Note: Where available, the PDF/Word icon below is provided to view the complete and fully formatted document
Pay equity and increasing female participation in the workforce

CHAIR (Ms Jackson) —This is the ninth public hearing of the House of Representatives Standing Committee on Employment and Workplace Relations inquiry into pay equity and associated issues relating to increasing female participation in the workforce. I welcome representatives of the Australian Bureau of Statistics to today’s hearing. Although the committee does not require you to give evidence under oath, I should advise you that this hearing is a legal proceeding of the parliament and therefore has the same standing as proceedings of the respective houses.

We have received a written submission to this inquiry from the Australian Bureau of Statistics which was accepted as evidence and authorised for publication earlier this morning, but you are welcome to present us with any additional submissions or to make an opening statement before the committee asks you some questions.

Mr Sullivan —I will make a very brief opening statement. As you say, you have our submission. Basically, what we have tried to present in that submission is a range of information to inform the committee on issues relating to pay equity—or, as it is more commonly known, the gender gap issue. We have covered that in quite a bit of detail in terms of the information sources we have available and how one might go about interpreting that information. We have also presented in the submission a range of other information specifically related to labour force participation and other statistics relevant to the terms of reference. In the appendix to the paper, we have also listed a range of articles and other statistics that might be of relevance to the committee.

CHAIR —Thank you very much for that. I have to say that this has been an area of the committee’s inquiry that has attracted much attention. Certainly in our discussions with academics, they have indicated—as they usually do, I must say, with great respect to them—the desire for more data and a general view that there is not sufficient data around generally to be able to drill down and begin to unpack some of the factors contributing to the causes of the gender gap. I would be interested in your response to that.

Mr Sullivan —We have prepared a presentation for the committee which we would be very pleased to go through, and we can perhaps cover all the issues there.

CHAIR —That would be fantastic.

Mr Sullivan —I will get Michael to walk us through that.

Ms HALL —Can I ask whether or not you might be prepared to give us a copy of the presentation.

Mr Sullivan —Absolutely—no problems at all.

Ms HALL —I would really appreciate that. I am sorry that I have to leave.

A PowerPoint presentation was then given—

Mr Gerrity —As Paul mentioned, we are going to just take you through a quick presentation outlining some of the measurement issues in looking at gender wage issues and issues related to female participation. Some of the estimates which we have presented in the submission we will re-present here and also some of the things to be aware of in analysing and interpreting the data. We will also look at some issues related to employment and the nature of employment which are useful in understanding the context of what the gender wage gap measures are telling us.

Basically, we have two different sources in terms of methodology for collecting information on earnings. We have business surveys and household surveys, so we collect information from a demand and a supply perspective, from businesses and from individuals. Business surveys give us more reliable dollar estimates of earnings because they are collected from employers’ payrolls. We get the composition of earnings—that is, ordinary time and overtime earnings—and hours as well, in some of the surveys, which I will go on to. We also get better industry estimates, because the businesses are coded to industry on the basis of information about them and their business registration.

In household surveys we get all the sociodemographic characteristics—age, country of birth, marital status and those kinds of things. We get some employment characteristics as well—that is, information about their job flexibility and stability, the hours they work and that kind of thing—but the earnings figures are less robust because they are collected from the people and that relies on their recall of the amount of earnings they get, and, in some cases, they are reporting on behalf of other people in the household as well, so it relies on them to know the earnings of those people.

Each methodology really has its own pros and cons, so we are going to give you a flavour of some data utilising those sources, and the pros and cons, throughout the presentation. Those are presented in the submission as well.

In the business surveys, we have the Survey of Average Weekly Earnings and the Survey of Employee Earnings and Hours, and for household surveys we have the Employee Earnings, Benefits and Trade Union Membership survey, which is a supplement to the monthly labour force survey. The Survey of Average Weekly Earnings is a quarterly survey—headline ABS level estimates of earnings. Information is collected at the business level, so we get aggregate wages for a business and an employment count, and we derive the averages that way.

The survey of Employee Earnings and Hours is two yearly, and it has extra information on the composition of earnings and the distribution as well. It is collected at the individual employee level. We survey businesses, and those businesses select a subsample of their employees, and then we collect information about the pay of those individual employees and other characteristics, which means we can get distributional data, which is really useful in this context. We can get information about their occupation and the methods used to set their pay—that is, whether they are award employees or whether they are paid by individual agreements or collective agreements. We can get information on the hours they are paid for as well.

The EEBTUM survey, as I will refer to it, the Employee Earnings, Benefits and Trade Union Membership survey, is, as I said, a household survey, so it has lots of demographic characteristics, which is where its value lies in this area. If you have any questions while I am going through, feel free to jump in. I am happy to take questions at any time.

Basically, those are the three key sources. How do they compare? This graph shows on the left-hand side columns for all employee total earnings, and you can see that the measures for AWE, EEH and EEBTUM are very consistent, around that 66 per cent. What we are showing here is the ratio of female earnings to male earnings. A lower ratio represents a larger pay gap—put it that way—but in the submission and this presentation we are focusing on the ratio of female to male earnings. So the first columns are for all employees and their total earnings. That includes overtime, for example, and ordinary time. The second column just looks at full-time employees who are paid the adult rate for their occupation, so we are removing the junior rates and removing the weekly earnings of part-time employees. Again, using both measures, the gap or the ratio is consistent across the three surveys, but one of the key points to make here is the difference in the ratio when you start to compare full-time adult employees as opposed to total employees. You remove the effect of part-timers and their variable hours. This will be a flavour throughout this—

Mr RAMSEY —Can I ask a question. Is there any industry specific in that? Does that take into account, to use a stereotype, that the blokes might be engineers and the girls might be nurses? Is there any allowance for that in the figure we are seeing there, or is this just bundling everyone up together?

