between the end of 2020 and start of 2021 captures the recovery of housing markets as the state enjoyed its greatest period of freedom since the onset of the pandemic.

Table 1. 6-month study periods with respect to important COVID lockdown/reopening dates.

Period 1 – Pre-COVID		Period 2 – Onset of COVID		Period 3 – Recovery
2019–2020		2020		2020–2021
Q4	Q1	Q2	Q3	Q4	Q1

First, overall rental market movements are analysed by comparing spatial and temporal trends in the GSyd and RNSW markets. Thereafter, the relationship between Airbnb listings activity and rental supply and price is examined. The analysis employs a measure of Airbnb listings activity³ as a proportion of total housing, Airbnb share, calculated for an SA2 area i in period t as: $\mathrm{AbbShar}{e}_{\mathrm{it}}={\displaystyle \frac{\mathrm{AbbListing}{s}_{\mathrm{it}}}{\mathrm{TotalResidencie}{s}_i}}$SA2 areas were binned into descending Airbnb activity deciles based on Airbnb share (i.e. 10 represents neighbourhoods with the greatest level of tourist activity). Trends for rental supply and price, based on Airbnb activity level, were analysed to establish the relationship, if any, between reduced tourist activity and LTR market movements. The case study markets of Bondi (Eastern Suburbs North SA3) and Byron Bay (Richmond Valley Coastal SA3), both iconic coastal communities with significant short- and long-term rental markets, are then examined as examples of the local impacts observed in the NSW rental market.

Finally, the interplay of changing rental yields and property investor activity during the first year of the pandemic, as measured by property sales that were subsequently rented out, as indicated by a rental bond record, within six months of settlement date, is investigated in selected high-price SA3 areas in Sydney and RNSW. So, for example, a property sold in P2 (Q2 2020–Q3 2020) would be classified as an investor purchase if it had a rental bond recorded at that address in ‘P2 + 6’ (Q4 2020–Q1 2021). To achieve this, APM property sales and RBB data were linked by matching the latitude/longitude of properties. Unfortunately, inconsistencies between the APM and RBB data formats for apartment numbers caused difficulties matching rental and sales data for units. Therefore, for the investor activity analysis, only houses were analysed. The omission of apartment data from the investor activity analysis is a limitation, however the exploration of houses still provides valuable insights.

Results

The geography of rental dynamics during the pandemic

• i) Long-term rental trends

The differential response of rental markets in NSW metropolitan and regional areas is illustrated in Figure 3 which shows quarterly rental price movements for GSyd and the RNSW, in essence the metropolitan and regional housing markets. The blue and orange lines represent the median rental prices of GSyd and the RNSW respectively. While any potential composition effects in the data were not investigated given the descriptive nature of the study, taking the median offsets this concern to a reasonable degree.

Figure 3. Quarterly Rental Price for GSyd and RNSW. Source: APM rental data.

In GSyd, rental prices had already softened as a result of potential oversupply prior to the onset of COVID. In the immediate post-pandemic period, rental prices were initially stable before dropping sharply, and this decline has continued in P3 before stabilising somewhat towards the middle of 2021. In regional areas on the other hand, rental prices were initially stable in response to the onset of the pandemic, however, have increased dramatically into 2021. These trajectories are consistent with the literature contrasting metropolitan and regional rental prices during the pandemic noted above.

Drilling down into the geography of LTR rental price changes at the SA3 level, Figure 4 visualises changes to rental prices between the two 6-month periods (P1-P2 and P2-P3), with blue representing falls in rental prices and red representing gains. This shows that in Sydney (Figure 4(a)), the most concentrated declines in rental prices in both periods were in traditional tourist/high-amenity locations in the inner suburbs close to the CBD, and the high value North Shore and Eastern Suburbs. In the middle and outer suburbs, while rents were more or less stable in Period 2, they rebounded in Period 3, especially in the outer west. It’s here that households seeking larger space may have added pressure on local rents during this period.

Figure 4. Rental price % change between periods, Q4 2019 to Q1 2021. (a) Greater Sydney: SA3 Areas; (b) Rest of NSW: SA3 Areas. Source: APM rental data.

