Economic Policy Uncertainty and Commercial Property Performance: An In-Depth Analysis of Rents and Capital Values

Abstract

Economic uncertainty has steadily increased in response to a series of unforeseen shocks, notably the Global Financial Crisis, Brexit, COVID-19, and the Russia–Ukraine war. This study examined the impact of economic uncertainty on rents and capital values in Australia’s office, retail, and industrial property sectors. The reactions of these performance indicators to national uncertainty shocks were assessed through reduced-form vector autoregressive (VAR) models, using quarterly data from 2001Q1 to 2022Q3. Overall, there is an inverse relationship between uncertainty and commercial property performance, with notable variations in magnitude and persistence across the different subsectors. Rents are more sensitive to external shocks across all three subsectors, highlighting their role as signals of short-term performance. Following one standard deviation shock in uncertainty, rents steadily declined for approximately three years in the office and retail subsectors. Industrial rents, however, exhibited muted reactions and recovered quicker, typically within five quarters. This resilience to external shocks displayed by the industrial subsector positions it as a compelling option for defensive investment strategies and portfolio diversification. Capital values are less reactive than rents, showing minimal responses to uncertainty shocks and little long-term persistence.

1. Introduction

The strength of the property market is generally considered a reliable indicator of the broader economy’s health, as evidenced by the strong correlations between Gross Domestic Product (GDP) growth, interest rates, employment rates, demand, and property prices (Baum 2009). These links make sustained property performance a topic of significant interest for researchers, policymakers, and investors. The property market is inherently complex, but the commercial property sector is even more intricate due to its unique financing structures, capital requirements, illiquidity, and information asymmetry (Clayton et al. 2009; Marcato and Nanda 2016). These inefficiencies make the commercial property space particularly susceptible to external economic conditions (Gholipour et al. 2022; Ling et al. 2010).

Over the past few decades, economic uncertainty has soared in response to a series of unexpected external shocks to the global economy, such as the Global Financial Crisis (GFC) of 2007–2008, Brexit, US–China trade tensions, the coronavirus pandemic (COVID-19), geopolitical concerns, supply chain disruptions, and the Russia–Ukraine war in early 2022 (Ahir et al. 2022; IMF 2022). Although these concerns are relevant across the global economy, uncertainty levels are further exacerbated in Australia due to a year of consecutive cash rate hikes by the Reserve Bank of Australia (RBA), rising inflation, and the fact that COVID-19 originated in the Asia–Pacific region (Allan et al. 2021; Gholipour et al. 2022; RBA 2023). In the immediate aftermath of these shocks, preliminary evidence suggests widespread implications for transaction volumes (Allan et al. 2021), vacancy rates (Gholipour et al. 2022), and reduced capital deployment in the commercial property industry (JLL 2023). Given the long-term nature of commercial property leases, the financial impact of these shocks will remain significant for years to come (Gholipour et al. 2022). The benefits of assets and geographical diversification for property portfolios are well researched (Eichholtz et al. 1995; Newell and Peng 2006), but recent market disruptions necessitate further examination, and the unique effects of the COVID-19 pandemic may only become apparent in the long term. Due to the varied effects of the pandemic and the subsequent capital reallocation from retail especially (Allan et al. 2021; Nanda et al. 2021), the findings of this study further highlight the temporal dynamics caused by economic policy uncertainty.

While the susceptibility of commercial property performance to external macroeconomic shocks is widely acknowledged, a notable gap exists in scholarly research regarding the extent of this impact on performance indicators. In the aftermath of the COVID-19 pandemic, research has increased on how rent levels, liquidity, and vacancy rates are affected (Allan et al. 2021; Gholipour et al. 2022; Ling et al. 2020). However, the existing literature lacks a comprehensive study investigating the relationship between uncertainty and crucial performance indicators that reflect the overall well-being of the commercial property market, such as rents and capital values. For instance, the increase in vacancy rates post-COVID-19 may directly result from lockdown restrictions, not necessarily economic uncertainty. Moreover, previous studies have predominantly concentrated on examining the impact of uncertainty on specific performance indicators within specific sectors, limiting the comparability and generalisability of their findings. There have also been divergent findings indicating no significant effects of uncertainty (Allan et al. 2021), or long-term implications for asset performance and portfolio construction (Che et al. 2023; Gholipour et al. 2021).

Motivated by growing concerns about economic uncertainty levels and the limited literature on how commercial property performance is impacted, this study addressed a notable gap by delving into the dynamic relationship between economic uncertainty and commercial property performance within the specific context of Australia. By investigating the response of rents and capital values to unit shocks in national uncertainty in the office, retail, and industrial subsectors, these findings offer further evidence of the dynamic nature of commercial property performance and the importance of subsector variations for overall investment strategies and portfolio management. For commercial property investors and portfolio managers, this focused examination also reflects the need for a more tailored understanding of the asset class, ultimately enhancing decision-making processes and contributing to the resilience and adaptability of commercial property portfolios in the face of economic uncertainty.

