Tenant Business Characteristics and Commercial Real Estate Portfolio Fundamental Performance

Abstract

This paper extends the literature on tenant characteristics and concentration on REIT performance. It fills in the knowledge gap of understanding how tenant business characteristics influence risk-return performance of directly held commercial real estate portfolios. Unlike the literature, it avoids compounding the peripheral information of financial leverage and capital market volatility into analyses for publicly traded real estate portfolios. We build a unique database of direct real estate by manually collecting fair value and rental value of investment real estate from annual reports of publicly traded real estate companies or trusts in New Zealand. Sharpe ratio and its significance on Z-statistic are analyzed and compared across twenty-four hypothetical portfolios sorted by geographic region, real estate sector and tenant business characteristic, respectively. It is found that certain tenant business industry can provide significantly superior performance during the entire sample period over other hypothetical portfolios. Specific tenant business industry and real estate sector show strong resilience to the 2007 global financial crisis. Tenant headquarter location signals commercial real estate risk-adjusted return performance. Our findings can facilitate researchers to extend their exploration of publicly traded real estate portfolios and help the understanding of market efficiency for commercial real estate and its related markets.

Keywords:

• Commercial real estate portfolio management
• tenant business
• sharpe ratio

This paper uses a novel and unique dataset to examine how tenants’ business can affect commercial real estate fundamental performance. It extends from the existent literature on how tenant diversification affects mortgage default (Ambrose et al., 2018) and the impact of tenant quality, concentration or earnings announcements on Real Estate Investment Trust (REIT) performance (Chacon, 2023; Chacon & Evans, 2023; Liu et al., 2019; Zheng & Zhu, 2021). The present study differs from the previous studies in that it analyses commercial real estate performance at the physical portfolio level, instead of mortgage or trust level. It treats aggregate physical commercial buildings as portfolios and investigates the risk-return performance based on portfolios consisting of real estate only (direct real estate investments). It avoids introducing peripheral information about operational expenditure, management efficiency or financial leverage into the risk-return analyses, unlike the previous studies on REIT performance (indirect real estate investments) at the level of business entity—trust (Chacon, 2023; Chacon & Evans, 2023; Liu et al., 2019; Zheng & Zhu, 2021).

The literature has analyzed how tenant characteristics affect a single real estate sector. The effect of tenant potential performance and characteristics has been examined in the context of retail real estate (Letdin et al., 2023; Liu & Liu, 2013). The findings of Stevenson et al. (2023) imply that tenant quality impact Australian office rent. The previous studies have yet to investigate how tenant characteristics influence commercial real estate portfolio performance at the physical building level across different real estate sectors.

This study fills in the knowledge gap in the literature and provides empirical evidence on how tenants’ business affect risk-return performance of real estate portfolios consisting of physical assets only. It provides additional insight to researchers and existent literature that tenants’ characteristics are important towards the portfolio performance based on physical assets. It sheds light on the rationale behind REITs’ investors response towards tenants composition or concentration. The detailed analyses based on various sorted portfolios using physical assets provide additional empirical evidence to the literature and highlight the importance of prudentially selecting tenants into commercial buildings at the sake of promoting real estate portfolio performance.

This research is conducted in a small economy, New Zealand, which allows us to exhaust the data of all individual commercial real estate that are held by publicly traded real estate companies or trusts. It has made it feasible for us to manually collect real estate fair value, income return and available tenant information regarding all individual buildings through published annual reports. The study spans four main real estate sectors—office, retail, industrial and medical sectors. The tenants are categorized with regard to their business industries, business operational area and head office location. We provide reasons on these categories in the Section of “Classification of Tenants”.

The present study extends the approach of analyzing diversification benefit for real estate portfolios from the literature, in order to examine the risk-return performance on hypothetical portfolios. Previous studies address the significance of diversification benefit for real estate portfolios, at the levels of indices and individual real estate (Baum & Colley, 2017; Cheng & Liang, 2000; Graff et al., 1997; Gyourko & Nelling, 1996; Hartzell et al., 1986; Lee & Byrne, 1998; Ling et al., 2022; Malhotra et al., 2020; Mattarocci & Scimone, 2020; Miles & McCue, 1982; Newell et al., 2005; Newell & Webb, 1996; Wolverton et al., 1998). Following a similar approach from the literature on portfolio analyses, we construct hypothetical real estate portfolios in line with three dimensions: geography, real estate sectors and tenants’ business characteristics. Tenants’ business is an additional dimension on top of those in the previous studies on sorting portfolios. We compare risk-return performance of the portfolios using Sharpe ratio. We analyze the statistical significance of different Sharpe ratio performance between each pair of hypothetical portfolios based on Z-statistic, following Jobson and Korkie (1981) and Lee and Stevenson (2005).

The findings show that selecting tenants of specific business industries or characteristics into real estate portfolios can promote portfolio risk-return performance. The research provides practitioners’ empirical evidence supporting their practice on selecting tenants regarding tenants’ business industries, operational area and head office location. The understanding of the importance of tenants’ business characteristics on real estate portfolio fundamental performance can facilitate researchers to conduct further analyses on direct and indirect real estate investments and financing decisions when the book value (fair value in annual reports) of real estate concerns (Alcock & Steiner, 2017; Tan, 2017).

This research contributes to the knowledge body and real estate portfolio management profession by building a bridge and rationalizing the connection between the understanding of tenant business characteristics and publicly traded real estate portfolio performance. The significant results of the study regarding tenant business characteristics on real estate portfolio fundamental performance, combined with existent findings of tenant quality/concentration on REIT performance (Chacon, 2023; Chacon & Evans, 2023; Liu et al., 2019; Zheng & Zhu, 2021), may imply certain level of market efficiency in indirect real estate investment markets.

The next section reviews relevant literature about risk-return performance of real estate portfolios and tenants’ characteristics. The classification of tenants is discussed in the subsequent section, followed by a section on total return and Sharpe ratio. Data source is then addressed. Results on Sharpe ratio and Z-statistic are presented and discussed. Further analyses on Sharpe ratio for each tenant business industry across each real estate sector are discussed. The final section concludes the research.

Literature Review

The Importance of Tenants’ Characteristics on REITs or Single Real Estate Sector Performance

Dated back as early as 2009, Smith introduces the relationship between tenants’ characteristics and commercial lease success. His study argues that tenant characteristics reflect the likelihood of tenant failure/default and the subsequent bank suffrage on overvalued assets. Later studies on tenants’ quality or characteristics have started to sprout in the literature.

Liu and Liu (2013) emphasize that, in the context of retail real estate, tenants’ own firm performance can be affected by other tenants’ contagion or competition in the same premise based on consumers’ complementary or comparison shopping. Tenants’ externalities in retail premise affect the fluctuation and performance of retail real estate, due to the percentage lease of retail real estate (Benjamin et al., 2000; Benjamin & Chinloy, 2004; Colwell and Munneke, 1998; Williams, 2014).

Lu-Andrews (2017) examines tenant quality measures—Altman Z-score and tenant size on REIT liquidity management—cash holdings and utilization of credit line. He finds that these two measures are inversely related to REIT total liquidity and unused credit lines. The findings vary across different property types. Ambrose et al. (2018) investigate how tenant diversification affects credit spreads for mortgage on retail real estate. They find that tenant diversification increases the credit spreads of mortgages.

Liu et al. (2019) analyze how tenant quality influences asset quality and REITs risk-return performance. They find that tenant quality increases the stability of long-term cash flows. Zheng and Zhu (2021) examine tenant concentration on the operational performance of REITs at the level of business entity—trust. They find that REITs with concentrated tenants are likely to face market valuation discount in the stock markets. Fisher et al. (2022) have stated the importance of tenants on the quality of buildings regarding the risk performance. Their findings also imply the effect of tenant quality on the systematic risk of buildings. Different from the early study on tenant concentration base, Chacon (2023) find that concentrated tenants lead to positive REIT profitability due to operational efficiency based on concentrated tenants.

Chacon and Evans (2023) test REIT investors’ reaction to earnings announcement of tenants of real estate held by REITs. They find that investors pay limited attention to negative surprises in material news announced by large tenants. Their results trigger our interest to investigate actual real estate portfolio performance by aggregating individual physical assets based on tenants’ potential quality.

Existent studies about the effect of tenants on REIT performance contain the volatility of financial market performance, as reflected by REIT price, market value of equity and REITs. The existent findings are prone to the fluctuation of financial market. The present study uses individual asset level data to analyze risk-return performance of commercial real estate portfolios under direct investments. The valuation data for individual real estate reflect asset-level value, instead of trust/company level value. Thus, the return information of our study does not depend on financial leverage or capital market performance.

Leung et al. (2023) argue the impact of specialty stores on the optimal tenant mix and rent revenue for shopping malls. They find convex curve of for total rental revenue based on the number of retail stores regarding the different levels of product substitution and marginal production cost. Their study implies the effect of agglomeration and competition among retail stores, including specialty stores, on the optimization of tenant mix and rent revenue. Their findings on product substitution can provide general implications on tenant concentration for retail shopping malls.

We are intrigued by Song and Liow (2023) findings of significant return premiums on their measure of industrial tail exposure risk for REITs. It is interesting to explore the relationship of extreme left-hand side return distribution between REITs and the twelve Fama-French industries. Our understanding of the return distribution and premiums for REITs come from systematic risk—macroeconomic drivers, capital market performance and unsystematic and idiosyncratic risk—locational choices and sectors of real estate on hold, REIT management and tenants. Each commercial real estate tenant fall into a specific business industry.

The literature has documented the importance of tenant characteristics, quality or concentration on REITs performance or specific real estate sector performance. It has not yet explore tenant business characteristics on the risk-return fundamental performance of commercial real estate portfolio. The present study extends from the earlier studies regarding the importance of tenant characteristics and the risk-return performance of real estate portfolios.

The Risk-Return Performance of Real Estate Portfolios

The literature on risk-return performance of real estate portfolios under direct investments focuses on the physical aspects of real estate assets. For example, earlier studies focus on geographical diversification by identifying several different regions based on its geographical boundaries, and real estate assets are assigned into one particular region (Byrne & Lee, 2000; Cheng & Liang, 2000; Eichholtz et al., 1995; Lee & Byrne, 1998; Lee & Stevenson, 2005; Miles & McCue, 1982, 1984). In US markets, there are several regional classification models such as the four-NCREIF-region model and eight-NCREIF-sub-region model. Economic diversification analyses the economic factors of each geographical location and re-groups these geographical locations into several economic regions instead of the traditional territorial boundaries (Hartzell et al., 1987; Heydenreich, 2010; Lee & Byrne, 1998; Mueller, 1993). Regarding real estate sector diversification, the real estate assets are classified into office, industrial, retail, residential, and hotel. This classification method is based on usage types or real estate functions.

The present study investigates risk-return performance of hypothetical portfolios when real estate assets are sorted into portfolios by the above geographical location, real estate sectors and tenants’ business characteristics. Tenants’ business characteristics reflect potential tenants’ quality. We argue that tenants quality affect rental income of real estate, based on the direct logic of from the literature (Grenadier, 1995; McConnell & Schallheim, 1983; Miller & Upton, 1976; Rubinstein, 1976; Smith, 2009) and indirect implications from the earlier studies (Chacon, 2023; Chacon & Evans, 2023; Liu et al., 2019; Liu & Liu, 2013; Zheng & Zhu, 2021).

