Models based on representative samples of the population are now widely used in the analysis of tax and social security systems, and they have played a central role in discussions of possible reforms. Although some of these models now incorporate behavioural responses in terms of changes in labour supply or other decisions, an important role is played by models which are purely arithmetical. It was these models, and particularly the model TAXMOD (now called POLIMOD) developed by Holly Sutherland and myself ( Atkinson and Sutherland 1988) as part of the ESRC Research Programme at the LSE, that the present chapter is concerned.
The TAXMOD model is in many respects like those employed in ministries of finance and treasury departments around the world. It makes use of data on the circumstances of a representative sample of individual families (obtained in our case from the regular household budget survey) to calculate the impact on net incomes of the current tax/benefit system, and compares it with that of the policy change. Models of this type have been widely used in government to examine the effect of changes in tax rates and allowances, or the impact of changing social security benefits, or of reforms that affect both taxation and social security.
What is different about our model is the emphasis we have placed on accessibility. A major aim of our research has been to bring this kind of model within the reach of those outside government. It is not locked in a security-conscious government department, nor is it the exclusive property of one particular research institute. The model is available to anyone who requests it, at marginal cost (in 1989 this was £75), and, just as important, it is written with the user in mind. The user is not required to have any computing expertise, only to know how