It would be an interesting analysis to do: how many current property owners are enjoying their property asset value gains through shrewd planning, and how many are enjoying it through sheer luck. Problem is, not many people would be willing to admit it.
For CoreLogic Executive, Software Engineering & Technology - Greg Dickason, property is now his professional subject matter and personal passion. Greg is sought after for his thought leadership as a renowned keynote speaker and regular contributor to the Australian Financial Review and industry news sites. Property, and the innovation surrounding it is now his life - but at the turn of the millennium, things were very different. Greg comments:
"I actually bought our first house because my brother-in-law told me to. Just 3 months after migrating to Australia, and with a 6-month-old daughter and a 1-month-old job, I had no idea at all about property or the market, but it was the year 2000 and turns out, it was the best investment I ever made. My brother-in-law just told me it was a good buy, luckily I listened and the bank agreed. At the time I had no idea how the bank assessed applications, how they valued a property, or put together all the figures to determine the risk before making a decision to give me the loan".
Across the ditch in NZ, it would be fair to say most buyers don’t know about the intricacies of a loan approval process until they’re right in amongst it themselves, but with a slowing property market and credit tight, more buyers are learning about lending because they’re having to navigate the mortgage process and ask the question: ‘Will the value of the property be enough to support my loan?’
Besides checking your credit history, and your capacity to pay, lenders check your actual loan collateral (the property you are buying), to make sure if they need to repossess and sell, they will recoup their money.
So how do lenders value your property?
In the normal process they will order an ‘Automated Valuation Model (AVM) Report, from a provider such as CoreLogic. This machine-generated report will contain an estimated value calculated by looking at recent comparable sales in the surrounding area. It will take into account relative land sizes, numbers of bedrooms and bathrooms plus other characteristics. The model will also compare the distance from the property of those comparables, and how far back in time they sold.
Modern AVM models even self-learn by using AI techniques to weight the different factors and become more accurate through ‘machine learning’. An AVM also estimates its own accuracy. In what is known as a ‘Forecast Standard Deviation’, it can tell the lender how much to trust the figure returned.
Lenders can then choose to use an accurate AVM and lend to you, which they will often do if you are providing a deposit of over 20% of the property’s value. If the AVM is inaccurate though, or your deposit is small, they may ‘cascade’ the valuation to a registered valuer working in your area.
Registered valuers are degree-qualified professionals who spend a considerable amount of time researching the area and walking through sold properties. Typically they will undertake a physical inspection of the property and provide a comprehensive report that meets Bank and Valuation Industry standards. On occasion, other forms of valuation may be carried out including a Desktop valuation.
This is why it can take your lender some time to make the decision. Although they can check your credit quickly, and in most cases your earnings and expenses (your capacity), the valuation part may need a physical inspection particularly if your deposit is low or the property is unusual.
Valuing a property ends up being a combination of machine learning models applied to property data, working in tandem with professional valuers that have deep local knowledge. Lenders want the process to be as quick as possible to give buyers the best experience, without sacrificing confidence in the valuation. This is driving rapid innovation, which is now Dickason’s specialty:
"At CoreLogic, we are building models that can actually be auto-tuned by the valuer themselves: a combination of people smarts and cloud-scale machine learning, and that’s pretty exciting.
As you can see, there’s a lot of background property intelligence going for mortgage approvals. CoreLogic is proud to partner with many of NZ’s major banks, insurers, and brokerages to make it all happen".
A version of this article first appeared in the Australian Financial Times.