On other hand, in regional areas (Figure 4(b)), while almost all areas have seen rental prices increase, it has been coastal and regional centres that have been impacted the most. In Period 2, rental prices around the regions varied quite markedly between coastal increases and inland declines. However, by Period 3 rental prices surged in nearly all areas. In particular, coastal retreats such as Byron Bay and the South Coast as well as rural amenity towns such as Mudgee and Bathurst have seen the largest increases.

The scenario described above, with metropolitan inner city and coastal locations bearing the brunt of the rental downturn following the onset of the pandemic while regional lifestyle sea- and tree-change locations, as well as outer suburban areas, responding to increased rental demand, confirms the more general commentary described in the literature above. Clearly, the pandemic played out very differently between these areas in NSW.

• ii) Short-term rental trends

Airbnb activity across NSW has been assessed by calculating the proportion of total dwellings in each SA2 area (derived from the NSW Valuer General’s property dataset) that were listed on Airbnb over the three study periods. Each SA2 area was then binned into deciles based on its Airbnb share to create an ‘Airbnb activity decile’ for each SA2 area. Figure 5 graphs the resulting quarterly Airbnb activity by SA2 Airbnb activity decile at the SA2 level for NSW split between GSyd and the RNSW.

Figure 5. Changes in Airbnb activity, Greater Sydney and Rest of NSW Q1 2019 to Q2 2021 (N = 73,937 total Airbnb listings). Source: AirDNA Airbnb data.

First, the figures provide an indication of the relative market share of active Airbnb listings at the SA2 level in both metropolitan and regional NSW. The distribution of median Airbnb share is comparable across both geographies, with over 10% Airbnb share (pre-pandemic) for the most active decile, over 5% for the second-most active decile, and a gradual decline thereafter. Second and more importantly, the figures indicate very different responses to the pandemic in Airbnb lettings activity – and by implication the short-let tourist market – for these two broad regions. Airbnb listings fell dramatically for Sydney from the onset of the pandemic and continued to fall throughout the study period, whereas it initially fell but then rapidly rebounded for the RNSW, particularly towards the end of 2020 and into 2021.

The relationship between short-term lettings activity and long-term rental supply

Was there any association between the levels of short-let activity noted above and the long-term rental market? In order to assess this, the total number of new rental bonds by SA2 was calculated to provide a measure of rental supply for GSyd and RNSW. Figure 6 compares changes to annual rental supply at the SA2 level for GSyd and RNSW by Airbnb activity decile to analyse how changes to Airbnb supply have interacted with trends in supply in the LTR market. The analysis has been split into two 12-month phases, which measures changes between Period 1 and Period 2 and Period 2 and Period 3. It should be noted that lower numbers of rental transactions in some rural and peri-urban SA2’s resulted in wider outliers. Therefore, we have used a box and whisker plot to capture the prevailing rental trends while preserving the outlier observations.

Figure 6. Rental supply changes by share of Airbnb lettings (deciles), Greater Sydney and Rest of NSW Q4 2019 to Q1 2021 (N = 73,937 total Airbnb listings). Source: APM rental data, AirDNA Airbnb data.

For the Greater Sydney region, there was a sharp decline in LTR supply across the city as the pandemic struck, but with greater losses in areas with lowest concentrations of Airbnb lettings, touching almost 50% in some areas. The only decile group where LTR supply increased between Period 1 and Period 2 was in the highest Airbnb activity decile, where LTR lettings increased by almost 10% on average. This result suggests that differences in LTR supply across Sydney during this period could indeed be explained by the possible influx of ex-Airbnb properties onto the mainstream rental market. Moving into Period 3, the data show that LTR supply in GSyd stabilised overall with most Airbnb activity deciles recording modest increases in lettings between the two periods. But these turned into moderate losses for SA2s in the highest Airbnb decile, possibly indicated properties were moving back into the STR sector.

For the RNSW, in the immediate aftermath of the pandemic rental supply was also down overall, but not as severely as in GSyd. But again, the losses were not as pronounced in the top two Airbnb activity decile, with lettings levels in some SA2s moving into positive territory. However, into Period 3, while LTR supply has recovered for some areas, it has fallen significantly in the highest Airbnb activity areas, suggesting that Airbnb activity may well have recovered strongly and surpassed pre-COVID levels.

