Economics: It's not often that long-term forecasts for the Irish economy are published, so when it happens it's worth paying some attention. Last week's NCB report on prospects for the Irish economy out to 2020, therefore, excited especially strong media interest, and is worth revisiting now that the dust raised by its launch has settled.

The NCB report sees GDP growing at an average rate of 5 per cent a year between now and 2020, implying another doubling of the size of the economy by then. The forecast is predicated on annual average growth rates of about 2 per cent in the labour force and 3 per cent in labour productivity.

This is a much more bullish outlook than that contained in the ESRI's last Medium-Term Review, published in December. In that review, the ESRI articulated two alternative trajectories for the economy, the one to which it ascribed the higher probability involving average annual growth rates of 1.3 per cent in the labour force, 2 per cent in productivity and just 3.3 per cent in GNP between 2005 and 2020.

The NCB economists defend their projection of 3 per cent annual labour productivity growth on the grounds that this is in line with the average recorded between 2000 and 2004, and is warranted by rising levels of educational attainment in the Irish labour force.

However, such productivity growth is about twice the rate that has become the norm in advanced western economies.

While abnormally large productivity gains were understandable in the context of the structural changes that occurred in the Irish economy in the 1990s, to expect much more of the same now that we have substantially completed the transition to a post-industrial services-oriented economy seems excessively optimistic.

In this regard it is striking that, while the average rate of productivity growth in the 2000-2005 period has been impressive, there has been a steep year-by-year deceleration as economic activity has shifted decisively away from manufacturing.

In 2005 it seems that GDP per person employed may not have risen at all.

NCB's labour force forecasts are based on their population projections, which have the population reaching five million in 10 years' time and exceeding 5.3 million by 2020 - an annual average increase of 1.7 per cent over the next 15 years. On the face of it, this is a startling prospect. However, population growth sustained at this rate is not implausible in itself.

Between 1985 and 2000, no fewer than nine states in the US (including California and Texas) recorded population growth at least as fast as this, and in some of these states (notably Arizona and Florida) population actually grew at a much faster rate.

The real question is: what's fuelling this growth? The key elements in any population projection are the birth rate, the death rate and the rate of migration, and one could argue with the NCB assessment in each case. However, the most critical element in the NCB forecast relates to migration patterns.

Here, it is projected that the inflow of new immigrants will remain at 70,000 per annum until 2010 (the same rate as in the 12 months to last April), before subsiding to 43,000 per annum by 2015, and that the outflow of emigrants will stabilise in the range 16,000-20,000 over the forecast period.

As a result, Ireland is seen gaining substantially in population terms on foot of net migration over the next 15 years and beyond, with immigrants comprising 20 per cent of the total population by 2020.

The problem with this kind of forecast is that there is precious little to guide the forecaster. Ireland's experience of immigration is too recent for there to be any sort of meaningful historical precedent, and our knowledge of the habits and intentions of our immigrants is too sparse.

As the NCB report itself states: "Assumptions about future migration levels and patterns are substantially conjectural . . . particularly so in a period of transition as is currently the case."

The basis for the NCB immigration projection is particularly problematical. It seems to lean heavily on the notion that Ireland will continue to draw a disproportionate share of migrants from the new EU accession states because all the other older member states (all, that is, except Sweden and the UK) that are currently restricting such inflows will continue to do so until 2011. But already Finland, Portugal and Spain have announced that they will lift such restrictions from next month, and a number of others have indicated that they are considering following suit.

There are other aspects of the NCB migration arithmetic that are highly debatable. For example, there appears to be an implicit assumption that the overall rate of emigration from the accession states will continue at something like its pace of the last couple of years until well into the next decade.

Perhaps it will, but there are strong counter-arguments. One is that the recent outflow may have owed a good deal to pent-up supply. Another is that, as the gap between living standards in the new accession states and the old member states narrows, the incentive to emigrate will diminish.

The NCB migration arithmetic is also based on the implicit assumption that most, if not all immigrants to Ireland are here to stay. I infer this from the fact that, even in the face of large projected inflows of migrants over the forecast period, the projected rate of outflow from Ireland is held broadly stable.

Again, this premise may prove to be correct, but our experience of immigration is much too recent to embrace such an assumption with confidence. It could be that the majority of recent and prospective immigrants to Ireland will stay only for a few years, in which case the large inflows may be followed, with a lag, by substantial outflows. We will just have to wait and see what pattern emerges.

Overall, the NCB migration projections/assumptions seem to me to be very ambitious. The same judgment applies, by extension, to their population and labour force projections.

Taken together with their optimistic assessment of prospects for productivity, this means that their projections of economic activity are, to my mind, pitched at the edge of what is possible.

The risk (probably small) that attaches to such forecasts is that they excite unrealistic expectations and unsustainable behaviours.

The great merit of such an exercise, on the other hand, is that it helps to focus light and attention on the future.

Jim O'Leary lectures in economics at NUI-Maynooth. He can be contacted at jim.oleary@ nuim.ie