Skip to main content Start of content

[Recorded by Electronic Apparatus]

Tuesday, May 30, 1995



The Chairman: Maybe we'll get started then. My name is Gar Knutson. I am the chair of the subcommittee.

I want to welcome Mr. Benjamin. As you are probably aware, we're looking generally into the issue of the economic impact of immigrants to Canada and our immigration policy. We're using as a starting point the book Diminishing Returns. So we want to invite you to make whatever comments you like, and not necessarily exclusively on your work on Diminishing Returns.

I don't know if you have appeared in front of a parliamentary committee before, but you can make a presentation. Normally they are 10 or 15 minutes long, and then we throw it open for questions and try to get some dialogue around any issues that need to be clarified or are contentious.

So thank you very much for coming.

Professor Dwayne Benjamin (Author, ``Labour Market Outcomes and the Participation of Immigrant Women in Canadian Transfer Programs''): Thank you.

What I would like to do is begin with a presentation of research that is addressed in the Diminishing Returns book, but also present a more broad set of research that I've been involved with.

There are handouts that essentially reproduce the overheads in case the overheads aren't easily seen.

What I'm going to talk about today is the participation of immigrants in transfer programs. The work I've been doing at University of Toronto is with Michael Baker, who is my co-investigator on this, and we have a chapter in the Diminishing Returns volume.

The work on immigrant transfer program participation is really only a small part of our more general interest in what we call immigrant assimilation, or more specifically the dynamics of immigrant adjustment in the Canadian economy.

The main focus is looking at immigrant earnings and how they function in the labour market and so on.


But an important offshoot of looking at the dynamics of earnings is to look at participation in the various income maintenance programs. Certainly performance in the labour market, or adaptation in the labour market and participation in the transfer program, is going to be totally interconnected as well as being independently interesting. As well, there seems to be a lot more policy interest or public interest in the transfer program participation than there is perhaps in the earnings performance.

What I'm going to do today, just to keep things focused a little bit, is review results from two papers that Michael and I have done. The first one is called ``The Receipt of Transfer Payments by Immigrants to Canada''. It's really the main set of results that we have in the area of transfer payments. That's forthcoming in the Journal of Human Resources. The paper that's in the Diminishing Returns volume is ``Labour Market Outcomes and the Participation of Immigrant Women in Canadian Transfer Programs''.

The main distinguishing feature between the two papers is really just the different units of analysis. In the main paper about the receipt of transfer payments we look at the UI behaviour of men, family receipt of transfer payments, and family receipt of grant subsidies, whereas in the other paper we focus exclusively on women, because there are a number of issues involved there.

By the way, if there are any questions along the way or if something is confusing, please just ask away. The first thing I want to do, because I think it's important to understand where I may come up with any suggestions or comments, is to give you an understanding of the methodology that labour economists might use to look at some of these questions. I think it's quite important to understand the limitations of what we know in Canada or anywhere about immigrant assimilation at all, let alone something like program participation.

I'm going to talk about results for really three very different programs: first of all, welfare or social assistance, which is the one that most people focus on in the media; second, unemployment insurance, which is a lot more complicated in some ways than looking at welfare; and finally, I'll address something we don't know very much at all about and is hardly ever looked at, the receipt of rent subsidies, which is another transfer program. At the end, I'll offer some conclusions to hopefully motivate some further discussion.

As for the methodology, as I say, our main focus is looking at dynamics. That means trying to look at the initial position that immigrants may enter Canada in, and then looking at their adaptation in the economy over time. So what we're really trying to estimate when we look at data on immigrant participation is what we call the assimilation profile, for which we're really just taking the wording out of the earnings performance literature. ``Assimilation'' really just means convergence to the mean outcome of the native-born outcome. Usually, as I say, we're talking about earnings.

