Location analysis and pricing of amenities: Portfolio construction in the era of data intelligence
14 / 12 / 22 - 5 minute read
We believe that the understanding of locations and identification of investment opportunities are nowadays not only based on our local experience and footprint. We believe that technology, intelligent algorithms and machine learning methods can help us confirm our knowledge and, most importantly, advise us about unknown patterns and market forces that are difficult to obtain. The way PATRIZIA incorporates machine learning algorithms in the portfolio construction is described in this report.
Urban development through the goggles of digital information
Urban development in European cities follows an increase in the supply of amenities. The growth in amenities is lower for cities that have already gone through substantial densification over the last years.
So far, our analysis reveals that the demographic expansion of European cities follows an increase in the supply of amenities that affect the wellbeing of households overall. Obviously, the larger the city in terms of households, the higher the supply of amenities per km2. However, results show that the growth in amenities is precisely limited for cities that have already seen substantial densification over the last few years. This observation allows us to conclude that the data captured through digital information systems such as OpenStreetMap and Google Maps correlates with, and is explainable through, official socioeconomic data, such as data from Oxford Economics. The opportunity to shape a specific city area in terms of (new) urban amenities is higher if a city is less densified
Transforming noise to signals
Modern location analysismeans combining the knowledge of local experts and opinions based on artificial intelligence. The PATRIZIA Amenities Magnet algorithm evaluates the attractiveness of a location based on the supply of amenities relative to the city. This is a benefit when understanding the drivers of a location and benchmarking locations within the city. With cities evolving continuously over time, the PATRIZIA Amenities Magnet Dynamic measures the change of amenities across years. It points out regions experiencing an upswing, and those that have become less attractive in terms of the amenities relative to the city.
The wellbeing of an inhabitant is proportional to the amenities accessible to them within a 15-minute bicycle ride. This is what is meant by the concept of 'chrono-urbanism'. The PATRIZIA Amenities Magnet 15 Minutes gives a disaggregated scoring to which degree a location will supply tenants with the seven basic urban needs: commuting, living, caring, working, educating, supplying and enjoying.
Adjusting signals into knowledge
'Chrono-urbanism' differentiates between seven well-being factors that affect specific aspects of urban life: commuting, enjoying, supplying, educating, working, caring and living. While each of these factors includes specific types of amenities inside the specific catchment area, the PATRIZIA Amenities Magnet 15 Minutes enables immediate benchmarking between cities across Europe. In other words, it benchmarks the supply of amenities an inhabitant can access in a normal catchment area.
Connecting signals to prices
Capturing the drivers of rental growth in a city can be challenging because they result from property characteristics, socioeconomic developments and public forces, or a combination of all these. One of the main advantages of machine learning algorithms is that they recognise the patterns and forces in data that affect rents. In Manchester, for example, the model identifies that between 2019 and 2021 rents grew on average by circa 3-4% p.a. However, rental growth is proportionally related to the distance to the city centre. Between 2019 and 2020, asking rents grew on average by 0.5 to 1.5% for properties 4-8 km from the centre and by 4.0 to 5.0% if they were 12-16 km away. The differences were even more pronounced in 2021, when rents in the outer ring of Manchester grew almost twice as fast as rents in the city centre.
Investment strategies in the era of data intelligence
The investment strategies that PATRIZIA is pursuing in the era of data intelligence are based on machine learning techniques combined with explainable artificial intelligence methods. These can identify value drivers and clusters of assets to consider in customised investment strategies.
About the authors
Dr. Marcelo Cajias
Head of Data Intelligence
Marcelo Cajias heads the Data Intelligence section, which is part of the Investment Strategy and Research team at PATRIZIA. In his role he is responsible for the global portfolio of analytical solutions and dashboards that support strategic investment decisions by means of observed and unobserved machine learning forecast models for various asset classes. Marcelo studied business administration at the University of Regensburg in Germany, majoring in statistics, econometrics and real estate economics. He received his doctorate for his thesis on the economic impact of sustainability on listed real estate companies.
His research has been published in various international journals and he has received the RICS Best Paper Award and the German Real Estate Research Prize.
Head of Global Investment Strategy, Research & Investment Solutions
Head of Investment Strategy & Research
Mahdi Mokrane joined the company in 2020 after six years at LaSalle Investment Management in London, where he was Head of European Research and Strategy and a Member of its European Management Board. Before joining LaSalle, he worked at AEW Europe, where he was Head of Research and Strategy and a member of the company’s European investment committee as well as the global securities allocation committee. He also worked closely and extensively on real estate debt and equity transactions in both UK and Continental European markets.
Associate - Data Intelligence
Anett joined PATRIZIA in November 2020 as data scientist. She is responsible for building and estimating statistical models to find hidden patters and signals in data. She employs sophisticated machine learning algorithms to be used in predictive and prescriptive research questions. Anett studied statistics at the University of Munich and received her doctorate at the University of Augsburg for her thesis on the measurement of investors' preferences regarding sustainable and responsible investments.