Average V Median wage Ireland?

Sam Russell

https://www.boards.ie/discussion/comment/118385280#Comment_118385280

I assume they have different definitions - such as wage vs income, or full time or not. Or could be over different time scales.

Geuze

https://www.boards.ie/discussion/comment/118385095#Comment_118385095

Benedict,

I am referring to the recent publication:

https://www.cso.ie/en/releasesandpublications/ep/p-eaads/earningsanalysisusingadministrativedatasources2020/

Earnings Analysis using Administrative Data Sources 2020

You mention "50 weeks min", I will look for that.

Geuze

Benedict,

I see now what you mean.

In the Annual Earnings section of this report, the annual data does refer only to workers who worked more than 50 weeks, yes.

This does not mean that they are FT workers.

I could work 15 hours pw, every week, week in, week out, so I am part-time, but my employment is still included.

Median annual earnings were €40,579 in 2020

Median annual earnings were highest in the Information & communication sector at €61,632 in 2020, with median earnings for males in the sector at €66,436 and females at €52,725. The lowest median annual earnings were in the Accommodation & food services sector at €24,298 where median annual earnings was €26,252 for males and €22,228 for females.

Mean annual earnings were also highest in the Information & communication sector in 2020 at €77,003, an increase of 7.2% from €71,832 in 2019 and an increase of 21.5% from €63,353 in 2015. The lowest 5-year increase in mean annual earnings was in the Education sector, which saw an increase of 8.0% between 2015 and 2020 from €47,817 to €51,644. The lowest mean annual earnings in 2020 was recorded in the Accommodation & food services sector, at €27,235. See Tables 8.1 to 8.15 and Figure 8.1.

Employees who worked for less than 50 weeks in the reference year are excluded from the calculations for annual earnings. This is done to improve comparability of the data over time. Typically the proportion of employments retained for annual earnings analysis is relatively consistent at both the total economy and economic sector level. However due to the impact of the COVID-19 pandemic on the labour market, the proportion of employments retained for annual earnings analysis in 2020 is lower than in previous years, specifically for those sectors most impacted by the COVID-19 pandemic. See Main Results chapter for further information on the impact of COVID-19 on the labour market and on earnings statistics.

Geuze

https://www.boards.ie/discussion/comment/118385095#Comment_118385095

Benedict,

I see your point.

The overall median for annual earnings is 40,579, that is 780 pw.

The median in the Distribution chapter is 629 pw.

One point is this: it looks to me that the Distribution chapter refers to all employments, rather than all workers.

Many workers have several employments during the year.

Also, I think the explanation is in here:

Employees who worked for less than 50 weeks in the reference year are excluded from the calculations for annual earnings. This is done to improve comparability of the data over time. Typically the proportion of employments retained for annual earnings analysis is relatively consistent at both the total economy and economic sector level. However due to the impact of the COVID-19 pandemic on the labour market, the proportion of employments retained for annual earnings analysis in 2020 is lower than in previous years, specifically for those sectors most impacted by the COVID-19 pandemic. See Main Results chapter for further information on the impact of COVID-19 on the labour market and on earnings statistics.

It looks like lots of workers are excluded from the annual earnings data, e.g. people who did 20 week jobs, then unemployed, then another job, etc. For example, tens of thousands of students.

Benedict

https://www.boards.ie/discussion/comment/118385791#Comment_118385791

The figures refer to earnings for those working 50 weeks or more. So, for example, people working sporadically would not come into this at all.

Also, if someone changes jobs so that his/her wages vary from week to week, it should not impact on the median. The median earnings figures per week week and per year should be reconcilable. But they do not appear to so be.

I had only looked at your last contribution - I've just now noticed the ones before that So, you make the point that 50 weeks plus does not imply full time as it could be a case of part time each week. Also, that one worker could have several different jobs simultaneously.

So perhaps the two figures are reconcilable after all - but perhaps the CSO is remiss in not clarifying this?

On the other hand does "Earnings" not clearly imply average earnings by the same individual?

Geuze

There are students and young people who have two employments at the same time, and who may also have several employments over the year, often with gaps in between.

These all seem to be excluded from the Annual data.

Or take my retired parent, with two pensions, he also has had years where he had two employments along with the two pensions. The two employments are maybe a few weeks each.

Benedict

https://www.boards.ie/discussion/comment/118388127#Comment_118388127

If the 629 relates to average earning for a specific job, the CSO would need to know what was earned each and every week from that job. If X missed 4 weeks in that job - but filled the gap in another job - the gap would skew the median figure in both jobs. If X x 100 worked 6 month in one job and 6 in another and earned 1k pw in both, then the median would be skewed.

The figures are a total mess and we don't know where we are.

https://www.boards.ie/discussion/comment/116283143#Comment_116283143

That wouldn’t be legal for personal income. Only pensions can be deducted.

https://www.boards.ie/discussion/comment/116285439#Comment_116285439

Interesting. It’s actually fairly bunched , particularly after tax.

interesting fact - most private sector workers start relatively poor*, most people earn above the median in their lifetime, and in income terms almost everybody in the private sector ends up poor. It’s a flow through life in many cases

* less than 70% of median.