Mr Gerrity —This is an aggregate sort of picture of it. That is one of the important points to make, I think: we are looking at aggregates, so you are looking at the different contributions, different occupational industry profiles and different earnings of industries as well. We can start to drill down, and we will do that through the presentation. You can see that there are some large differences across industries in particular.

We are highlighting the importance of looking at population groups—whether you are looking at all employees or full-time employees and also the kind of earnings series you use. We are going to focus on some ordinary time series as well as total earnings, because overtime is obviously a more variable type of earnings. Men tend to have a lot more overtime than women as well, so that is something that needs to be taken into consideration.

Which source do I use? It depends on the analysis you are undertaking, as I have alluded to. Each survey has a specific purpose and so it is useful to think of what you are trying to achieve in your analysis and then which source you would want to use. In terms of changes over time, that is where the AWE survey, the average weekly earnings survey, has its value. It has a quarterly time series. It goes back to 1983 in its current survey form. The data goes back another 20 or so years as a payroll administrative base source. It is not consistent, but it does provide a really good time series at that aggregate level. The other two sources let you start to drill down a bit more.

This graph shows a time series from AWE. The bottom line is the all employees total earnings, which is the first column we had on the previous graph. The grey line is full-time adult total earnings. Again, that is the second column on the previous graph. The third line is full-time adult ordinary time earnings. Here we are limiting it to ordinary time by removing overtime. You can see the effect of that. The gender gap is lower again once you look at ordinary time earnings only. Another point to make from this graph is the two full-time adult series. The ratio has gone up slightly so the gender gap has gone down. In the total earnings series it is about the same as it was, if not slightly lower. But all the series do exhibit some variability over time. In all three series the gap has increased marginally in the last three years, according to AWE figures.

Drilling down a bit, there are some other compositional factors. We can look at occupation and industry, for example. There is more disaggregation in the submission as well, with different kinds of employment, and I commend that to you. I have ordered these from lowest to highest, or the largest gap at the top and the lowest gap at the bottom. You can see that for managers and professionals, which are the two highest skilled occupation groups, the gender wage gap is highest. In some of your more lower skilled occupations, the gap is lower. That is going to be a theme that comes through in a lot of these things. It is those higher earning industries and occupations where the gap tends to be higher while the ratio is lower.

There are large variations, though, within these occupations. You are looking at quite different jobs in those broad level occupations. Once you start to drill down you will start to see some quite large differences as well. This is from the EEH survey. We also have occupation data from EEBTUM, the household survey. We can drill down in both those sources to the two digit and three digit occupation classifications to get information about more specific occupations, but the sample sizes start to limit what you can do when you start to disaggregate the data too much.

Looking at industry, again ordered from the industry with the highest wage gap or the lowest ratio at the top down to the smallest gap at the bottom, you can see finance and insurance, mining, property, business services. In some of the industries with higher average earnings the gender wage gap is higher, and in some of the industries with lower earnings the gap is lower. You will also see that things like accommodation, cafes and restaurants are more award related industries, and they will come through in a moment too. Similarly, there are quite different jobs in the same industry, and that is also something to be aware of. Looking at methods of setting pay, looking at the gender wage gap from a wage negotiation perspective, this data is also from an EEH survey. I should have added before that this is May 2006 data.

CHAIR —But you do have 2008 data now, don’t you?

Mr Gerrity —It is being been run, with the results to be released in April. That is the expectation. We will be able to update all these figures at that stage. I guess the thing I want to highlight here is that, if you look at award only employees, those people who are employed at the award rate—or at not more than the award rate, I think is the technical definition—there basically is no gap. The gap in that measurement is about 0.995.

CHAIR —On ordinary time earnings?

Mr Gerrity —Yes, at full-time adult ordinary time earnings the ratio is about 0.995. In 2004 it was actually over one, so the gap was the other way—women earned more than men on average for award only employees. You can see in terms of collective agreements that the ratio is about 0.992 and individual arrangements are a bit lower again. Another point in terms of that individual arrangements category is that it includes over award employees. These are people who are probably still fairly award reliant but who are receiving a bit extra above the award. If you removed those, I would suggest that in that individual arrangements category the ratio might be even lower, if you look at where the award only employees are.

We move to our household survey, our EEBTUM survey, to look at some more demographics. We can look at gender wage gap by age. You can see that that gap is higher or, in this case, that the ratio is lower in your prime working age groups, which are also the age groups with higher average earnings.

Ms BIRD —Does that reflect the movement through the promotional positions?

Mr Gerrity —Yes, the career progression issue. That is some of the key analysis we have done. Now I will move on to looking at some other approaches to analysing the gender wage gap. These are really important in understanding and getting a good idea of what the age gap really is. It is very useful to look at hourly earnings instead of weekly earnings and also to look at other measures of distribution from the arithmetic average or the mean.

You might want to limit your analysis to full-time employees. To get more of a picture of like with like, you remove the part-timers where there is a really large variability in the hours they work and therefore the weekly earnings they are going to get. But if you do analyse just full-time employees you are actually removing a large section of the workforce as well, so it may become less representative in terms of who you are comparing. Using hourly earnings is a really good way of taking all those factors into consideration. Hourly earnings allow the impact of part-time employment to be reflected in these comparisons. As I mentioned, a higher proportion of women work part time, and we will see that later in the presentation. The hours that part-time women workers do are quite different to the hours that part-time men workers do as well, so the gender wage gap can be quite different when we look at that. We have hours data collected in the employee earnings and hours survey as well as in the EEBTUM survey. We get it from the business and household survey source.

CHAIR —That is essentially based on the number of hours according to someone’s payroll records?

Mr Gerrity —Correct.