2. Literature Review

2.1. Property Investment Decision-Making

Property investment decision-making is a complex multi-stage process that requires investors to assess the current market performance and future expectations to maximise returns (Roberts and Henneberry 2007). Due to several distinct features and market inefficiencies, accurate performance assessment is more challenging in the commercial property markets than in residential or listed property markets (Ling et al. 2010; Marcato and Nanda 2016). Longer leases, low liquidity, higher transaction costs, lack of transparency, and information asymmetry all represent significant sources of inefficiency (Baum 2009). According to Marcato and Nanda (2016), these market inefficiencies increase search costs and encourage speculative purchasing behaviours, particularly under suboptimal market conditions and uncertainty. The performance of commercial property investment is also influenced by subjective factors such as investor sentiment and market trends, which often contribute to periods of sustained mispricing and market distortions in response to an external shock (Clayton et al. 2009; Heinig and Nanda 2018). These inefficiencies and distinct features make commercial property performance highly susceptible to broader macroeconomic forces (Nayar et al. 2023).

Investment decisions rely on future expectations, making it the most volatile component of aggregate demand in response to a significant exogenous shock (Carrière-Swallow and Céspedes 2013; Sah et al. 2010). Due to the forward-looking nature of property investment decisions, uncertainty about the future movement of the economy further complicates the decision-making process (Chmielewska et al. 2020). Knight (1921) originally distinguished uncertainty from risk by explaining that the former represents a situation where decision-makers have no probabilistic information on how external events may influence the outcomes of their decisions. According to Jackson and Orr (2019), significant economic shocks lead to fluctuations in perceptions and activity, ultimately impacting economic activity and other asset classes, such as commercial properties. Uncertainty can impact asset performance through various transmission channels. In the case of the property market, there is a well-established connection between macroeconomic uncertainty and performance due to the intrinsic linkages between the two domains (Akinsomi and Mkhabela 2016; Gabauer and Gupta 2020).

2.2. The Threat of Rising Uncertainty on Property Performance

Long-standing economic principles indicate that all investment decisions are invariably made under some level of uncertainty (Hargitay and Yu 1993). However, recent trends and events have steadily increased uncertainty levels (Ahir et al. 2022; Bloom et al. 2022). From 2016 to 2022, the global economy experienced a series of significant disruptions that steadily increased uncertainty: Brexit, the 2016 US presidential election, trade tensions between China and the US, the outbreak of COVID-19, and the Russia–Ukraine war (Ahir et al. 2022; Bloom et al. 2022). On a global scale, the International Monetary Fund (IMF) recognises the growing threat of uncertainty and the exacerbating effects of events such as the Russian invasion of Ukraine (IMF 2022). In addition to globalisation, improvements in information transmission channels, and an increased reliance on data-heavy decision-making models, these events have made uncertainty a primary concern of property investors (Baker et al. 2016; Christiaens and Macharis 2021).

The interplay between broader economic movements and property performance has been the subject of an extensive body of research, the consensus highlighting an inverse relationship between uncertainty and performance (Bird and Yeung 2012; Gholipour 2019; Jackson and Orr 2019; Nguyen and Lee 2021). High levels of uncertainty reduce investor confidence and impact decisions related to property investment, development, and financing (Jackson and Orr 2019; Wu et al. 2020). The sensitivity to uncertainty may vary across different sectors, for instance, with office markets potentially experiencing more pronounced effects compared to industrial or retail markets (Gholipour et al. 2022; Milcheva 2022). Additionally, the global nature of EPU shocks implies that shifts in major economies can have notable repercussions on international commercial property markets. Understanding the nuanced relationship between economic uncertainty and commercial property performance is crucial for stakeholders navigating these dynamic markets.

Although the threat of uncertainty is pervasive globally, the unique composition of each country’s property sector and diverse policy responses to these shocks lead to varying impacts (Schätz and Sebastian 2009; Wu et al. 2020). In the Australian context, several factors have elevated uncertainty levels for private markets like the commercial property space. First, the fact that COVID-19 originated in the Asia–Pacific region and lasted so long means that countries in this region have experienced more intense investor reactions (Allan et al. 2021). These concerns have been further exacerbated due to recent monetary policy decisions by the RBA and rising inflation (Chong 2023). Between May 2022 and May 2023, the RBA implemented a consistent series of cash rate target increases to address rising inflation, leading to notable impacts on interest rates and uncertainty levels (Chong 2023; RBA 2023). Consequently, uncertainty has emerged as the primary concern for most commercial property investors. JLL’s (2023) survey of top investment managers worldwide showed that over 70% of investors regard economic uncertainty as the biggest challenge to capital deployment in the commercial property sector.

