The June Live Register figures have been released today. While no doubt the journalists will mainly focus on the new peak in the numbers “signing on”, I think it is worth drawing attention to the rate at which the increase in the seasonally adjusted numbers (both males and females) is falling.
.
Author: Brendan Walsh
The number of births recorded in Ireland reached a twentieth century peak of 74.0 thousand in 1980 and fell to an all-time low of 48.2 thousand in 1994.
Births began to increase in the second half of the 1990s and the rate of increase has been rapid over the past five years. The latest figures released by the CSO (Vital Statistics Third Quarter 2008) raise the possibility that the 1980 peak may soon be surpassed, although of course this would still imply a significantly lower birth rate as the population is now some 30 per cent above its 1980 level. The TPFR (total period fertility rate – roughly the number of births per woman over her child-bearing life span) declined from 3.23 in 1980 to a low of 1.85 in 1994 but was back up to 2.03 in 2007.
These huge swings in the birth rate have serious implications for resource allocation, especially in the health and education sectors. The Dublin maternity hospitals are now bursting at the seams, while the educational system will continue to feel the impact of growing numbers of school-age children for years to come.
Will the number of births continue to grow? Population projections are notoriously uncertain, at best serving to illustrate the implications of alternative assumptions about key demographic variables. The projections for 2011-2041 published by the CSO in April 2008 assumed that the TPFR would decline to 1.78 by 2011, which seems a high given the continuing buoyancy in the birth rate.
But there are reasons for believing that the birth rate could suddenly take a nosedive, as happened in the 1980s. The peak in the number of births recorded in 1980 was quite dramatic and the subsequent rate of decline steep. (The decline began shortly after Pope John Paul II’s visit to Ireland in September 1979, when he preached against contraception and abortion.) More significant is that it coincided closely with the sharp rise in unemployment as Ireland entered the deep recession of the 1980s. The following graph shows the correlation between the number of births (four quarter moving average) and the seasonally adjusted numbers on the Live Register, lagged three quarters. The coincidence between the peak in the number of births and the floor in the unemployment rate is striking.
The second graph is the comparable picture for the years 2005 to date (using the standardised unemployment rate in place of Live Register figures).
The unemployment rate began to increase dramatically in the second half of 2008, so it should begin to affect births registered from the second quarter of 2009 onwards. The latest available data are for 2008Q3, but it is interesting to note that the graph shows the number of births plateauing over the period 2007Q4-2008Q3. If the experience of the early 1980s is anything to go by, this graph will nosedive over the coming years. Watch this space!
As readers will be well aware, the Irish unemployment has soared since the end of 2007. Most of the short-run commentary focuses on the monthly Live Register (LR) figures, which we know contain many landmines of interpretation. The Quarterly National Household Survey (QNHS) data are based on more economically meaningful (ILO) definitions, but these too need to be handled with care. (For example, anyone working for pay or profit for one hour a week or more is classified as employed.)
The survey data allow us to look at the employment rate – that is, the proportion of the adult population in employment – and this is probably more meaningful as a current economic indicator than the unemployment rate. (The employment rate is the product of the labour force participation rate and (one minus) the unemployment rate.)
A look back at the employment rate over the past twelve years is interesting. The male employment rate has fallen by five percentage points – from 70.5% to 65.5% – since the third quarter of 2007. (N.B. These figures are not seasonally adjusted, but I do show the four-quarter moving average.) 
This brings it back to where it was in the late 1990s. The female employment rate dropped by only two percentage points – from 52.7% to 50.7% – over the same period. 
This leaves it where it was in 2006. The overall rate fell by three and a half percentage points, from 61.5% to 58.0%, so it is back to here it was in 2004.
Female participation held up well in 2008, but male unemployment has risen, and participation fallen, faster.
The continuing relatively high participation rates is one hopeful sign in an otherwise gloomy landscape. The forthcoming QNHS for the first quarter of 2009 will probably show further rises in unemployment and falls in participation, but perhaps later this year there will be signs of stabilisation.