Benedict

https://www.boards.ie/discussion/comment/118390060#Comment_118390060

If most people earned above the median, it wouldn't be the median. This appears to be a contradiction in terms.

Besides, nobody seems to be quite sure what the median ft wage is. The CSO have given us two different figures for 2020 and we are not even sure if the figures relate to total earnings of a person or does the figure relate to the job. And where a worker has more than one job, these 2 figures could be different.

https://www.boards.ie/discussion/comment/118391849#Comment_118391849

there was a qualifier, "in their lifetime". The median wage is a snapshot in time, but as I said people earn more at certain points in their life and less at other points. I've earned below and above the median.

Benedict

https://www.boards.ie/discussion/comment/118391906#Comment_118391906

Oh, yes, I see what you mean.

But you seem to know what the median figure is? Nobody else seems to know. Not trying to be smart or anything, I'm being serious.

https://www.boards.ie/discussion/comment/118391927#Comment_118391927

No. I agree with you that the figures are all over the place.

Benedict

If calculating the median ft income is not possible then the public should be made aware that it is not possible. But publishing two very different figures cannot be in anyone's interests - especially when the method of calculation is not clear. For example, is the income per job or per person? And can jobs be placed end to end for the purpose of getting the total weeks to 50 weeks? And can simultaneous employments be included in the calculation?

If the CSO haven't a bull's notion of what the median wage for a full time JOB is, then for goodness sake can they not just say so.

Benedict

I note that Willie O'Dea has been arguing lately that the State Pension should be increased - ideally to 34% of what he's has called the "average" ft earning. This is another example of how the confusion surrounding the terms average and median worms its way into everything. We know that the average earning has very little to do with what most ft workers actually earn. In fact the State Pension is currently well above the 34% of what most ft workers earn. I know that Willie means well and I hope he achieves a raise in the pension - but does he realise, I wonder, that the State Pension is already more than 34% of what most ft workers earn?

Peregrinus

https://www.boards.ie/discussion/comment/118432248#Comment_118432248

Willie O'Dea isn't noted for realising things. So I kind of doubt that he realises this.

However, with the state retirement pension, I'm not sure that either the mead or the median full-time earnings is the most relevant comparator. The state pension kicks in at retirement, as a substitute for your earnings which are now ceasing. So I think the most meaningful comparison is not with (any kind of) average earnings for the workforce as a whole, but with some kind of average earnings for full-time workers in their 60s.

Benedict

Iirish Independent Jan 3rd (John Downey's column):

"Mr O’Dea said pensions were supposed to be 34pc of average industrial earnings.

But by that measure there was a gap of €8.55 per week between pension rates and average earnings."

Pension currently E253 plus E8.50 is E261.

Mr O'Dea uses the term "average". According to Leo, the average is more than E900 p/w - so presumable Leo would agree that the pensioners should get more than E300 p/w.

The desperately need to start synchronising their figures because at the moment, they're all over the place.

Podge_irl

Willie O'Dea is not in government so he can believe what he would like.

Benedict

https://www.boards.ie/discussion/comment/118451241#Comment_118451241

There is so much confusion when it comes to median ft wage. Nobody knows what it is. Even the CSO have published two completely different figures for the same year - as we have seen in recent CSO quotes on this thread.

One of the main problems is that there is widespread misunderstanding of the difference between median and average. When Leo announced that the average ft worker earned over 900 pw most people thought most people earned that much. In fact - no matter which of the CSO figures you choose to believe, any ftw earning 800 pw is actually earning more than most ftw - but they think they're earning a lot less! Surely this must lead to employees seeking rises because they feel underpaid - while maybe their pay isn't too bad.

Peregrinus

https://www.boards.ie/discussion/comment/118467584#Comment_118467584

I think it's unlikely that many people base their thinking on pay rises on either the mean or the median wage, nationally. They would base their thinking on their understanding of typical or appropriate earnings for the position, and in the field, in which they themselves work, and they won't be using information from the CSO to tell them what that is.

https://www.boards.ie/discussion/comment/116285152#Comment_116285152

The days of been completely swamped in work now are nearly over in a lot of companies.

There paying you for for your expertise and look to keep there best employees happy. They higher paid employees pass off work to the lower paid teams

The days of been appointed meaningless tasks or tasks below your paygrade or given to cheaper employees hence in IT you have Level 1, Level 2, Level 3 support, junior programmers, intermediate programmers and senior programmers, junior sys admin, Intermediate sysadmin and senior sysadmin.

The people on big salaries pass off as much work as they can to the lower level teams. Your paid for your expertise and not on how busy you are.

Benedict

https://www.boards.ie/discussion/comment/118471747#Comment_118471747

You seem to be suggesting that knowing the median ft wage is not important to people. That would be handy because nobody in Ireland knows what it is.

However, I feel sure that the are many, many people out there who would like to know what most ft workers earn if only because it lets them know where they stand in relation to other workers.