CHAIR —It does not necessarily reflect, for example, unpaid overtime that is being performed but it does reflect paid overtime?

Mr Gerrity —That is right. Hours paid for data is only collected for non-managerial employees. I can go into the definition of that, but basically managerial employees are those who are considered to have strategic responsibilities in the conduct of the business and who usually do not have an entitlement to overtime. It is historically meant to separate those people who are paid by the hour and those who are not, although that is probably becoming a lot more grey now they it has ever been. What you are doing is probably removing most of the managers’ occupation group. There is a large crossover between that managerial status and occupation, so that will probably remove some of the effects of that unpaid overtime, because they are the people who are more likely to work the unpaid overtime. I think your point is still valid—it does exclude those sorts of things.

We can do an hourly rate from EEBTUM, the household survey supplement, on hours actually worked. You can get more of an effective hourly rate, but I guess both have their advantages and disadvantages, depending upon how you want to look at it. The data I have here is from the EEH survey, so it is hours paid for of non-managerial employees, and the ordinary time earnings are on the left. You can see the ratio is quite different when you compare weekly and hourly earnings. With hourly earnings that gap is hardly a gap at all for non-managerial employees’ ordinary time earnings. Even for full-time adult ordinary time earnings the ratio is slightly higher when you look at hourly earnings as compared to weekly earnings, which shows that the impact of the different kinds of hours that full-time people work still can have a difference in terms of the measurement of the gender wage gap.

We will look at medians and means as well. Particularly for things like earnings analysis, the mean is not necessarily the most representative measure of average. Because you are adding up the total earnings of people and dividing by the number, if you have a selection of people with very high earnings that will pull the mean up quite considerably. But, because the median represents the midpoint of the distribution, it does not get pulled up quite as much. So the mean is skewed upwards compared with the median, and I guess that is particularly relevant in this sort of context, given that many more men work in those very high-paying jobs. If you want to look more at the average person, the median is probably a better measure to look at.

The use of means or medians can have a large impact on earnings analysis. This figure I have here of 65 per cent is for a particular earnings series, not for the all-employee series. I am using that to highlight how representative the median is compared with the mean if you can see that 65 per cent of employees are fewer than the mean. That is not for the all-employee series; that is for one series—I forget which one it is now—but it is a good example of the difference the measures can give. EEH and EEBTUM both have medians available because they are collected at the individual level.

What difference does it make? This is a time series from EEH showing median and mean earnings for full-time adult, non-managerial ordinary time earnings. You can see that the ratio is higher for all years in this time series for the median. In fact, with regard to the median between 1996 and 2006 the ratio has gone up slightly, whereas the mean has gone down slightly.

CHAIR —Particularly between 2002 and 2004.

Mr Gerrity —There was a bit of a change there. I am not sure what caused that.

Mr RAMSEY —A new survey officer.

Mr Sullivan —The issue of advanced sample survey could contribute to that. It is over two-year intervals.

Mr Gerrity —Yes, that is right.

CHAIR —Do you mean change in the size of the survey?

Mr Sullivan —Just the business that we actually sample of each survey that we have.

Mr Gerrity —The samples are independent, survey to survey. Finally, in terms of earnings, another way you can look at the distribution is looking at deciles. The decile represents breaking up the distribution into 10 equal-sized groups. Decile 1 represents the point at which 10 per cent of employees earn less than that figure and decile 2, 20 per cent less. So decile 9 represents the top 10 per cent of earners. You can see again that the wage gap increases or the ratio decreases as you move up the earnings scale, so it is up closer to one for the lower earners and getting towards 0.83 or so for higher-earning people.

Mr RAMSEY —Does that actually tell us anything different?

Mr Sullivan —No, it is just another way of putting it.

Mr Gerrity —It is just another way of representing it. We can look at the industries and occupations and say, ‘These are the ones with the highest earnings,’ but this is just representing it distributionally. That is it for the earnings section, so I can take questions now or we can save them for later.

Mr RAMSEY —You had a slide up there about the difference between weekly earnings and hourly earnings. That would suggest that females on hourly earnings are receiving very close to 100 per cent. Would you explain to me what that means—what that is telling you in a practical sense.

Mr Gerrity —In the practical sense it is showing that weekly measures are not as representative as the hourly measures. When you look at an hourly perspective, the gap is lower. It probably tells me that there is less of a gap than the weekly measure would indicate.

Mr RAMSEY —Is it telling you that it is more to do with the structure of the workforce than it is about pay rates and equity? Is it about the way people choose to work and structure their life?

Mr Gerrity —Yes, I think so.

CHAIR —We have had quite considerable evidence about this and anecdotally. Our problem is that you can identify multiple contributing factors of what causes the gender pay gap. One of them, you can argue, is the industry people choose to work in or do in fact work in, whether it is by choice or otherwise—whether it is hours of work. I guess it is being able to drill down into that, because we know all these factors, but really all that tells you is that on ordinary time earnings, when you just look at hours, the gap is not that great. But that is not the practical reality of everyday working lives for men and women across industries, because men get paid more overtime. It is illustrative, but it is not definitive.

Mr HAASE —Men work more overtime.

CHAIR —Men get paid more overtime—they have more paid overtime.

Mr HAASE —Both are facts.

Mr Gerrity —Yes, I think that is right. The question is whether it represents a wage discrimination or whether it represents aspects to do with the kind of work people are doing, career progression, life choice and all those things. That is what the next section is going to touch on.

CHAIR —And all of the above!

Mr Gerrity —That is what the next section is going to touch on a little bit: looking at the kind of employment that men and women choose or the kind of work they do, the changes that have happened, looking at participation and looking at superannuation at the end as well in a sort of life cycle perspective. This provides a bit of a context, I suppose, for those measures of the gender wage gap, to help understand what it really means.