The concept of “global warming” or, more vaguely “climate change”, is now deeply embedded in the public’s consciousness and indeed in the economic agenda of most developed countries. Expensive policy responses are being put in place to avert possible future environmental damage. However, as has been pointed out by Richard Tol, “the impact of climate change on Ireland is [likely to be] moderate”, although he argues that as a constructive contribution to the global problem we should introduce a carbon tax.
This Post is prompted by my puzzlement at the contrast between predictions that our climate is heating up and the prospect of the third dismal summer in a row, coming on the heels of a severe winter and “broken” spring weather.
Here is a summary of the evidence for warming in Ireland provided by a group studying climate change at NUIM:
The Mean annual temperatures in Ireland have risen by 0.74°C over the past 100 years (McElwain and Sweeney, 2007). This increase largely occurred in two periods, from 1910 to the 1940s and from the 1980s onwards, with a rate of warming since 1980 of 0.42°C per decade. In Ireland, 6 of the 10 warmest years have occurred since 1995 with the warmest year within this period being 1997.
From this quotation it is clear that the trend in average temperature in Ireland was quite erratic over the twentieth century. Issues such as trend breaks, autocorrelation, and statistical significance, need to be addressed, as does the relevance of jumps in temperature early in the twentieth century for trends in the twenty-first century.
The Environmental Protection Agency’s press release (27th April 2009) based on this report states
The projections show that average temperatures will rise by 1.4°C to 1.8°C by 2050, and be in excess of 2°C relative to the 1961-1990 baseline by the end of the century.
The basis for this fairly precise projection is unclear. Of course, a lot of very sophisticated climate change modelling is underway in Ireland and world-wide. To give a flavour of what is involved, consider the following summary of the modelling underway at Met Éireann:
Technically, it dynamically downscales the relatively coarse-grained information produced by global models to tease out the finer details over a smaller area. This approach – regional climate dynamic modelling – is unique in Ireland. The work is done in collaboration with the Meteorology and Climate Center at UCD, and more recently, with the Irish Center for High-End Computing (ICHEC).
At a simpler level, a visit to the CSO Database under Environment, Climate opens up fifty one years of monthly data on temperature, rainfall, sunshine, and wind speeds for 15 weather stations across the Republic. This wealth of easily-accessible data could help take your mind off the banking crisis during the wet summer months. More seriously, I thought it was worth mining these series to see if the effects of climate change can be discerned in this record of the last half century of weather.
I selected for analysis the mean monthly air temperature at Dublin Airport over the period January 1958 through April 2009 in the belief that this is a meaningful indicator of the climate affecting the largest concentration of population in Ireland. This simple approach raised some interesting issues.
This first Chart shows the twelve-month moving average of Dublin’s temperature over the period January 1958 to April 2009.
Even when seasonal effects are removed, the series is very erratic. Periods of cooling have been abruptly followed by warming periods. For example, temperatures fell sharply between 1983 and 1986, but then there was a period of noticeable warming from mid-1986 to late-1989. 1986 was the second coldest calendar year in the 51-year period, but 1989 was the warmest. It is especially striking that the series is very erratic and trendless over the last twenty years. The average temperature (9.4⁰ C) for the most recent twelve-month period, May 2008 to April 2009, was below the mean (9.8⁰ C) for the whole fifty-one year period and almost the same as that for the year 1958 (9.5⁰ C).
The observations centered on 1963 reflect the exceptional winter of 1962-63. According to one account
The winter of 1962/1963 was savage, the coldest for more than 200 years outstripping even ‘white 1947’ for bitter temperatures . . . It began freezing on Christmas Day in 1962 and barely relented until March. By early January 1963 much of Britain and the eastern part of Ireland were blanketed in snow.
This exceptionally cold period is still included in the baseline (19861-1990) for average temperatures used on the Met Éireann site.
However, selective illustrations do not prove anything. The Chart shows the trend line through seasonally-adjusted data. The trend is positive and significant, but not impressively so. Moreover, it is not stable, as is shown in the following Chart for the 25-year period 1983-2008. Over this period the positive trend is not significant.