The risk in real estate investment is directly influenced by a tenant’s decisions. A commercial lease is the governing document that determines the relationship between a landlord and tenant. A tenant periodically pays the rental and other recoverable expenses to a building’s owners in exchange for the rights to occupy real estate asset for a definite term. While a building is tenanted, real estate investors face a potential for a tenant to default on lease contract. At the time a lease expires, a tenant is able to re-negotiate the terms for space and rental price. A landlord is exposed to the risk of vacancy periods and searching costs to find an alternative tenant. It is argued that tenants play an important role in the determination of the return and risk profile for real assets because tenants are the ultimate source of income for real assets (Frew & Jud, 1988; Sivitanides, 1997; Wheaton & Torto, 1988). Tenants’ default has significant impact on the income stream from commercial real estate.

The Foundation of Tenant Quality on Real Estate Asset and Portfolio Performance

Through tenants’ lens, it is argued that tenants prefer to remain in their present locations although they are not permanently attached to a particular real estate asset. There are four reasons behind this observation. Firstly, long-term lease contracts are more preferred by landlords because they can serve as a strong signal of the certainty of future cash flows generated from real estate assets. Lease expiries or break dates provide the possibility of a tenant’s relocation and affect the future income of the building. Most lease contracts have a few infrequent rental review dates. This implies that landlords compromise the flexibility of adjusting rental timely in response to market movement in order to ensure certainty.

Secondly, significant transaction costs and/or opportunity costs are imposed by the imperfect allocation of real estate assets. Under bargaining theory as developed by McAllister and Tarbert (1999), landlords who have poor-quality tenants are constrained from replacing poor tenants with better-quality tenants. Landlords are often prepared to stay in a compromised position due to searching costs and unwanted premise vacancy periods. These costs are described as variable costs by McAllister and Tarbert (1999) and depend on unpredictable market conditions. As argued by Crosby and Murdoch (1997), tenants often receive an incentive package after they threaten their landlords that they want to relocation to an alternative premise, in particular during periods when the demand-side of markets is weak. This phenomenon is more evident in small economies where landlords’ choice of finding replacement tenants is limited. When a landlord fails to maintain high quality tenants, the optimal value of real estate assets cannot be realized, due to the inferior quality of tenants.

Thirdly, the potential costs associated with the business relocation process impose significant barriers to both tenants and landlords. In reality, business relocation is a complicated, painful, and time-consuming procedure. Tenants prefer focusing on their main business activities rather than going through this procedure without a real incentive. Negus (2007) analyzes the leasing behavior of restaurant tenants and cites that restaurant tenants are more stable than other types of retailers because of the immovability of high valued restaurant fit-outs. Negus (2007) also implies that tenants’ adhesiveness depends on the nature of tenants’ business and is reflected by the structure of lease agreement.

The last but not the least, economic region theory suggests that the level of business activities within a region has a significant impact on the performance of a real estate asset through influencing the demand for real estate assets within the region. In consideration of market imperfection, it is argued that tenants’ business activities affect the tenants’ affordability and security that are reflected in building value in terms of rental income and risk associated with the income. It is anticipated that there is a certain level of heterogeneity for the performance between buildings attributing to tenants in different business industries. Further, it is implied that there is potential of reducing risk and improving risk-return performance for buildings that attract tenants in different business industries. Crosby et al. (2006) suggest that the unique lease structure for small businesses is tailored to suit their specific risks, to which large tenants are not significantly exposed.

It is, therefore, argued that the ex-ante quality of tenants determine the risk and return characteristics of asset performance. The expected future performance of a tenant could have direct impact on the fluctuation of future rental income and thus real estate value. Due to the existence of potential barriers on tenant adhesiveness, such impact can continue during or even beyond their lease terms until a tenant finally moves out of premise.

Figure 1 summarizes the logic on how tenant quality impacts risk-return performance for real estate portfolios. The quality of tenants determines the risk profile of a lease contract and influences the valuation of lease contract. The valuation of lease contract affects the total return of real estate assets, which is an important component of the real estate asset performance. Hence, tenants’ quality influences real estate asset performance. Consequently, tenants’ quality impacts managerial strategy for real estate portfolios.

Figure 1. The impact of tenant quality on real estate asset and portfolio performance.

Classification of Tenants

Classification by Tenant Business Industry

Based on the above literature review, the present study argues that tenants’ business performance can influence the rental income and potential vacancy rate of investment real estate (Frew & Jud, 1988; Sivitanides, 1997; Wheaton & Torto, 1988), besides the elements of fundamental demand and supply regarding real estate space (Archer & Ling, 1997) and economic shocks along economic cycles (Dong & Li, 2012). The quality of real estate portfolios under management is affected by the cycles of tenants’ business industry, the nature of their business and specific industry sector performance. The industry spillover effects and systematic risk of industry performance have been studied by Li et al. (2023). The information of tenant business characteristics is anticipated to be important in determining the performance of real estate portfolios.

Lu-Andrews (2017) and Zheng and Zhu (2021) measure tenant quality using financial performance information of tenants who are publicly traded companies. However, this approach is not feasible for the present study. New Zealand is a small economy (Dong & Li, 2012) with a small scale stock market. The main board of New Zealand Stock Exchange has only 184 listed securities and a pooled market capitalization of around US$ 104 billion. There are only limited number of tenants who operate internationally. Selecting publicly traded companies/tenants will substantially reduce the sample size of the study and will bias the analytical results of commercial real estate performance.

Smith (2009) initiates the concept connecting tenant industry to commercial real estate leasing success. His study argues that tenant industry reflects uncontrollable macroeconomic drivers that indicate systematic risk a tenant exposed to. His argument aligns with the context of the likelihood of tenant failure/default and the subsequent bank suffrage on overvalued assets.

Stemming from Smith’s (2009) concept, the present study uses the classification of tenants’ business industry to distinguish tenants’ potential default risk. This approach is supported by Li et al. (2023) that the contribution and spillover of business sectoral systematic risk constitute the overall market systematic risks. It means that business sector/industry signals the systematic risk (combined through its contribution and spillover effect) that a specific business industry is exposed to.

Empirical findings on publicly traded real estate companies in Europe justify the importance of signaling effect of tenant business industry on these companies performance (Muckenhaupt et al., 2023). Muckenhaupt et al. (2023) find that the systematic risk of tenants’ industries is captured in real estate equity return. Zhang et al. (2023) also use a categorizing code—SBI classification as a proxy of tenants activities in their examination of retail rents in high street districts.

Based on the above literature, the present study uses tenants business industry classification code—Australian and New Zealand Standard Industrial Classification (ANZSIC) to categorize tenants when sorting hypothetical portfolios alongside their business industries. ANZSIC has been jointly developed by the Australian Bureau of Statistics and Statistics New Zealand. It is similar and common as the SIC code in US markets. New Zealand government agencies, for example New Zealand Accident Compensation Corporation, New Zealand Inland Revenue, and Statistics New Zealand, have all adopted ANZSIC to describe a business.

ANZSIC uses a top-down method based on the predominant activity of each individual business unit and assigns each business firstly to a division and then to a subdivision within that division. There are 20 main business industries identified by ANZSIC. Each of them represents a single sector that has a unique group of businesses and associated business activities. However, ANZSIC has too many classification categories. The main concern for the present study is the potential reduction of the number of observations in each category when there are a large number of classification categories.

Therefore, it is necessary to reduce the number of classification categories without sacrificing the internal homogeneity of each category. Business categories which have similar features are re-grouped and re-named in the present study. The ANZSIC 20 business categories are amended as shown in the top section of “ANZSIC Classification Categories” Table 1. The same table also presents other types of tenant classification based on tenant business operation area and head office location. The discussions on these two classifications are shown in the following two sub-sections.

Table 1. Classification tenant business characteristics and corresponding portfolio code.

Description		Portfolio code
ANZSIC classification categories	Modified tenant business categories
Agriculture, forestry, & fishing; mining; electricity, gas, water, services	Natural resources	MSA
Financial & insurance services	Financial services	MSB
Rental, hire, & real estate services; professional, scientific, & technical services; administrative, & support services; information, media, & telecommunications	Professional & advisory services	MSC
Manufacturing; construction	Manufacturing & construction	MSD
Transportation, postal, & warehousing	Logistics & transportation	MSE
Public administration & safety; healthcare & social assistance; education & training; waste services	Government	MSF
Wholesale trade	Wholesale retail	MSG
Retail trade; food services	Specialty retail	MSH
Accommodation; arts & recreation services; other services	Other services	MSI
Tenant business operation area
Business operates in local/one region only		OS1
Business operates in New Zealand nationwide		OS2
Business operates in Australasia region		OS3
Business operates worldwide		OS4
Tenant headquarter location
New Zealand		HQ1
Australia		HQ2
The rest of the world		HQ3

The Natural Resources category stands for tenants whose main income comes from the demand of resources and energy in the primary sector. The category of Financial Services represents mainly financial institutions that are capital intensive and have access to significant financial resources. Professional & Advisory Services consists of tenants that are information and/or intelligence or knowledge intensive. Manufacturing & Construction are tenants who focus on large-scale manufacturing activities and substantial production of goods and products. These tenants are labor, capital, and technology intensive. The main business of tenants in the Logistics & Transportation category is to store, distribute, and transport goods and products from one location to another. Government tenants are either directly or indirectly financially supported by the government. This category includes all government departments and institutions supported by the government, for example education, healthcare, and public services. Wholesale Retail are large types of retail tenants such as department stores and provide a wide variety of goods and products to retail customers. Specialty Retail are smaller types of retail tenants who are specialized in one or a few types of goods and products for retail customers.

Classification by Tenant Operation Area

Besides the above classification approach, this study also classifies tenants under another attribute—the geographic coverage of a tenant’s business/operation. It is a qualitative variable in relation to the location of business operation for each tenant. The area of a tenant’s operation markets is an effective indicator of the scale and the stability of the tenant’s business operation. A tenant who targets only a local market or a single geographical region is more vulnerable to local economic fluctuations than a tenant who operates in multiple markets or regions.

By expanding operation into multiple economic regions, tenants can possibly enjoy non-synchronous demand cycles across regions and increase their business stability. Therefore, we propose that tenant business operation area signals potential risk that a tenant is exposed to. It also indicates a tenant’s quality. It is considered that entering into a different market is an intensive action in terms of a firm’s capital, labor, knowledge and intelligence. Hence, it is quite difficult for a business to change its business geographical coverage in the short or medium term. The present study proposes that the tenant’s business coverage is one of the classification attributes.

The above Table 1 shows the classification by tenant operation area. The first category (OS1) represents tenants whose main business activity is predominantly concentrated in one region only. This is the lowest tier of tenant categories in terms of tenants’ operational area. As the area of a tenant operation expands, Category OS2 represents New Zealand nationwide businesses operating in multiple locations for different target markets. Category OS3 stands for a group of larger businesses operating in the Australasia region that includes New Zealand, Australia, and the Pacific Islands.

The last and biggest operation area is represented by Category OS4 which contains multi-national global corporations operating across different continents. This tenant classification method captures different tenants’ reactions in response to various scale of economic fluctuations. An unexpected downturn in New Zealand economy is likely to have less impact on tenants who focus on overseas markets. Hence, real estate occupancy by tenants in Categories OS3 and OS4 is less likely to be negatively affected by economic fluctuations in New Zealand. Tenants under these two categories are expected to provide stable rental income and protects capital returns to landlords who own relevant commercial real estate.

