Bridging Watch: AVMs don’t have mood swings

Fairfax_Chris-(Jan16)

Automation of valuations within bridging and other specialist sectors is inevitable – and can bring benefits to lenders 

Last month the Association of Short Term Lenders reported that bridging lender sentiment had “bounced back considerably” since the EU referendum. Consequentially, a significant percentage of lender members think bridging business volumes will grow over the next six months.

While any news of growth is welcome, data must be handled with some sensitivities in mind.
The research asked respondents about their feelings, which are of course subjective and heavily influenced by mood and agenda.

To predict more accurately the future of the bridging market, we need more in-depth qualitative and quantitative data, and to combine this with high-quality statistical analysis of the state of play.

My concern is that lenders, seeing increases in business volumes, agreements in principle, applications and completions, will associate this with a growing confidence in the economy, whether this is the case or not.

Incongruence is rife. Nationwide reported a 0.6 per cent rise in asking prices in August despite the National Association of Estate Agents reporting a significant fall in registered househunters. Data must be analysed and understood before drawing reliable conclusions.

Those involved in bridging need to shift focus from achieving organic growth to initiating changes to make the sector more accessible. Fortunately, this is already happening. The sector has gained from huge transformation in recent years in terms of interest pricing, and progressive action has been taken to further improve costs of entry by permitting automated valuation models.

Automation of valuations within bridging and other specialist sectors is inevitable. Here is why:

Accuracy: AVMs work by finding near-target properties and assessing characteristic similarity. They then use powerful algorithms incorporating local market evidence to estimate the value of each property. The results have proprietary analytics applied, giving a valuation estimate and confidence level for the property.

Cost: A standard full-valuation inspection costs £0.75–£1 per £1,000 value. Therefore, valuation of a £500,000 property would cost £375–£500. Bridging valuations are often more comprehensive and expensive, especially if they need to be delivered within a tight deadline. However, an AVM costs £20-£100 – a significant saving.

Risk: Traditional valuations are conducted using the comparable method. AVMs are little different from any other tool used by surveyors for analysis and estimate of value. In fact, AVMs can analyse more data than can reasonably be achieved manually by a valuer. But valuers have the advantage of reviewing the condition and features of the property. Clearly, modelling is more accurate with higher levels of data, so AVMs are best suited to high-density urban areas. A prudent lender would compare AVM result variability with actual sales or full valuations.

Competition: Progressive digital challenger banks are embracing technology where possible. This reduces operational and transactional costs, enabling them to lower their product cost and be more competitive. Assuming risk is managed and accuracy high, adoption of AVMs is essential to retain a competitive advantage.

Speed: Whereas full inspections and reporting can take weeks, AVMs are immediate.

A Rics paper entitled ‘Automated Valuation Models (2013)’ opens by describing property valuations as “both an art and a science…
a science because it is possible to establish trends and analyse how they are interpreted by buyers, including the value placed on particular property characteristics”.

Over the past decade vast increases in property data have led to significant improvement, robustness and reliability of AVMs, and usage within the mainstream market has increased. Specialist residential lenders have been less quick to adopt them in areas other than secured loans.

I hope more lenders embrace automation and the power of quantitative modelling to ease consumer costs and overall experience. AVMs are reliable, unlike human feelings.

Chris Fairfax is managing director at Positive Lending