While we are unable to directly relate these changes at the address-level, these results nevertheless demonstrate the likelihood of a cause-and-effect relationship between movements in the STR market, as indicated by Airbnb activity levels and LTR supply. Moreover, these data also highlight the distinctive geography of these changes.

• i) Rental vacancy rates

What is the association, if any, between STRs and LTR vacancy rates? As STRs were converted back to the long-term market, it might be expected that this would have an immediate impact on rental supply leading to an increase in vacancy rates (as noted by Evans, Rosewall, and Wong 2020), and vice versa, where LTRs are converted to STRs. Figures 7 compares the moving averages of vacancy rates taken from Real Estate Institute of NSW data and the Airbnb share of the rental market over the study period for GSyd and the RNSW. To simplify the analysis and to provide a degree of geographical disaggregation, we have selected exemplar SA3 sub-regions in both areas. In GSyd these are Inner, Middle and Outer. For the RNSW, we have focused on five regional sea-change and tree-change hotspot areas: Northern Rivers, Coffs Harbour, Mid-North Coast, South Coast, and Central West. The full list of LGA’s included in each sub-region is detailed in Appendix A.

Figure 7. Airbnb share of rental market (left hand) and vacancy rates for LTR’s (right hand) for selected metropolitan and urban vacancy rate areas. (a) Greater Sydney; (b) Rest of NSW (selected hotspots). Source: REINSW, AirDNA Airbnb data

In Inner Sydney (Figure 7(a)), Airbnb activity clearly declined from around 8.5% of the dwelling stock in 2019 to below 7% mid-2021, with two-season holiday period ‘blips’. The effect on vacancy rates is striking: as expected, as Airbnb share has fallen by nearly 2%, vacancy rates have increased by roughly 2% over the same period albeit starting to fall back marginally at the end of the period. The pattern of vacancies for middle suburban Sydney is comparable. In Outer Sydney where the STR market is much less prominent, vacancy rates fell to a low point of around 2.5% by the end of the study period. This may well reflect increased demand for suburban housing for those seeking more space and amenity, noted above.

For the Rest of NSW (Figure 7(b)) we see this relationship in reverse. For the selected regional tourist hotspots, Airbnb activity increased from the onset of the pandemic. Simultaneously, LTR vacancy rates have dropped by as much as 2% for some areas to below 1% for all regional hotspot areas. With vacancy rates this low in the LTR market, demand for new rental listings is likely to push up rental prices.

• ii) Rental price

Trends in rental prices are consistent with the predicted impact of differentiated Airbnb listing rates shown above, indicating a dynamic relationship between Airbnb and LTR markets, as illustrated in Figure 8. While this analysis does not employ rigorous empirical methods to isolate the impact of Airbnb, it nonetheless provides descriptive trends that reveal a close and contrasting association between fluctuations in Airbnb activity and rental prices across NSW. In GSyd, there is seemingly a negative association between LTR rental price and Airbnb listing rates in both the earlier and later COVID periods, whereby rental prices fell further in high-activity Airbnb neighbourhoods. In the most active Airbnb decile, the median rental price drop was 8% in the immediate aftermath of the pandemic, and a further 4% into 2021.

Figure 8. Long-term rental prices changes by share of Airbnb lettings (deciles), Greater Sydney and Rest of NSW Q4 2019 to Q1 2021 (N = 73,937 total Airbnb listings). Source: AirDNA Airbnb data

In Regional NSW, while there are no clear trends for the immediate COVID period (P1), there was an increasing upward trend in rental prices with increased Airbnb activity for the latter period. In decile 10 (highest Airbnb decile) the median and Q3 rental price spikes were 5% and 15%, respectively. Evidently, restrictions in both rental supply and vacancy rates to below 1% were associated with dramatic spikes in regional rental prices where Airbnb was active.

