• Dženeta Schitton

Financial risk management in CRE industry - from old to innovative

In commercial real estate risk management system, when it comes to digitalisation and digital transformation, it is clear that that solutions "one fits all" is not possible.


In digitalisation and automatisation of the risk process, there are practical hurdles on the way many companies are not able to overcome.

However, more importantly, there is one more crucial missing link in the commercial real estate risk management and that is the absence of advanced data analytics due to the "low data frequency" problem.


What is the data problem commercial real estate industry has?


Commercial real estate industry does not have huge amounts of data which are usual in other industries, such as financial. Though, in the wider real estate frame, there are those subcategories such as residential real estate where lots of data points are available and it is easier to do data analytics with more general and non-custom made solutions. In the commercial real estate sector on the other hand we have to use advanced data analytics, such as statistical modelling in order to retrieve value helping us to overcome the business challenges.


As the big data sets have their challenges, especially demanding are situations where the data frequency is low - in commercial real estate sector market and portfolio data are being in the best case updated once per month. So when it comes to analytics we have to think out of the box in order to overcome this obstacle and to turn it into an advantage.


The low frequency data problem might be the reason risk management in the commercial real estate sector did not experience so much of the innovation which already started to happen in the other industries. Some companies are there where they were 5o years ago while some others are mostly focused on gathering data in data warehouses providing some basic business analytics.


We are still missing companies taking a new leap into the innovation and engaging with predictive analytics technologies.


In this blog post we will highlight an overview of the possible risk management solutions, starting with the old systems and moving to the newer ones, partly already present in the market.


At the end we will highlight the possibilities available when it comes to introducing advanced analytics and overcoming the low frequency data problem present in the commercial real estate investments industry.




1. Companies with the "old" risk management systems



During the "old times" when doing investments projects in commercial real estate sector, the risk management systems were based on very detailed risk and performance parameters, but offered only basic traditional analytics. When talking about expectations there were more or less only subjective ones, covering the next quarter or so. The fact is that some real estate companies and investment funds are still using the old risk management tools.


In these systems, uncertainty about future developments is mainly covered by sensitivity analysis which is done by changing main parameters within a certain range (e.g. interest rate +/- 5%, office rent +/- 10%). Here the goal is to see how some performance or risk parameters are reacting to changes of those market parameters (e.g. yield goes down by 10%, what does this mean for the expected equity yield...).


If we analyse this approach in the light of today's possibilities, it is clear that changing one (or even two or three) parameters while keeping others untouched is not sufficiently dynamic in order to be able to demonstrate any real market behaviour.

Furthermore this change range (e.g. +/- 10%) is rather not derived from real market behaviours but is again the result of subjective expectations of a risk manager resp. is a trial and error based system which helps to get a very general and approximate risk assessment calculation.


Again, needless to say, that this kind of system does now allow taking into account interconnections and associations between different asset classes and/or regional markets in a dynamic way. It is rather a "manual" approach which includes only a limited range of data within single a asset class silo or market silo.


If we take the simple Cash Flow calculation of a real estate project as an example we can illustrate how outdated this old system is.


A Cash Flow calculation within a more traditional frame would take into account initial market and project parameters of a possible investment (i.e. current market rent, your purchase price, current vacancy of the building vs the market etc) and would derive an expected performance metrics of this project. This calculation is usually done for a certain period of time (i.e. the investment term) and in general keeps the parameters unchanged, unless there is some really big change happening.


But, small and continuous changes in the market environment will not be reflected in this old calculation frame. Maybe somewhere "in the middle" of the project term, some market parameters like rent price, maybe vacancy... will be adapted at fixed points in the time line. Though, these amendments again reflect subjective estimates as it is the case with any exit yield incorporated at the end of the investment term. And this will be the basis for any further reporting to decision makers or shareholders regarding the project/ portfolio.


One has to be aware that this whole calculation framework is based on nothing more than an educated guess. We just "expected" something. In this construct, market behaviour was hardly embedded.


As said, there are still companies which are coping with this old systems, while handling huge portfolios.


If they are still successful in the market this can be attributed to their experienced management boards and the fact that up to 2021 a big part of the real estate business was still done in the these analog circumstances.


In the new post pandemic era the global economy is simply switching to a new "frequency", which is digital, so the fact that something is still working today does not mean it will stay this way.


Companies which are still ahead should react in time before they start to lose the race with the competition. Those who are already losing it, for sure have nothing to wait for!


2. Companies which already did the next step


Another fact is that many of the commercial real estate companies decided to digitalise/digitally transform even before the pandemic in order to stay competitive and gain advantage in the market.


When it comes to the financial risk management approach, most of the efforts were invested into the process of data gathering and organising big data warehouses.


This includes portfolio data, accounting data as well as digital access to market databases of providers such as CBRE or similar. These data sets are then being combined into reporting systems with some business analytics results.


Reports are mostly standardised and can be provided in a fast way. And this is one of the problems with these kinds of solutions, as they prove to be very inflexible when risk parameters have to be added or the need for additional risk metrics arise.


