Residents of the Thamarrurr region have a number of potential sources of cash income. These range from wage labour in CDEP or in other more mainstream forms of work, unemployment benefit and other benefit payments from Centrelink, agreed payments to traditional land owners, and private income from the sale of art and craft works. Set against these, of course, are routine deductions from income, such as those for house rent and power charges, much of which is now debited at source via Bpay.

Accurate data on income levels, and employment and non-employment sources of income, are notoriously difficult to obtain due to a variety of conceptual problems. For one thing, most available data on income refer to period of time, such as annual or weekly income, whereas the flow of income to individuals and households within the region is often intermittent. Census data, for example, are collected for all sources of income in respect of a ‘usual week’ and then rounded up to annual income. What might constitute ‘usual weekly’ income in many Aboriginal households is difficult to determine. On the credit side, there is the likelihood of intermittent employment and windfall gains from sources such as gambling, cash loans, and agreed payments. These sources of income combine with debits, for example due to loss of employment and sometimes welfare payments, to create a highly complex picture even over a short space of time, and one that standard methods of data gathering are likely to misrepresent.

Even if adequate questions were asked regarding income, high levels of population mobility would make it difficult to establish a consistent set of income recipients over a period of time. This is further complicated by job mobility with individuals often employed on a casual or part-time basis and moving into and out of longer-term jobs. As for the circulation of cash between individuals and households, information on this is non-existent. Also lacking are data on expenditure, although a common pattern reported from similar communities is one of cash feast and famine against a background of high costs for essentials such as food and transport (Beck 1985: 89; Taylor & Westbury 2000).

The most comprehensive source of personal income data for the region based on a consistent methodology is available from the census. It should be noted, however, that census data report income in categories, with the highest category left open-ended. Consequently, actual incomes have to be derived. In estimating total and mean incomes, the mid-point for each income category is used on the assumption that individuals are evenly distributed around this mid-point. The open-ended highest category is problematic, but it is arbitrarily assumed that the average income received by individuals in this category was one-and-a-half times the lower limit of the category.

Also, the gross income reported in the census is intended to include family allowances, pensions, unemployment benefits, student allowances, maintenance, superannuation, wages, salary, dividends, rents received, interest received, business or farm income, and worker’s compensation received. Whether all such sources are reported in Thamarrurr, or elsewhere for that matter, is unknown. One distinct advantage of census data, however, is that it provides a means by which one estimate of dependence on income from welfare can be derived. This is done by cross-tabulating data on income with labour force status as a basis for distinguishing employment income from non-employment income, the latter being considered a proxy measure of welfare dependence.