This issue is further unpacked in Figure 9 which plots the relationship between nominal changes in Airbnb listings and rent levels over the study period at the SA2 level for NSW. The data have been disaggregated into three cohorts: the 25 SA2s with the highest numbers of Airbnb listings at the start of the period (which between them accounted for 40% of all such listings), other SA2s located in the Sydney metro area (264 SA2s and 28% of listings), and other SA2s located in the rest of NSW (299 SA2s and 32% of listings). The chart supports the proposition that locations with substantial concentrations of Airbnb activity also saw the greatest movement in rents over the period, both positive and negative. Rent generally fell in metro SA2s, which also experienced losses of Airbnb listings (bottom left-hand quadrant) with Airbnb ‘hotspots’, such as coastal Bondi Junction and inner-city Waterloo and Chippendale, showing notable rental declines. Non-metro SA2s are largely clustered in the upper right-hand quadrant with modest positive changes in Airbnb listings and moderate rent rises. But noticeably, the few Top 25 Airbnb locations in non-metro tourist ‘hotspots’, such as coastal Byron Bay and Ulladulla and the popular ski-resort of Jindabyne, saw significant increases in both listings and rents. The computed trend lines show little relationship for the other Metro and Non-Metros SA2s but a significant 0.48 R² for the Top 25 SA2s.

Figure 9. Comparison of changes in Airbnb listings against changes in LTR rental prices, January 2019 to April 2020.

In aggregate, then, looking at changes in rental price and supply against Airbnb activity level suggests a strong relationship between STR activity and LTR market responses over the pandemic period for those locations with concentrations of short-let listings. While the relationships are complex and may not be uni-causal, it can be concluded that in metropolitan tourist hotspots, where Airbnb activity significantly fell, LTR supply increased and rents went down. In regional tourist hotspots on the other hand, where Airbnb recovered strongly due to local demand, LTR supply and vacancy rates declined, and rents increased. Hence, this analysis evidences both the association between declining rental prices amid reduced Airbnb activity, and vice versa for increasing Airbnb activity, indicating a symmetrical nature of the ‘Airbnb effect’ that has been documented extensively elsewhere in the literature (e.g. Horn and Merante 2017; Thackway et al. 2022). Nevertheless, it is also apparent from Figure 9 that substantial rent increases were prevalent in a large number of largely in non-Metro SA2s over the period. The pressure on these local housing markets in part reflecting an upswing in housing demand in these areas is clearly evident, contrasting to the experience in many metro SA2s where rents remained depressed.

Exploring the local impacts: case studies of Bondi beach and Byron bay

By way of an illustration as to the divergent trajectories of local housing markets in NSW that resulted from the pandemic, the examples of the northern Eastern Suburbs SA3 of Sydney – broadly centred on the iconic tourist location of Bondi Beach – and the equally iconic Byron Bay in the Richmond Valley Coastal SA3 region of northern NSW, offer examples of the most extreme responses. Both are epicentres of Airbnb activity but exemplify how the interaction between the short-let and long-let rental markets interacted in divergent ways through the pandemic.

In Sydney’s Eastern Suburbs (Table 2(a) and Figure 10(a)) where Airbnb lettings fell significantly following the onset of the pandemic, market rental prices declined by upwards of 10% in the six months after the onset of COVID and falling by 16.5% in Paddington. Into Period 3, while Airbnb lettings continued to decline, market rents in areas closer to the coast such as Bondi and Bronte began recovering through to the first quarter 2021. However, for the more central suburbs of Paddington and Woollahra, rental prices continued to decline well into 2021, suggesting these suburbs had yet to see any return of international migrants, including students. Figure 10(a) confirms that rents in the LTR sector in the Eastern Suburbs North SA3 around Bondi fell significantly during the onset of the pandemic but then recovered slowly into mid-2021. In contrast LTR letting numbers increased, peaking during mid-2020, indicating a possible influx of ex-Airbnb lettings into the mainstream housing market as well as tenants moving to take advantage of rent reductions.

Figure 10. Rental price and vacancy rate quarterly movements in case study areas. (a) Eastern Suburbs North. Source: NSW Rental Bond data; (b) Byron Bay. Source: NSW Rental Bond data.

