As the September referendum approaches, the rival campaigns have been selecting apparently incompatible statistics to back their competing claims for Scotland's future.

As a consequence, many people discount all statistical analysis as unreliable. There are, however, two aspects of the future that most commentators appears to agree on: the number of elderly people in Scotland is increasing and North Sea oil production is declining.

In the context of this social and economic change, a key question is whether the decline of Scottish oil will result in a lowering of the number of the tax-paying working-age population. The future of this relationship between population structure and government income and expenditure is arguably more important for the future of fiscal sustainability in Scotland than the loss of tax revenues from oil production itself. That is why it is important to understand the influence of the predicted economic change on demographics and how this will affect fiscal sustainability.

The response of the Yes campaign to any suggestion of a problem is to say an independent Scotland will have the power to make policies to create the new jobs for an increased number of workers, including in-migrants who will benefit from more liberal policies on immigration than in the rest of the UK.

If future Scottish Governments could improve Scotland's economic performance, increase the size of the working age population or increase labour market participation, tax receipts would be boosted and welfare costs reduced. The No campaign dismiss this as the politics of wishful thinking.

It is agreed the changing population size and structure through variations in the birth, death, fertility and net migration rates will have an impact on government revenues and expenditures. These changes will have an impact on the volume and structure of the tax take, as well as the spending on education, health, pensions and other services. The key drivers of changes in the medium- and long-term fiscal balance are changes in the demographic composition of the population. The conclusions of the Scottish and British Governments therefore depend on the demographic assumptions they have made.

The economists in Edinburgh and London have ignored a fundamental difference between projections and forecasts. Demographers make projections about population change based on past demographic trends. Economic forecasts can take account of anticipated or merely vaguely possible future change. Demographers do not take into account possible impacts of future policies or uncertain social and economic changes, which are important in the case of Scottish oil.

Although almost everyone agrees production of North Sea oil will decline, no one has accommodated this in their population forecasts, and hence in their fiscal projections for Scotland in 20 or 50 years' time. It is assumed by all rival economists that the future trends in the population of Scotland will continue as in the past, despite the positive employment impact of the oil industry in the recent past and the potential for job losses in the future.

The Office of National Statistics (ONS) has developed three different migration scenarios for Scotland. The Scottish Government chose the most optimistic scenario to support its future fiscal forecast, but none of the scenarios consider the possibility of employment decline and a return to the out-migration of the young workers whose taxes are needed to support Scotland's ageing population.

As North Sea oil activities decline, net migration trends for Scotland in the next decade are unlikely to be the same as the last decade, irrespective of whether the government is located in Edinburgh or London. All other things being equal, a decline of in-migration is inevitable and a return to net out-migration is the most likely scenario. It is surprising that the fiscal implications of a net out-migration scenario have not been examined by anyone.