2.3. Uncertainty and Commercial Property Performance: Market Implications

Previous studies in other developed property markets, such as the UK and the US, affirm that sustained periods of economic uncertainty increase property price volatility (Andre et al. 2017), reduce returns (Gholipour et al. 2021), and limit investment activity (Meinen and Roehe 2017). Uncertainty is also a key determinant of foreign direct investment (FDI) volumes because investors generally prefer to allocate funds to countries exhibiting lower levels of uncertainty (Lieser and Groh 2014). In the first six months of 2020, property transaction volumes in the Asia–Pacific region dropped by 32% as investors reeled back spending due to increased risk perceptions and the volatile landscape of the property market (Allan et al. 2021). High levels of uncertainty often lead to tighter credit conditions and increased borrowing costs, further constraining investment activity and dampening property prices (Bloom 2009).

Gholipour et al. (2022) examined the impact of COVID-19 on the Australian office market, reporting a positive relationship between uncertainty and vacancy rates that remained significant for at least three years after the initial shock. COVID-19 also impacted the resilience of the retail asset class through changes in spending patterns and the rise of online retail (Che et al. 2023; Nanda et al. 2021). Lang et al. (2022) also found causal links between uncertainty and excess returns in the commercial property market, reporting a significant positive relationship, which implies that investors demand higher premiums due to increased risk perceptions amid conditions of uncertainty. Similarly, Newell and Marzuki (2023) showed the impact of COVID-19 on global capital cashflows to the commercial property space by examining risk dynamics due to rising inflation concerns, geopolitical concerns, and supply chain issues. Beyond the performance metrics, the existing literature has also linked uncertainty to changes in investor attitudes and risk perceptions, which further influence investment strategies and capital deployment (Abdallah et al. 2020; Jackson and Orr 2019).

The outbreak of COVID-19, geopolitical tensions, and changes in monetary policies have all contributed to increased uncertainty in the Australian commercial property market, impacting investor sentiments and property performance. Despite reported negative responses of performance indicators to rising uncertainty, there remains a limited body of literature examining commercial property performance under these conditions. As uncertainty emerges as a primary concern for commercial property investors, the need for in-depth research into its effects on investment decisions and market outcomes is pertinent. Thus, this study addressed this gap by assessing how performance indicators such as rents and capital values react to exogenous shocks across different subsectors of the commercial property sector: office, retail, and industrial. These findings highlight the growing impact of economic uncertainty on commercial property performance and the varied responses in different subsectors to provide valuable insights into effective investment strategies for navigating periods of high uncertainty.

3. Results and Discussion

This study examined the impact of perceived economic uncertainty on commercial property performance in Australia, specifically focusing on rents and capital values in the office, retail, and industrial sectors. Using reduced-form VAR models, the discussions draw on the IRFs to trace the responses of commercial property performance indicators to shocks in national uncertainty levels.

3.1. Impact of Uncertainty in the Office Sector

As shown in Figure 1a,b, generalised IRFs were examined to trace the responses of office rents and capital values to one standard deviation (SD) in national uncertainty shocks. Overall, the findings show that office rents react negatively to positive national uncertainty shocks in the short term. Specifically, one SD shock in the EPU results in significant rental declines after the second quarter. This finding aligns with the delayed responses observed in the commercial property sector due to exogenous shocks such as the GFC of 2007–2008 and COVID-19 (Antonakakis et al. 2015; Gholipour et al. 2022). The decline continues for three years, following which rents reach a steady state and begin recovering. Gholipour et al. (2022) reported similar results based on office vacancy rates, although the authors observed that vacancy rates only react after a year. The discrepancy between rents and vacancy reactions indicates that rents are more sensitive to national uncertainty shocks in the short term, possibly due to the annual rental reviews that make rents more volatile to market disruptions, whereas vacancy rates are inherently tied to long-term commercial leases.

Practically, the sensitivity of office rents to external uncertainty shocks sheds light on the implications of broader economic movements for property investment strategies and portfolio management. The observed negative reactions of office rents to positive national uncertainty shocks underscore the importance of a dynamic investment strategy in the short term. Investors and portfolio managers should be prepared for significant rental declines in the aftermath of such shocks, with the impact becoming notably pronounced after the second quarter and persisting for three years (Lang et al. 2022). As the decline stabilises over the three-year period, opportunities emerge for strategic portfolio adjustments and potential repositioning to take advantage of higher returns as the market recovers (Gholipour et al. 2022; Newell and Marzuki 2023). This nuanced understanding of the temporal dynamics can inform investors and portfolio managers in tailoring strategies that balance short-term resilience with long-term recovery prospects, thereby enhancing the adaptability and resilience of commercial property portfolios.