To test the stability of the trend more formally, I fitted a linear trend and monthly dummies to the temperature data for the whole period and two sub-periods. The following results were obtained:
|
Period |
Trend (t-ratio) ((upper – lower bound)) |
Extrapolated change in average temperature over a century o C |
|
Full sample: 1958-2008 |
.0009 (3.5) ((.0004 – .0014)) |
0.48 – 1.68 |
|
First half: 1958-1982 |
.0009 (1.3) ((-.0005 – .0023)) |
-0.6 – 2.76 |
|
Second half: 1983-2008 |
.0004 (0.6) ((-.0009 – .0017)) |
-1.08 – 2.04
|
These results illustrate uncertainties about global warming in the Dublin area. While all the estimated trends are positive, they are at best weakly significant and provide a wide range of estimates of the pace of warming. The most recent data provide the least support for the warming hypothesis.
Changes in summer and winter weather are probably more economically significant than changes during the transitional seasons of autumn and spring, so it is worth looking at the evidence by season.
Dublin readers may take little convincing that our summers have not been getting much warmer. The graph shows that the city has not experienced a really warm summer since 1983. The time series show that while over the 51-year period 1958-2008 there has been a small positive trend, it is not statistically significant (R2=.068, P=.0648). If we confine our attention to the second half of the period – 1983-2008 – the trend is negative (but not statistically significant).
Climate change is often said to be about extremes, rather than averages. The CSO site gives a time series on the maximum temperature recorded each month. This series behaves like that for average summer temperatures, exhibiting a weak upward trend over the whole period and a weak downward trend for the past 25 years. (The highest temperature recorded at Dublin Airport over the last half century was 28.7o C in August 1990, whilst the highest recorded in 2008 was only 22.3o C.)
A common perception is that our winters have been getting milder, but it is possible that we are still influenced by memories – or accounts – of the exceptionally severe winters of 1947-8 and 1962-3. The statistical evidence for milder Dublin winters is weak. While over the 51-year period there is an upward trend, but its statistical significance is low (R2 = 0.0462, P=0.14). Over the second half of the period, the trend is negative, but the R2 is a non-significant 0.0001.
Well, perhaps we are enjoying warmer /mellower autumns? Not significantly. With R2s of 0.0096 and 0.03, the trends in autumn temperatures over the 51-year period and the second 25-year period are not significantly different from zero.
Warmer /earlier springs, perhaps? Maybe. The trend coefficient for the full 51-year period is almost significant at the 0.05 level, implying an increase of about 1.2⁰ C a century in Dublin’s average spring temperature. However, the trend over the second 25-year period, 1983-2008, is not statistically significant (the R2 falls to 0.0445).
Overall, then, the data for Dublin’s temperatures suggest some weak evidence of a slow upward trend over the 51-year period 1958-2008, but none over the 25-year period 1983-2008. There has been no warming in summer or winter, but perhaps during the transitional seasons.
Of course, temperature is only one dimension of climate and change may be occurring on other dimensions. The most frequently mentioned possibility is that the warmer Atlantic Ocean will lead to stormier and wetter weather across Ireland. However, the evidence for Dublin does not support this view – although very variable, Dublin’s rainfall shows no trend over the last half century.
As mentioned above, extremes are important in the climate change literature. However, extremes – especially of rainfall and wind speed – can be very local. The heaviest downpours tend to be produced by thunder storms confined to small areas. The highest rainfall recorded over a 24-hour period (184 mm) in Ireland was measured during a thunderstorm in Mount Merrion, County Dublin, on 11th June 1963. (For this and other nuggets see The Climate of Ireland by P. K. Rohan, The Stationery Office, Dublin, 1975).
The series on the CSO website for “Most Rain in a Day” provides rainfall extremes for the 15 weather stations. Over the the 1958-2008 period the evidence for Dublin is of considerable variability but no trend. The wettest day in the whole period at Dublin Airport was in June 1993, when 82.3 mm were recorded. But wettest day in the following year, 1994, had only 21.8 mm of rainfall. (The only month over the half century in which no rain was recorded was April 2005).
Of course, the Dublin area could be affected by warming occurring elsewhere in the world, most importantly by the impact of the widely-predicted rise in sea levels, which would presumably increase the incidence of coastal flooding. Here is Met Éireann’s take on this topic:
Estimates of sea level rise from satellite observations around Ireland are consistent with the global picture: increases of 2.3 to 4.7 mm/year since 1993. At current rates of change, mean sea levels in Dublin, Sligo Bay and Slea Head will be 25, 44 and 40 cm respectively, above present day levels by the end of the century.
