Kempen Real Estate Update: What determines if your shopping centre will survive?
The barter economy features very early on in the annals of human history – the exchange of goods and services has shaped human society throughout the ages and that trend is no different today. The modern retail property owner spends much of their time looking at what makes a great retail environment work, and yet we find that little attention has been given to the factors that drive the rents in those properties. With our big data approach, back-tested with data published by public real estate companies, and with the help of several listed companies and non-listed funds, we have decided to add to the existing knowledge and perform our own investigation on the factors that influence retail footfall and rents.
To do that, we analysed 1,115 shopping centres and street retail units owned by 16 listed real estate companies and 5 non-listed funds specialized in retail property management. All of the assets are based in Europe (both in the UK and on the Continent). Our study aimed to use a bottom-up approach to identify which factors have a statistically significant ability to explain rent levels and growth. As we believe that the ability of a shopping centre to draw customers is one of the foremost determinants on rentals, we looked both at the local demographic data, and centre specific data. By the end of the our research assignment, an incredible amount of information has been processed and stored in our framework, and we then analysed it in the following order.
Location, location, location
The first difference between shopping centres is topographical: where does the store sit? Historic and densely populated urban centres that we find across Europe create their own retail art forms, and attract both local shoppers and tourists. Some studies assert that rents are expected to be high where there is dense populations, big footfall and high visibility. But suburban shopping centres in the right demographics can thrive too, if the built environment properly reflects the nature of those who use them. The first step in our analysis was therefore to rank the shopping centres on the basis of the quality of their location. For that, for every shopping centre and street retail unit that we analysed, we assessed demographic factors, economic and labour market statistics, as well as other aspects of the local environment. We looked especially closely at factors like population growth and density, employment rate and growth, GDP per capita, level of education, disposable income, tourism arrivals, crime rates, online purchase data, and many more.
What do existing tenants tell us about the quality of a shopping centre?
Today’s shopping centres that want to survive need to provide a compelling reason for shoppers to leave their houses. Tenant quality and variety is a significant factor in boosting sales, so a desirable retail mix is expected to contribute positively to the synergies within the centre, as well as to the rent level. Having identified 160 major retail tenants active in Europe that we feel either enhance or detract from a centres attractiveness, we were able to locate 7,000 sites rented by those tenants in our shopping centres and applied a direct score to the tenant’s presence. The criteria that we considered to assess the tenant quality concerns the ability of retailers to attract consumers in the physical store and generate sales, and therefore improve rental income. The exact score is assigned by considering factors like credit rating, financial condition of tenants without a credit rating, the tenant’s business line, and other indicators either sourced from the press or the landlords themselves. For example, tenants such as Apple or Tesla received a high tenant score based on their ability to attract consumers in the physical store and pay high rent. Tenants that are known to have recently filed for bankruptcy or are closing stores, received a low tenant score.
When it comes to tenant type, we looked closely into the following categories across the entire tenant dataset: “Fashion” (apparel and shoes), “Health” (pharmacies, drugstores, opticians and doctors’ offices), “Services” (travel agencies, laundries, real estate agents, bank branches etc.) and “Leisure” (restaurants, cafes, art galleries, amusement parks etc.). We then focused on identifying which tenants appear to be resilient to the online competition and growing their offline presence, and which are not. The easiness to replace a tenant type with a tenant representing another use has also been evaluated. However, we note that in the dynamic world of retail, the relation between the tenant quality today and historical rents may not meet the forward looking approach because the profitable tenants of the past are not always the same ones that are expected to be successful in the future. Hence, our final tenant score has been defined on the basis of the forward looking expectations about which tenants will make a retail asset profitable, and the inclusion of the variables related to the tenant mix aims to examine whether the forecasts are already in motion.
“Our added value lies in our ability to go beyond merely collecting data, but also to make educated decisions about what the data means to value of portfolios, and how to be best positioned in terms of our holdings.”
Most interesting findings
All of the factors mentioned above come into play simultaneously, forming a complex matrix of data and behaviours that must be understood if any retail environment is to adapt itself successfully to the needs of consumers. The next step for us was therefore to examine which of the factors are the most significant in predicting rent levels and growth, and to establish how to weigh the considered factors and parameters to come to the final Property Score of every retail asset under consideration. To do that, we first looked at the rent per square meter between 2015 and 2019 of the sample of 100 shopping centres on which public data is available. The second variable employed in our research was annual footfall, on which we retrieved public data for 81 out of the 100 assets considered in the rent analysis. We subsequently regressed the two variables on the locations variables and on the variables related to the tenant mix to examine whether the considered factors are consistently and significantly influencing the shopping centres profitability and popularity.
First, from the whole set of demographic variables that we took into account, we found that population density was the factor with the highest positive correlation with both rent levels and footfall. This variable is followed by GDP per capita and disposable income, both proved to have a positive impact on predicting rent levels. This makes intuitive sense. When a shopping centre is built near a residential area, especially with high income and purchasing power, it is expected to attract a large volume of consumers and thus increase sales. An increase in population density is also expected to positively impact retail sales and the level of rents. Areas with high GDP per capita should command higher rental rates than those with low level. Finally, the change in personal income has resulted to be significant and positive in explaining the rent levels.