Office capital values also respond negatively to unit shocks in economic uncertainty, although the magnitude is more significant than rents in the short term. Following one SD shock in perceived uncertainty, office capital values decline for up to six quarters, followed by a near-return to steady-state levels by the end of the fourth year. Given the dynamic relationship between yields and capital values (Baum 2009), the declines in capital values suggest that national uncertainty represents a substantial risk factor for which investors demand a premium (Tsolacos et al. 2018). This premium could be reflected in yields, through which capital values are impacted in the aftermath of an unexpected market disruption. Although both rents and capital values decline in the short term in response to national uncertainty shocks, the magnitude of the reaction and recovery reveals practical implications for investment strategies amid conditions of uncertainty. Capital values begin recovering after approximately six quarters, distinctly shorter than rents, which keep declining for up to three years. Allan et al. (2021) reported similar effects across the Asia–Pacific region, providing further evidence that rents are more volatile in the short term. Considering the susceptibility of rents and capital values to uncertainty shocks in the short term, investors and portfolio managers could pivot to long-term strategies to exploit opportunities that arise as markets recover after an unexpected exogenous shock (Allan et al. 2021; Milcheva 2022; Newell and Marzuki 2023).

3.2. Impact of Uncertainty in the Retail Sector

Following recent disruptions such as COVID-19, lockdown restrictions, and the rise of online retail, Australia’s retail sector has experienced significant capital reallocation and reduced performance (Allan et al. 2021; Nanda et al. 2021). In the immediate aftermath of the pandemic and subsequent lockdown restrictions, rents dropped by up to 32% in the Asia–Pacific region (Allan et al. 2021; Nanda et al. 2021), the resilience capabilities of retail assets were altered (Che et al. 2023), and capital was reallocated to different asset classes (Allan et al. 2021; Newell and Marzuki 2023). The IRFs presented in Figure 2a,b indicate why the post-pandemic transformation of the retail landscape is a crucial consideration for investment strategies and long-term capital allocation (Bitterman and Hess 2021).

In response to national uncertainty shocks, retail rents exhibit minimal reactions in the first quarter. After two quarters, there is a substantial decline in retail rents, which continues until an initial levelling off by the fourth quarter. In the subsequent two years, there are alternating periods of rises and declines, but retail rents notably do not return to a steady state over a five-year horizon. These declines reach a nadir by the tenth quarter, after which a gradual recovery begins. Notably, the declines in retail rents were lower than office and industrial assets, indicating resilience to market disruptions (Allan et al. 2021). Despite the well-reported effects of recent pandemic-related disruptions in this asset class, the diverse mix of essential services could explain this muted reaction through consistent demand even after significant market disruptions and uncertainty. Although some previous studies found no significant impact of uncertainty on retail performance (Allan et al. 2021), Che et al. (2023) suggested different retail clusters exhibit varying levels of resilience, in part due to their adaptability.

This pattern underscores the need for investors and portfolio managers to anticipate a delayed but significant impact on retail rents, necessitating strategic adjustments to mitigate potential losses (Allan et al. 2021; Jackson and Orr 2019). The subsequent two years, marked by alternating periods of rises and declines, indicate a prolonged period of volatility and emphasise the challenge of achieving stability within the retail sector in the aftermath of uncertainty shocks. Despite this volatility, Che et al. (2023) suggested that different sub-classes may exhibit varying levels of resilience to exogenous shocks. Portfolio managers should consider incorporating flexibility and diversification strategies to navigate the extended periods of volatility in the retail sector, particularly as opportunities for strategic repositioning become apparent after the initial uncertainty shock.

In contrast to rents, retail capital values display less volatility in response to perceived uncertainty shocks. This discrepancy can be attributed to the nature of rents, which tend to reflect short-term market signals, while capital values incorporate expectations for the future. Specifically, following a national uncertainty shock, retail capital values experience a decline for three quarters. However, after this initial period, they begin to recover and return to pre-shock levels within approximately three years. The resilience of capital values to these external shocks demonstrates retail’s role in strategic decisions and the muted reactions to exogenous shocks (Allan et al. 2021; CBRE Research 2021).

This divergence suggests that investors and portfolio managers may find retail capital values to be more resilient indicators of long-term value, allowing for a more strategic and measured approach to decision-making (Allan et al. 2021). The real options theory becomes particularly relevant in this context, as the resilience of retail capital values allows investors to exercise flexibility in their choices, considering the option to wait and observe market dynamics before committing to irreversible capital allocation decisions (Bird and Yeung 2012; Gholipour et al. 2022; McDonald and Siegel 1986).