Classification by Tenant Headquarter Location

The third key attribute is headquarter location, which is also a qualitative variable. The location of head office primarily determines tenants’ business strategies and corporate governance, which in turn affects tenants’ potential costs. In an early study, Leif (1984) introduces factors determining a firm’s location of headquarters. They are location of parties that need frequent and time-consuming fact-to-face contact, business connection and requirement of negotiations with other firms in the same city, accessibility to various services including financial, lawyers, consultancy, libraries, restaurants and etc., convenient transportation facilities, and other amenities that fit into executive’s and key personnel’s family needs.

Lunnan et al. (2019) analyses negotiation costs and information costs between headquarters and subsidiaries regarding to their locational distance, coordination mechanism, relationship atmosphere and social integration. They find that locational distance increases negotiation costs whereas headquarter and subsidiary relationship atmosphere reduces both negotiation and information costs. Contrary to their hypothesis, their results show that social integration increases bargaining costs. Their study implies the effect of head office location on costs to tenants who are either headquarters or subsidiaries.

Ambos et al. (2020) show the challenges behind the logic of local—global and social—commercial for multinational organizations. It indicates the importance of headquarter location regarding an organization’s operational regions, which affects the tension between headquarters and subsidiaries. The level of such tension implies the smoothness of tenants’ business operation and potential costs to tenants.

Overall, the literature indicates that tenants’ headquarter location influences tenants’ day-to-day business operation and their various costs. Thus, the present study also classifies tenants based on their headquarter location. Due to the small size of the New Zealand economy (Dong & Li, 2012), multinational corporations do not generally consider New Zealand as one of their main markets. Their operations in New Zealand are normally considered as an extending part of their global offerings. Except for a few New Zealand-based corporate tenants, multinational corporations view on New Zealand markets is dramatically different to that of New Zealand based companies who mainly focus on New Zealand markets. It implies that these two different types of tenants take different perspectives on making strategic decisions, such as business expansion or contraction. It is anticipated that New Zealand based companies have competitive advantage over overseas based companies in New Zealand local markets that they are familiar with.

Australian based companies are classified as a separate category. Australian based companies are expected to act differently from either New Zealand companies or other overseas companies. Australian companies are technically classified as oversea companies who are governed by a completely separate legal system from New Zealand; but they are strongly integrated into New Zealand economy. Australian and New Zealand economies are integrated (Conway et al., 2013). It suggests that Australian based companies will encounter similar boom-bust phases of economic cycles to New Zealand based companies, which leads to correlated commercial space demand and financial affordability. Similar to global corporate tenants, Australian companies also face some barriers when they expand their business across Tasman to extend their services or offerings to New Zealand. Table 1 in an earlier sub-section shows the classification of tenants’ headquarter location.

It is believed that tenants with headquarters in “The rest of the world”, HQ3, are likely to have business operates in worldwide, OS4 during the sample period. However, there are observations in the sample that tenants with headquarters in “The rest of the world” have business operating in local only, nationwide in New Zealand, or Australasia, especially from 2004 to 2008 and in 2012. It is interesting to see that tenants only choose “The rest of the world” as their headquarter when they have worldwide operating business during a few years after the 2007 global financial crisis.

Despite the above observation, it is hard to conclude that tenants having business operating in worldwide usually have their headquarter in “The rest of the world”. As shown in the sample, tenants having business operating in worldwide have chosen to locate their headquarters in New Zealand, Australia or “The rest of the world”. Figure 2 presents the yearly percentage of tenants who have business operates worldwide (OS4) based on their headquarter location (HQ). It shows that on average 26% worldwide operating tenants have chosen to locate their headquarters in New Zealand or Australia during the sample period. This observation implies that HQ3 and OS4 are not definitely correlated. Z-statistic on Sharpe ratio for sub-periods also tell their differences in the Section of “Z-statistic on Sharpe Ratio for Hypothetical Portfolios”.

Figure 2. Percentage of worldwide operating tenants based on their headquarter location (2004–2012).

It is noted that most commercial real estate assets have more than one tenant. This research adopts an equal-weighted approach to capture the performances of those real estate assets that have more than one disclosed major tenant because the information of tenants’ occupying area is not available. The real estate asset is considered to be occupied solely by a single tenant if only one tenant is disclosed in the annual reports. The details of major tenants will be used to categorize them according to the three classification approaches as discussed in this sub-section. The appraisal-based annual capital return of real estate assets is arguably anchored on the quality of major tenants.

Real Estate Total Return, Sharpe Ratio and Z-Statistic

The annual return for a real estate asset is calculated as a percentage by using the formula as follows: (1) $$R_{i,t}=\frac{V_{i,(t+1)}+I_{i,t}}{V_{i,t}+C_{i,t}}-1\mathrm{ }$$(1) where R_i,t represents the total return in percentage for real estate i during period t. $V_{i,t}$ represents reported fair value at the end of the financial period t; whilst $V_{i,(t+1)}$ denotes reported fair value at the end of each financial year (t + 1).

$I_{i,t}$ is the total recurring income received from the real estate during the year from (t) to (t + 1) recorded in legal documents (leases, licenses etc.) rather than the actual income received. It is possible that the contractual income is different from the actual income. It is also argued that the certainty of receiving contractual income is related to tenants’ credibility and financial stability. $C_{i,t}$ stands for any capital improvement that occurred during year t. Periodic income returns and fair value of real estate assets will be collected and used to assess mean-variance characteristics across different portfolios.

Total return is the sum of income return and capital return. Capital return is calculated relatively straightforward because each publicly traded property portfolio has properly disclosed the most recent valuation of fair value for each real estate in an annual report. During the sample period, New Zealand Equivalent to International Accounting Standard 40 Investment Property (NZ IAS 40) requires firms to choose either fair value model or cost model when reporting investment real estate. Publicly traded property portfolios follow NZ IAS 40 and have chosen to report fair value for each real estate.

This study uses financially reported fair value, instead of transaction price of commercial real estate, because reported fair value is the only available regular annual information for each commercial real estate. The literature has found that appraisal based valuation information may bring in smoothing effect on capital returns of real estate (Adams & Venmore-Rowland, 2000; Edelstein & Quan, 2006; Newell & Lee, 2011). Following the literature, we acknowledge the smoothing effect from fair value. We argue that the problem of smoothing effect is mitigated and will not affect our general findings, especially when capital return of all real estate in our sample is calculated from fair value and the performance of all hypothetical portfolios is compared based on the same type of valuation and approach.

The Sharpe ratio is calculated based on the following formula: (2) $$S_j=\frac{R_j-R_f}{{\sigma }_j}\mathrm{ }$$(2) where $S_j$ represents the Sharpe ratio of portfolio j. The higher the Sharpe ratio is, the more efficient the portfolio is. $R_j$ is the expected return of portfolio j. $R_f$ is the corresponding risk free rate during the period. The risk-free return is determined by the average 90-day Treasury bill rate. ${\sigma }_j$ represents the estimated standard deviation of the excess returns of portfolio j.

The present study also analyze Z-statistic in order to investigate the significance of difference for Sharpe ratio between each pair of hypothetical portfolios. Following Jobson and Korkie (1981) and Lee and Stevenson (2005), Z-statistic on Sharpe ratio is calculated as follows: (3) $$Z=\frac{{\sigma }_c{(R}_r-R_f)-{\sigma }_r{(R}_c-R_f)}{\sqrt{\Theta }}\mathrm{ }$$(3) and (4) $$\Theta =\frac{1}{\mathrm{T}}\left[2{\sigma }_c^2{\sigma }_r^2-2{\sigma }_c{\sigma }_r{\sigma }_{\mathit{\text{cr}}}+\frac{1}{2}{R}_c^2{\sigma }_r^2+\frac{1}{2}{R}_r^2{\sigma }_c^2-\frac{R_c{R}_r}{2{\sigma }_c{\sigma }_r}\left({\sigma }_{\mathit{\text{cr}}}^2+{\sigma }_c^2{\sigma }_r^2\right)\right]$$(4) where ${\sigma }_c$ represents estimates of standard deviation of the excess return (based on total return) for a portfolio in a corresponding column. Note that Z-statistic for each pair of hypothetical portfolios is presented in a matrix for each period (as shown in the tables in appendix). The value of Z-statistic in each cell denotes the relative performance of a pair of hypothetical portfolios in corresponding row and column. Likewise, ${\sigma }_r$ stands for estimates of standard deviation of the excess return (based on total return) for a portfolio in a corresponding row. $R_r$ is the total return of a portfolio in a corresponding row, and $R_c$ denotes the total return of a portfolio in a corresponding column. $\Theta$ is a term defined in Equation (4)(4) $$\Theta =\frac{1}{\mathrm{T}}\left[2{\sigma }_c^2{\sigma }_r^2-2{\sigma }_c{\sigma }_r{\sigma }_{\mathit{\text{cr}}}+\frac{1}{2}{R}_c^2{\sigma }_r^2+\frac{1}{2}{R}_r^2{\sigma }_c^2-\frac{R_c{R}_r}{2{\sigma }_c{\sigma }_r}\left({\sigma }_{\mathit{\text{cr}}}^2+{\sigma }_c^2{\sigma }_r^2\right)\right]$$(4) according to Lee and Stevenson (2005). It captures the quadratic value of total return and variance of excess return for each portfolio and cross terms of standard deviation and covariance of excess return between each pair of portfolios. It is normalized by the number of observations in each period.

A statistically significant and positive Z score indicates that a portfolio in a corresponding row outperforms a portfolio in a corresponding column in a matrix, based on risk-adjusted return. On the contrary, a statistically significant and negative Z-value means that a portfolio in a corresponding column outperforms a portfolio in a corresponding row in a matrix of Z-score.

Data Source

During 12 years from 2001 to 2012, 394 real estate (buildings) are disclosed in annual reports published by all publicly traded real estate portfolios. For any real estate acquired during a financial year, the first return is calculated based on the first year value and its annualized income divided by the acquisition price. This study uses the information for all real estate in annual reports for the publicly traded real estate portfolios from 2003 to 2012. This sample period gives us the return data from 2004 to 2012. This sample period covers the entire period before, during and after the 2007 global financial crisis. There are 1,831 of year-return observations in total. The frequency of data is yearly.

This research uses the information in annual reports of publicly traded real estate portfolios for the list of tenants for each commercial building. Additional information on tenants is obtained by manual research on tenants’ business and operation.

Hypothetical Real Estate Portfolios and Their Performance

Hypothetical Portfolios—Sorted by Geographic Region, Real Estate Sector or Tenant Business Characteristics

Twenty-four hypothetical portfolios are constructed for comparative analyses on their risk-return performance. Following the literature, eight hypothetical portfolios are formed in accordance with two conventional sorting categories—geographic region and real estate sector. The additional dimension of this study to categorize hypothetical portfolios is on tenants’ business characteristics. Sixteen hypothetical portfolios are constructed based on three types of tenants’ classification as explained in the above Section of “Classification of Tenants”.

Following the literature, the present study keeps a consistent approach dealing with potentially inter-connected performance between real estate and tenant when all hypothetical portfolios are structured. Real estate return and risk connect to tenants’ business characteristics and performance when information disclosed in annual reports is used for the measure of real estate performance, for example rental return and capital return for individual real estate. Like the literature, hypothetical portfolios structured along one sorting category contain potential performance information from the other sorting category (Dong & Li, 2012; Hartzell et al., 1986; Lee & Byrne, 1998; Lee & Stevenson, 2005). In other words, the risk-return performance of a hypothetical portfolio structured for one single geographical region contains the information of specific real estate sectors and tenant business characteristics. The reason is that the total return of individual real estate contains the information of region, sector and tenants, especially when rental income and appraisal value information is used. Following the literature, the above consistent approach of structuring portfolios is appropriate for the comparison of hypothetical portfolios along different sorting categories.