So there are three features we're going to try to focus on here when it comes to looking at program participation. The first of these is what we call the assimilation rate, which is really simply looking at the question of how immigrant participation transfer programs change over time in Canada; specifically, whatever the initial conditions that immigrants face when they come here, do immigrants become less reliant on transfer programs with their years in Canada. I am just going to use the computer variable I call YSM, which stands for ``years since migration'', so I may revert to the shorthand of calling it YSM. But it really is just looking at this effect over time of immigrant participation.

The second thing we look at is specifically these initial conditions, what we call cohort effects. I'll show a diagram on that in a second. The basic idea when we're looking for cohort effects is to see whether or not there's evidence that the initial conditions for immigrant program participation are changing over time and in particular are they deteriorating over time, since in the U.S., where this literature was initiated, that was the main concern. They called it ``cohort quality'', which is a value-laden term, but it really just means how are immigrants performing and are they starting off worse than they used to.

The third important feature of the assimilation profile we want to examine is the differences between immigrants and natives. It doesn't seem to make sense to me to talk about immigrants completely in isolation in terms of any behaviour unless you have some kind of a benchmark, and we tend to use a native-born Canadian benchmark. People can argue about whether that's the right benchmark or not, but that's what we think is at least an interesting starting point.

Mr. Hanger (Calgary Northeast): Why would somebody consider that to be a questionable benchmark and not relevant?

Prof. Benjamin: The main reason for that would be - and this is argued in the U.S. to some extent - that perhaps it's the case that we don't really care if immigrants are actually less reliant on transfer programs than native-born; all we care is that they use any resources at all. So instead of looking at cost-benefit, we say if they impose any cost why admit immigrants? That's the extreme case as opposed to saying, well, if they're no worse than natives....


For each of these three questions there's another overriding thing that we address in our research. This is of course the hard question, which we can never really nail down, but we get some information on it. That is how much of the differences between immigrants and native-born Canadians or between immigrant groups over time could be accounted for or ideally explained by differences in labour market outcomes. So we're really interested in linking these two things throughout, or at least seeing how much differences in take-up rates can be explained by differences in labour market outcomes, which of course leads more directly to questions about admission criteria and so on.

Let me show you this diagram I promised. This is really just to show you what we're doing. It's actually a very simple exercise if you have the data to do it. What we want to do in the best of all possible worlds is track a group of immigrants at a given point in time, follow them through the years in Canada. That's what you'd like to do, but the data rarely permit it. Most of the initial studies, and in fact some of the studies in the Diminishing Returns book, look at just one point in time and try to infer what the dynamics are like.

So you can imagine looking at two groups of immigrants that differ only in the fact that one's been in Canada five years longer than the other, and you look at them at a point in time in 1985. Just imagine that this vertical axis is earnings. You would infer that the difference in earnings between those two groups was the effect of being in Canada for five years, because that's the only difference between those two groups; one's been in Canada five years longer.

That of course assumes that this immigrant group started out here and their earnings grew like that, that this immigrant group will start here and their earnings will grow at the same rate and everything. Or you could just take this graph and relabel it and call it program participation.

But that assumes that the initial position of an immigrant group is not changing over time. If you're going to infer anything about dynamic behaviour from a single point in time, you have to make a lot of assumptions. Of course, even for what we're doing you have to make assumptions, too.

Just to illustrate what a cohort effect is, if you have this gap here, earnings between people of two different vintages in Canada in a single year, it may be the case in fact that this group's earnings didn't grow as much and they didn't start quite as low as this group here. The gap is the difference in the assimilation, what we call the five-year assimilation, plus this initial condition, this lower position they are in to start with. So in order to actually measure and disentangle the growth or assimilation part from the initial conditions, you really need to use more than one cross-section, and that's why we have to use several different data sets.

Mr. Nunez (Bourassa): Could you not speak too fast? I can't get the translation.

The Chairman: The other thing you have to remember is that you're not dealing with people who are skilled in economic terminology.

Prof. Benjamin: I understand. This is the only really technical part. It's just to explain why it's hard to learn something about dynamics from a cross-section and why we had to jump through the hoops we did.