3.3. Impact of Uncertainty in the Industrial Sector

The findings regarding industrial rents, as illustrated in Figure 3a,b, highlight the resilience of the industrial property subsector compared to office and retail, carrying significant practical implications for investment strategies, portfolio management, and repositioning. The recent literature highlights the growing significance of the industrial property subsector and last-mile logistics, attributable primarily to lockdown restrictions and the surge in online retail activities (Carson et al. 2021; Nanda et al. 2021). In response to one SD uncertainty shock, industrial rents immediately decline, but the recovery commences after approximately three quarters. Rents return to their steady-state levels by the fifth quarter, fully rebounding from the initial shock and displaying minimal long-term persistence.

This resilience aligns with the evolving landscape of the industrial property subsector, which has gained increased significance due to lockdown restrictions and the surge in online retail activities, making industrial assets pivotal components within investment portfolios (Allan et al. 2021; CBRE Research 2022). The muted reactions of the industrial assets also suggest that they could play a key role in enhancing portfolio resilience during periods of uncertainty. Investors and portfolio managers may consider allocating a proportion of their portfolios to industrial assets to mitigate the impact of exogenous shocks on overall portfolio performance. Allan et al. (2021) reported similar flight-to-safety trends amid the COVID-19 pandemic, as shown by significant capital reallocation from retail to industrial assets in the Asia–Pacific region. Similarly, Jackson and Orr (2019) reported higher resilience in the industrial sector in the UK compared to the retail and office sectors. Milcheva’s (2022) investigation of the US commercial property space also aligns with these findings, noting that the COVID-19 pandemic exacerbated the struggles of some subsectors, such as retail, and provided new opportunities for industrial and residential assets. It is telling that many commercial investors in the Australian commercial property space pivoted from other assets to industrial and logistics in the immediate aftermath of the COVID-19 pandemic, thus increasing the industrial market share (Allan et al. 2021; CBRE Research 2022).

Similarly, industrial capital values exhibit minimal sensitivity to national uncertainty shocks. Capital values decline in response to national uncertainty shocks, reaching their lowest point around one year after the shock but subsequently recovering and returning to pre-shock levels within six quarters. From an investment strategy perspective, the minimal sensitivity of industrial capital values implies that investors may consider allocating a portion of their portfolios to industrial assets as a hedge against external shocks (Milcheva 2022). This allocation can contribute to overall portfolio resilience and mitigate the impact of uncertainty on the investment portfolio’s value. The observed recovery within six quarters suggests that industrial capital values have the potential for relatively swift rebounding, providing opportunities for strategic adjustments and repositioning. Considering the increased importance of industrial and last-mile logistics assets in commercial property portfolios (Carson et al. 2021; Lashgari and Shahab 2022), this resilience also provides a basis for more defensive strategies during periods of economic uncertainty.

4. Materials and Methods

4.1. Data

With a total capital value of AUD 226.6 billion, Australia’s commercial property sector is composed of three main subsectors: office (45.1%), retail (28.8%), and industrial (19.9%) (MSCI 2023). This study examined the impact of economic uncertainty on Australia’s office, retail, and industrial sectors using quarterly data from 2001Q1 to 2022Q3.

Table 1. Variable Description.

Variable	Description	Source
Off_Rent	Prime office net face rents	CBRE Research
Off_Cap	Prime office capital values
Ret_Rent	Regional retail net face rents
Ret_Cap	Regional retail capital values
Ind_Rent	Super-prime industrial net face rents
Ind_Cap	Super-prime industrial capital values
EPU	Economic policy uncertainty index (Australia)	Economic policy uncertainty www.policyuncertainty.com (accessed on 8 April 2024)
CPI	Consumer price index (quarterly % change)	Australia Bureau of Statistics
GDP	Gross domestic product (quarterly % change)
UNEM	Unemployment rate (expressed in %)
HP	Real house price index, deflated by CPI

briefly describes the variables and data sources.

Proprietary data on net rents and capital values were provided by CBRE, a global leader in commercial real estate services and investment. The dataset includes aggregated national data on premium offices, regional retail centres, and super-prime industrial properties. Prime office spaces mainly refer to landmark office buildings in major CBD markets with first-grade design and functionality, regional retail comprises shopping complexes serving a distinct city or state, and the super-prime industrial subsector includes facilities of superior design, functionality, and location (Property Council of Australia 2023). The values presented in the dataset are based on per square meter measurements, ensuring a standardized comparison across the different property types. This coverage allows for a comprehensive analysis of prime real estate markets across major cities in Australia.