Return and Risk of Hypothetical Portfolios

Table A1 (in Appendix) presents descriptive statistics of total return for each hypothetical portfolio over the study period. It is interesting to note that for portfolios sorted by geographic region, Auckland (AKL) shows the highest total return 12.64%, income return 4.26% and capital return 8.38%. Christchurch (CHCH) has the second highest total return 9.67%. The portfolio of Wellington (WEL) presents the highest risk on total return 4.34% and has the worst performed total return commercial real estate −22.21%.

Regarding portfolios sorted by real estate sectors, the portfolio with Industrial real estate (IND) performs the best in terms of return and risk. Its total, capital and income return is the highest among these portfolios sorted by sectors, with annual return 14.98, 6.03 and 8.96% respectively. The risk of total return for this portfolio is the lowest among the four portfolios, with only 1.30%. Office portfolio (OFF) has the second highest total return 10.78%. However, its risk 3.81% for total return is also higher than other portfolios under this sorting category.

The best performed portfolio sorted by tenant business industry is Logistics & Transportation (MSE) with the highest annual total return of 15.74% and the lowest risk on total return, 1.04%. Its capital return of 6.70% and income return of 9.04% are also the highest among all hypothetical portfolios. The second highest total return in this sorting category is shown by Natural Resources (MSA) with 12.32%. The demand for these two industries is expected to have low elasticity during the span of the sample period. The lowest total return observed in this sorting category is Government (MSF) while the risk for total return is moderately high, 3.37%. It is interesting to observe that the capital return for Government (MSF) portfolio is the lowest with 0.78% among all tenant business industries.

Commercial real estate with tenants’ headquarters in “The rest of the world” (HQ3) presents around 2% higher annual total return than those with tenants’ headquarters in New Zealand (HQ1). Real estate with tenants’ headquarters in “The rest of the world” (HQ3) has the highest capital return, 4.63% among this sorting category. Real estate with tenants’ headquarters in Australia shows the highest risk of total return in this sorting category, 3.15%.

Real estate for tenants with worldwide (OS4) or Australasian (OS3) business operation has similar total return, capital return and income return performance, with their total return around 12.30%. Commercial real estate for tenants with Australasian (OS3) business operation shows slightly higher risk (0.18%) on total return than those for tenants with worldwide operation (OS4). Real estate for tenants with New Zealand nationwide operation presents the lowest total return and capital return among this sorting category, 10.66 and 2.40%, respectively.

Tables 2 and 3 present the total return and risk performance for the hypothetical portfolios with regard to five-year rolling periods: Period 1 (P1: 2004–2008), Period 2 (P2: 2005–2009), Period 3 (P3: 2006–2010), Period 4 (P4: 2007–2011), Period 5 (P5: 2008–2012). The separation of rolling periods is consistent with the argument and findings in previous studies about asymmetric systematic risk (Lee et al., 2008), the impact of different economic phases (Dong & Li, 2012) and varied effects from capital markets before and after the global financial crisis (Lee et al., 2016).

Table 2. Total return performance for each hypothetical portfolio during each sub-period.

	Total return
Portfolio	P1 (%)	P2 (%)	P3 (%)	P4 (%)	P5 (%)
AKL	16.74	12.93	11.19	11.20	9.90
CHCH	13.28	9.93	9.32	6.77	4.42
WEL	19.49	14.34	10.53	8.50	4.79
OTH	15.41	8.58	7.37	6.77	6.06
OFF	16.94	13.54	10.84	9.33	6.51
IND	18.17	13.72	12.53	13.52	12.63
RET	15.85	9.32	7.86	7.10	7.08
MED		6.61	8.43	8.22	7.73
MSA	17.95	14.59	10.60	9.51	7.74
MSB	17.96	14.11	11.14	9.62	6.80
MSC	17.22	12.60	11.05	9.77	7.52
MSD	14.59	11.84	9.75	10.14	9.25
MSE	19.69	16.54	15.20	14.38	13.33
MSF	16.35	12.59	10.22	9.30	5.97
MSG	16.19	11.16	9.33	9.08	8.98
MSH	15.75	9.46	8.15	8.66	7.73
MSI	15.59	13.70	11.49	11.61	9.86
HQ1	16.52	12.11	10.02	9.34	7.96
HQ2	17.26	12.66	11.22	9.79	8.47
HQ3	16.96	13.42	11.29	11.51	8.93
OS1	16.76	12.86	11.04	10.27	8.74
OS2	16.14	11.50	9.64	8.94	7.71
OS3	18.43	13.64	11.87	10.29	8.50
OS4	16.79	13.04	10.78	10.74	8.63

Table 3. Risk of total return for each hypothetical portfolio during each sub-period.

	Total risk
Portfolio	P1 (%)	P2 (%)	P3 (%)	P4 (%)	P5 (%)
AKL	1.38	3.55	3.44	3.17	2.45
CHCH	1.69	3.21	3.40	4.18	2.98
WEL	3.47	6.52	5.41	4.95	3.24
OTH	2.42	4.66	3.56	2.86	2.21
OFF	2.52	5.31	5.25	4.92	3.49
IND	2.10	1.84	1.33	1.20	0.83
RET	1.48	4.68	3.67	3.12	2.45
MED		1.33	1.99	2.06	2.09
MSA	2.19	4.59	4.02	3.48	2.96
MSB	3.50	6.53	6.19	6.27	4.10
MSC	2.95	5.95	6.05	5.25	3.64
MSD	1.65	2.00	1.64	1.49	1.26
MSE	1.02	1.14	0.92	1.11	0.94
MSF	2.75	4.62	4.06	3.69	2.36
MSG	1.50	4.23	3.23	2.84	2.23
MSH	1.76	4.21	3.49	2.93	2.34
MSI	1.00	1.55	2.27	2.13	1.97
HQ1	1.78	3.75	3.44	3.14	2.20
HQ2	1.98	4.98	4.43	3.93	2.93
HQ3	2.00	4.04	3.95	3.50	2.61
OS1	1.80	3.45	3.31	2.99	2.15
OS2	1.94	4.07	3.70	3.35	2.31
OS3	2.20	4.93	4.52	3.95	2.90
OS4	1.80	4.08	3.79	3.44	2.60

For hypothetical portfolios for real estate regions, Auckland (AKL) portfolio maintains close to or above 10% annual return. Auckland is the only portfolio having achieved positive capital returns in Period 5 (2008–2012). It suggests that Auckland portfolio has the greatest value adhesiveness since the global financial crisis occurred in 2007. The total return of all other portfolios (i.e. Christchurch (CHCH), Wellington (WEL), and Other regions (OTH)) are clustered together in Period 4 and Period 5. Auckland portfolio also achieves the highest income return throughout all five periods. For portfolio risk, the general trend is that risk spikes significantly in Period 2 and gradually declines in the last three periods with the exception of Christchurch portfolio, which shows another increase in risk in period 4.

For real estate sectoral portfolios, Industrial portfolio (IND) outperforms other portfolios across all five rolling periods, in terms of value-weighted annual total returns. The gap between Industrial portfolio and other portfolios expands over time, particularly in Period 4 and Period 5. The Industrial portfolio also consistently shows the smallest volatility, which gradually declines over time. By contrast, Office (OFF) and Retail (RET) portfolios show a significant increase in portfolio risk in Period 2. The risk for both portfolios tends to decline afterwards but Retail portfolio achieves a far greater reduction on risk after Period 2 than Office portfolio.

Regarding hypothetical portfolios sorted by tenants business characteristics, Logistics & Transportation (MSE) consistently shows the highest total return than all other hypothetical portfolios over the five sub-periods. On the other hand, Industrial portfolio (IND, sorted by real estate sector) presents the second highest total return over Periods 3 to 5. Other portfolios show similar total return trend over the five sub-periods, with total return differences falling into around 5% range among them. Interestingly, real estate with tenants in Financial Services (MSB) has the highest risk on total return over the five sub-periods. Real estate with tenants in Professional and Advisory Services (MSC) has the second highest risk on total return from Periods 3–5. Commercial real estate with tenants in Logistics & Transportation (MSE) has sustained the best performance on risk, and it has the lowest risk on total return from Periods 1 to 4 and slightly higher risk, 0.11% higher than the risk on total return for Industrial portfolio (IND).

Sharpe Ratio for Hypothetical Portfolios

Table 4 shows the Sharpe ratio of all hypothetical portfolios for the entire sample period and each sub-rolling period. The order of portfolios is ranked by their Sharpe ratio performance over the entire sample period. The rank of Sharpe ratio is also visualized in Figure 3. Different colour of columns denote different sorting categories for hypothetical portfolios. Blue columns show portfolios sorted by tenant business characteristics—business industry, operation area or headquarter location. Red columns represent portfolios sorted by real estate sectors. Orange columns denote portfolios sorted by geographic region.

Figure 3. Sharpe ratio of total return for all hypothetical portfolios for the entire period (2004–2012) (Blue columns represent portfolios sorted by tenant business characteristics. Red columns stand for portfolios sorted by real estate sectors. Orange columns denote portfolios sorted by geographic regions.).

Table 4. Sharpe ratio of total return for each hypothetical portfolio for the entire sample period and each sub-period (ranked by Sharpe ratio for the entire sample period).

		Sharpe ratio
Code	Portfolio\Period	Entire	P1	P2	P3	P4	P5
MSE	Logistics & transportation	2.26	4.22	2.00	2.59	2.84	3.50
IND	Industrial sector	1.66	2.23	1.09	1.57	2.32	2.94
MSD	Manufacturing & construction	1.42	2.16	1.00	1.02	1.63	2.00
MED	Medical sector	1.33		0.56	1.50	1.96	2.02
MSI	Other services	1.28	2.22	1.43	1.19	1.84	2.29
AKL	Auckland	1.22	3.11	0.93	0.71	0.89	1.22
HQ3	HQ-rest of the world	1.10	2.69	0.85	0.64	0.83	1.01
OS4	World wide operation	1.06	2.75	0.84	0.61	0.77	0.98
MSG	Wholesale retail	1.05	2.94	0.81	0.61	0.71	1.26
MSA	Natural resources	0.94	2.02	0.96	0.61	0.64	0.73
OS3	Australasia operation	0.93	2.62	0.84	0.65	0.60	0.97
HQ2	HQ-Australia	0.90	2.41	0.77	0.62	0.58	0.95
HQ1	HQ-New Zealand	0.89	3.28	0.81	0.56	0.62	1.00
OS1	Local operation	0.86	2.62	0.87	0.64	0.67	1.23
OS2	Nation wide operation	0.86	3.17	0.72	0.51	0.57	0.90
MSC	Professional & advisory services	0.75	1.87	0.63	0.48	0.49	0.60
MSB	Financial services	0.73	1.83	0.71	0.46	0.43	0.46
MSH	Specialty retail	0.72	2.55	0.44	0.35	0.55	0.85
OFF	Office sector	0.70	1.99	0.78	0.50	0.47	0.50
RET	Retail sector	0.65	2.65	0.46	0.34	0.36	0.74
CHCH	Christchurch	0.63	1.46	0.57	0.56	0.27	0.22
MSF	Government	0.52	1.93	0.78	0.51	0.53	0.64
WEL	Wellington	0.48	2.39	0.79	0.45	0.37	0.23
OTH	Other regions	0.40	1.91	0.26	0.21	0.25	0.44

The top performed portfolio on risk-adjusted return is the one sorted by tenant business industry—Logistics & Transportation (MSE) portfolio with a Sharpe ratio of 2.26, followed up by the portfolio sorted by real estate sector—Industrial portfolio (IND) with a Sharpe ratio of 1.66. Note that MSE and IND do not represent similar portfolio composition, although both of them are ranked the highest on Sharpe ratio. For example, tenants with business in Logistics & Transportation (MSE) occupy space in three main real estate sectors—office buildings, retail centers and industrial space. Meanwhile, similar to office sector, industrial sector hosts tenants from all nine tenant business industries (nine ANZSIC categories based on classification on tenant business industry). Therefore, the coincident high performance of these two hypothetical portfolios does not mean they have similar portfolio composition. The significant Z-statistic on Sharpe ratio (shown in the subsequent section) between the two portfolios for the entire sample period and three sub-periods proves their difference. Similar observation is noted for Wholesale Retail (MSG), Specialty Retail (MSH) and Retail (RET) portfolios. Although they sound plausibly similar, however the composition in these hypothetical portfolios is different. Wholesale Retail (MSG) and Specialty Retail (MSH) show higher Sharpe ratio than Retail portfolio (RET). The significant Z-statistic on Sharpe ratio between Wholesale Retail (MSG) and Retail (RET) portfolios supports this point.