We also found tourism arrivals positively and significantly correlated with rent levels, which was also in line with expectations. The relation of the crime rate, as a proxy for the safety of a given region, and rent levels was a question mark for us at the beginning. On the one hand, people desire to live in locations with a low crime rates. On the other hand, the most populated areas often happen to be the ones with increased level of crime. In the end, we were able to confirm a negative correlation between crime rate and rent levels, which means that shopping centres in areas not considered to be safe find it hard to charge the top dollar rent.
When it comes to tenant type, we found that tenants in the “Leisure” category were the most positively correlated with footfall, so their extensive presence in a shopping centre should be considered a strength. The proportion of tenants in the “Fashion” category is positively correlated with rent levels. However, the same category proved to be negatively correlated with footfall. This could be explained by the fact that fashion tenants, which have been the core reason for many shopping centres to open in the first place, are undergoing a significant rationalization of space today, report declining sales, and may not be a good indication of rent levels going forward. As for other categories, we found that a significant presence of “Health” tenants is negatively correlated with rent levels. This is a rather new retail category, introduced in shopping centres vastly in response to the shrinking space occupied by “Fashion” tenants. It appears that tenants in the “Health” category pay relatively low rent. Additionally, windows of pharmacies and health clinics are boring, nobody likes to visit such premises out of pleasure, and if you are a tenant that needs to be near them, you need to make sure that you make your window extra interesting to slow shoppers down.
We then explored the correlation among the explanatory variables and found that many of the demographic and economic factors are positively correlated with each other. In cases where the coefficients suggested multicollinearity, we removed one of the variables or introduced a third variable averaging the correlated two. What partially came as a surprise was the positive correlation between online purchasing and most of the demographic and economic factors that positively impact both the rent levels and the footfall. Online purchasing is especially highly correlated with employment rate and disposable income. We have indeed found research showing than affluent people are more likely to make a purchase online than the non-affluent, significantly contributing to online sales. As a consequence, we needed to conclude that the more online shopping takes place in an area, the higher rent levels and footfall should be expected in the nearby shopping centres. This came as counter intuitive given the fact that e-commerce is believed to be one of the major drivers of the retail industry downturn. However, we still think that over the longer term the increasing willingness of all demographic groups to shop online could be a significant threat to the brick-and-mortar retail environment, as once shopping rituals are broken, it is difficult to re-establish them.
Finally, we analysed retail sales per square metre publicly disclosed by one of the companies. Our findings clearly suggested that retail sales can explain the passing rents per square metre. This is great news because it means for us that the inclusion of variables affecting retail activity is a good predictor of rents a shopping centre can charge.
Based on all our analysis, we assigned weights to all the factors considered under Submarket Score and the Tenant Score in the way to best explain the rent levels and footfall numbers, while omitting factors that proved not significant. In this way we came to the Property Score at the asset level for all the 1,115 shopping centres we analysed. Subsequently, we aggregated the data on a country level to obtain a portfolio score for all the European REITs, UK REITs and Dutch non-listed property funds. We then were able to estimate the corresponding long-term rental growth rate per company per country, which – when implemented in our models – ultimately lead to adjusted NAV numbers that impact our investment decisions.
The science of shopping is a hybrid discipline: part social science, part physical science, and partly art. Retail environment is also changing rapidly, and many of the trends we just wrote about might already be decelerating or advancing quickly. Because the science of shopping is being invented as we go along, we could never be sure what we would find until we found it, and we often needed to stop to understand what we were seeing. The fact that shopping patterns as well as demographics are constantly evolving, we know that we can still get surprised, and we are prepared to update and re-run the data on a regular basis. The overwhelming lesson that we have learned is that amenability and profitability of a shopping centre are totally and inextricably linked. In other words, if you build and operate a retail environment that fits the highly particular needs of shoppers, your rent levels and rent growth will be superior to those of your peers that don’t.
The findings are useful for us to better understand the performance of retail property sector, to assign portfolio growth rates in our models, to model capital expenditure needed to keep up with evolving trends, and eventually to better understand the value per share that every analysed retail portfolio should bring. Our added value lies therefore in our ability to go beyond merely collecting data, but also to make educated decisions about what the data means to value of portfolios, and how to be best positioned in terms of our holdings.
The views expressed in this document may be subject to change at any given time, without prior notice. Kempen Capital Management N.V. (KCM ) has no obligation to update the contents of this document. As asset manager KCM may have investments, generally for the benefit of third parties, in financial instruments mentioned in this document and it may at any time decide to execute buy or sell transactions in these financial instruments.
The information in this document is solely for your information. This article does not contain investment advice, no investment recommendation, no research, or an invitation to buy or sell any financial instruments, and should not be interpreted as such. This document is based on information that we consider reliable, but we do not represent that it is accurate or complete, and it should not be relied on as such.
The views expressed herein are our current views as of the date appearing on this document. This document has been produced independently of the company and the views contained herein are entirely those of KCM.
KCM is licensed as a manager of various UCITS and AIFs and to provide investment services and is subject to supervision by the Netherlands Authority for the Financial Markets.