Most of this originally appeared on the TwentyEA blog.
One of the things that we’ve been interested in here at TwentyEA is how the asking price and the final sale price relate to each other and how the two of those relate to the “fair value” of a property.
There are a number of different strategies employed by agents; some value high (perhaps to win listings) and some value low (perhaps to protect a reputation for selling properties quickly). Some discount from that initial price and some don’t, holding the line. All of this means that its quite hard to work out what is going on and whether an agent ends up achieving a high final sale price or a low one.
To look at this more scientifically we need to have an idea of a “fair value” for a property. To do this we have what is known as an AVM (or Automated Valuation Model). This uses complicated maths to allocate a price to every property in the country based on what other similar or close-by properties have sold for and how prices have moved in an area over time. These models (and in our case here the one built by TwentyEA) are pretty accurate at an individual property level – so much so that this is the type of technology that banks use to base mortgage offers on. When an average of a large number of properties is used to (e.g. the average value of properties in a street) the numbers that AVMs produce are typically very accurate – to within a fraction of a percent.
What we have done here is that the original listing price of the properties listed by an estate agency business has been compared to the price that the property finally sold for and the “fair value” that the AVM has produced. This shows the difference in outcome say between an estate who starts high then discounts against an agent who starts low and discounts by the same amount. Whether you are a buyer or a seller you’ll appreciate that these are not the same things at all.
To make the comparison between agents even easier the averages per agent are then applied to a typical house price which means we get to the sort of statement where we can say – backed up by data – things like “on average agent X will sell a typical property of value £250,000 for £5,000 more than the average in their market”.
So what are the criticism of this analysis? It seems that there are two common things that people comment on when looking at this type of work. The first is whether this means this (”sells for £xk more”) happens every single time and the answer of course is no. We are dealing with averages here and so whilst we can’t be certain that it happens in every single case we can say that this is what happens on average. The second criticism is the question of whether the fair values are exactly that. They are generated through the use of very sophisticated maths and, again, are not spot on predicting the exact price in every single case. They are though very accurate again when looking at a collection of properties through the use of averages. We are happy that this is an entirely reasonable (and industry standard) way of looking at prices.
So when we say ““on average agent X will sell a typical property of value £250,000 for £5,000 more than the average in their market” are we happy that we have the data to back this up? Absolutely. Can we guarantee that this will happen every time in the future? No – but in fairness this isn’t what we are saying…[Get your branch results here]
Pricing and discounting strategy is a complicated subject and different agents have completely reasonable reasons for behaving in different ways. What buyers and sellers should do though is ask their agency what their strategy is and whether this is going to result in the optimum outcome for any particular client.