The first graphic up here shows the labour force participation rate over the last 20 years, from 1988 to August 2008. The key things we see from that are that the participation rate for men is considerably higher than for women, although participation amongst women has increased quite a lot over those 20 years, whereas for men it has decreased a bit.

CHAIR —It also tells us, given that one of our terms of reference is looking at recommendations for increasing women’s participation, that we need to do a little bit more than just allow the natural rise to continue, because on that kind of projection we will be here for quite a few years before it naturally increases.

Mr HAASE —Whilst you are at that point, it might be a good time to explain the foundation of that percentage. What is it a percentage of?

Mr Gerrity —It is the percentage of employed and unemployed people—it is a proportion.

Mr RAMSEY —Within an age group.

Mr HAASE —How do you determine somebody who is a member of the population that is available for work, and what do you do for those who are not available for work? I want to understand the stack. What is labour force participation rate? You are telling me that, in August 2008, it is damn near 60 per cent for women. Sixty per cent of what?

Mr Gerrity —Sixty per cent of the civilian population aged 15 and over are either employed or—

Mr HAASE —Fifteen and over—until?

Mr Gerrity —Just 15 and over.

Mr HAASE —Fifteen and over?

Mr Gerrity —Yes, there is no cut-off. They are either employed or unemployed.

Mr HAASE —So somebody who has never been in the workforce, never sought work, never been dismissed and never been on welfare would be included in that stat as a member of the total female population within that wage gap and therefore would be considered to be unemployed?

Mr Gerrity —They would not be unemployed; they would be not in the labour force. But, yes, they would be—

Mr HAASE —You just said ‘in the labour force, employed or unemployed’.

Mr Sullivan —No, sorry. The labour force consists of those employed and those unemployed.

Mr HAASE —The labour force, okay. Good show.

Mr Sullivan —And outside that are those persons who are not in the labour force, and that would be your person—

Mr HAASE —So the person I spoke of is not anywhere in that stat.

Mr Gerrity —They are part of the denominator.

Mr Sullivan —They are the difference between 100 per cent and the percentage shown.

Mr Gerrity —Yes.

Mr HAASE —All right: 15 and over, males and females—the same age?

Mr Sullivan —The same definitions.

Mr HAASE —Beautiful, thank you.

Mr Gerrity —I guess that shows an overall sense of participation. But, for the people who are working, what kind of work are they doing? We can look at the proportion of male and female people who are working part time. Again, you can see quite a large difference in the proportion of men and women who work part time. Almost half of the women work part time; it is about 45 per cent. For men, it is about 15 per cent. But both have been increasing over the last 20 years. I guess that is a lot of people, but, again, ‘full time’ is people who usually work 35 hours or more a week, so you could have people who are not necessarily working five or 10 hours but could be working 20, 25 or 30 hours. They are part time.

Mr HAASE —I will stop you there again, if I may. On your definition of part time: if accumulated hours are up to 35, that is part time?

Mr Gerrity —Yes.

Mr HAASE —Over 35 hours is full time?

Mr Gerrity —Yes.

Mr HAASE —So they might have three part-time jobs and be working over 35, and therefore they would not show in this stat?

Mr Gerrity —You would be full-time then, because it is the hours that you work in all jobs.

Mr HAASE —That is good. We looked at a stat some time ago, didn’t we, where we were concerned that actually it might be an additional job? There was something that was quite confusing.

Mr Gerrity —We have information about the number of hours people work in their different jobs. A lot of the information we collect is as supplements to this labour force survey—that EEBTUM survey is about the main job, and the employer surveys are about jobs, so if you had more than one job it would count twice because it is about the job, not about the person.

Mr HAASE —Yes, I think I am happy with that.

Mr Gerrity —Casual employment, as well, is another issue that is fairly topical and fairly relevant here. This shows a time series of the proportion of female and male employees without paid leave entitlements, so that is without holiday and sick leave. I guess that is an ABS proxy for casual employment. There are other measures of casual employment that we have, and this is one that we have a particularly long time series for. It is a particularly good measure because it is less subjective than some of the other ones. It is exactly what it says—that is, people who do not have leave entitlements—which is a good way of looking at a casual.

Mr HAASE —When you say that, of course, you know that the casual rate includes the entitlement for leave as a proportion of the hourly rate.

Mr Gerrity —That would be the expectation.

Mr HAASE —Yes.

CHAIR —But it is not necessarily the case.

Mr Gerrity —Yes. And collecting that kind of information about the casual rates, if you are looking at earnings, is sometimes hard because those casual rates are rolled into people’s salary in some situations.

CHAIR —So this is making no judgement about how they got to be working without paid leave entitlements; it is just a fact: no paid leave entitlements.

Mr Gerrity —Yes. We do collect information about people who receive a casual loading, but, again, they may not know whether they get one or not. We also collect information about people who consider themselves to be casual. There is a really big overlap between those three, as you would expect. I think we have referenced an article in the submission too which goes into more detail about the various measures of casual employment.

CHAIR —A quorum has been called in the House.

Mr HAASE —You will probably have to go, won’t you? Does it make me behind the times when I say that part-time employees do not get paid leave?

Mr Gerrity —They may well get paid leave.

Ms Taylor —They can get paid leave. You can be part time, but it does not mean necessarily that you are casual.

Mr Gerrity —Most casuals probably are part time, but they are not mutually exclusive. You can have full-time casuals and part-time casuals.

Mr HAASE —So if somebody is working a number of hours and they consider themselves to be part time—years ago, it used to be over 15 per week, and they got 25 per cent loading—

CHAIR —Can I just interrupt you, Barry, and ask if we can form a subcommittee.

Mr HAASE —Yes, I so move.

CHAIR —As there is no objection, it is so ordered. I will join you again shortly.