Under normal circumstances, you would expect the CSO to have figured out what the median fw is. As we know from this thread, the CSO have published two very different figures for median ft wage.

Is it 780 pw or 629 pw? Leo suggested the "average" ft worker earns over 900 pw (most people took this to mean the normal or median was a whopping 900 pw).

So the question remains - what is the median ft wage in this country?

What's the problem? Why is it a mystery?

Podge_irl

many people out there who would like to know what most ft workers earn

The median wage does not tell you this.

Peregrinus

https://www.boards.ie/discussion/comment/118475886#Comment_118475886

There's two factors here.

The first has been discussed at much length in this thread - it's much, much more difficult to calculate a median than a mean - you need a much larger and fuller dataset. As far as calculating the median full-time wage, the necessary dataset doesn't exist, so we are down to estimating the median full-time wage from incomplete data. There are various techniques for doing this, and at a guess if the CSO has published two different figures that probably is the result of applying two different techniques. And if the two figures are very different, well, that may tell you something about the limitations of the techniques in terms of the kind of accuracy and reliability they can deliver.

The other factor has also been mentioned. You say that you feel sure that "there are many, many people out there who would like to know what most ft workers earn". In the first place, the median wage doesn't tell you want most workers earn. Most workers do not earn the median wage; in fact, very few workers earn the mediam wage. In the second place, I'm not sing much evidenced to support your belief that there is great demand for the figure. All the evidence in this thread suggests that very little attention is paid to it - the CSO has published two estimates but neither of them seem to feature much in public discourse.

Benedict

https://www.boards.ie/discussion/comment/118487541#Comment_118487541

Obviously I know that if the median ftw wage is established, it doesn't mean most earn exactly that amount - but it definitely is a benchmark figure and it gives workers an indication of whether they are earning more or less than most workers.

You seem to think that the two conflicting figures for median ftw is justified by the fact that they may have used different methods of calculation? So if they use a third method, do they get a third figure?

If they get two different answers to the same question, then something is wrong. There's only one right answer to a sum. Two and two makes four no matter what "method" you use - unless there's something wrong with the method.

CrookedJack

https://www.boards.ie/discussion/comment/118488483#Comment_118488483

Two plus two is a silly example. It's not the sum that is open to interpretation its the data that is prepared before the "sum" is calculated.

If you wanted a very simple example you could think about the sample size. If they want to find the median they might take a representative sample and find the median there, but they have to decide what size sample would be most accurate. They might go for 1000 earners, or 10000, or 100 and each might give slightly different answers to the "sum" as you describe it. That is just an example there are much more complex factors involved in properly preparing an accurate algorithm and more than one way to approach this.

I know you have a political axe to grind but as much as you seem to want this to be a simple sum and easy to find, that just is not the case.

Benedict

https://www.boards.ie/discussion/comment/118488966#Comment_118488966

I have no political axe at all. And even if I did have, it would have nothing to do with the reality which is that nobody knows what the median ftw wage is. CSO have given two completely different figures - are you saying there can be two different correct answers to the question "What is the median ft wage"?

The answer is either "Yes" or "No".

CrookedJack

No. I'm saying that no one can answer that question for the variety of reasons that have been explained to you ad nauseum on this thread.

The closest we can get is "What is a reasonably accurate estimate of the median ft wage (for a given understanding of what ft means)?", to which there can be multiple differing answers.

Just insisting on a yes or no answer does not suddenly simplify the question.

Peregrinus

https://www.boards.ie/discussion/comment/118488483#Comment_118488483

They haven't used "two different methods of calculation" — they haven't used any method of calculation. They cannot calculate the median wage because the don't have the dataset that would be needed to do so. That dataset doesn't exist - nobody has it.

So all they can to is estimate the median wage based on such data as they do have. There are various ways of doing this. If two estimates, derived under different estimate methods, are very different, that suggests that the estimation methods are pretty crude, and not very reliable, given the available data. This isn't a justification for the differing estimates - it's an explanation of it.

"Something is wrong", but not necessarily wrong with the methods used. The methods used may be giving the best estimate they can, given the data available. Something else that may be wrong is that people might be assuming that an estimate based on limited data can always have the accuracy and reliability of a calculation based on full data. Often, it can't. What's wrong in this scenario is not the estimation methods, but people's unrealistic expectations of them.

The advantage of publishing the result of both estimations is that the gap between them signals to you the limited value of the estimation methods in this context. If they had just published the one figure, you might think it was as solid as a calculation - you might be given a false sense of certainty about what the median wage is.

Benedict

Crooked Jack's "ad nauseam" comment is condescending and silly and gets us nowhere. So I will move on to your comments.

As far as I am aware, there is only one method of calculating a median figure. If the CSO have arrived at an "estimate" which they consider to be the best estimation of the median then they must have based that estimate on a set of figures which they consider to be the best available to them.

The median figure should not vary unless the data varies - and there cannot be two "best estimates". They cannot accept two different bodies of data and consider both of them to be the "best".

It may be the case that the CSO have decided to use two different definitions of a ftw and this would give two different medians (if the two groups had different incomes). But if that is the case, they should have pointed this out.