Because our Quarterly National Household Survey of employment was initiated only six years ago, it has only recently become possible to calculate seasonal fluctuations in employment, writes Garret FitzGerald.
However, we now know that just under 40,000 additional people work here during the summer tourist season, and about 15,000 less than normal are employed in the earlier months of the year. In other words there is an annual seasonal variation in employment of about 55,000, or 3 per cent.
By using these new seasonally-adjusted data on employment we can get a picture of the underlying quarterly trend of employment. It is encouraging to find that, after a brief setback last summer, since then there has actually been an increase each quarter in the underlying level of employment, adding up to a cumulative rise of no less than 36,000, or more than 2 per cent.
And during this nine-month period seasonally-adjusted unemployment, as defined for the purpose of this survey, has remained virtually unchanged at 85,000, or 4.6 per cent.
However this recent employment increase was exactly offset by a similar reduction in hours worked, with the result that the total number of hours worked remained unchanged over this period. And as output, measured in terms of GNP, was somewhat higher in the second quarter of this year than it had been nine months earlier, there appears to have been a rise of about 2 per cent in national productivity - output per hour worked - over this period. That is not at all a bad outcome for our economy during what may prove to have been the closing stages of a European recession.
It has to be said that there can be a good deal of public confusion about employment and unemployment figures. Data from different official sources for April last year yielded employment figures, unadjusted for seasonality, ranging between 1,642,000 and 1,750,000, and unemployment figures for that month in 2002 ranged between 77,000 and 159,000 - not to speak of a Live Register figure of 171,000.
Why are there such wide variations between these figures? Basically because all these figures represent answers to quite different questions, asked by different methods - some, like the Census data, involving replies to a written questionnaire, whilst others are replies to questions put by an interviewer.
As to the Live Register figures, the CSO constantly points out that these are not designed to measure unemployment, for they include people working part-time up to three days a week, as well as seasonal and casual workers who may be entitled to Unemployment Assistance.
The Census in April 2002 of all households in the country required members of the labour force to fill out for themselves a form that included questions as to whether they were working for payment or profit, or were unemployed, or looking for their first job. The Census total for the labour force was 1,801,000. And 1,642,000 of these said they were working for payment or profit, whilst 159,000 said they were unemployed or looking for their first job.
The Quarterly Household Survey, referred to earlier, showed that in the March-May period of last year there were a total of 1,796,000 people in the labour force - a figure very close indeed to the Census figure of 1,801,000. However, out of that total the number admitting to have been at work was 40,000 more, and the number claiming to be unemployed was 40,000 less, than was recorded in the Census carried out in the middle of this period.
Why this difference? I suspect because, whilst this Sample Survey was carried out by interviews in each household, the Census involved people simply filling up a form themselves. And a certain number of people gave a more honest - or perhaps one should say a more complete - answer when faced with an interviewer! Even more interesting, however, are the replies to a different and much more detailed second question also asked by the interviewers in that same March-May 2002 Household Survey - a question designed by the International Labour Office in Geneva with a view to eliciting precisely comparable data on each country's labour force.
This question asked whether in the immediately preceding week the respondent worked for one hour or more for payment or profit - or, if the respondent was ill or on holidays, whether, but for that, he or she would have been working on the basis of this definition?
For the purpose of this question the definition of "unemployed" is that the person concerned has been without work but has been available for work, and has taken specific steps during the preceding four weeks to find work.
This more precise question yielded 30,000 more people in the labour force than had emerged from the earlier "leave it yourself" question in the very same survey. And within this larger total, about 40,000 less people emerged as unemployed, (77,000 as against 115,000), whilst an additional 70,000 more people emerged as having been at work!
This ILO definition of employment, namely a minimum of one hour's work in a week, may seem a very low standard indeed by which to judge whether people were employed or not, but this question is designed to cover part-time as well as full-time work, and the answers to it, as published, distinguish between those who worked for 1-9 hours, 10-19, and so on up to 40-44 hours, and 45 hours and over.
Thus we know from the grossing up of the data in respect of March to May last year that 27,000 people worked for 9 hours or less during this period, 88,000 for 10 to 19 hours, and so on up to a figure of almost 240,000 people who worked for 45 hours or more. There is also a residual figure that includes those whose hours are variable, as well as some who did not reply to this question. The average hours worked in that period by all full-time and part-time workers was just under 38 hours per week - 42 hours in the case of men and 32.5 in the case of women.
Clearly this more precise question provides a more reliable figure of the actual scale and intensity of employment, including part-time and short-time working, as well as of unemployment, than do questions, (whether posed by an interviewer or given in response to a written questionnaire), which leave it to the respondents themselves to decide whether they are part of the labour force, and whether, if so, they are employed or unemployed.
It is these ILO-defined figures that provide the basis for the employment increase data cited at the outset of this article.