ACTING CHAIR (Mr Haase) —It used to be that you had part-time classification and you got 25 per cent loading in your hourly rate, and that covered your sick pay and holiday pay. If you had casual loading, it was below 15 hours per week and you got a 50 per cent loading. That was your consideration for sick leave and holiday leave. Neither of those two classifications would have paid leave. So my part-timers would have been in your casuals.

Mr Gerrity —They would, but not all part-timers would be casual.

ACTING CHAIR —Because some part-timers—

Mr Gerrity —Have that leave.

ACTING CHAIR —receive an amount of pay per week and, when they take leave, they are paid that amount during their absence. Is that what you are saying?

Mr Sullivan —Yes.

Mr Gerrity —They might have a pro-rata amount of four weeks annual leave and they might still have a couple of days sick leave.

ACTING CHAIR —Yes. I am interested to know how you finally differentiate between the two. Who do you ask—an employer or an employee—for this survey?

Mr Gerrity —This is based on household surveys. We ask people—

ACTING CHAIR —So the employee is answering this question?

Mr Gerrity —The individual is, and we ask about whether, in their main job, they have paid leave entitlements. We ask them whether they have paid holidays and paid sick leave.

ACTING CHAIR —And does the respondent understand sufficiently to determine that if they get it paid in their hourly rate on a weekly basis—or if they are absent and paid the weeks they are absent with a wage packet that week? Are they sufficiently sophisticated to answer that correctly? I would doubt it.

Mr Gerrity —Yes. The data that we get is of sufficient quality, I think, to look at—

ACTING CHAIR —If I am getting paid 15 hours a week and there is a 25 per cent loading in there, and if I do not turn up for work for a week and I get paid nothing, I am going to say that I am a casual employee.

Mr RAMSEY —Barry, there is a division.

ACTING CHAIR —I am sorry, folks; we have to go. We will suspend the hearing until such time as we can reconvene. Please excuse us.

Proceedings suspended from 11.50 am to 12.05 pm

ACTING CHAIR (Mr Haase) —Carry on, Michael. Or was I carrying on, perhaps, at the time? I have had a long discussion with my colleague, and we will resolve it internally. But I understand the basis of your statistics. I will discuss it with the chair.

Ms Taylor —I would like to point out that there is a quarterly publication, ABS catalogue No. 6105.01, that is referred to in the submission. There is also an article referred to in that submission which talks about this issue of casual employment. In the article which we referred to, we have done some analysis of people who say ‘yes’ or ‘no’ to this question and whether they receive a casual loading or not. So you can pick up those people who say—

ACTING CHAIR —So you do use the term ‘loading’?

Ms Taylor —In a different question, yes.

ACTING CHAIR —Do you use the term ‘part-time loading’?

Ms Taylor —No, it is ‘casual loading’.

ACTING CHAIR —You are not in industrial relations, I know, but to your knowledge is there still a category of employee who works what is termed ‘part-time hours’ whereby they are paid a loading for each hour. This is in consideration of annual leave and sick pay. They receive a weekly pay packet for each week they attend their workplace but when they go on leave and are able to go on leave—because that is the determination of their employment—they do not receive a wage packet, and so they would say, ‘No, I do not go on paid leave.’ Therefore, they would fall into that stat, even though it is technically correct that they are receiving paid leave because that is the award. That is the basis of their engagement because they are paid a 25 per cent or whatever it is these days loading. Sharryn, you know far more about this than I do.

Mr McColl —The interpretation there by them would be that they do not get paid leave; they are actually paid compensation for not being able to take paid leave. If they get sick, they do not get paid. If they do not get sick, they are effectively being paid an average.

Mr HAASE —I would like a stat that said they received paid leave, because they are receiving it for every hour they are working under an award classification and determination. It makes it a messy stat is all I am saying.

Mr McColl —That cross-over is in the article.

Mr HAASE —I will read the article.

Ms Taylor —It is not an easy thing to measure, which is why we have three measures trying to look at it.

Mr HAASE —Lies, bloody lies and statistics.

CHAIR —I apologise that there may be further divisions but we will see how we go.

Mr Gerrity —I mentioned before that we could look at some occupation and industry profiles to see where men and women work. This graph from our labour force survey shows the proportion of males and females by occupation. There is quite a large difference between certain occupations. When we look at, say, technicians and trade workers, we see that a quarter of men are technicians and trade workers and it is about five per cent of women. Conversely, if you look at clerical and admin workers, sales workers and community and personal services workers, you see a much higher proportion of women in those occupations.

CHAIR —The only surprising static there is for professionals.

Mr Gerrity —There are a lot of nursing professionals in that statistic.

CHAIR —Given that that is the second-highest pay gap by the same grouping, it is interesting that that is the only group where the number of women employed are greater.

Mr RAMSEY —This is probably some stupid misreading of the stat: if five per cent of technicians and trade workers are women and 25 per cent are men, what are the other 70 per cent?

Mr Gerrity —Five per cent of employed women are technicians and trade workers. The figures for men add to 100, and the figures for women add to 100. It does not represent the proportion in that occupation.

Mr RAMSEY —I knew it was some stupid way of looking at it.

CHAIR —Nurses and teachers are professionals.

Mr Gerrity —We can drill down further to display more of the details of occupations in industries as well. That is possible from the survey. If we look at industries, again there are a few industries with some considerable differences. I guess not surprisingly, a far greater proportion of men work in the manufacturing and construction industries. Education, heath care and so on have a higher proportion of women. That ties into the occupations we looked at before.

Mr HAASE —Once again, just to reinforce Rowan’s question, just about five per cent of the women making up the female workforce of Australia are in other services?

Mr Gerrity —That is right.

