Agents say auctions get higher prices. The research says it depends.
The belief that auctions achieve higher prices shapes which properties get listed by auction and which do not — in Melbourne, roughly 60% of residential properties sell by auction, compared to much lower rates in Brisbane, Perth, and Darwin where dominates . This step explains what the academic evidence actually says about the , why the original findings have been challenged, and what the unresolved debate means for you as a buyer evaluating which sale format suits your situation.
That 9.5% figure comes from a study of over 536,000 transactions in Melbourne, Sydney, Brisbane, Adelaide, and Perth. But a different 2012 study using rigorous selection-bias controls on Sydney and Christchurch data found no significant auction premium at all . The research is contested — and that contestation matters when an agent tells you auctions "achieve the best price."
The premium traces back to a single 1996 study — and its acknowledged flaw
The idea has a specific origin in the academic literature. In 1996, Kenneth Lusht published a study examining 163 properties sold by auction and 58 sold by private treaty in Melbourne, covering January 1988 to March 1989. Using a — a statistical method that controls for property characteristics like size, location, and condition — Lusht found that auction sales commanded a premium of approximately $19,326, or around 8% of the mean sale price at the time .
That 8% premium became an influential data point, widely cited and built into the conventional wisdom that auctions maximise sale prices. But Lusht himself acknowledged the study did not control for in the choice of sale method. Two specific selection effects operate in the Australian market.
Effect 1 — Higher-quality properties go to auction. Agents tend to recommend auctions for properties in high-demand areas with multiple potential buyers. These properties would likely achieve strong prices regardless of sale method. Comparing their sale prices to the broader private treaty market compares premium stock to a wider pool.
Effect 2 — Failed auctions contaminate the private treaty sample. A property that goes to auction, does not meet its reserve price, and subsequently sells via private treaty contributes a low-price result to the private treaty data. This artificially depresses the private treaty average. It is not a fair private treaty sale — it is a failed auction dressed as one .
Both effects inflate the apparent auction premium in a simple comparison.
Correcting for selection bias splits the research in two
The most direct challenge to the Lusht finding came from Alex Frino, Maurice Peat, and Danika Wright in a 2012 paper published in Accounting and Finance. Using a larger dataset covering properties in Sydney and Christchurch, and applying two more rigorous methods — a and a matched sampling procedure — they found no significant difference between prices achieved at auction and those achieved by private treaty once selection bias was controlled for . When you compare like with like, the auction premium disappears.
That finding was itself challenged. Lepone, Frino, Mollica, and Vassallo (published 2010, using data from 2005 to 2009) examined over 536,000 transactions across five Australian capital cities. After applying a two-stage model to control for self-selection, they found that houses sold by auction generally still commanded a higher price across all cities studied — approximately 9.5% for houses .
Their findings were more nuanced for units. The auction premium for units only appeared in cities where auctions were less common — specifically Perth and Adelaide. In Melbourne and Sydney, where auctions are prevalent, no premium was found for unit sales. Where auction is the default method, the premium weakens.
Finding | Evidence |
|---|---|
A raw auction premium exists in simple comparisons | Consistent across studies — but reflects property selection, not mechanism |
Correcting for selection bias eliminates the premium | Frino, Peat & Wright (2012) — no significant premium on Sydney/Christchurch data |
A premium may survive correction for houses in some markets | Lepone et al. (2010) — ~9.5% premium for houses across five capital cities |
Unit sales show no premium in high-auction markets | Lepone et al. (2010) — premium absent for units in Melbourne and Sydney |
More recent review work on behavioural factors notes that the design of the auction mechanism — competitive, time-pressured, publicly visible — generates conditions that can push buyers past their private reservation price . But whether that effect survives when you compare equivalent properties sold through each method remains contested.
The premium, if real, is a cost to you — not a gift
As a buyer, you are on the other side of this debate. The "auction premium" is not something you receive. It is an additional cost. The evidence has three practical consequences.
Do not assume the auction price is the fair price. The final auction price reflects what the most motivated buyer in the room was willing to pay on that day, under competitive conditions, without a cooling-off period or conditions for exit. That is not necessarily the same as the property's underlying market value.
Competitive bidding creates specific psychological pressure. The public, real-time nature of bidding — watching a competitor raise their paddle, knowing the deadline is immediate — activates what economists call the : the tendency for the winning bidder in a competitive auction to have overestimated value relative to the next bidder. The more bidders in the room, the more acute this effect becomes .
Private treaty gives you information and time. A private treaty listing usually has an asking price, which anchors your expectations. You can make an offer subject to conditions. You have time to think. None of those apply at auction. Neither format is automatically better — they create different decision environments.