Income Predicts Crime Better Than House Price — Sydney's Data Proves It

Our earlier analysis found that house price is a surprisingly weak predictor of suburb safety — it explained just 13.5% of the variation in crime rates. So we asked the obvious follow-up: is there a better predictor? We ran the same analysis using median household income instead of price. The answer is clear.

The headline finding

Across 311 NSW suburbs with a population above 3,000, median weekly household income correlates with residential crime at r=-0.53 — a moderate-to-strong negative relationship, and a stronger one than price managed.

More importantly, income explains 28% of the variation in crime rates between suburbs. House price explained 13.5%. Income does roughly twice the work.

To make the comparison airtight, we re-ran both correlations on the exact same set of 277 suburbs that have both a reliable median price and Census income. Head to head:

• Household income: r=-0.51, explains 25.9% of crime variation.
• House price: r=-0.37, explains 13.5% of crime variation.

Income wins by a 12-point margin on identical data. If you want a single number to estimate a suburb's safety, what households earn tells you far more than what their homes cost.

The pattern that never plateaus

The chart above shows the average safety score climbing steadily across income bands. The crime numbers underneath it are even more striking, because unlike price, they keep falling the whole way up.

• Suburbs where households earn under $1,000/week average a residential crime rate of 2,540 per 100,000 residents.
• $1,000–$1,500: 1,663.
• $1,500–$2,000: 1,104.
• $2,000–$2,500: 823.
• $2,500+: 507.

That is a clean, monotonic decline — every step up in income buys real safety. This is the key contrast with price. Our price study found safety plateaued above $1.5M: paying an extra million bought no additional safety, because the premium was buying harbour views and architecture, not lower crime. Income has no such ceiling. The richest band records a fifth of the crime of the poorest. Income tracks the thing that actually drives the difference — the socioeconomic stability of the people who live there — rather than the desirability of the postcode.

Which crimes track income most closely

Not every category responds to income the same way:

Domestic assault has by far the strongest link (r=-0.56, explaining 31% of variation). Domestic violence is the crime most tightly bound to financial stress, unemployment and housing instability. As household income rises, it falls faster than any other category — the same pattern we saw with price, only stronger.

Motor vehicle theft (r=-0.44) and non-domestic (street) assault (r=-0.44) follow. Higher-income suburbs have more garaged, secured parking and less of the late-night street activity that drives public-place assault.

Break and enter (r=-0.43) sits just behind.

Steal from dwelling (r=-0.33) and robbery (r=-0.30) are the weakest links — the same two laggards as in the price study. Robbery in particular follows people and foot traffic, clustering around transport hubs and entertainment strips regardless of how much the surrounding households earn.

The high-income suburbs that break the rule

A strong correlation still leaves exceptions, and the exceptions are revealing. Among the highest-earning third of NSW suburbs, a handful still record high crime — and they share an obvious profile.

Surry Hills tops the list: households earn a healthy $2,308/week, yet its residential safety score is just 11 out of 100, with a crime rate of 2,761 per 100,000. Merewether ($2,287/week, safety 12) and Alexandria ($2,654/week, safety 32) follow the same script: affluent, dense, inner-city suburbs wrapped around entertainment precincts. Forest Lodge ($2,437/week, safety 47) rounds out the pattern.

These are not poor suburbs with a crime problem. They are wealthy suburbs where the crime reflects the environment — nightlife, foot traffic and a transient day-and-night population — rather than the residents. It is the same confound that muddied the price analysis, and it is the single biggest reason neither income nor price ever explains *most* of the variation.

The affordable suburbs that are quietly safe

The other side of the coin is the more encouraging story: lower-income suburbs that are nonetheless among the safest in the state.

Bowral in the Southern Highlands earns a below-median $1,535/week yet scores a perfect 100 safety, with a crime rate of just 260 per 100,000 — the lowest in the entire dataset. St Georges Basin ($1,113/week, safety 100), Bossley Park ($1,542/week, safety 100), Merimbula ($1,146/week, safety 100) and Banora Point ($1,268/week, safety 94) tell the same story.

The pattern in the safe-and-affordable suburbs is consistent: lower-density regional and coastal towns, stable and predominantly owner-occupier, with limited entertainment precincts. Income is a strong signal, but community fabric and density clearly matter too — these suburbs punch well above their income on safety.

What this means for suburb research

• Income is a better safety proxy than price. If you are using a single number to gauge a suburb's safety, the median household income tells you more than the median house price — and it is free Census data.
• The income effect doesn't plateau. Unlike price, safety keeps improving across the whole income range, because income tracks the resident population rather than the desirability of the location.
• Density is the great confound. Wealthy inner-city suburbs near entertainment precincts buck the trend in both analyses. Neither income nor price can see nightlife.
• Domestic assault is the most income-sensitive crime. If family safety is the priority, household income is an especially informative signal.

Better still: don't rely on any single proxy. Look up the actual crime data, safety score and full profile for the suburb you care about.

Methodology

Income data: ABS 2021 Census median weekly household income. Crime data: BOCSAR (Bureau of Crime Statistics and Research) annual suburb-level data, latest year. Population: ABS 2021 Census. Residential crime rate is the sum of the five residential categories — break and enter dwelling, steal from dwelling, motor vehicle theft, domestic assault and robbery — per 100,000 residents. Non-domestic assault is analysed as a supplementary category and is not part of the residential rate.

Analysis covers 311 NSW suburbs with population above 3,000 that have both income and crime data; the price comparison uses the 277 of those with a reliable median house price (at least 5 recorded sales). Pearson and Spearman correlations agree in direction and rank throughout. Full methodology and data sources.