Mr HAASE —That is of the workforce, isn’t it—not the ones who are employed today but of the workforce?

CHAIR —Yes, the female labour force.

Mr Gerrity —Female employed people—currently employed.

Mr HAASE —This is not our old labour force?

Mr Gerrity —No, it excludes the unemployed people. Yes, almost one-fifth of all employed women are in healthcare system and another 12 per cent are in education and training.

Mr RAMSEY —For instance, if you are an administrative clerk in a hospital, are you in the healthcare or in the public administration system?

Mr Gerrity —It depends where the business is coded. But, if it is a hospital, it is coded under the hospital and would be in health care. It is based on what the business is doing, so it would more than likely be health care.

We can now look at hours worked as well. I mentioned before the differences in the hours that men and women work, and this highlights that. The hours range is on the bottom—one to 15 hours a week, 16 to 21 hours a week and so on. This is hours usually worked as well. You can see the proportion of women who usually work one to 15, 16 to 29, 30 to 34—the part-time hours—is much higher than the proportion of men. If you look at your more traditional full-time hours—35 to 39—there is not really much of a difference. But, as you move into those longer hours, there is a much greater proportion of men than women working 40, 50 and 60-plus hours.

CHAIR —Paid hours?

Mr Gerrity —No, this is usual hours worked. It will include paid and unpaid hours. It is likely that a lot of those longer hours may well be unpaid—managerial type people. We will look now at bringing in some more family related data. This shows hours worked by mothers. Again, we can look at the age of the child on the bottom there—nought to four years, five to nine years, 10 to 14 years—and then the hours range. If you look over here, this is less than 16, 16 to 24, 25 to 34 and 35 or more.

CHAIR —Does it tell you the number of children or is it just by the youngest child and that is it? So you could have six kids but—

Mr Gerrity —The age of the youngest child, yes.

Ms Taylor —It is usually the youngest child which determines the mother’s hours of work.

Mr Gerrity —Not unexpectedly, the higher proportion of working mothers with younger children works a lower number of hours. As the age of the youngest child increases, there is obviously a higher proportion working full time. Looking at duration with employer is also interesting. This shows the continuous duration with the employer. One of the key things from this graph is the proportion of people whose continuous duration is 20 years or more. It is almost double for men than for women. If you think about the effect that career breaks can have on earnings, this highlights the differences in job tenure.

Let us look at moving into the retirement phase and at the main source of income at retirement. This is for those people who have retired and for retired people whose last job was less than 20 years. It is not for those people who have been out of the workforce for a long time. There is a higher proportion of men than women whose main source of income is a pension or superannuation and a much higher proportion of women with no income. But bear in mind that ‘no income’ means there is no personal income but it does include a partner’s income and living off savings and things like that. This reflects the lower levels of participation and so on in previous years to build up superannuation and so on.

We can also look at expected source of income at retirement, which is perhaps more in useful in that it looks at the expectations of people who are currently in the workforce or of working age. Again, there are a higher proportion of men than women with superannuation as their expected main source, but with government pensions there are now a high proportion of females whose expected main source of income is a pension. The no income category is still much lower for both men and women but, again, there are still a higher proportion of women than men. If you looked at this kind of data by age you would probably see that the older age groups probably look more like that previous graph, and for the younger age group that no income group would probably be a lot lower and the superannuation a lot higher.

Superannuation coverage: what that means is that you either have accounts in the accumulation phase or you are actually drawing on super and you have received a lump sum or you are drawing on super and you are receiving a superannuation pension. It shows that a higher proportion of men than women have accounts in the accumulation phase, although there is not a particularly striking difference, and a slightly higher proportion of women have no superannuation coverage.

We can also look at superannuation balance as well. This is for people who are in the accumulation phase. They are currently contributing to superannuation. Key things to look at in this graph are that there are a much greater proportion of women than men who have balances of less than $10,000—so right at the low end—and almost twice the proportion of men who have superannuation balances of $100,000 or more.

CHAIR —Thank you very much for that. That was extremely useful in putting your submission into some sort of context. We have a few questions—and I suspect we may have more and we may come back to you. In part, you have answered one of my questions—that is, with regard to the last statistic we looked at about women with dependent children, is that the only kind of information you gather for the number and age of dependent children that might shed some more light on these career breaks?

Ms Taylor —No, we have a survey—for which I am not responsible so I would need to look at it in more detail—the Pregnancy and Work Transitions Survey, which looks at information about people’s breaks in their work, how much leave they take and when they go back into the workforce. There is that kind of information. We also have information from our Labour Force Survey about various types of families within the labour force and what the labour force status is of each of the partners in that as well. So we do have a range of information available.

CHAIR —One of the things I think has been a live issue for us is that the working life experience of women is quite different, and that has probably the greatest impact on something like superannuation, for example. Do you have that sort of information longitudinally—the average trace of a woman’s working life? Intuitively, and by common practice, I think most of us know that we work for—and it seems like your stats say this—at least 10 years or thereabouts and then we have a career break, normally for a first child, and then return to the workforce in varying capacities.

What isn’t explained is why the part-time employment, for example, continues once women are in their 50s. We had evidence—I think it was in South Australia—from one group of women in particular who indicated that, as older women, they would be prepared to work more because this is the time when their kids are off their hands and they are ready, willing and able to undertake greater hours. But for some inexplicable reason they are not doing that.