To resolve potential issues of data smoothing, which is prevalent in appraisal-based indices for private investment, the capital values were desmoothed using the first-order autoregressive (AR1) process proposed by Geltner (1993) and adopted in similar studies testing commercial property performance (Bond and Hwang 2007; Hoesli and Oikarinen 2016; Lee 2008; Lee et al. 2022). Smoothing originates from temporal aggregation bias and the anchoring bias of appraisers acting in an imperfect market characterised by limited transactions and incomplete information (Bond et al. 2012; Geltner 1993; Quan and Quigley 1989). Although different desmoothing parameters have been used in previous studies, Clayton et al.’s (2001) findings provided an empirical basis for adopting 0.8 to represent the weight given to new information by appraisers. As such, the capital values were unsmoothed with an alpha value of 0.8 to generate the true capital values using Equation (1), as proposed by Geltner (1993):

$$r_t={\displaystyle \frac{1}{1-\alpha }}\left(r_t^{\ast }-\alpha r_{t-1}^{\ast }\right)$$

(1)

where $r_t$ denotes the actual underlying return, $r_t^{\ast }$ is the reported valuation-based return at t, and $\alpha$ represents the smoothing parameter—a weight given to information about the prior valuation, $\alpha \in \left(0, 1\right)$. In this study, 0.8 was adopted as the desmoothing parameter. Variations of this parameter within the range of ±10 percentage points were also tested.

Further tests were also conducted to verify the consistency of results due to the potential sensitivity of results to variations of the desmoothing parameter. Following the recommendations of Hoesli and Oikarinen (2016), the direction and magnitude of impact remained relatively unchanged, even with desmoothing parameters of 0.7 or 0.9.

The Economic Policy Uncertainty (EPU) index for Australia was adopted as a proxy for perceived uncertainty. This index is constructed following the methods outlined in “Measuring Economic Policy Uncertainty” by Baker et al. (2016) (available at https://www.policyuncertainty.com; accessed on 8 April 2024). The process counts the number of articles containing the term ‘uncertainty’ and its variants in eight Australian newspapers: Daily Telegraph, Courier Mail, The Australian, The Age, The Advertiser, Mercury, Sydney Morning Herald, and The Herald Sun. These raw counts are then scaled by the number of all articles in the same newspaper and month, standardised, averaged by month, and multiplicatively rescaled to a mean of 100 from January 1998 to December 2012. Although the index is published monthly, the end-of-quarter values were taken to maintain consistency with other variable frequencies.

Despite the availability of alternative uncertainty measures, such as stock market volatility and forecast dispersion, the EPU index is advantageous because it incorporates topics of fiscal, monetary, regulatory, and trade policy relevance (Moore 2017). The text-based methodology also captures the prevailing perceived uncertainty among market participants and is regularly updated (Gholipour et al. 2022; Moore 2017). Consequently, the EPU index has become an established and popular measure of economic uncertainty in the extant literature (Andre et al. 2017; Gholipour 2019; Gholipour et al. 2022; Jackson and Orr 2019; Wang et al. 2020; Wu et al. 2020).

The final set of control variables includes macroeconomic indicators and the house price index to account for the influence of broader economic factors and the systemic effects on commercial property performance (Hoskins et al. 2004; Ling and Naranjo 1997). These macroeconomic variables include quarterly CPI, GDP growth, and unemployment rates (UNEM). Given the intrinsic links between house prices and commercial property performance (Baum 2009; Lan 2019), the real house price index also proxies the performance of the Australian housing market. The coverage of this index contains new and existing dwellings in the eight state capitals across Australia, deflated by CPI.

Table 2. Descriptive Statistics.

	Statistics					Normality			ADF Test
Variable	Mean	Median	Min	Max	Std. Dev.	Skew.	Kurt.	JB-Stat	Prob.	Lags	Level
Off_rent	6.430	6.487	6.017	6.796	0.245	−0.376	−1.078	6.144 **	0.014	3	I (0)
Off_cap	9.225	9.215	8.558	9.890	0.401	−0.053	−1.147	4.782 *	0.063	3	I (0)
Ret_rent	7.260	7.295	6.906	7.404	0.144	−1.075	0.107	16.413 ***	0.026	1	I (0)
Ret_cap	10.108	10.153	9.580	10.484	0.215	−0.960	0.513	13.758 ***	0.085	0	I (0)
Ind_rent	4.615	4.667	4.317	4.909	0.143	−0.964	−0.068	13.756 ***	0.024	0	I (1)
Ind_cap	7.314	7.298	6.652	8.224	0.351	0.375	−0.006	1.853	0.000	0	I (1)
EPU	4.533	4.581	3.338	5.743	0.542	−0.156	−0.243	0.665	0.013	1	I (0)
CPI	2.603	2.500	−0.300	7.300	1.258	1.243	2.676	51.026 ***	0.004	3	I (1)
GDP	0.708	0.700	−6.700	3.900	1.096	−3.152	25.419	2132.99 ***	0.001	3	I (0)
UNEM	5.403	5.391	3.486	6.962	0.739	−0.128	0.106	0.226	0.079	1	I (0)
HP	4.580	4.573	4.061	4.955	0.199	−0.322	−0.157	1.153	0.022	3	I (0)

