Real Estate markets dependencies revealed -Regional Market Clusters
Updated: Mar 3, 2021
Uncertainty in Real Estate Markets
It is the beginning of 2021. The COVID-19 pandemic has a firm grip on the economy in Europe and there is an ongoing discussion among real estate professionals, how the current situation will have an impact on the real estate markets. In some segments the further development seems clear. In other asset classes, like office, there is much more uncertainty where things will head to.
Talking about office markets, thereis a wide range of opinions how markets will develop. On the one hand, we have the doomsday prophets forecasting the big crash. On the other hand, there are those market participants who see the resilience of markets confirmed in any little positive hint the market is sending. And then there are those ‘ordinary’ folks admitting that there could be a negative impact coming along with this pandemic. Albeit, so far nothing really crucial happened. In fact, there is no clear picture yet and one will have to wait how the market finally will turn out.
In short, there is a lot of uncertainty in the market. More interestingly, in a lot of circumstances the discussion, or shall we say the ‘opinion making’, runs exactly along the boundaries of local markets. Experts’ analysis reflect a respective local market in all its colours but in most cases neglect the connections and potential dynamic relationships in the wider frame and among other real estate markets. Sure enough, there are market reports handling not only one local real estate market but the DACH region as a whole, or the biggest Western European cities or similar. Nevertheless, those reports reflect several markets in parallel but do not draw the connections and interdependencies between main parametrs of those markets.
In this article we will show why is this approach insufficient and may lead to wrong conclusions. And we will elaborate how Networks Dynamicscan give a hand in significantly reducing the ongoing uncertainty about possible future developments while embedding and precisely quantifying effects different markets can have on each other. As a showcase, we will demonstrate example of the prime office market in Vienna/ Austria.
Vienna Office Market - On The Surface
Since the last financial crisis, the Vienna prime office market has taken a spectacular development. A compressing investment yield and rent prices rising to a much higher level made it a good deal for any investor stepping in early enough. The following graphs strikingly buttress this statement:
The year 2020 with the glooming pandemic in its back, obviously did not show any significant impact on this prime office market. The take-up of rental space was even higher than in the year 2019. Vacancy levels did not only remain on a convenient level but found a new low with 4.2 % by the end of the year 2020. Rent prices also were quite stable.
Only the investment volume in the office segment for the first 9 months in 2020 stayed behind the levels of the previous periods. Anyhow, this did not hinder the investment yield in this asset class — after a slight ‘hick-up’ in the middle of the year — to stay at the all time low of 3.25 % p.a. (by end of the third quarter).
All seems good, does it? First commentators and spectators are already celebrating the resilience of the market even during those extraordinary times.
Vienna Office Market - Behind The Curtains
Is the prime office market in Vienna really such resilient as it looks like? Let's consider two topics why conclusions cannot be drawn so easily.
Looking a bit deeper into the Vienna office prime market reveals that up to 30 % of the rental take-up in this segment was accounted for by just one single deal. And the tenant is a state-owned company. Was this a lucky-punch for 2020? A deal with such a size (over 50,000 sqm) needs pre-preparation and was hardly originated in 2020. Having this, the current take-up level in Vienna should not be taken as sign of market resilience as missing out on just one deal would change the picture decisively.
Much more interesting in this local market segment is the development of the investment yield given the prevailing investment volume. Here is an overview of how much is usually invested in the office market segment in Vienna per quarter:
Examining the empiric data on investment volume from 2005 to 2020, most commonly amounts around MEUR 300 were invested per quarter. In the last several years, volumes reaching MEUR 750 can be observed more frequently.
The interesting part starts with the notion that the more investment capital flows into a market segment the more pressure will be put on the investment yield, meaning rising asset prices are to be expected. Though, this is not a general rule in this asset class in this local market. In fact, watching the development of the last two years it gives the impression that it almost does not matter if the investment volume is large or small, like MEUR 660 or just MEUR 180 in a period. There is no overly strong signal for the direction of the investment yield.
Having an investment volume in one period which is three to four times higher than in another period and the investment yield repeatedly reacts the same way in both volume ranges does not make sense. In this context, the Vienna prime office market behaves irrational — as long as it is observed in an isolated way.
Let's see why.
In a nutshell, the Vienna office segment is a relatively small market. This becomes quite visible when being compared to a market like Frankfurt/ Germany. Overall office space in Vienna is appr. 6 millions sqm. Frankfurt has appr. 15.4 millions sqm. Things get even more obvious when comparing dynamic market parameters, like the range of rental take-up per quarter:
or the range of investment volume per quarter:
The small size of a local market leads to two major consequences:
Small markets are much more influenced by extraordinary events. In case of Vienna, just one big deal ‘stabilised’ the rental take-up in the difficult year 2020 which was mostly seen as a sign of resilience of the market during times of crisis. In fact, it seems this was just a “lucky-punch” happening by chance exactly at this time. The fact is that larger markets are not this vulnerable for those kind of events.
