The five-yearly census is an enumeration of population and housing. It provides a range of details regarding the number and structure of dwellings and it is possible to classify these according to Aboriginal or non-Aboriginal occupancy and other housing-related variables. Table 6.1 shows the number and type of Aboriginal and non-Aboriginal dwellings in the Northern East Kimberley with the former classified as such if one or more adults in a dwelling are Aboriginal.

In 2001, a total of 2,065 dwelling units were recorded in the Northern East Kimberley.¹ Of these 717 (35%) were Aboriginal dwellings, most of which were separate houses, although 13% were improvised dwellings and caravans. Thus, the proportion of improvised Aboriginal dwellings had fallen only slightly (from 17%) since 1981. While most non-Aboriginal dwellings were also separate houses, the key point of distinction is the much higher average occupancy rate in Aboriginal dwellings—more than twice that recorded for non-Aboriginal dwellings (5.0 compared to 2.3), although it should be noted that, away from the towns of Kununurra, Wyndham and Halls Creek, the average Aboriginal occupancy rate increases to 7.0. As a benchmark, it is interesting to compare this rate with the average of 3.7 persons per Aboriginal dwelling recorded for Western Australia as a whole in 2001. Once again, however, these rates are based on actual census counts. Using the 2001 ERP as the base produces a revised overall Aboriginal occupancy rate of 5.7, and a figure of 8.0 for non-urban communities.

Thus, in the 20 years since mining commenced at Argyle, the overall regional occupancy rate for Aboriginal dwellings has been reduced, but only slightly from 7.0 to 5.7 persons per dwelling, and it remains substantially above the State average. Aboriginal people in the region are also now marginally more likely to be accommodated in conventional housing, but the indication is that provision of such housing has only just kept ahead of increased demand due to new household formation, with only limited impact on the aggregate level of overcrowding.

While continuing high Aboriginal occupancy rates partly reflect larger Aboriginal household size and a cultural preference for extended family living arrangements, it is also a measure of the inadequacy of housing stock available to accommodate the regional population. To acquire a better sense of the adequacy of housing, occupancy rates must be set against dwelling size and one measure of this is provided by the ratio of available bedrooms to the population in dwellings. Overall, in the Northern East Kimberley, the census recorded a total of 5,208 bedrooms in 2001. Of these, 1,836 (35%) were in Aboriginal dwellings (645 in communities and 1,191 in towns). Using the number of persons per dwelling inflated to match the ERP, this produces an average figure of 2.2 persons per bedroom in Aboriginal dwellings, with no difference apparent between town and country. The equivalent figure for non-Aboriginal dwellings is less than 1.0.

More refined measures also include an indication of housing affordability as well as functionality from an environmental health perspective. Applying basic overcrowding measures, Jones (1994) identified the Wunan ATSIC Region as ranked 5th highest out of 36 ATSIC Regions across the country in 1991 in terms of the size of its unmet housing need calculated on the basis of additional bedrooms required to meet an accepted occupancy rate. At that time, almost 60% of families were found to be in overcrowded dwellings (Jones 1994: 54) and the total extra bedroom need to overcome this was estimated at around 1,000. By 1996, the number of overcrowded households in the Wunan Regional Council area was still high at 45% and the region ranked 8th highest in the country (Jones 1999: 24), with a total extra bedroom need of 756 (Jones 1999: 38). In both years, the East Kimberley region stood out as having the greatest Aboriginal housing need of all Western Australian regions. While similar calculations have not been made using 2001 Census data owing to their complexity, the simple fact that the overall Aboriginal occupancy rate in 2001 is only slightly lower than recorded in 1981 suggests that overcrowding levels remain little altered.

This is supported by data from the 2001 CHINS, and a Wunan ATSIC Regional Council survey of community housing conducted in March 2003. The combined findings from these sources are presented in Table 6.2 which shows the number of dwellings, the estimated maximum population, and derived occupancy rates for each community and outstation in the Northern East Kimberley. The overall occupancy rate was 7.1 persons per Aboriginal dwelling, which is very close to that recorded by the 2001 Census for non-urban communities in the study region. It also accords exactly with the findings of a housing survey conducted in Warmun community in March 1998. However, this masks considerable diversity of circumstance, with occupancy rates ranging from 28 persons per dwelling in some instances to just one in others. Of particular note, however, is the fact that the larger communities of Warmun and Kalumburu have occupancy rates above the regional average, as do a number of town-based communities such as Mirima in Kununurra, and Mardiwa Loop in Halls Creek.

Of course, these data reveal nothing of the quality of housing stock. The 1997 WA Environmental Health Needs Survey (EHNS) provided some measure of this allowing for a more refined (and meaningful) measure of occupancy based on persons per functional dwelling (defined against minimum environmental health criteria). This re-calibration produced some excessively high occupancy rates in Northern East Kimberley communities—Mardiwah Loop (14.0), Koongie Park (12.5), Warmun (11.3), Kalumburu (11.1), Oombulgurri (10.4), Flying Foxhole (35), Cockatoo Springs (13.5), Hollow Springs (27), and Nulleywah (40). Unfortunately, the lack of a follow-up survey by the time of writing prevents any assessment of how these rates may have changed in the interim.

More recent measures of the quality of housing stock are provided by the 2001 CHINS that included an assessment of the condition of dwellings owned or managed by Indigenous Housing Organisations. Such dwellings are categorized according to the extent of repairs needed in the following way:

The 2001 CHINS identified a total of 368 permanent dwellings managed by an Indigenous Housing Organisation in the study region. Of these, 85 (23%) were declared in need of replacement, 58 (16%) were in need of major repair, and all of the remainder were found to be in need of minor repair. These figures for major repair and replacement are in line with estimates made by the Wunan ATSIC Regional Council community housing survey of March 2003 which identified an immediate requirement for 23 new houses, and 125 to be upgraded, although it should be noted that these needs were also seen as contributing to an increase in housing stock to assist in relieving overcrowding. In effect, the indications are that a major component of the Aboriginal-owned and managed housing stock in the region is in need of significant upgrading, while additional dwellings are also required to reduce overcrowding. According to the ATSIC survey, the greatest concentrations of need for extra housing, or housing upgrades are at Warmun (2 new houses required and 30 to be upgraded), Kalumburu (5 new houses required and 24 to be upgraded), and Oombulgurri (1 new house and 39 to be upgraded). The other new housing need identified is scattered across the region in smaller communities such as Woolah, Bell Springs, Emu Creek and Goose Hill.