Note: This table presents the descriptive statistics of all variables adopted for this study. The dataset contains 87 observations (2000Q1 to 2022Q3). Normality assumptions were tested with skewness (Skew.), kurtosis (Kurt.), and the Jarque–Bera test (JB-stat). The p-values of the JB-stat are denoted by asterisks; ***, **, and * indicate statistical significance at 1%, 5%, or 10%, respectively. Stationarity assumptions were also tested with the Augmented Dickey–Fuller (ADF) test. The ADF test reports the p-value of the statistic for the first lag with a significant value (as suggested in the ‘Lags’ column, determined by the Akaike Information Criterion). I (0) is integrated of order zero, denoting that the variable is stationary at level. I (1) is integrated of order one, denoting that the variable is stationary after taking the first difference.

presents the descriptive statistics of all the variables.

4.2. VAR Model

The responses of commercial property performance to national uncertainty shocks were tested using reduced-form vector autoregressive (VAR) models. Without an established theory of causal relationships, VAR estimations are ideal because all variables in the system are treated as endogenous (Sims 1980; Stock and Watson 2001). This model specification overcomes the limitation of imposing strict relationships between independent and dependent variables (Marcato and Nanda 2016). In a reduced-form VAR specification with two variables and one period, the time path of one sequence, y_t, is influenced by the current and past (one period lag) realisations of a different sequence, x_t, and vice versa (Gholipour et al. 2022; Stock and Watson 2001). Different VAR systems were specified for each performance indicator—office rents and capital values, retail rents and capital values, and industrial rents and capital values. Each of these systems tested the response of the commercial property performance indicators to unit shocks in national uncertainty levels. Equation (2) shows the matrix representation of the VAR model:

$$\left[\begin{array}{c}{y}_t\\ x_t\end{array}\right]=\left[\begin{array}{c}{a}_1\\ a_2\end{array}\right]+\left[\begin{array}{cc}{b}_{11}& b_{12}\\ b_{21}& b_{22}\end{array}\right]\left[\begin{array}{c}{y}_{t-1}\\ x_{t-1}\end{array}\right]+\left[\begin{array}{c}{u}_t\\ v_t\end{array}\right]$$

(2)

This model specification is particularly well-suited for the study due to its ability to capture the dynamic interrelationships among multiple time series variables without imposing restrictive theoretical assumptions (Pesaran and Shin 1998; Stock and Watson 2001). This approach has been effectively employed in similar studies within the property market context. For instance, Marcato and Nanda (2016) also leveraged the VAR model to explore the interplay between economic conditions and property market performance, further validating its appropriateness for this type of analysis. Gholipour et al. (2022) also applied a VAR framework to analyse the effects of economic uncertainty on vacancy rates and liquidity in the commercial property sector, demonstrating the model’s robustness in capturing temporal dynamics. By adopting a VAR model, this study aligns with established methodologies in the field, ensuring a rigorous and comprehensive examination of the dynamic relationships between economic uncertainty and commercial property performance in Australia.

These results were interpreted through generalised impulse response functions (IRFs), which represent the dynamic response of a system to an instantaneous input signal, providing insights into the system’s behaviour over time (Pesaran and Shin 1998). According to Pesaran and Shin (1998), impulse responses in a VAR model provide a more scientifically rigorous framework for interpretation than individual coefficients. These IRFs show the direction, magnitude, and significance of uncertainty shocks to the property market over a 5-year time horizon after the first-moment shock. The estimating method can be expanded to include a variety of variables and multiple lags. In practice, choosing pertinent factors and the appropriate lag length must be aided by economic theories (Marcato and Nanda 2016; Sims 1980; Stock and Watson 2001). As such, the lag length in these model specifications was determined via the three primary information criteria for model selection: Akaike, Schwarz–Bayesian, and Hannan–Quinn, in addition to Akaike’s final prediction error (FPE) procedure and the sequential modified Likelihood Ratio (LR) test. The LR test examined the hypothesis that the coefficients on a predefined lag are collectively zero, utilising a X2 (Wald) statistic (Marcato and Nanda 2016). Starting from the maximum lag, the lag length was gradually reduced by one period until rejection at the 5% significance level was validated.

