Light at the end of the Tunnel?

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.
.

19 replies on “Light at the end of the Tunnel?”

Strip out the first 3 months of the year – lets say for whatever reasons these were freakish outliers – the graph then looks pretty steady, with mild inclines and declines.

Second derivatives do not a recovery make. More people signed on in June 2008 than in June 2007.

@Niall O Sullivan: It is worth examining critically the details of the graph, but deleting 25% of the total data set based on “let’s say for whatever reasons” is not justified. I suspect that quarter 1 2009 was the peak of the world recession, and that is what those high numbers reflect within Ireland. Ireland will remain stagnant longer than most countries, but the INCREASE in unemployment is likely slowing, and soon reversing? Maybe too soon to say that (Enda notes we are talking about second derivatives) but other evidence indicates that at least on a global basis the recession is easing? Admittedly Ireland will lag world recovery substantially.

So I agree that 3 data points removed from the graph changes its nature, but the graph only has 12 data points, and your choice of deletion does not seem well-founded.

@ Gregory O’Connor: I think you took me up wrongly. Time will tell, but what I am suggesting is that a prolonged m-o-m increase of 10-15K is whats happening (base point:Jan 2008) – and the only “blip” in this long running painful trend may well be the 3 months of Q1 09. (This 10K-15K seasonally adjusted increase has been observable in most all of the 12 months of 2008 – or as good as – and in the last 3 months of 2009).

Using Dec 08 as the base point and stating that the rate of increase is shooting downwards may well prove to be technically accurate – for a period of time – but misleading.

The average monthly change each quarter in the numbers on the LR over the past couple of years or so is this – maybe there is light at the end of the tunnel

Q1’07 100
Q2 1367
Q3 1067
Q4 2467
Q1’08 9133
Q2 5433
Q3 9533
Q4 16233
Q1’09 26567
Q2 13567

(June 11,400)

@Niall O’Sullivan
@Gregory O’Connor
@Michael Crowley

I have edited my graph to show the monthly increases back to the beginning of 2007.
The point is that the tsunami [to borrow a phrase] of registered unemployment that welled up mid-2008 may now be subsiding.

By Post Author

Surely we are currently living in a sort of buffer zone and it does not make sense to draw any conclusions until we have taken all the required remedial action and this has had a chance to work its way through the system.

Er, didn’t the rate of increase slow down in Q2 2008 aswell? Could it be seasonal effects?

@yoganmahew

I didn’t want to complicate things by discussing the outliers for March (+) and April (-) 2008. You will recall that Easter fell on 23rd March in 2008 – almost as earlier as it can. I think the effect of this on the LR figures was an uptick associated with the end of the school term occurred in March and was reversed in April. Normally this would occur within the month of April and not show up in the monthly LR data.

By Post Author

Plotting the original data (number of persons) should give some indication of a trend – for the data as found. However, using any trend shown to prophesy a future situation is not such a good idea. I would be skeptical about the data itself – it seems to be filtered and somewhat biased. There are differing interpretations about ‘un-employed’. You also need to ‘sectorize’ the different categories of ‘un-employed’ persons. This economic downturn is significantly different to previous episodes, so considerable caution is required when analyzing any data.

The way to deal with this matter is to observe what happens over a minimum of a six-month span. If the trend is basically level, then you may have a bottom”. Now you wait a further three months to see if the trend has changed. Patience!

Brian P

I think it is far too soon to be making any comments regarding any slowing of the level of the increase. The detailed breakdown figures will be available next week and we can then see whether the numbers joining the register is in fact decreasing. It is likely that the decline is down because of external issues such as emigration.

There was actually a decrease in the number of males claiming Jobseekers Benefit (234) during the month while there was an increase in the number of females claiming JB (4264). Most of the increase was in people claiming Jobseekers allowance. If someone with 2 years of PRSI payments loses their job they are entitled to JB. The period of time it can be claimed has been reduced from 9 months to 6 months claimants move more quickly from JB to JA. Self employed people who are not entitled to JB can claim JA after being out of work of 12 months. This is consistent with a lot of self employed contractors in the construction sector having lost their jobs lat year but are only now entitled to JA. Put simply the rate of new job losses as measured by JB is slowing but the claimant count is still rising as people migrate / become entitled to JA.

Are Brian Cowen and Brian Lenihan statistically-illiterate? As a statistician myself, I’m increasingly irritated by the number of statements these two gentlemen are making that appear to have little statistical basis, and which are actually having the effect of damaging confidence in the economy. I’m being quite serious about this. Its a major deficiency in a politician at the present time to have no grasp of statistics. I’m not condemning them outright if they are (actually, I vote FF) as they have other talents. Both are solicitors and presumably very knowledgeable about legal matters, whereas I, a statistician, know nothing about legal matters.

My reason for asking is that Brian Cowen is quoted in the Irish Times today as ‘accepting that the unemployment rate will be 15.5%’ by the end of 2009. I really can’t see this at all, unless there is a major deterioration in the world economy. The unemployment rate in June was 11.9%. To reach 15.5% by December it would have to increase by 0.6% monthly. But, as Professor Walsh has pointed out, the monthly increase is now down to 0.3%. If it continues to increase at this rate, the unemployment rate will be 13.7% by December, not the 15.5% Brian Cowen predicts.
However, I’ll be surprised and disappointed if the monthly rate of increase does not fall further. I’m not saying 15.5% by December is impossible, merely that, as things stand at the moment, its very unlikely.

Coincidentally, the ‘consensus’ forecast for the unemployment rate at the end of 2010 is 15.5%! Perhaps BC got the year mixed up!

I like how it takes about nine months to call a recession, but nine seconds to call the recovery.

@John, I’m a bit disappointed. Not just by the FF vote! As you know, it takes two numbers to make a percentage. If you think one of them is not going to rise, do you think the other one might fall? Participation rates anyone?

Comments are closed.