Family size and relative need

This thesis examines three questions concerned with the relative income needs of familiesof different sizes - often summarised by indices known as equivalence scales. The first is the extent to which researchers and policy makers should offset the costs offamily composition (eg the expenditure costs of children) with the benefits associated withdemographic choice (eg the joys of parenthood). Chapter 2 concludes that there aredemographic and financial market constraints that will often make a narrow focus onexpenditure costs appropriate for distributional research and tax/transfer policies. However,this will not always be the case. One implication of this result is that it may be reasonablefor distributional research to use different equivalence scales for adults and children in thesame household.Part 2 of the thesis introduces a new method for the estimation of the within-householdincome distribution and expenditure costs of different family types. This is based upon thehousehold welfare model of Samuelson together with Laus method for modelling the jointconsumption of household goods. In Chapter 4, this method is applied to the estimation ofequivalence scales for older singles and married couples. The estimation is based upon adetailed set of assumptions about the extent of joint consumption for 17 differentcommodity groups. The main conclusions are that: the theoretical model fits the observedbehaviour well (with the exception of some home production effects); that aged couplesshare their income relatively evenly; and that the relative rate of pension for aged singlesin Australia is probably too low.In Part 3, the thesis examines how changes in poverty can be estimated when there isuncertainty about the equivalence scale. The thesis proposes a new method which permitsa set of upper and lower bounds for the equivalence scale to be assumed, which in turndetermine upper and lower bounds for the increase in poverty. This method is applied tomeasure the change in poverty in Australia during the 1980s. Equivalence scales can befound that imply either an increase or a decrease in poverty