So when a certain change happens, due to the changed market or portfolio circumstances, it is very hard to change the parameters in the reports or to get other analytics a management board might need.

This inflexibility problem can be very well solved with introducing modular solutions, which are usually SaaS based or in the form of APIs - this way a company can extract the data which is needed and do the necessary analytics outside of the IT system and data warehouse settings prevailing within the company staying flexible and responsive to changing market and portfolio circumstances.


With these "risk as a service" solutions the company is flexible enough to:


- often change the risk parameters as business needs are demanding

- to change and combine different markets and portfolios

- to apply different kinds of analytics in order to answer to different risk questions


This kind of automatisation is enabling decision makers (or any other risk and other professionals using the tool within the company) to, no matter location, have 24/7 updated overview of the markets of interest and to change them as often as they like in a fast, efficient and reliable way.


This is something which is easy to implement and it in fact automatises the risk process in a focused and efficient way allowing companies to move fast and stay competitive.


On the other hand, when we talk about introducing advanced analytics to commercial real estate financial risk management the progress is more on the modest side.



3. Companies which choose to digitally transform the risk management


However attractive the automatisation part sounds, in the global economy undergoing digital transformation, risk management systems have to be able to offer more than a pure business analytics, no matter how seamlessly it might be working.


Commercial real estate companies and investors have to be aware that almost every industry is starting to use advanced and predictive analytics in order to make the most use out of their data and to gain competitive advantage. Not using these tools means you are not fighting with the equal means against competition and eventually "leaving money on the table".


As mentioned before, the problem with the commercial real estate industry is in partly the low frequency of the data flow, which is disabling the standard use of machine learning and AI tools. We have to work on the implementation of the advanced analytics, to take these limitations and turn them into advantages.


The fact that there is a general wrong perception that more data means better results so many companies are mostly focused on data gathering and often fail to get conclusive and reliable results out of them. This is partly due to insufficient and inappropriate analytical tools applied.


In the case of commercial real estate financial and risk data, we have a lot of possibilities offered within statistical modelling frameworks in order to get the most out of the data available - we are able to retrieve the connections between the market and portfolio parameters, to analyse their mutual impacts in real time and predictively. These "value enriched data" we are then able to use also for data hungry tools - such as machine learning.



And this is the missing link in the commercial real estate risk management, which is going to become the next development stage outgrowing mere automatisation improvements.

Within this system we overcome a lot of the limitations of the old system, as well as those experienced by the systems where some kind of digitalisation is already occurring:


1. In depth financial risk analysis of the existing project, portfolio in relation to the market i.e. revealing hidden patterns, digging out disguised associations between regional real estate markets, or between asset classes using the data we have available 2. Based on this knowledge, mirroring the market behaviour in a dynamic way: market parameters are changing according to their shown behaviour any time and in parallel with all other parameters while taking the association among parameters into account.


3. In-depth predictive analytics (short and mid-long term), implementing a cluster behaviour, fat tail behaviour and analysis of market impacts in case of extraordinary events (e.g. COVID-19). 4. Simulating and synthesising market and portfolio data, to be able to incorporate data-hungry machine learning algorithms, while keeping GDPR or data secrecy concerns in line.


5. Creating self-learning systems - with every market update and every development, the framework learns from its mistakes and self-improves (feature engineering, supervised/ unsupervised learning on updated data, steady re-evaluation of dependency structures, goodness-of-fit tests help to improve those advanced analytical techniques on a constant basis.)


The results of this market and portfolio "mirroring" and the way they interact with each other, in the real time and predictively, can be smoothly implemented into the performance and risk metrics of a real estate company - up-do-date, with simple and 24/7 access.


Conclusion

Commercial real estate industry is facing many challenges with the new post-pandemic world, which just highlighted how deeply transformative changes companies need to undertake in order to switch to the digital realities surrounding us in the global business environments.


The competition is becoming fierce and data driven decisions are not something nice to have anymore - they are becoming the new standard. This makes the old system of decision making obsolete and almost contra-productive if we want to remain competitive. On the other hand, we have to be aware that commercial real estate industry faces the limitations which come with low frequency data flow which needs to be addressed.


Implementing the right advanced data analytics tools can offer to the companies in the commercial real estate industry that competitive edge which they need to navigate their risks in data driven world and to be able to assure the most accurate predictions of future potential developments/performance.

We have to be able to get real time notifications about any change, any move, any deviation in the market and portfolio, which is derived from the current market behaviour and goes beyond mare . +/- 10% benchmark. In 2021 we have to take a step away from reading the market data and making best/ worst/ banking cases.


We have to be able to give any decision maker a quantified basis of potential future developments backed by objective, advanced and data based estimates, how likely is it for those developments to materialise.


In data driven world we need more transparent and objective tools for the decision making. In order to achieve this we need to use the industry knowledge to set the important business goals and then to combine the right kinds of analytical methods, such as statistical modelling, machine learning, networks analytics or structural equation models into custom made risk algorithms addressing these goals.

These complexed analytical tools are in fact self-improvable systems, most of our decision making will have to rely on in order to overcome limitations of the old and new, but incomplete financial risk solutions in order to keep a competitive edge.