Short Summary
A widely-held belief says property booms are contagious — one boom makes the next more likely, so you should chase the momentum. HtAG Analytics put that idea on trial across 6,229 house markets and 1.44 million monthly price observations (2007–2026) using the mathematics of self-exciting processes. The verdict: booms are not contagious in time. What looks like a 10× contagion signal is price momentum in disguise; once distinct booms are counted properly they arrive at random, about 4.4 years apart, with no memory of the last. The edge is not timing the ripple — it is selecting the right market on fundamentals.
Ask around any property circle and you will hear it: booms are contagious. One suburb catches fire, the next ignites, and if you are quick you can ride the wave. It is intuitive, it is everywhere — and, on a rigorous test, it is wrong. According to HtAG Analytics, Australian house-price booms do not feed on themselves in time. This article explains how we know, in plain English, and what it means for how you actually choose where to buy.
Here is the nutshell answer. Are property booms contagious? No — not in time. Across 6,229 Australian house markets from 2007 to 2026, distinct booms in a suburb arrive independently, roughly 4.4 years apart, with a statistical fingerprint indistinguishable from pure randomness. A recent boom tells you nothing about when the next one is due.
In 30 seconds
The myth: booms are contagious — a boom makes the next boom more likely, so chase the momentum.
The test: 6,229 house markets, 1.44M monthly observations, the same maths used for earthquake aftershocks.
The verdict: not contagious. The “10×” signal is momentum; distinct booms arrive at random, ~4.4 years apart.
So what: “overdue” is a fallacy and chasing a running suburb is chasing noise. The edge is market selection on fundamentals.
Table of Contents
- What “contagion” actually means
- The seductive illusion: booms look 10× contagious
- Counting fires, not days
- The verdict: booms are memoryless
- Seeing it in live data: six Brisbane neighbours
- The edge that survives: market selection
- Surface this data inside your AI agent
- From data signal to portfolio decision
- Key takeaways
- FAQs
What “contagion” actually means
To test a claim you first have to define it. In science, a contagious process is a self-exciting one: the occurrence of an event temporarily raises the rate at which further events occur. Earthquakes are the classic case — a big quake throws off aftershocks. The mathematical tool that captures this is the Hawkes process, and its single most important number is the branching ratio: the average number of new events each event directly triggers.
In plain English
The branching ratio is the number of new booms each boom sets off, on average. Zero means booms never trigger other booms. One means every boom triggers exactly one more — a self-sustaining epidemic. Our whole question is: is that number above zero?
If the branching ratio is zero, booms are not contagious — they are simply the visible peaks of independent, fundamentally-driven cycles. That is the hypothesis this research set out to test against the full HtAG transaction record.
The seductive illusion: booms look 10× contagious
At first glance the data screams contagion. In the month after a boom, the odds of another boom in the same suburb are about 9.8 times the normal rate, and they stay elevated for a year. Taken at face value that implies a branching ratio of 2.60 — each boom apparently triggering more than two others. That is not merely contagion; it is a physically impossible, explosive epidemic. Australian house prices manifestly do not behave that way — and that contradiction is the first clue the signal is measuring something else.
The culprit is momentum. A modelled monthly price series is highly persistent: across the 6,229 markets the median month-to-month price autocorrelation is 0.83, so a shock takes about 3.8 months to fade to half its size. In other words, a single boom mechanically produces a run of elevated months. The “10× excitation” is very largely the same boom counted again and again — not one boom igniting the next. Strip the ordinary momentum out and the one-month excitation collapses below baseline.
Counting fires, not days
The decisive test sidesteps momentum modelling entirely. Think of a boom as a bushfire. Counting boom-months is like counting the days a fire burns — a long fire looks like many events. Counting episodes is like counting separate fires. Contagion is about whether one fire starts another, so you have to count fires, not days.
HtAG therefore collapses each suburb’s boom-months into distinct episodes — a new boom only counts when it is separated from the last by at least six clear months. That yields 14,941 distinct booms across 6,161 suburbs, and 8,780 gaps between one boom and the next. The question becomes simple: do those distinct booms bunch together in time, or arrive independently?