Diagnostic tests of stationarity and white-noise disturbances were conducted before the interpretation of outputs to maintain the basic assumptions of the VAR model. First, the stability of each VAR model was checked with the Inverse of AR Characteristic Polynomial test. As shown in Appendix A, all models are stable since the roots have moduli less than one and lie within the unit circle. The Augmented Dickey–Fuller (ADF) test was also conducted to confirm the stationarity of the variables, utilising the Akaike Information Criterion (AIC) to identify the suitable lag length, followed by the differencing of non-stationary variables. The results of these unit root tests are presented in Table 2. Further, Lagrange Multiplier (LM) tests were conducted to determine the presence of autocorrelation in all model residuals, and are presented in Appendix B. The tests returned p-values above 0.05, which were jointly rejected to establish the reliability of the estimations and the absence of omitted variables (Johansen 1995). Thus, the residuals of all VAR estimations exhibit no systematic patterns or dependence over time (Gholipour et al. 2022; Johansen 1995; Sims 1980). Appendix C, Appendix D and Appendix E also detail the variance decomposition of all the VAR systems, showing the contributions of different variables in explaining the rents and capital values over the 5-year horizon after the first-moment shock in national uncertainty (Pesaran and Shin 1998).

5. Conclusions

This study examined the impact of national uncertainty shocks on the performance of the commercial property sector in Australia, specifically focusing on the office, retail, and industrial subsectors. The primary motivation behind this research stems from the unprecedented levels of uncertainty triggered by various shocks such as COVID-19, Russia’s invasion of Ukraine, disruptions in global supply chains, and geopolitical tensions. Within the Australian context, the surge in inflation and recent monetary policy decisions by the RBA to consistently increase the cash rate target has further compounded the issue of uncertainty, resulting in significant declines in investment volumes and a reduction in the overall performance of the commercial property sector.

The response functions of rents and capital values were tested through reduced-form VAR models, using quarterly data from 2001Q1 to 2022Q3. To present comprehensive insights into these relationships in the Australian commercial property market, the response of these performance indicators to national uncertainty levels was estimated in all three major subsectors: office, retail, and industrial. In addition to including several control variables, including quarterly changes in CPI and GDP, unemployment rates, and the real house price index, several diagnostic and robustness checks were conducted to ensure the validity of the results. These findings were then discussed based on IRFs, which assessed the response of rents and capital values to unit shocks in the EPU.

Rents emerged as the most sensitive indicator to uncertainty shocks in the Australian commercial property sector, indicating that rents provide performance signals in the short term. Both rents and capital values exhibited muted reactions in the initial two quarters following the shocks, confirming the long-standing notion of a lagged response of the property sector to external disturbances. In the case of office rents, national uncertainty shocks lead to declines starting after two quarters and persisting for three years. This finding contrasts with recent research on other indicators like vacancy rates, which only begin to react after a year. Uncertainty also affects capital values in the short term, which suggests that investors in office, retail, and industrial properties demand a premium to compensate for the significant risk associated with uncertainty. Across all three subsectors, capital values exhibit minimal reactions to uncertainty shocks; office and retail capital values recover after three years, while industrial capital values recover within a year.

Across the different subsectors, distinct reactions to uncertainty affect investment strategies, portfolio management, and repositioning. Office and retail assets are more sensitive to unexpected uncertainty shocks, demonstrating the relative resilience of the industrial sector. While all sectors initially react similarly within the first year following an exogenous shock, the industrial sector’s responses are smaller compared to office and retail. Moreover, the industrial sector recovers faster for all three performance indicators, typically starting within a year. In contrast, office and retail require approximately three years to return to pre-shock levels. The industrial property sector’s recovery speed and relative stability make it an attractive choice for defensive strategies, risk-averse investors, and capital reallocations during uncertain economic conditions, providing a potential flight-to-safety destination.

These insights also emphasise the need for a nuanced approach to portfolio management, recognising the varying sensitivities of different property subsectors to uncertainty shocks. In particular, the changing landscape of office and retail habits requires agile strategies to navigate short-term volatility. Overall, understanding the increased resilience of industrial assets provides valuable insights for investors seeking to navigate uncertain market conditions and optimise the performance of their portfolios. Not all property subsectors respond similarly to external uncertainty shocks, and the distinct characteristics of the industrial sector contribute to its resilience. Investors may leverage this understanding to optimise their portfolios, tailoring their investments to align with the specific dynamics of different subsectors.

Regardless of these insights into the relationship between uncertainty and performance in Australia’s commercial property sector, certain limitations inhibit the generalisability of the results. First, employing a larger dataset or panel data would enhance the robustness of the findings and allow for a more comprehensive examination of the relationship between uncertainty and commercial property performance. Notably, this study underscores the need for future research to delve into investor repositioning strategies under conditions of uncertainty, exploring how stakeholders adapt and optimise their portfolios in response to dynamic economic environments. Moreover, an intriguing avenue for exploration lies in assessing whether economic uncertainty has become an increasingly influential factor driving commercial property performance over time, considering the evolving global economic landscape. The limitations of the EPU index, which measures the prevalence of uncertainty in the news, could also be addressed by adopting alternative uncertainty proxies such as stock market volatility and forecast disagreement. A more comprehensive understanding of these dimensions can contribute to refining real estate investment strategies and bolstering the resilience of portfolios in an ever-changing economic climate.