The other two portfolios sorted by tenant business industry—Manufacturing & Construction and Other Services are ranked the third and fifth respectively, based on the Sharpe ratio. Medical portfolio (MED) sorted by real estate sector is ranked the fourth. The best performed portfolio among those sorted by geographic region is Auckland portfolio (AKL). It is ranked the sixth on Sharpe ratio. The other three portfolios sorted by geographic region are in the tail of the low range side of the Sharpe ratio, with Other Regions (OTH) and Wellington (WEL) portfolios showing the lowest Sharpe ratio, 0.48 and 0.40 respectively. Office (OFF) and Retail (RET) portfolios have the sixth and fifth lowest Sharpe ratio among all hypothetical portfolios.

Overall, portfolios sorted by tenant business characteristics perform generally better than most portfolios sorted by geographic region and real estate sector. Three out of the five top performing portfolios and five of the ten top performing portfolios among all hypothetical portfolios are business industry portfolios. The finding implies that choosing tenants in well-performed business industry helps promote the pure real estate portfolio performance in a small economy, like New Zealand.

Table 4 illustrates Sharpe ratio performance for each sub-period. Regarding portfolios sorted by tenant business industry, Logistics & Transportation (MSE) portfolio generally sustains the best performance on Sharpe ratio for each sub-period. Although this portfolio experiences the most contraction in its Sharpe ratio in Period 2, it subsequently shows a leading position and robust recovery from Period 3 onwards. Other Services (MSI) portfolio achieves only fourth position in Period 1. It subsequently shows the least contractions in Period 2 and has remained the second best performing portfolio from Period 2 to Period 5.

Manufacturing & Construction portfolio (MSD) generally presents similar trend to Other Services (MSI) portfolio. MSD does not achieve a superior risk-adjusted return in Period 1, and it has maintained the third position from Period 2 to Period 5. The other six portfolios sorted by tenant business industry are generally clustered closing to each other. Financial Services (MSB) and Professional & Advisory Services (MSC) are highly correlated with very similar Sharpe ratio results throughout the entire sample period. Compared with other portfolios, MSB, MSC, and Government (MSF) portfolios generally provide lower risk-adjusted returns, especially in Periods 4 and 5. Wholesale Retail (MSG) and Specialty Retail (MSH) portfolios have a medium level of Sharpe ratio in Period 1. Their Sharpe ratio mostly reduces in Periods 2 and 3.

Regarding portfolios sorted by tenant headquarter location, the portfolio with tenants’ headquarter in the rest of the world portfolio (HQ3) slightly outperforms the other two portfolios throughout the entire study period except in Period 4. However, the portfolio with tenants’ headquarter in New Zealand (HQ1) outperforms HQ3 in Period 1. Among portfolios sorted by tenant operation area, the portfolio with tenants local business operation (OS1) has marginally superior Sharpe ratio in all periods. The lowest Sharpe ratio performance is for the portfolio with tenants business operating in nationwide in New Zealand (OS2). Interestingly, OS2 performs worse than worldwide operation (OS4), Australasian operation (OS3) and local operation (OS1).

Auckland portfolio (AKL) demonstrates a clearly better performance not only for the entire study period but also for each of the individual rolling periods in comparison with other regional portfolios. Since the global financial crisis in 2007, Auckland portfolio shows the largest contraction of the risk-adjusted returns. The gap between the Auckland portfolio and other three regional portfolios narrows in period 2 (2005–2009) and period 3 (2006–2010). These results imply that the global financial crisis may have had a greater impact on Auckland portfolio than other regional portfolios. Since period 4, the Sharpe ratio for Auckland portfolio improves at the earliest and fastest pace, which shows that Auckland remains the most attractive investment destination for office space in New Zealand.

Other portfolios sorted by geographic region shows a similar risk and return profile to Auckland portfolio; but their overall performance is not so good as that of Auckland portfolio for sub-periods. Christchurch (CHCH) and Wellington (WEL) portfolios show a different risk and return pattern from that of the Auckland portfolio, and their Sharpe ratio continuously declines in Periods 4 and 5. Christchurch portfolio (CHCH) shows the strongest resistance in response to economy downturns in Period 2 and 3; but its performance has dropped to the worst position in Period 5.

Regarding portfolios sorted by real estate sector, the Sharpe ratio of Industrial portfolio (IND) has the least contraction in Period 2. Industrial portfolio shows a quick and strong recovery in Period 3 with further improvements on Sharpe ratios in Periods 4 and 5. Retail portfolio (RET) achieves its best risk-adjusted performance in Period 1; but its Sharpe ratio remains the lowest in all subsequent periods until Period 5. The Sharpe ratio for Office portfolio (OFF) continuously declines during the entire study period. Office portfolio (OFF) becomes the least attractive portfolio in the final sub-period (2008–2012). The Sharpe ratio for Medical hypothetical portfolio (MED) is not available for Period 1 because the performance data for Medical real estate only becomes available from 2008. Since Period 2, the Medical portfolio (MED) shows constant performance results, with a Sharpe ratio just below that of the Industrial portfolio (IND) but significantly outperforming Office (OFF) and Retail (RET) portfolios.

Z-Statistic on Sharpe Ratio for Hypothetical Portfolios

The present study conducts tests on Z-statistic to compare equality of Sharpe ratio performance between each pair of hypothetical portfolios. The analyses are done for the entire period and each sub-period. Tables A2–A7 (shown in Appendix) present matrices with Z score for each pair of hypothetical portfolios.

As presented in Table A2 for the entire sample period, hypothetical portfolios sorted by tenant business industry (portfolio code starting with “MS”) significantly outperform most portfolios sorted by geographic region, except for the region of Auckland (AKL). For example, most portfolios sorted by tenant business industry perform significantly better on risk-adjusted return than Wellington portfolio (WEL), except for portfolios with Government (MSF) and Specialty Retail (MSH) tenants. Six out of nine portfolios sorted by tenant business industry significantly outperform Other Regions portfolio (OTH) regarding Sharpe ratio. The six portfolios are Natural Resources (MSA), Manufacturing & Construction (MSD), Logistics & Transportation (MSE), Wholesale Retail (MSG), Specialty Retail (MSH) and Other Services (MSI). Four portfolios sorted by tenant business industry show significantly better performance than Christchurch portfolio (CHCH). These four portfolios consist of tenants from Manufacturing & Construction (MSD), Logistics & Transportation (MSE), Wholesale Retail (MSG) and Other Services (MSI), respectively. The only portfolio significantly outperforming Auckland (AKL, sorted by geographic region) is the one sorted by tenant business industry—Logistics & Transportation (MSE). Real estate with tenants of Wholesale Retail (MSG) significantly outperforms the portfolio of Retail (RET) real estate on risk-adjusted return.

Comparing portfolios sorted by tenant business industry and real estate sector, five out of nine portfolios sorted by tenant business industry significantly perform better on risk-adjusted return than one to three portfolios sorted by real estate sector. Portfolios of Logistics & Transportation (MSE) tenants significantly outperform three main real estate sectors—Office (OFF), Industrial (IND) and Retail (RET) at 1% to 5% significance level. Portfolios consisting of Manufacturing & Construction (MSD) tenants present significantly better performance than Office (OFF) and Retail (RET) portfolios. Two portfolios with tenants in Professional & Advisory Services (MSC) and Other Services (MSI) outperform Office portfolio (OFF) at 10% significance level. Wholesale Retail (MSG) performs better than Retail (RET) portfolio at 1% significance level. No significance is found between Medical portfolio (MED, sorted by real estate sector) and other hypothetical portfolios. Comparing Logistics & Transportation (MSE) portfolio with all other hypothetical portfolios, it is found that MSE significantly performs better on risk-adjusted return than all other portfolios except for Medical portfolio (MED), for the entire sample period.

Summarising the above, a number of hypothetical portfolios sorted by tenant business industry significantly outperform portfolios sorted by geographic regions and real estate sectors, regarding to Sharpe ratio for the entire sample period. It implies that prudentially selecting tenants with specific business industry can help build a better performed real estate portfolio than omitting consideration on tenant business industry. The findings may rationalize the results from the existent literature about tenant industry on publicly traded real estate portfolios (Muckenhaupt et al., 2023).

Regarding portfolios sorted by tenant headquarter location, it is found that all three portfolios (HQ1, HQ2 and HQ3) significantly outperform Wellington (WEL), Other Region (OTH) and Office (OFF) portfolios. The three portfolios sorted by tenant headquarter location also significantly perform better than portfolio with Government (MSF) tenants. All four portfolios sorted by tenant operation area (OS1, OS2, OS3 and OS4) significantly perform better than Wellington (WEL) and Other Regions (OTH) portfolios. Portfolios with tenants under Australasian (OS3) and Worldwide (OS4) operation significantly outperform Office portfolio (OFF). However, all seven portfolios sorted by tenant headquarter location or operation area significantly underperform Industrial portfolio (IND). The findings imply that generally selecting tenants according to their headquarter location or operation area can help structure a portfolio that presents superior performance than other types of real asset selecting strategy, which depends on specific geographic location and real estate sector in an economy.

The following highlights key findings on Z score for the five sub-periods (as presented in Tables A3–A7, respectively in Appendix). There is a pattern that the 2007 global financial crisis significantly hit three portfolios—Specialty Retail (MSH, sorted by tenant business industry), Retail portfolio (RET, sorted by real estate sector) and Other Regions portfolio (OTH, sorted by geographic region). Specialty Retail (MSH) significantly underperforms most other portfolios in Period 2 (2005–2009) the period just before and across the 2007 global financial crisis. During the period, Retail portfolio (RET) significantly underperforms the majority of other hypothetical portfolios. The same phenomenon is observed for Other Regions portfolio (OTH) that significantly performs worse than most other portfolios during the period.

The pattern of underperformance is found for Other Regions portfolio (OTH) in Period 3 (2006–2010) and Period 4 (2007–2011), when the global financial crisis went in-depth. Its underperformance eases very slightly in Period 5 (2008–2012). Specialty Retail (MSH) continues to significantly underperform in Period 3, and its underperformance gradually eases from Period 4 to Period 5. In Period 5, it significantly outperforms four portfolios while underperforms five portfolios. The four portfolios that it outperform are Wellington (WEL), Other Regions (OTH) and Office (OFF) portfolios and the portfolio with Financial Services (MSB) tenants. The underperformance of Retail portfolio (RET) eases slightly during Period 3, and its performance becomes worse in Period 4, just the period at and after the global financial crisis. Period 5 witnesses the slight ease of its underperformance regard risk-adjusted return for Retail portfolio (RET).