In plain English
The “coefficient of variation” measures how bunched-up events are in time. A value of 1.0 is pure randomness — the signature of a memoryless process like radioactive decay. Above 1.0 means events cluster into bursts (the fingerprint of contagion). The distinct-boom gaps come in at 0.98 — random, not clustered.
The verdict: booms are memoryless
Measured on the 8,780 gaps between distinct booms, the coefficient of variation is 0.98 (95% confidence interval 0.97–0.99) — statistically indistinguishable from a memoryless, random process, and nowhere near the value self-excitation would produce. The average gap is 52.9 months — about 4.4 years. Booms arrive like the ticks of a random clock, carrying no memory of the last one.
According to HtAG Analytics, once booms are counted as distinct episodes the branching ratio falls from an illusory 2.60 to effectively zero — a self-sustaining boom epidemic would require a value near one. Australian housing sits at roughly zero.
The result holds across stricter and looser definitions of a “boom”, so it is not an artefact of one arbitrary choice. The only meaningful departure from pure randomness is a mild excess of very long gaps around 13–14 years — the echo of the broad national price cycle moving all suburbs together, which is the opposite of localised, suburb-to-suburb triggering. This test looks only at booms travelling through time; whether they travel through space is the subject of our companion study, the property ripple effect on trial.
Seeing it in live data: six Brisbane neighbours
The finding becomes concrete with live HtAG data. Take six adjacent house markets in northern Brisbane. All six rode the same 2020–21 upswing and, as at 30 June 2026, all six sit at the same cycle phase — decelerating together. That shared movement is the national cycle, not one suburb igniting the next. And crucially, the suburb that ran hardest over the past five years is not the strongest market going forward.
| Suburb (adjacent, same LGA) | Typical Price | Cycle (GRC) | Actual 5-yr growth | RCS Overall |
|---|---|---|---|---|
| Zillmere | $1,175,556 | (+)Decreasing | +90% (most) | 47 |
| Boondall | $1,199,266 | (+)Decreasing | +75% | 81 |
| Wavell Heights | $1,848,773 | (+)Decreasing | +72% | 85 |
| Chermside | $1,395,270 | (+)Decreasing | +48% (least) | 27 |
Source: HtAG Analytics, houses, as at 30 June 2026 (High confidence). Growth is total realised house-price growth over the five years to June 2026.
Read it and the myth dissolves. The hottest recent runner (Zillmere) scores a middling Relative Composite Score of 47; the well-known “premium” name Chermside scores just 27; and quieter Boondall and Wavell Heights score 81 and 85. Recent heat is not the forward signal. Chasing it would also have pointed you at the group’s weakest structural base: Zillmere sits in the lowest socio-economic band of the six (IRSAD decile 3) and carries the strongest “running hot” reading — a warning, not a green light. What separates these neighbours is fundamentals — and that is exactly what booms being memoryless implies: you cannot time your way in, so you have to select your way in.
The edge that survives: market selection
If booms were contagious, the rational move would be reactive — detect a running suburb and chase the ripple. This finding removes the statistical basis for that in the time dimension. Three practical consequences follow:
- “Overdue” is a fallacy. Because booms are memoryless, a suburb that “hasn’t moved in years” is not building pressure that must release. There is no clock counting down.
- Recency is not a signal. That a suburb boomed recently tells you nothing about its next move; reacting to a boom already in progress is chasing noise.
- Fundamentals are where the repeatable edge lives — supply tightness, demand depth, affordability, income and infrastructure. Booms are the surface; the drivers are underneath, and they are what a disciplined process reads before the cycle arrives.
This is not a counsel of despair — it is the opposite. It says the edge is cross-sectional: choosing the right market on its fundamentals, not timing the wrong one. That is precisely what HtAG’s Property Intelligence layer is built to do, and it is what our 14-year Dex backtest put on trial against the entire market — with the results tracked openly in the Evidence Portal. The same logic drives our findings on what really drives suburb growth and the regional-versus-metro question. Our study of LGA versus suburb data shows exactly how much a council-level average can hide. You can map the fundamentals yourself on the GeoDex heatmap, and read the spatial half of this research in the ripple effect on trial.