In light of what seems like an estimated 45 mm rise in sea levels over the last 15 years, it is strange that there does not seem to have been much reporting of increased coastal flooding over recent years.
This note is about the Dublin area. There is considerable variation in climate – and possibly in climate change – within relatively small geographical areas. There certainly is a lot of variation in the weather across Ireland – for example, summer 2007 was quite good in the West but terrible in Dublin. A sampling of data from other Irish weather stations provides stronger evidence for warming – with much higher correlations and more consistently significant positive trends. However, even in the south and west the estimates of trends are quite unstable. For example, the data for Shannon Airport show a strong positive trend in temperature over the 51-year period 1958-2008, but no significant trend for the most recent 20-year period and a negative (and significant) trend over the past 10-year period. There is not much support for the view that things have been getting stormier – in Belmullet there has been no trend in maximum wind gusts since 1958 and no trend in rainfall over the past twenty years.
To conclude: The implications of climate change – and of our reaction to the fear of climate change – are too important to be ignored by economists. In the face of uncertainty there is a strong argument for erring on the side of caution, so that even weak evidence of warming might be justification for strong policy responses. But reliable estimates of past trends are nonetheless essential for the projections of future trends on which these responses should be based.
The title of Colm McCarthy’s Post Goodbye to All That is Robert Graves’ account of his experiences in the Great War. This evoked in me some nostalgie de la boue so I had a look at a graph of Ireland’s unemployment rate.
Looking at the early years in the graph, it seemed that the surge in unemployment in the first half of the 1980s got built into the structure of the economy. Indeed, an influential comparative study of unemployment in OECD countries estimated that the Irish equilibrium or “natural” unemployment rate had risen from 9 per cent over the period 1969-79 to 13.1 per cent between 1980 and 1988 (Layard, Nickell, and Jackman, Unemployment: Macroeconomic Performance and the Labour Market, Oxford University Press. 1991). As if to vindicate this claim, after a faltering improvement in the early 1990s, the unemployment rate was headed for 16% in 1994, some five years after growth had resumed.
There was, of course, a lot of debate about how genuine the unemployment – especially the Live Register (LR) – figures were. In 1996 the CSO undertook a special inquiry that showed less than half of those registered as unemployed were unemployed in the ILO (Labour Force Survey) sense of the term. As a result after September 1998 all those who have been unemployed for six months were called for interview to assess whether they suitable for an existing vacancy or in need of training. To cite the OECD, “nearly half either failed to attend the interview or refused intervention, and 28 per cent were struck off the rolls . . .”. This helped close the very large gap that had emerged between LR unemployment and unemployment as measured in the Labour Force Survey. However, it would not affect the ILO estimates shown in the Graph and its timing does not coincide closely with the rapid fall in unemployment that began in 1994.
Once it started, the decline in unemployment was spectacular. By 1999 the rate had fallen to 4% and it remained close to this level for the next six years. At the same time our emigration was reversed and large-scale immigration became a feature of our economy. Behind all of this was a rate of employment growth that has few parallels in any country.
I hope Colm will permit some nostalgia for these years.
It would be nice to be able to draw lessons from this period that would serve us now that we are experiencing such an extraordinary reversal of fortunes on the labour market front. I tried to do so in a chapter over-optimistically entitled “When Unemployment Disappears: Ireland in the 1990s” [Chapter 8 in Martin Werding, (ed.), Structural Unemployment in Western Europe: Reasons and Remedies, MIT Press, 2006]. My conclusion at the time was that “The exceptional performance of the Irish labour market during the 1990s was not triggered by radical structural reforms”. This despite the fact the a new emphasis on active labour market policies did make some contribution. Note that these policies have not averted the unprecedented rise in unemployment over the last two years.
A Keynesian story does better as an explanation – increases in aggregate demand fuelled by global growth, inward FDI, competiveness etc.
These conclusions point to a gloomy prognosis for our unemployment rate as the macroeconomic conditions of the late 1990s are unlikely to reappear for some time.