The findings show that portfolios consisting of real estate with tenants, real estate sector or geographic region that are vulnerable to negative market shocks tend to be sensitive to crises, regarding their risk-adjusted return performance. The particular vulnerable tenant business, real estate sector and geographic region in New Zealand are Specialty Retail tenants, retail real estate sector and regions other than large to medium size cities.

Table A5 shows that Industrial portfolio (IND, sorted by real estate sector) significantly outperforms most other hypothetical portfolios, except that it significantly underperforms Logistics & Transportation portfolio (MSE, sorted by tenant business industry) in Period 3 (2006–2010). Industrial portfolio (IND) continues to perform significantly strong on risk-adjusted return during Periods 4 to 5 and its comparative performance with Logistics & Transportation portfolio (MSE) becomes insignificant during these two sub-periods. Across the five sub-periods, Logistics & Transportation portfolio (MSE) significantly outperforms Industrial portfolio (IND) from Periods 1 to 3 and for the entire sample period, based on Z score for Sharpe ratio. The significant findings further indicate that the composition of these two portfolios are not similar. Like mentioned in the Section of “Sharpe Ratio for Hypothetical Portfolios”, Logistics & Transportation (MSE) tenants are presented in all three main real estate sectors—office, industrial and retail. Industrial portfolio (IND) contains tenants from all nine business industries.

There are some interesting findings on the significant comparative performance between the portfolio with tenants having headquarters in “The rest of the world” (HQ3) and the portfolio with tenants operating worldwide (OS4) in Period 3 (2006–2010) and Period 4 (2007–2011). The portfolio with tenants having headquarter in “The rest of the world” (HQ3) significantly outperforms the portfolio with tenants operating worldwide (OS4) during Periods 3 and 4. Such (out)performance is insignificant in Periods 1, 2 and 5. Based on significance in findings, it can be noted that HQ3 and OS4 do not contain similar portfolio composition. While it is believed that most tenants with headquarters in “The rest of the world” (HQ3) operate worldwide, tenants operating worldwide (OS4) have their headquarters in New Zealand (HQ1), Australia (HQ2) and “The rest of the world” (HQ3) (as discussed earlier in Figure 2). The above implies that surrounding the period of the 2007 global financial crisis, having headquarters aligned with tenants’ globally operating markets can provide competitive advantage to risk-adjusted return on real estate portfolios over portfolios with other tenants who operate globally but have their headquarters in New Zealand or Australia. The result and implication may support the argument on headquarter-subsidiary tension and various costs in the literature on organizations (Ambos et al., 2020; Lunnan et al., 2019).

Extended Analyses on Sharpe Ratio for Tenant Business Industry in Each Real Estate Sector

We also analyze Sharpe ratio separating different real estate sector for each tenant business industry, in order to develop a big picture of how tenant business industry performs across different real estate sectors. Figure 4 presents Sharpe ratio of total return for portfolios under each tenant business industry and for each real estate sector over the entire sample period. The Sharpe ratio is ranked from high to low. It is fascinating to see that industrial sector real estate with tenants across different business industries generally has higher Sharpe ratio than office, retail and medical sector properties. Retail sector and office sector with a few tenant industries are on the low tail side for Sharpe ratio. The top Sharpe ratio is observed for office buildings with Wholesale Retail (MSG) tenants. Most office real estate with tenants of different business industries sits in the lower middle trail for Sharpe ratio.

Figure 4. Sharpe ratio of total return for portfolios by tenant business industry separated in each real estate sector for the entire sample period (2004–2012). Symbol “R” stands for retail sector. “I” represents industrial sector. “O” indicates office sector. “M” represents medical sector.

Figure 5 separates Sharpe ratio performance of total return for the five sub-periods (P1–P5). Sharpe ratio is generally higher in P1 (2004–2008) than other sub-periods. It shows that real estate across different sectors with various tenant business industries generally has better Sharpe ratio performance for most of years before the 2007 global financial crisis in the sample. Industrial real estate with tenants of three business industries—Financial Services (MSB), Professional & Advisory Services (MSC) and Government (MSF) shows higher Sharpe ratio in P5 (2008–2012) than earlier sub-periods. Office sector with Wholesale Retail (MSG) tenants presents very high Sharpe ratio in P5. The above findings imply inverse demand for industrial real estate with these tenants businesses after the 2007 global financial crisis. Note that the Sharpe ratio is not reported in P5 for office sector with Financial Services (MSB) tenants and in P1 for medical sector with Government (MSF) tenants because extremely low number of observations in these two sub-periods.

Figure 5. Sharpe ratio of total return for portfolios by tenant business industry separated in each real estate sector for the five sub-periods from 2004 to 2012. Symbol “R” stands for retail sector. “I” represents industrial sector. “O” indicates office sector. “M” represents medical sector.

Figure 6 presents Sharpe ratio of income return across the five sub-periods. It is interesting to observe that Sharpe ratio of income return generally increases from P1 to P5 among all tenant business industries and for each real estate sector. This phenomenon can possibly be attributed to the adhesiveness of income return due to lease contracts. Real estate lease terms and its structure can delay tenants’ and landlords’ response on rent towards instantaneous market fluctuations. On the contrary, Sharpe ratio on capital return over the five sub-periods looks in good order as shown by Figure 7. The variation of Sharpe ratio for capital return across the sub-periods is smaller than that for income return. Sharpe ratio for P1 is generally higher than that for other sub-periods. Industrial real estate with Government (MSF) tenants displays higher Sharpe ratio in P4 and P5 than other sub-periods. The Sharpe ratio for capital return signals valuers consideration on dynamic market conditions over different sub-periods.

Figure 6. Sharpe ratio of income return for portfolios by tenant business industry separated in each real estate sector for the five sub-periods from 2004 to 2012. Symbol “R” stands for retail sector. “I” represents industrial sector. “O” indicates office sector. “M” represents medical sector.

Figure 7. Sharpe ratio of capital return for portfolios by tenant business industry separated in each real estate sector for the five sub-periods from 2004 to 2012. Symbol “R” stands for retail sector. “I” represents industrial sector. “O” indicates office sector. “M” represents medical sector.

Overall, the results on Sharpe ratio for different tenant business industries and in each real estate sector show different real estate sectoral performance even for tenants from the same business industry. The findings also imply varied Sharpe ratio performance over different sub-periods between income and capital return.

Conclusion

This research extends from the literature on the importance of tenant characteristics and composition on REIT performance. It fills in the knowledge gap of connecting tenant business characteristics with directly held commercial real estate portfolios at the physical asset level. The study avoids potential noises from financial leverage and capital market volatility that are prone in existent REIT analyses for tenants.

The distinct Sharpe ratio performance of portfolios across different tenant business industries rationalizes the importance of incorporating tenants’ business industry and other business characteristics into the understanding of directly or indirectly held commercial real estate portfolios. Specific tenant business industry, for example, Logistics & Transportation (MSE) significantly outperforms all other hypothetical portfolios and shows great resilience to the 2007 global financial crisis. Industrial real estate portfolio also performs better than other portfolios in a small economy, like New Zealand.

Our findings on the impact of tenant business characteristics on commercial real estate portfolio Sharpe ratio performance can imply how active tenant management may improve total assets. It may subsequently help with reducing nominal operating costs based on economies of scale (Malhotra et al., 2020). Our general findings also support the argument of how actively managed real estate funds enjoy positive evaluation from the markets. It implies the importance of ensuring overall asset quality through active management of buildings in the context of real estate portfolio management (Mattarocci & Scimone, 2020). The significant findings on Z-statistic for tenant headquarter location against their worldwide operation over sub-periods may provide additional implications to studies on headquarter location choice and multinational organizations (Ambos et al., 2020; Leif, 1984; Lunnan et al., 2019).

While real estate assets are illiquid, fund and lease managers may face challenges on actively shuffling real estate assets. The management lease terms for a portfolio can be prudentially strategized and discretionally exercised. Fund and lease managers are advised to take into account potential tenants’ business industries and the cyclical pattern of these industries when negotiating with tenants and liaising with marketing team for marketing strategy.

Further studies can be extended to propose new research methods and explore orthogonal effects separating the three sorting categories—geographic region, real estate sector and tenant business characteristics. Subsequently, the interrelation between real estate and tenant characteristics can possibly be measured and clarified in risk-return analyses. Studies in this direction may facilitate to improve the fair value reporting process for commercial real estate, with limited comparable investment real estate information for valuation. Other studies can also be conducted on analyzing the probability of tenant default based on their business characteristics when historical data on tenant default for commercial real estate can be obtained.

Acknowledgments

We would like to thank anonymous referees for their highly valuable comments. We appreciate YiZheng Li’s assistance in calculating Sharpe ratio for portfolios of tenant business industry in each real estate sector for the section of Extended Analyses.

Disclosure statement

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

Appendix

Table A1. Descriptive statistics of total return for each hypothetical portfolio (entire sample: 2004–2012).

	Total return						Capital return		Income return
Portfolio	Return (%)	Risk (%)	Min (%)	Max (%)	S.D. (%)	Median (%)	Return (%)	Risk (%)	Return (%)	Risk (%)
AKL	12.64	2.47	−16.47	68.00	9.32	11.98	4.26	2.40	8.38	0.15
CHCH	9.67	3.44	−8.14	49.64	9.40	10.91	1.48	3.34	8.19	0.20
WEL	9.55	4.34	−22.21	35.91	7.54	11.01	1.59	4.29	7.97	0.26
OTH	8.40	2.83	−7.68	31.20	7.44	9.00	0.34	2.76	8.06	0.29
OFF	10.78	3.81	−22.21	68.00	8.02	10.93	2.62	3.73	8.16	0.23
IND	14.98	1.30	−10.27	48.13	8.59	11.88	6.03	1.24	8.96	0.11
RET	9.30	2.73	−12.27	49.64	8.40	8.57	1.67	2.65	7.63	0.18
MED	7.73	2.09	3.12	12.22	2.75	8.29	−0.97	1.90	8.69	0.24
MSA	12.32	3.13	−4.58	52.62	9.90	11.91	4.01	3.08	8.31	0.21
MSB	11.86	4.42	−3.68	27.85	5.60	11.74	3.97	4.30	7.89	0.26
MSC	11.63	4.04	−14.30	65.50	8.47	10.81	3.32	3.94	8.31	0.23
MSD	11.87	1.54	−5.27	52.62	7.83	12.68	3.31	1.46	8.57	0.14
MSE	15.74	1.04	−0.44	68.00	10.27	13.81	6.70	1.00	9.04	0.09
MSF	9.08	3.37	−22.21	33.54	6.88	9.73	0.78	3.37	8.31	0.23
MSG	11.36	2.59	−0.79	65.50	10.23	10.80	3.42	2.51	7.94	0.24
MSH	9.98	2.55	−16.47	48.13	9.89	10.44	2.09	2.46	7.89	0.15
MSI	11.63	1.80	−9.53	36.27	10.16	9.91	2.65	1.72	8.98	0.15
HQ1	10.91	2.58	−22.21	68.00	8.93	11.07	2.54	2.53	8.36	0.15
HQ2	11.52	3.15	−12.27	42.99	7.32	11.01	3.56	3.09	7.97	0.15
HQ3	12.95	2.79	−14.30	65.50	9.89	12.13	4.63	2.72	8.32	0.24
OS1	11.17	2.48	−16.47	49.72	8.64	10.36	2.62	2.44	8.55	0.15
OS2	10.66	2.72	−22.21	68.00	8.92	11.33	2.40	2.66	8.25	0.16
OS3	12.29	3.09	−12.27	52.62	8.61	11.21	4.17	3.02	8.12	0.15
OS4	12.30	2.81	−14.30	65.50	9.33	11.86	4.07	2.73	8.23	0.22

Table A2. Z-statistic of Sharpe ratio of total return for each pair of hypothetical portfolios (entire sample: 2004–2012).