Ms Taylor —There are probably a few things going on. I would like to draw your attention to a survey that was released last Friday. It is one that is run every six years. We have a new module in that Survey on Employment Arrangements, Retirement and Superannuation whereby we try to look at how people balance their work and caring responsibilities. We have taken a broadbrush approach to caring. We are looking at people caring for children, what kinds of arrangements they use if they work or whether they do not work because of those caring arrangements. We have also had a look at people’s caring responsibilities for ill people within the household, for ill people outside of the household and for older parents—all that kind of thing. It does not give you a longitudinal picture but what it does give you is a bit of an idea for the previous six months and for the ongoing six months how that caring responsibility is likely to affect their working arrangements or whether they are going to employment at all. That should shed some light on that information. We can get back to you on the pregnancy and work transitions in terms of whether there is anything in there which might give an idea of the breaks in employment.

Mr HAASE —With the collection of statistics in relation to the equity gap, I have been quite amazed at the number of submissions we have received from government agencies all condoning compilations of slightly different statistic sets. I know that these submissions are all making a contribution to your basic statistics, but does the ABS make an effort to facilitate the avoidance of unnecessary, wasteful duplication in the collection of those statistics? Is there any effort made by the ABS to attend to that issue? Those submissions all reflect a slightly different view, I would suggest, as a result of their slightly different approach to the same issue—the gap in wage equity.

Mr McColl —The bureau does make an effort and has a legislative requirement to coordinate official statistics. We make submissions to cabinet on other departments’ submissions about collecting information so as to try and ensure that we do not duplicate. We have that role for official statistics. We have another role for business surveys. For people collecting that sort of information from businesses, as our Employee Earnings and Hours survey does, the government has set-up a statistical clearing house. That was one of the responses to red-tape provider load, particularly for small businesses. So, anybody who is going to get Commonwealth funding for a business collection that would be duplicating something else or, alternatively, anybody who is going to be collecting data that is different and necessary but will be of poor quality and not useable for government purposes, we try and make sure that that does not happen. Even if a government department contracts it out to a private researcher, it still has to get through that process. That does not reach into what the private sector will do.

Also, state and territory governments could decide that they want to continue with a collection because there are different employment arrangements under federal and state governments. They may have particular collections associated with monitoring and evaluating their employment legislation, as well as trying to evaluate the circumstances of people in their state who may be under either federal or state awards. So there are some legitimate needs around the country for somewhat different statistics in that space. We definitely try and make sure that we stop duplication where there is Commonwealth funding involved. When we are aware of duplication elsewhere, we try to prevent it.

Mr HAASE —You would have access to our published submissions; however, I am not sure whether you have taken the time to access some of them to make comparisons. If you have accessed them, I think you would have been a little overwhelmed by the number and variety of statistical sets. Thank you for the answer.

CHAIR —I warn you again that I suspect we will be coming back to you with a significant number of questions in regard to the potential implications, if we intend to make recommendations about greater data gathering. I appreciate that it is likely to have resource impacts on the bureau. Thanks to media reports in other areas, I am conscious of these things. So we should be clear about what the implications of our recommendations are.

One of the things that have become apparent to us—at least from anecdotal information and some academic work that I think your own information backs up—is that, under individual employment arrangements, women were—for want of a better description—more disadvantaged. The gap appears to have increased or to have become greater under those arrangements. We are particularly interested in the impact of government legislative reform. We are conscious of the reforms that occurred post-96 and then in 2005 with Work Choices. Can you provide us with more information about what happened as a result of those 2005 legislative reforms?

I will now turn to the future. Given that we are likely to see new industrial relations legislation introduced certainly before the end of the year and, one would hope, in place shortly thereafter, how long do you think it would be before those reforms could start to be measured? Do you have a mechanism whereby you sit down in your branches or areas and say: ‘Right. This is a pretty significant reform. How are we going to put measurements in place to pick this up?’ How does this sit with the current regular survey reports? That is probably a huge question. Do you want me to break it down into three or four parts?

Mr Sullivan —No. I was going to suggest that the 2008 Employee Earnings and Hours survey, which we currently have in the field, will inform on some of those legislative reforms relating to 2005 in terms of pay arrangements—how people negotiated their salary or otherwise. That will capture some of that, and that will be coming out in April.

On the general issue of whether ABS look at significant change in the environment and try to work out how we can best respond to that, we certainly do go through that process. Clearly, in the current environment, for us to do something new we have to stop something current; it is a priority-setting approach, which we do go through. That is an issue that we tackle.

The only other observation I was going to make was on the individual arrangements. Our data shows that the gap is 20 per cent at individual arrangements. But again, probably what we need there to inform on that better is the occupation, the industry and things like that. That is just a raw number and will include the highest paid executives and individuals. It reflects all those other elements.

CHAIR —We have seen better data than that. I think EWR’s pay equity audit process with some of the private sector companies has been much more informative in terms of what is actually happening in a workplace with issues associated with pay equity—not the least of which is surprising employers that they have a pay equity gap. They insist that they do not have one until they have it audited. For us to be able to come to the point of view that, let us say, women are worse negotiators of individual wage arrangements than men, your raw data is never going to be able to provide that sort of information.

Mr Sullivan —Inform on that fully.

CHAIR —Yes. That is what I thought.

Mr McColl —I was going to add something on your question of how we can be in a position to start measuring the impacts of changed legislation. We have an expectation that, in drafting the legislation and putting it to parliament, the responsible agency will have thought about detail and actually provided estimates of the cost of the evaluation of whether it is working and what impacts it is having. They should have talked to us about how they are going to do that measurement, rather than assuming that it will somehow magically happen. So hopefully we get engaged in that process at an early stage so we can advise on how it could be measured and what it would cost, rather than after the fact the legislation is changed and we have to see what happened and there is nothing in place. So it gives us some lead time to plan for that and for government to respond to what it might take to evaluate it. We try to anticipate that and get engaged at an early stage.

CHAIR —Except that when you are operating on the basis that you believe there have been some substantial impacts from legislation to the detriment of many in the community, you want to act quickly to address that. You are not going to sit down and wait until you plan for how you are going to evaluate whether or not you have succeeded.