Surface this data inside your AI agent
The HtAG Developer Portal exposes the suburb-level cycle, momentum and composite-score data behind this article through MCP (Model Context Protocol) connectors. Investors and buyers agents using Claude, Perplexity, Manus AI, ChatGPT (via custom connectors) or any other MCP-compatible AI agent can query it directly — and get answers grounded in HtAG’s data rather than the market myths an ungrounded model will repeat.
Browse the endpoint catalogue at developer.htagai.com and submit the HtAG Developer Portal application — approved members receive an API key and an MCP setup guide for their preferred tool.
From data signal to portfolio decision
If booms cannot be timed, selection is everything — and selection is what the HtAG platform is built for. The Relative Composite Score, cycle position and the rest of the metrics in this article are live across every Australian market, so you can compare markets on fundamentals instead of chasing last month’s headline.
Start your HtAG Analytics membership → · Apply for Developer Portal access →
Key takeaways
- Australian house-price booms are not contagious in time — a boom does not raise the odds of the next boom in the same suburb.
- The apparent “10× contagion” is price momentum (median month-to-month persistence 0.83) — one boom counted many times.
- Counted as distinct episodes, gaps between booms have a coefficient of variation of 0.98 — memoryless — with a mean of ~4.4 years.
- “Overdue” is a fallacy and recency is not a signal; the branching ratio is effectively zero.
- The only repeatable edge is market selection on fundamentals — validated over 14 years in the Dex backtest.
FAQs
Are property booms contagious?
Not in time. Across 6,229 Australian house markets (2007–2026), HtAG Analytics found that distinct booms arrive independently — about 4.4 years apart, with a coefficient of variation of 0.98, indistinguishable from randomness. A boom does not raise the odds of the next boom in the same suburb.
Is a suburb ever “overdue” for a boom?
No. Because booms are memoryless, there is no building pressure that must release. How long ago a suburb last boomed carries no information about when — or whether — it will boom again.
Why does the raw data look like contagion?
Because a modelled monthly price series is highly persistent (median month-to-month autocorrelation 0.83). A single boom produces a run of elevated months, so counting boom-months naively makes one boom look like many — manufacturing an illusion of clustering.
If I can’t time booms, what should I do instead?
Select markets on their fundamentals — supply, demand depth, affordability, cycle position and risk — before the shared driver arrives. HtAG’s Relative Composite Score and Property Intelligence tools are built for exactly this cross-sectional comparison, and the approach is validated in the 14-year Dex backtest.
How do I access HtAG cycle and momentum data inside Claude or Perplexity?
Through the HtAG Developer Portal. Browse the catalogue at https://developer.htagai.com/ and apply at https://links.htag.com.au/widget/form/GFVegAaXzeTUH7QzRl1T. Approved members get an API key and an MCP setup guide so any MCP-compatible AI agent can query HtAG suburb data directly.
Citation: HtAG Analytics — Property Boom Contagion whitepaper (self-exciting point-process test; 6,229 house markets; 1,443,971 monthly observations; January 2007 – May 2026).
The conceptual framework behind this research is published openly for transparency and education. Its proprietary implementation — the calibration, weighting, validation and underlying data behind HtAG’s market-selection scores — remains the confidential intellectual property of HtAG Analytics.
This article forms part of the HtAG Property Intelligence Reference Library — a structured knowledge base documenting the concepts, metrics and methodologies used to analyse Australian residential property markets.
Reference Standard PI-BOOMCONTAGION · Version 1.0
Disclaimer: This article is for educational purposes only and does not constitute financial advice. Statistical findings describe historical patterns in the data analysed and are not predictions of future price movements. Past performance is not indicative of future results. Always conduct your own due diligence and consult a qualified professional before making investment decisions.