Table 2. Airbnb activity and rental price movements in case study area.

a) Eastern Suburbs North.
	Eastern Suburbs North
	Airbnb Share (%)			Median Price ($)			Median Price (% Change)
SA2 Name	P1	P2	P3	P1	P2	P3	P1 – P2	P2 – P3
Bondi – Tamarama – Bronte	24.7%	20.4%	19.1%	750	680	700	−9.3	2.9
Bondi Beach – North Bondi	32.1%	27.7%	25.7%	750	670	695	−10.7	3.7
Bondi Junction – Waverly	13.2%	11.0%	9.5%	730	650	650	−11.0	0.0
Double Bay – Bellevue Hill	7.5%	6.5%	6.0%	745	665	660	−10.7	−0.8
Paddington – Moore Park	9.8%	8.8%	7.7%	695	580	510	−16.5	−12.1
Woollahra	7.9%	6.6%	6.2%	750	675	645	−10.0	−4.4
b) Richmond Valley Coastal.
	Byron Bay
	Airbnb Share (%)			Median Price ($)			Median Price (% Change)
SA2 Name	P1	P2	P3	P1	P2	P3	P1 – P2	P2 – P3
Bangalow	23.6%	21.8%	22.9%	745	750	850	0.7	13.3
Brunswick Heads - Ocean Shores	15.2%	15.4%	14.7%	540	550	795	1.9	44.5
Byron Bay	40.4%	39.6%	41.4%	750	700	950	−6.7	35.7
Lennox Head - Skennars Head	10.9%	10.2%	12.5%	620	590	700	−4.8	18.6
Mullumbimby	13.3%	12.2%	12.4%	630	590	695	−6.3	17.8
Bangalow	23.6%	21.8%	22.9%	745	750	850	0.7	13.3

In the Byron Bay area (Table 2(b) and Figure 10(b)), at the onset of the pandemic, while rental prices were initially stagnant or negative, with the largest falls in the tourist hot spots of Byron and Mullumbimby, the proportion of lettings accounted for by Airbnb only fell back marginally. But as pressure on the area increased by internal migration and local tourism noted above (ABS 2021) and Airbnb lettings recovered, market rents increased dramatically towards start of 2021. Byron Bay itself and nearby Brunswick Heads recorded unprecedented rental prices increases upwards of 36% and 45%, respectively in the latter period.

Rental yields and investor activity

Finally, in the context of variable rental growth, the overall property price increases noted above and the interplay between the supply of STR and LTR sectors, what does the data tell us about the activities of mainstream LTR investors in terms of the level of investment or disinvestment in rental property during this period? With rents falling into the pandemic period, it might be expected that investors would have pulled back from the market at this time. To investigate this, the relationship between changing rental yields and property investor behaviour was investigated in a sub-sample of ten local harbourside and coastal GSyd and coastal RNSW locations selected to reflect higher value markets where capital growth and rental returns might be attractive to investors. As mentioned previously, the analysis was only possible for the purchase of houses for investment during the period, not apartments. The analysis was conducted at the SA3 level to obtain sufficient investor data for stable and reliable results. Properties bought in the period were denoted as ‘investor buying’ if they were rented out in following 6 months.

• i) Changing rental yields

Rental yields (Table 3(a)) for houses have displayed divergent trends in the selected locations. Rental yields fell in most areas in the immediate aftermath of the COVID-19 pandemic (P1 to P2) due to declining rental prices and rising property prices. In GSyd by P3, as housing prices continued to rise while rental prices remained stagnant, rental yields fell by a further 0.2--0.3%. Although housing prices also grew in RNSW, a strong recovery for rental prices led to rental yields stabilising in these regional areas, and even rising by 1% in the South Coast.

• ii) Investor buying

Table 3. Summary of rental yield and investor activity in selected higher-value GSyd and RNSW areas

