Explaining Seeming Anomalies in the Indigenous Voice Referendum Vote in Greater Sydney

In Australia’s 2023 Indigenous Voice Referendum (see the two previous posts), a significant number of solid Labour electoral divisions voted “no,” some by a significant margin, even though the Labour Party strongly supported the measure. It was a different story on the other side of Australia’s political divide. Only one division that voted for a party in the country’s conservative alliance in 2022 voted “yes” in 2023. That division was Bradfield in northern Sydney, a traditionally strong Liberal (center-right) constituency. But a handful of historical Liberal strongholds that had bucked the party for “teal” independents in 2022 also supported the referendum. We will examine the rise of Australia’s teal politicians – who are on the left regarding socio-cultural and environmental issues but are center-right economically, in a later post. For today, we will take a closer look at the seemingly anomalous electoral divisions in the Indigenous Voice referendum in the greater Sydney area. As always, we shall do so through cartographic analysis.

The first map shows the “yes” vote in the 2023 referendum. As can be seen, support was highest – by a significant margin – in the central urban area (the divisions of Sydney and Grayndler). The other “yes-voting” divisions are concentrated in the northeastern suburbs. (Reid, to the west of the urban core, is mapped as having supported the measure by a thin margin, reflecting the Wikipedia table used to make this map; many other sources, however, map Reid as having voted “no.”) The second map shows the local results of Australia’s 2022 parliamentary election. As can be seen, the “teal” divisions are also located in Sydney’s northeastern suburbs, all of which voted “yes” in the referendum. In contrast, Labour’s strongholds in 2022 were located in central Sydney and in the western suburbs, as has been the historical norm. The third map takes information from the second map and overlays it on the first, outlining the “anomalous” parts of Sydney in this election (“no-voting” Labor divisions and “yes-voting” Liberal (and teal) divisions. The same overlay will be employed in the remaining maps in today’s post.

As other sources have noted, level of education was a good predictor in the Indigenous Voice referendum. The next map cartographically examines this correlation. As can be seen, rates of college education are much higher in eastern Sydney than in the rest of the region. The most highly educated divisions either supported the measure or narrowly turned it down, with strongly Liberal Mitchell being the only significant exception. In Labour-voting divisions, those with high levels of education supported the referendum whereas those with low levels voted “no.” (The data used to make this map, like the rest others in this sequence, come from the 2021 Australian Census.)


Not surprisingly, the map of median household income correlates closely with that of educational attainment. Centrally located Sydney and Grayndler, however, have lower income levels than would  be expected from their educational profiles. This is partly due to their young populations; at 33, the median age in the Division of Sydney is tied for the lowest in the region mapped. The general message conveyed by this map is that relatively poor Labour-voting areas in suburban Sydney voted “no,” as did relatively poor Liberal-voting areas in the western exurban fringe. Wealthy electoral divisions, in contrast, either supported the measure or narrowly turned it down, with Mitchell again forming an exception. It is not coincidental that the two wealthiest divisions in Sydney form the epicenter of the “teal rebellion” against the Liberal Party.

Religion was another factor in the Indigenous Voice Referendum. As can be seen, the yes-voting areas of the center and northeast all reported high levels of “no religion” in the 2021 Australian census. Strikingly, the most irreligious divisions, Sydney and Grayndler, had the highest percentage of “yes” votes. In contrast, the divisions with the lowest percentages of non-religious residents all supported Labour in 2022 and all voted “no” in 2023.

The religiosity map loosely correlates with the map of people born in Australia. The divisions with the highest percentages of foreign-born residents are concentrated in central Sydney and in the relatively poor western suburbs. Those located in the west rejected the referendum, some by substantial margins. Many of these areas have large Muslim populations. Peripheral districts with high percentages of Australian-born residents also voted “no,” whereas those in the wealthier and more highly educated center and northeast voted “yes.”


The final map, showing marital status, is not as revealing. It does show, however, that divisions with low percentages of married people tended to vote “yes,” although one with a high marriage rate, Bradfield, voted “yes.” (Bradfield was Australia’s only Liberal-voting division in the 2022 election that supported the referendum). Intriguingly, the wealthiest divisions of Sydney have some of the region’s highest and lowest rates of marriage. It is perhaps not coincidental that Mitchell, the richest “no-voting” division, also has the region’s highest marriage rate.

These various patterns and correlations can tell us a lot about both Sydney’s electoral geography and the (changing?) ideological positions of Australia’s major parties. Before delving into these important issues, however, we will examine greater Melbourne through the same comparative-cartographic lens employed in today’s post.