Guest Article by Nicholas Tuan, Associate, European Capital Markets & Cross Border Valuation, Savills. Nicholas discusses the need for the industry to embrace innovation to harness a data-driven approach that will improve advanced analytical techniques.
The Commercial Real Estate industry is one where many skills and professions play a part.
I cover the valuation sector but am intending to look at this through a slightly different lens. The valuation profession has long been deemed as an “art” given the hidden assumptions, implicit growth rates and subjective interpretation of a building’s qualities. The increasing sophistication of the industry requires money managers to report quarterly findings in an increasingly detailed manner, drawing on information from multiple sources and similarly, begs the question - how do I put a value to my real estate?
With increasing access to data, real estate valuation businesses have sought to incorporate the use of advanced analytical and calculation software systems, most notably in the residential sector with Automated Valuation Models (AVMs). These services utilise algorithms which dynamically trawl databases of existing properties and past transactions to produce best estimates of current value based on given search criteria. From a valuation business perspective, this topic would seem to put the role of a valuer at risk; there is no doubt that it is important to acknowledge the rise of big data, machine learning and artificial intelligence that can re-shape the way how we might think about valuing property. However, in commercial and high value residential real estate in particular, there are often low volumes of truly comparable transactions and the heterogeneous nature of such assets means that AVMs are likely to be less reliable.
An AVM cannot physically survey the property being valued. It works by the operator making an assumption of the condition and relies on data being reliably captured on the condition of comparables for the modelling to work. This inevitably means minute details and variables of an individual property’s conditions will be ignored. Viewed purely as a datapoint in a lake of information, limitations exist to making a judgement call and this is where the role of the valuer is key. AVMs are likely to lag the market in both directions and they cannot incorporate sentiment or, in fact, the information gained through access to agency colleagues with real-time access to bidding processes. Every investor has a different risk tolerance and motivation for transacting and this is not felt by a computer system, which can only act in accordance with the rational way it has been set up.
AVMs should however find their place in the market as it only takes a few seconds to generate valuations and requires little manual effort, not only lowering labour intensity, but also lowering the risks made by human error in traditional valuations along with the risk of fraud. Thus, AVMs are still useful for valuation purposes and can be used for checking purposes part-way through mortgage terms to take changes in property value into account. The role of the valuer should not be underestimated – high quality valuation advice remains sought after – but access to AVMs can help the professional arrive at a data-driven conclusion.
Through our experience of working with institutional funds and lenders, there exists a pain point to be considering multiple avenues for bringing the process of valuation to finalisation from both a risk and analysis perspective. Valuation reports are becoming more concise and the approach to valuation is clear and justified. It is evident that there is a need to harness the efficient and proper use of data management and handling and a data-driven approach leads to a better understanding of the valuation conclusions for the reader.
As recommended by the RICS Independent Review of Property Investment Valuations (2021), the knowledge and application of valuers should be improved in respect of advanced analytical techniques. Innovation in the real estate sector is continually being moved forwards by data science. I am increasingly excited to see how this space will evolve as innovation brings forward structural changes to how we think about analysing, presenting and working with real estate data.