a) Rental yield movements.
	Rental Yields – Houses
		Median Price						Rental Yield
		Rental (Annual)			House			%			% Change
	SA3 Name	P1	P2	P3	P1 (‘000)	P2 (‘000)	P3 (‘000)	P1	P2	P3	P1 – P2	P2 – P3
METRO	Eastern Suburbs – North	1,350	1,200	1,200	3,020	3,200	3,600	2.3	2.0	1.7	−0.4	−0.2
	Sydney Inner City	850	800	780	1,580	1,575	1,674	2.8	2.6	2.4	−0.2	−0.2
	Ryde – Hunters Hill	630	650	620	1,675	1,609	1,800	2.0	2.1	1.8	0.1	−0.3
	Manly	1,395	1,295	1,400	2,525	2,725	3,080	2.9	2.5	2.4	−0.4	−0.1
	North Sydney – Mosman	1,200	1,095	1,100	2,910	3,000	3,350	2.1	1.9	1.7	−0.2	−0.2
REGIONAL	Byron Bay	593	595	680	770	820	910	4.0	3.8	3.9	−0.2	0.1
	Coffs Harbour	450	465	510	520	560	620	4.5	4.3	4.3	−0.2	0.0
	Kiama – Shellharbour	500	500	525	669	715	744	3.9	3.6	3.7	−0.3	0.0
	South Coast	470	500	700	530	550	630	4.6	4.7	5.8	0.1	1.0
	Port Macquarie	430	450	480	550	560	609	4.1	4.2	4.1	0.1	−0.1
b) Investor buying activity.
	Investor Buying – Houses
		Properties Bought						Investor Buying
		Total			Listed as Rentals			%			% Change
	SA3 Name	P1	P2	P3	P1 + 6	P2 + 6	P3 + 6	P1	P2	P3	P1 – P2	P2 – P3
METRO	Eastern Suburbs – North	693	655	892	98	75	124	14.1	11.5	13.9	−2.7	2.4
	Sydney Inner City	373	424	490	170	168	199	45.6	39.6	40.6	−6.0	1.0
	Ryde – Hunters Hill	538	543	738	74	67	111	13.8	12.3	15.0	−1.4	2.7
	Manly	224	214	231	29	22	22	12.9	10.3	9.5	−2.7	−0.8
	North Sydney – Mosman	334	287	413	69	62	73	20.7	21.6	17.7	0.9	−3.9
REGIONAL	Byron Bay	592	604	766	114	86	138	19.3	14.2	18.0	−5.0	3.8
	Coffs Harbour	693	655	892	98	75	158	14.1	11.5	17.7	−2.7	6.3
	Kiama – Shellharbour	811	780	899	116	80	149	14.3	10.3	16.6	−4.0	6.3
	South Coast	617	697	992	82	83	162	13.3	11.9	16.3	−1.4	4.4
	Port Macquarie	730	738	959	104	89	163	14.2	12.1	17.0	−2.2	4.9

Tables 3(b) investigates the response of property investors to fluctuating rental yields in metro and regional NSW. A significant dip in investor buying, as indicated by the numbers listed as rentals six months later, is evident for both metropolitan and regional areas as rental yields dropped in the immediate response to the pandemic. After the initial shock as rental yields declined in Sydney, investor buying also generally declined, notably in Sydney Inner City, although there was a partial recovery into P3. Conversely, in regional locations, while investor activity fell back between P1 and P2 it rebounded significantly as rental yields stabilised, surpassing pre-COVID levels in all SA3’s other than Byron Bay.

Collectively, contrasting the strategies of property investors in Sydney and the regions illuminates the highly variable relationship between rental yields and investor activity. In the Sydney SA3s, where rental yields fell in both periods on the back of rising property prices and falling rents, investors consistently pulled back from the property market. In the Regional SA3s, while rental yields dropped in the initial shock period, recovering rental prices catalysed broad re-entrance into the market by investors. While this does not prove a comprehensive causal link between rental yields and investor purchasing, it nonetheless contributes a strong part of the story of metropolitan and regional NSW’s diverging rental markets during the pandemic.

Conclusions and policy implications

There is clear evidence from this study that local housing submarkets function in distinctly varying ways. In NSW, metropolitan and regional rental markets have displayed markedly varying responses to the pandemic, often moving in almost perfect opposition at times, driven by differential local economic, lifestyle and migration trends. This is a potentially important finding. Too often housing markets are treated as a simple ‘black boxes’ which are expected to respond to simple policy interventions. Australia’s housing policy approaches largely ignore the geography of housing markets (Pawson, Milligan, and Yates 2019). As the geographically differential responses of metro and regional rental sub-markets to the COVID pandemic have illustrated, this study highlights the need for spatially differentiated housing policies that consider the dynamics of local housing submarkets.