Mr McColl —What we are suggesting is that there should be some plan of how it will be evaluated before the legislation comes forward, because I would have thought the representatives of the people would want to know how we will know what success will look like and therefore we should be part of the planning process for the measurement in advance as well.

CHAIR —Are you in a position to give a comment about the AWIRS survey, which many people have said to us today they missed greatly and think should be reinstated, or is that asking you to say something beyond your positions?

Ms Taylor —We know there is keen interest in it, and I know there is a keen interest in having the ABS run it, but the issue for us is funding and competing priorities. For example, we have a labour statistics advisory group and several members on that are fairly keen to see another AWIRS survey, but when you produce lists of priorities, while it is important, it is not at the top of the list of priorities either. So we are aware of that but we have no immediate plans to do anything in that regard because we do not have the resources to do that.

CHAIR —Are the labour advisory group’s recommendations about priorities something that is publicly available?

Mr Sullivan —They are an advisory group. Essentially the process is that we present to them our intended work program and where we are going and they advise us on that. Those papers are certainly freely available. They are not on our website or anything but there are no issues about the circulation of those. They also have the opportunity to inform us on data needs that we are currently not meeting. So the AWIRS information base is the sort of data they would be interested in. Whether it can fully inform on pay equity, I am not sure. I think it is very much another dimension of EEH, in some respects. It really would be a welcome addition to the dataset, I suppose, to inform on pay equity, but, if your expectation is that AWIRS will give a full insight into it, I do not know that that would be the reality yet.

CHAIR —No, I do not think we are at the stage yet where we will ever get a full explanation for it, but it would be nice to be able to identify the most significant contributing factors. If you are going to try to address it in the policy sense, I think we are at the stage where there is an acceptance that the gap is there. For a variety of different reasons, are there any social responses we should have to that in the interests of equity and bringing it closer together?

Mr Sullivan —The only other observation about AWIRS is that it was conducted by, I think, the department of industrial relations at the time. It was not an ABS survey. We assisted them in the process of selecting the sample. But I am not fully across the issues of—

CHAIR —I think that is where Mr Haase’s question was going. DEEWR collected a whole range of information and data that they are using for various purposes, which does not seem to sit that comfortably with the APS’s information that it is collecting about its own workforce—which certainly does not sit that comfortably with the information that the Community and Public Sector Union had. With one workforce, the Australian Public Service, you would think the Commonwealth would be able to give pretty good statistics that gave a picture of the nature of its workforce, the hours and the income differentials across the board. We were not able to achieve that, so maybe there is a bit more work for you to do out there in collaboration with your own Commonwealth agencies. I am very conscious of time commitments.

Mr RAMSEY —Do you have the ability to break these figures down—although I do not know how important it is to this inquiry—into a rural and regional focus at all?

Mr Gerrity —The household survey data we can.

Mr RAMSEY —Is that difficult to do? I do not want to give you a huge job.

CHAIR —It would be very interesting to the committee if the gender pay gap is significantly different in regional and rural areas as opposed to metropolitan areas.

Mr RAMSEY —That is what I represent. There are so many differences in the way the workforce is structured.

Mr Gerrity —Table 11 in the submission charts capital city versus balance of state and territory. It shows the gap is not particularly large, but that does not cover the different states.

CHAIR —We have already had evidence that, for example, Western Australia’s pay gap was the greatest of any state. Arguably that is because its largest industry was mining, and the rates of pay for men in mining were the highest rates of pay.

Ms Taylor —We could do a capital city figure and a balance of state figure. If you want us to go down to individual regions then our samples get so small that you could not do that. We could do some kind of aggregation for you but not detail.

Mr RAMSEY —That would be good, if it is possible without overloading the system.

Mr McColl —Madam Chair, in relation to your question about the evidence from Adelaide about older women wanting to work more hours, the survey that Sue talked about captured preferences for working hours. About 15 per cent of women, given how much they were paid, would have preferred to work more hours than they currently were. That 15 per cent is not cross-classified by age or numbers of hours currently worked at the moment, but it can be. We will have a look at what we can pull out to give you a look at that as well.

Ms Taylor —We asked the underemployed and those people who are not in the workforce but would like to be: what do you think is stopping you? You can pull out that kind of information.

CHAIR —That might be useful. We have heard that in South Australia there are over 60,000 women who could be working more hours. If you need that specific reference I can try and find it and get the committee to pass that on to you.

Mr McColl —Australia wide, our survey found 673,000 women of all ages would prefer to work more hours. Table 11 cross-classifies it a lot of ways, but it does not do that one.

CHAIR —One of our briefs is, with the skills shortage, how can we increase that participation? If we have 673,000 who would be happy to work more, it is a question of how we get them where we need them.

Mr RAMSEY —That means it is 7,000 in Western Australia and 60,000 in South Australia. That is certainly a concern!

CHAIR —I am not one of those statistics!

Mr McColl —That was 670,000 not 67,000.

Mr RAMSEY —Sorry, I thought you said 67,000.

Mr McColl —No. My apologies.

CHAIR —Thank you very much for your attendance here today. I know we have asked you for some additional information and I would be very grateful if you could forward that to the secretary. I think I have forewarned you a number of times that there may be additional questions. They will be forwarded to you by the secretariat for your response. You will be sent a transcript of your evidence today, to which you can make corrections of grammar and fact. Hansard may wish to check some details concerning your evidence, so please check with them as you leave here today. On behalf of members of the committee, I would like to thank you very much for your attendance here today. It is most appreciated. I especially thank you for the presentation.

Resolved (on motion by Mr Ramsey):

That this committee authorises publication, including publication on the parliamentary database, of the transcript of the evidence given before it at public hearing this day.

Committee adjourned at 12.40 pm