Small markets get much more influenced by external events, i.e. developments in similar but larger markets. A local market may seem to react irrationally, e.g. in terms of investment volume/ investment yield development as we saw in the Vienna prime office market. Though, behind the scene this local market may follow a trend impacted by larger and potentially similar markets.
The second issue is especially important. It shows that events in a smaller local market are more noisy and may be even exaggerated by rare events. The real signal for this local market comes from an external source, i.e. a related real estate market or a cluster of markets. That is why a local market may seem to behave irrational and that is why the sole focus on a local market is not always sufficient.
Regional Real Estate Clusters
In certain ways, different local markets have the potential to be related and to influence each other. In this context, there is a higher chance of larger local markets (“local hubs”) causing the development of smaller markets. The challenge is to connect those local markets into regional real estate clustersin a structured and quantitative way by using advanced statistical tools and combining network analytics modelling.
Once more, let’s have a look at the local office markets Vienna and Frankfurt we did for our client at D-DARKS.
The checking of the prime office investment yields of both markets unveils a very strong association between both markets in recent years with a correlation amounting to 0.94 (yield … investment yield Vienna, yield.f … investment yield Frankfurt):
It does not stop there. A check-up on some of the biggest office markets in Germany (Munich, Hamburg, Frankfurt, Berlin and Düsseldorf) reveals a quite strong association as well. Correlation in all of these cases is close to 1. See the graphic below:
Hence, there is a qualitative suspicion that we see a regional real estate cluster in this asset class with local markets interacting with each other in terms of captial flow and investment yield development. And, there is even more suspicion that the Vienna office market as the much smaller market is influenced by the German cluster.
This makes sense simply by having arbitrage considerations in mind. Markets which are similar in their political, legal and economic framework will adjust their price levels to some extent. Otherwise, arbitrage activities in those market will go on until the structural price differences are levelled out. Question is now, how to make use out of this cluster from a predictive analytics point of view.
Dynamics Networks Models
In the beginning of this article, I have mentioned that there is a lot of uncertainty regarding the future development of the office market due to possible impacts of the COVID-19 pandemic. Well, predictive analytics aims to reducing this uncertainty and — in this case — making use of those regional real estate clusters by means of dynamics networks models.
Basic approach of a dynamics networks model is to build up a framework which takes the interactions of market parameters within a local market into account, but additionally incorporatess the dependencies and causal effects between different real estate markets within a regional real estate cluster. By embedding the empiric data of those markets into the model structure, one can generate a very powerful predictive analytics tool.
This instrument helps to check on potential future developments in a market (expressed in terms of conditional probabilities) and to define market scenarios which have the highest probability to materialise. With this properties in place, it is a strong tool to re-evaluate the assumptions of real estate valuations, to assess the risk position of an investor’s real estate portfolio in specific markets or to gauge the risk/ opportunities for investments in new markets.
Another very important properties of this instrument are the ability to constantly adjust on newly evolving market behaviours and that the whole approach is quite stable even in case of low data availability. Here is a simplified model of the prime office market Vienna being linked up with the regional cluster of Germany’s biggest cities (mainly represented by Frankfurt):
The results are stunning. The behaviour of the investment yield development in the prime office market in Vienna is significantly less irrational when allowing for the investment yield development in Frankfurt to have an impact. Forecasts are getting much more accurate.
As I mentioned, conditional probabilities and highest probable market scenarios: Conducting dynamics networks does not stop with evaluating possible investment yield developments. In fact, the potential development of any market feature implemented in the model can be requested. Requests can be done in the form:
"What is the probability of the rent price rising above a certain benchmark?" or
"What is the probability of the rent price rising above a certain benchmark provided that the vacancy rate gets below a certain benchark?” or
"What is the probability of the investment yield in Vienna to fall below a certain benchmark provided the investment volume for the Vienna office market is above a certain benchmark and provided the investment yield for the Frankfurt office market stays beneath a certain benchmark?"
You get the point. This makes dynamics networks such a powerful and flexible tool for re-evaluation, assessment and simulation of potential future market scenarios and the risks coming along with those developments.
Networks Dynamics — this is a powerful predictive analytics tool, very adjustable and especially useful in rougher times as we have them right now. It helps to break down the ‘silo viewing’ of local real estate markets and is therefore an indispensable risk management tool for real estate investors.
But, there are no free lunches. Accuracy of the instrument depends on appropriately establishing the model structure. Saying this, a bunch of professionals’ expertise is necessary in order to create the right framework.
And do not make any mistake! Not all of the markets can be connected in that way shown above. That is why it is called regional real estate clusters. As an example, office markets in London (here West End and Central London) seem to miss a clear connection to the ‘German cities cluster’. To start with, both parts show a low association in terms of investment yield development with a correlation of just 0.35 on average.
As we are continuing to build up our “heat map” showing the risk clusters around the globe, we are excited to find those still missing connections together with our clients.