The research also sheds light into possible drivers of the divergent rental trends observed by investigating the transfers between the LTR, STR, and property investor markets. First, this analysis provides evidence to the potentially symmetrical nature of Airbnb’s relation to rental markets. While the analysis does not statistically test this relation and does not consider other drivers of STR trends such as informal leases or the accelerated uptake of property maintenance/renovations during low travel periods, the seemingly positive association between Airbnb listings activity and LTR rents and supply is evident in both the macro analysis and local case studies. While previous studies have documented the link between Airbnb growth and increasing rental prices (Horn and Merante 2017), this study provides one of the first examples of rental price reductions amid declining Airbnb activity where the latter have added to the local letting pool. This is a particularly important finding in the context of Airbnb policy aimed at controlling total Airbnb listings. Increasingly, tourist hotspots are introducing locally targeted regulation to counter the negative impacts STLs arguably have had on both LTRs and tourist accommodation in such areas. This includes Byron Bay council, which in June 2022 proposed a 90-day annual limit for Airbnb listings, subsequently over-ruled by the NSW State government who passed the issue to the NSW Independent Planning Commission for a decision. The results of this adjudication were awaited as this paper was finalised (Convery 2022 MacKenzie 2022;; Rachwani 2023). This research therefore contributes to the wider discussion around Airbnb regulation, particularly in regional areas where rental supply is already stretched. Second, the analysis on investor activity highlights a clear and consistent response from property investors to changing rental yields. This underscores the power of Australia’s highly financialised housing market to influence investor decisions, leaving renters to the vagaries of investor sentiment and strategies.

Finally, the research illustrates the power of linking multi-sectoral unit-level housing datasets to analyse housing market trends that transcend the division of traditional housing market sectors and also avoid the pitfalls of analysing relationships between spatially aggregated data sets. Linking data at the unit level between the STR, LTR, and property buyer markets can provide near real-time insights into rental housing market drivers that can aid policymakers to better respond to local and macro shocks. Ultimately, the findings of this study help understand the contrasting responses of metropolitan and regional areas to the COVID-19 pandemic, and potentially shed light on necessary future policy directions to better manage and regulate Australia’s housing market.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Long-term rental: Rental lettings for 6 months or more let on standard tenancy agreements and subject to the NSW Residential Tenancies Act 2010.

Short-term rental: Holiday lettings subject to the NSW regulatory framework for short-term rental accommodation under the Code of Conduct for the Short-Term Rental Accommodation Industry and registered with the NSW Department of Planning and Environment under the Environmental Planning and Assessment Regulation 2000.

2 The Australian Bureau of Statistics’ Statistical Area 2 generally have a population between 3,000 and 25,000 with an average of about 10,000 people. Statistical Area 3 cover areas which are broadly functionally similar with populations ranging from 30,000 to 130,000 people

3 The Airbnb data include all available lettings listed on the Airbnb website, including those with resident landlords. There is an argument to only include investor landlord Airbnb listings, given their more transient interaction with the LTR market. However, the difficulty in defining investor landlords through a combination of high frequency and whole property listings, given inconsistencies in definitions (e.g., Gurran and Phibbs 2017; Cocola-Gant and Gago 2021) and web-scraped data, have led researchers to generally include all listings in studies of Airbnb’s housing market impacts (e.g., Horn and Merante 2017; Barron, Kung, and Proserpio 2020; Thackway et al. 2022). It should also be noted that Evans, Rosewall, and Wong (2020) state that around three quarters of all Airbnb listings are for whole houses or apartments.

Appendix A

Vacancy Rate Areas.

Table

Study Area	Vacancy Rate Area	LGA’s
Greater Sydney	Inner Sydney	Ashfield Botany Bay Lane Cove Leichhardt Marrickville Mosman North Sydney Randwick Sydney Waverly Woollahra
	Middle Sydney	Auburn Bankstown Burwood Canterbury Canada Bay Hunters Hill Hurstville Kogarah Ku-ring-gai Manly Parramatta Rockdale Ryde Strathfield Willoughby
	Outer Sydney	Baulkham Hills Blacktown Blue Mountains Camden Campbelltown Fairfield Gosford Hawkesbury Holroyd Hornby Liverpool Penrith Pittwater Sutherland Warringah Wollondilly Wyong
Rest of NSW Hotspots	Northern Rivers	Ballina Byron Kyogle Lismore Richmond Valley Tweed
	Coffs Harbour	Bellingen Clarence Valley Coffs Harbour
	Mid-North Coast	Gloucester Great Lakes Greater Taree Kempsey Nambucca Port Macquarie
	South Coast	Bega Valley Eurobodalla
	Central West	Bathurst Regional Blayney Cabonne Lithgow Mid-Western Regional Oberon Orange