	AKL	CHCH	WEL	OTH	OFF	IND	RET	MED	MSA	MSB	MSC	MSD	MSE	MSF	MSG	MSH	MSI	HQ1	HQ2	HQ3	OS1	OS2	OS3	OS4
AKL
CHCH	−1.47
WEL	−2.40	−0.62
OTH	−2.27	−1.02	−0.32
OFF	−2.57	0.22	1.62	1.09
IND	0.94	1.99	2.22	2.76	1.82
RET	−1.70	0.11	0.65	1.87	−0.15	−2.17
MED	0.11	0.69	0.83	0.90	0.61	−0.31	0.66
MSA	−1.25	0.81	1.56	1.58	1.00	−1.70	0.81	−0.38
MSB	−2.30	0.33	1.58	1.25	0.66	−1.71	0.29	−0.58	−0.77
MSC	−2.47	0.38	1.69	1.24	1.49	−1.71	0.35	−0.56	−0.79	0.38
MSD	0.46	1.57	1.88	2.34	1.47	−2.73	1.74	0.08	1.22	1.35	1.35
MSE	1.85	2.51	2.72	2.97	2.42	1.73	2.71	0.85	2.37	2.28	2.30	2.46
MSF	−2.26	−0.46	0.65	0.47	−1.20	−2.12	−0.48	−0.79	−1.31	−1.22	−1.31	−1.76	−2.66
MSG	−0.51	1.70	1.67	2.87	1.10	−1.39	3.54	−0.28	0.30	0.99	0.88	−0.89	−2.20	1.56
MSH	−1.73	0.33	0.91	2.57	0.09	−2.34	0.55	−0.60	−0.68	−0.05	−0.13	−1.86	−2.73	0.73	−2.08
MSI	0.15	1.29	1.87	1.81	1.41	−1.07	1.28	−0.06	1.17	1.21	1.27	−0.42	−2.34	1.75	0.45	1.25
HQ1	−1.79	1.07	2.12	2.12	1.35	−1.71	1.25	−0.43	−0.16	0.97	0.86	−1.28	−2.46	1.98	−0.74	1.00	−0.99
HQ2	−1.95	0.94	2.28	1.79	2.24	−1.50	0.98	−0.42	−0.16	1.59	1.42	−1.10	−2.25	2.07	−0.53	0.74	−0.92	0.05
HQ3	−0.96	1.28	2.52	2.31	2.20	−1.30	1.47	−0.22	0.71	1.89	1.95	−0.82	−2.13	2.34	0.17	1.60	−0.50	1.35	1.24
OS1	−1.48	1.00	2.45	1.88	1.15	−1.61	0.97	−0.46	−0.24	0.79	0.63	−1.20	−2.38	2.70	−0.74	0.68	−0.99	−0.32	−0.29	−1.22
OS2	−1.91	0.95	1.81	2.07	1.05	−1.78	1.17	−0.46	−0.31	0.75	0.63	−1.36	−2.46	1.63	−1.00	0.86	−1.04	−0.90	−0.35	−1.41	−0.06
OS3	−1.69	1.12	2.68	2.06	2.28	−1.53	1.17	−0.39	−0.03	1.55	1.44	−1.10	−2.29	2.53	−0.45	0.95	−0.90	0.45	0.60	−1.18	0.71	0.62
OS4	−1.62	1.19	2.42	2.16	2.21	−1.35	1.36	−0.26	0.59	1.85	1.95	−0.89	−2.18	2.20	0.04	1.39	−0.61	1.17	1.19	−0.99	1.05	1.11	1.04

Note that a significant negative value indicates that a portfolio in the corresponding column outperforms a portfolio in the corresponding row, vice versa. Red cells denote 1% significance level. Green cells denote 5% significance level. Yellow cells denote 10% significance level.

Table A3. Z-statistic of Sharpe ratio of total return for each pair of hypothetical portfolios (Period 1: 2004–2008).

	AKL	CHCH	WEL	OTH	OFF	IND	RET	MED	MSA	MSB	MSC	MSD	MSE	MSF	MSG	MSH	MSI	HQ1	HQ2	HQ3	OS1	OS2	OS3	OS4
AKL
CHCH	−1.18
WEL	−1.33	0.88
OTH	−0.77	0.93	−0.38
OFF	−2.67	0.32	−1.48	0.06
IND	−0.56	0.69	−0.12	0.30	0.19
RET	−0.28	1.44	0.18	1.21	0.46	0.30
MED
MSA	−1.83	0.23	−0.47	0.09	−0.07	−0.29	−0.95
MSB	−1.44	0.39	−1.37	−0.07	−0.76	−0.32	−0.59		−0.20
MSC	−1.61	0.40	−2.00	−0.04	−1.15	−0.29	−0.54		−0.17	0.17
MSD	−0.59	0.57	−0.17	0.21	0.13	−0.30	−0.33		0.12	0.25	0.22
MSE	0.56	1.47	0.98	1.28	1.21	1.67	0.82		1.36	1.30	1.27	1.76
MSF	−1.61	0.53	−1.85	0.01	−0.23	−0.23	−0.51		−0.10	0.35	0.22	−0.17	−1.23
MSG	−0.10	1.63	0.35	1.51	0.61	0.48	1.27		0.59	0.74	0.69	0.50	−0.66	0.67
MSH	−0.35	1.79	0.11	1.70	0.40	0.26	−0.19		0.38	0.54	0.49	0.29	−0.89	0.48	−0.77
MSI	−0.63	0.58	−0.14	0.23	0.18	−0.01	−0.29		0.31	0.31	0.28	0.06	−1.41	0.23	−0.47	−0.22
HQ1	0.40	2.28	2.09	0.99	2.19	1.23	0.63		1.67	2.47	2.28	0.92	−0.43	2.65	0.34	0.83	0.65
HQ2	−1.21	0.90	0.06	0.39	1.12	0.14	−0.16		0.40	1.35	1.17	0.18	−0.96	1.20	−0.34	−0.10	0.15	−1.06
HQ3	−1.03	1.05	0.67	0.57	1.04	0.34	0.03		0.80	1.06	1.19	0.37	−0.80	1.28	−0.15	0.10	0.40	−0.73	0.45
OS1	−1.06	1.50	1.17	0.58	1.68	0.59	−0.17		0.96	2.34	1.82	0.24	−1.08	2.16	−0.46	−0.06	0.27	−2.25	0.97	−0.10
OS2	0.05	1.69	0.77	0.97	1.07	0.58	0.38		0.79	1.21	1.13	0.61	−0.53	1.36	0.16	0.52	0.58	−0.46	0.80	0.50	0.94
OS3	−0.94	1.13	0.54	0.57	1.23	0.28	−0.02		0.63	1.37	1.29	0.32	−0.84	1.56	−0.21	0.05	0.30	−0.98	0.91	−0.16	0.00	−0.68
OS4	−1.30	1.03	0.82	0.59	1.17	0.36	0.07		0.86	1.14	1.30	0.40	−0.77	1.30	−0.11	0.14	0.43	−0.61	0.59	0.35	0.17	−0.41	0.28

Note that a significant negative value indicates that a portfolio in the corresponding column outperforms a portfolio in the corresponding row, vice versa. Red cells denote 1% significance level. Green cells denote 5% significance level. Yellow cells denote 10% significance level.

Table A4. Z-statistic of Sharpe ratio of total return for each pair of hypothetical portfolios (Period 2: 2005–2009).

	AKL	CHCH	WEL	OTH	OFF	IND	RET	MED	MSA	MSB	MSC	MSD	MSE	MSF	MSG	MSH	MSI	HQ1	HQ2	HQ3	OS1	OS2	OS3	OS4
AKL
CHCH	−2.01
WEL	−1.96	1.33
OTH	−2.49	−1.16	−2.48
OFF	−3.78	−0.41	0.85	2.15
IND	0.35	0.99	0.60	1.75	0.56
RET	−3.02	−0.44	−1.78	1.18	−1.79	−1.21
MED	−0.14	0.00	−0.08	0.10	−0.08	−0.18	0.03
MSA	−0.49	0.50	0.90	2.35	0.65	−2.22	0.90	0.14
MSB	−1.21	0.72	−0.75	2.02	−1.00	−0.65	1.39	0.05	−0.64
MSC	−3.73	0.55	−1.02	2.68	−0.96	−0.86	1.91	0.03	−1.01	−0.51
MSD	0.13	0.76	0.39	1.41	0.38	−0.45	0.97	0.15	0.12	0.47	0.66
MSE	1.46	1.69	1.46	2.11	1.42	2.53	1.93	0.59	1.74	1.44	1.60	2.20
MSF	−1.11	1.80	−0.05	1.80	0.02	−0.55	1.31	0.08	−0.49	0.56	1.52	−0.37	−1.40
MSG	−2.90	0.93	0.09	1.87	0.13	−0.56	2.45	0.10	−0.49	0.41	1.01	−0.36	−1.62	0.10
MSH	−3.33	−0.64	−2.35	1.55	−1.94	−1.45	−0.24	−0.04	−2.06	−1.47	−4.64	−1.13	−2.01	−1.53	−1.90
MSI	0.84	1.26	0.97	1.72	0.92	2.46	1.39	0.30	1.10	0.97	1.15	2.12	−1.38	0.92	0.95	1.54
HQ1	−1.23	0.83	2.62	2.12	1.95	−0.70	1.58	0.09	−0.07	2.29	5.23	−0.85	−1.98	0.82	−1.13	2.38	−0.98
HQ2	−0.73	1.53	−0.08	4.00	−0.04	−0.64	1.89	0.08	−0.73	0.41	0.60	−0.43	−1.55	−0.06	−0.19	2.34	−0.99	−0.17
HQ3	−1.00	1.42	0.53	2.47	0.44	−0.58	1.90	0.10	−0.49	0.67	19.93	−0.33	−1.54	0.37	0.22	2.66	−1.00	0.43	0.49
OS1	−1.67	1.14	2.03	2.11	2.03	−0.56	1.65	0.11	0.02	2.68	13.35	−0.70	−1.82	1.17	−0.81	2.04	−0.80	0.27	0.85	0.13
OS2	−2.09	0.94	−0.94	1.94	−0.69	−0.74	1.57	0.06	−0.80	0.11	0.77	−0.53	−1.60	−0.48	−0.53	1.91	−1.09	−2.08	−0.25	−1.40	−1.35
OS3	−0.34	1.38	0.24	26.22	0.28	−0.57	1.89	0.10	−0.76	0.74	0.81	−0.33	−1.54	0.29	0.16	2.02	−0.96	0.14	0.21	−0.03	−0.13	0.48
OS4	−0.91	1.35	0.45	2.43	0.39	−0.59	2.01	0.10	−0.51	0.64	9.93	−0.34	−1.56	0.31	0.19	2.71	−1.00	0.39	0.40	−0.15	−0.19	1.46	−0.01

Note that a significant negative value indicates that a portfolio in the corresponding column outperforms a portfolio in the corresponding row, vice versa. Red cells denote 1% significance level. Green cells denote 5% significance level. Yellow cells denote 10% significance level.

Table A5. Z-statistic of Sharpe ratio of total return for each pair of hypothetical portfolios (Period 3: 2006–2010).

	AKL	CHCH	WEL	OTH	OFF	IND	RET	MED	MSA	MSB	MSC	MSD	MSE	MSF	MSG	MSH	MSI	HQ1	HQ2	HQ3	OS1	OS2	OS3	OS4
AKL
CHCH	−0.79
WEL	−1.72	−0.58
OTH	−2.05	−1.47	−1.66
OFF	−3.11	−1.77	1.48	1.71
IND	1.67	1.81	1.81	1.97	1.74
RET	−2.06	−1.00	−0.64	1.02	−1.03	−1.89
MED	0.53	0.64	0.69	0.87	0.66	−0.05	0.78
MSA	−1.80	−0.52	2.80	1.69	2.80	−2.83	0.74	−0.59
MSB	−1.54	−0.52	0.22	2.08	−0.73	−1.71	0.90	−0.68	−1.11
MSC	−1.65	−0.37	1.08	1.68	−0.63	−1.75	0.88	−0.66	−1.24	0.50
MSD	0.78	0.97	1.17	1.45	1.07	−3.02	1.29	−0.34	0.96	1.08	1.09
MSE	2.15	2.09	2.11	2.24	2.11	1.50	2.33	0.67	2.13	2.08	2.10	2.19
MSF	−1.53	−0.35	0.76	1.50	0.18	−1.81	0.90	−0.65	−1.39	0.53	0.34	−1.07	−2.10
MSG	−0.67	0.23	0.73	1.85	0.54	−1.71	2.34	−0.62	−0.03	0.73	0.57	−0.91	−2.29	0.43
MSH	−2.53	−1.19	−1.04	1.32	−1.58	−1.98	0.11	−0.77	−1.81	−1.51	−1.35	−1.36	−2.29	−1.28	−1.84
MSI	1.14	1.18	1.43	1.56	1.40	−0.90	1.47	−0.21	1.36	1.34	1.41	0.45	−1.84	1.40	1.09	1.54
HQ1	−2.65	−0.72	2.98	1.80	3.03	−2.79	0.82	−0.63	−0.50	3.05	3.05	−2.47	−3.03	1.58	−1.50	1.85	−1.36
HQ2	−0.93	0.43	1.51	2.33	1.39	−1.62	2.49	−0.59	0.10	1.61	1.29	−0.84	−2.09	1.15	0.11	3.54	−1.13	1.18
HQ3	−0.92	0.49	1.93	2.18	1.49	−1.76	1.75	−0.58	0.43	1.45	1.56	−0.94	−2.11	1.33	0.21	2.67	−1.20	1.57	0.21
OS1	−1.31	−0.53	2.61	2.09	2.68	−2.64	1.13	−0.57	0.18	3.00	2.64	−2.23	−2.98	1.96	−1.25	1.98	−1.14	1.56	0.68	0.03
OS2	−2.73	−0.27	0.80	1.67	0.27	−1.88	1.22	−0.66	−1.55	0.60	0.41	−1.14	−2.21	0.05	−0.58	1.96	−1.43	−2.78	−1.90	−2.72	−2.01
OS3	−0.63	0.69	1.85	2.36	1.88	−1.58	2.33	−0.56	0.51	1.89	1.68	−0.78	−2.05	1.86	0.30	3.35	−1.10	2.00	1.35	0.12	0.33	2.72
OS4	−1.46	0.30	1.64	2.11	1.24	−1.79	1.72	−0.60	−0.02	1.27	1.30	−1.00	−2.15	0.97	0.02	2.69	−1.25	1.19	−0.13	−1.74	−0.37	2.57	−0.51

Note that a significant negative value indicates that a portfolio in the corresponding column outperforms a portfolio in the corresponding row, vice versa. Red cells denote 1% significance level. Green cells denote 5% significance level. Yellow cells denote 10% significance level.

Table A6. Z-statistic of Sharpe ratio of total return for each pair of hypothetical portfolios (Period 4: 2007–2011).

	AKL	CHCH	WEL	OTH	OFF	IND	RET	MED	MSA	MSB	MSC	MSD	MSE	MSF	MSG	MSH	MSI	HQ1	HQ2	HQ3	OS1	OS2	OS3	OS4
AKL
CHCH	−1.37
WEL	−1.77	0.45
OTH	−2.27	−0.07	−0.79
OFF	−1.93	0.71	1.14	1.44
IND	1.96	1.92	2.00	2.22	2.01
RET	−2.15	0.31	−0.08	0.78	−1.05	−2.11
MED	0.73	1.16	1.07	1.16	0.99	−0.24	1.10
MSA	−2.15	1.03	1.36	1.75	1.40	−2.05	1.51	−0.89
MSB	−1.96	0.53	0.65	1.63	−0.78	−2.04	0.60	−1.02	−1.45
MSC	−1.98	0.71	1.09	1.69	0.56	−2.02	1.09	−0.98	−1.37	1.47
MSD	1.34	1.49	1.57	1.84	1.55	−2.94	1.71	−0.23	1.52	1.60	1.56
MSE	2.20	2.17	2.23	2.34	2.24	0.64	2.41	0.56	2.24	2.22	2.23	1.66
MSF	−1.34	1.28	1.59	1.29	0.50	−1.90	1.13	−0.97	−0.68	0.67	0.29	−1.42	−2.17
MSG	−0.88	1.30	1.57	1.98	1.21	−1.90	2.58	−0.88	0.32	1.28	1.09	−1.39	−2.33	0.88
MSH	−2.54	0.73	0.88	1.93	0.55	−2.20	1.16	−0.96	−0.76	0.89	0.51	−1.74	−2.32	0.12	−0.85
MSI	1.69	1.78	1.86	1.94	1.89	−0.67	1.96	−0.08	1.91	1.86	1.87	0.33	−1.50	1.79	1.66	1.87
HQ1	−1.66	1.15	1.69	1.92	1.59	−1.97	2.27	−0.91	−0.23	1.47	1.31	−1.47	−2.27	0.78	−0.69	0.53	−1.83
HQ2	−1.59	1.12	1.96	1.98	1.79	−1.95	2.35	−0.94	−0.59	1.55	1.22	−1.46	−2.22	0.55	−0.89	0.18	−1.80	−0.99
HQ3	−0.66	1.27	1.66	2.43	1.73	−2.04	1.97	−0.77	1.57	1.92	1.83	−1.44	−2.14	1.14	0.59	3.24	−1.61	1.24	1.32
OS1	−1.04	1.52	2.20	1.92	1.85	−1.85	2.44	−0.88	0.21	1.59	1.43	−1.32	−2.18	1.82	−0.26	0.63	−1.72	0.82	1.56	−0.74
OS2	−1.90	1.01	1.48	1.81	1.24	−2.02	2.02	−0.94	−0.76	1.25	0.97	−1.53	−2.29	0.39	−1.05	0.17	−1.87	−2.96	−0.14	−1.54	−1.51
OS3	−1.48	1.21	2.08	2.02	2.21	−1.93	2.17	−0.92	−0.41	1.78	1.59	−1.43	−2.19	0.86	−0.65	0.35	−1.78	−0.41	1.21	−1.21	−1.16	0.73
OS4	−1.58	1.20	1.61	2.40	1.67	−2.04	2.01	−0.81	1.27	1.85	1.75	−1.49	−2.21	1.01	0.35	3.46	−1.69	1.10	1.20	−1.62	0.53	1.47	1.05

Note that a significant negative value indicates that a portfolio in the corresponding column outperforms a portfolio in the corresponding row, vice versa. Red cells denote 1% significance level. Green cells denote 5% significance level. Yellow cells denote 10% significance level.

Table A7. Z-statistic of Sharpe ratio of total return for each pair of hypothetical portfolios (Period 5: 2008–2012).

	AKL	CHCH	WEL	OTH	OFF	IND	RET	MED	MSA	MSB	MSC	MSD	MSE	MSF	MSG	MSH	MSI	HQ1	HQ2	HQ3	OS1	OS2	OS3	OS4
AKL
CHCH	−1.41
WEL	−1.93	0.02
OTH	−1.98	0.42	0.80
OFF	−2.34	0.55	1.06	0.23
IND	1.98	2.05	2.20	2.30	2.18
RET	−1.37	1.13	1.76	1.22	1.19	−1.98
MED	0.66	1.50	1.48	1.30	1.21	−0.66	1.06
MSA	−1.69	0.88	1.00	0.73	0.82	−2.12	−0.03	−1.04
MSB	−2.35	0.46	0.85	0.07	−0.48	−2.24	−1.23	−1.24	−1.00
MSC	−2.33	0.71	1.19	0.66	1.44	−2.17	−0.63	−1.13	−0.57	1.85
MSD	1.37	1.64	1.86	1.99	1.82	−2.77	1.56	−0.01	1.66	1.90	1.79
MSE	2.14	2.30	2.61	2.46	2.48	0.47	2.42	0.98	2.14	2.42	2.37	1.42
MSF	−1.10	0.75	1.68	0.48	0.40	−1.92	−0.28	−1.17	−0.14	0.45	0.10	−1.47	−2.73
MSG	0.12	1.80	2.22	2.14	2.10	−1.62	2.67	−0.63	1.22	2.04	1.81	−1.01	−2.19	1.39
MSH	−1.79	0.96	1.58	1.81	1.38	−2.27	0.40	−0.96	0.36	1.67	1.12	−1.90	−2.40	0.48	−1.19
MSI	1.49	1.78	2.00	1.81	2.00	−0.64	1.65	0.22	1.77	1.94	1.90	0.32	−1.30	1.79	1.15	1.67
HQ1	−0.97	1.31	2.02	1.70	2.20	−1.91	1.46	−0.84	0.77	1.96	1.72	−1.40	−2.45	0.98	−1.33	0.63	−1.62
HQ2	−1.11	1.34	2.05	1.72	2.49	−1.90	1.64	−0.87	0.73	2.15	1.91	−1.41	−2.35	0.83	−1.55	0.40	−1.60	−0.68
HQ3	−0.97	1.09	1.73	1.82	1.58	−2.24	0.70	−0.83	0.70	1.80	1.39	−1.77	−2.31	0.76	−0.59	1.37	−1.57	0.03	0.18
OS1	0.02	1.56	2.55	1.74	1.94	−1.56	1.55	−0.66	0.89	1.77	1.51	−0.95	−2.56	2.65	−0.10	0.94	−1.32	0.92	0.95	0.51
OS2	−1.50	1.22	1.74	1.52	2.05	−1.99	1.04	−0.91	0.64	1.89	1.66	−1.53	−2.37	0.63	−1.75	0.21	−1.67	−1.03	−0.95	−0.33	−0.98
OS3	−1.06	1.31	2.15	1.74	2.56	−1.90	1.42	−0.86	0.73	2.15	1.91	−1.40	−2.41	0.96	−1.27	0.53	−1.63	−0.59	0.38	−0.12	−1.01	0.76
OS4	−1.45	1.11	1.77	1.92	1.78	−2.19	0.77	−0.85	0.73	1.99	1.59	−1.75	−2.34	0.74	−0.78	2.15	−1.61	−0.08	0.12	−0.31	−0.62	0.33	0.04

Note that a significant negative value indicates that a portfolio in the corresponding column outperforms a portfolio in the corresponding row, vice versa. Red cells denote 1% significance level. Green cells denote 5% significance level. Yellow cells denote 10% significance level.