So there's the simple explanation of cohort effects. What we're really trying to do is just measure those lines. What we actually do is we have data from the Survey of Consumer Finance, which is a micro-data set available from StatsCan. It's a random sample survey of about 60,000 households, and we're going to follow samples of cohorts of immigrants who arrive in Canada at different points in time and track their use of transfer programs from 1985 to 1990.

The Chairman: How do you ask someone whether they've used a rent subsidy?

Prof. Benjamin: Actually, that's a good question. There's a different survey than the Survey of Consumer Finance; it's called the Household Income, Facilities and Equipment Survey. They ask a whole bunch of things, such as whether you have a VCR, and a bunch of other questions about the stock of appliances. One of the questions they ask is whether your rent is subsidized. They ask if you're paying rent or if your building is such that you pay rent below the market rent. It's self-reported.

The Chairman: They're asking these people where they were born?

Prof. Benjamin: They ask people whether they're an immigrant to Canada and where they were born. In the census they ask you the details of what country you were born in, and they also ask you when you arrived. So the information we're using is whether they are immigrants to Canada and when they arrived. Then we group people into intervals based on when they arrived. Those are the cohorts we're following.


The Chairman: Right. And you surveyed how many households?

Prof. Benjamin: This is a Statistics Canada survey. They surveyed 60,000 per month. There's a labour force survey that is conducted every month. That's how they compute the unemployment rate. Every year the Survey of Consumer Finance asks a lot of detailed questions about where families received their income that year, including labour earnings, self-employment earnings and transfer program participation.

It's a valid question too as to what the quality of the data are. Statistics Canada has spent a fair bit of time evaluating the transfer income component in particular, since that's pertinent to what we're doing, and looking at underreporting of unemployment insurance and welfare. It is certainly the case that unemployment insurance is underreported by about 15%. Transfer income levels are underreported by about half. The incidence, which is whether or not you receive this form of transfer payment, is all we're all looking at and is less poorly measured than the actual dollar amount of income.

Mr. Hanger: On that particular point, wouldn't it have been easier to go to income tax rather than going through the collection of data at this point and having certain questionable assumptions to make? I know they have all that data available and then you would clearly see if there were unemployment insurance -

Prof. Benjamin: Correct.

Mr. Hanger: Even it's a check and balance.

Prof. Benjamin: That's what Statistics Canada uses to check theirs. They know the actual administrative record. That means they know exactly how much. They know what the sample is and they can inflate it up and see exactly how much unemployment insurance it should add up to. That's how they check it.

Ideally, of course, it would be nice to have access to the income tax records and all kinds of stuff. The data would be much better. There is another data set used by one of the other studies in this book, but it has its own limitations. We were talking about this earlier, because at this point it's not available in a micro-data form like this is, where I can actually look at the individual or at a family and look at every activity the family did, including their education levels. I also have native-born and immigrants. I don't have immigrants from only the 1980s, like that other data set. This has its own problems, but the other data set will have its own problems as well.

The Chairman: Professor, if you'd like to finish, we'll throw it open to questions.

Prof. Benjamin: The story is actually very simple in terms of pictures, but of course, as you know, with graphs you can always be a little misleading, so I do want to spend a little bit of time interpreting it.

First I'm going to talk about social assistance and welfare participation. The first unit analysis is families, because this is really the main unit upon which social assistance is evaluated. It depends on family income and so on. This looks a little bit complicated, but all the graphs are going to be the same. Only the variables are different.

I'll show you how to read it. What we have here is plotted. Let's look at figure 2, the incidence of receipt as a percentage of the group that receives social assistance. The first group is Canadian-born. The dark line refers to the percentage of receipt in 1985. The next line is 1990. We're interested in the change from this year to that year, because we don't want the contamination of cohort effects. You can go through each of these groups, these immigrant cohorts - and 56P just means prior to 1956 - and the 1956 to 1965 span shows the immigrants who arrived in that time period, etc. Of course most of the interest is really in the more recent immigrant cohorts.

We're interested in a bunch of different things. First, did the reliance or use of the transfer program for a given cohort diminish with time or did it go up? Of course we want a benchmark, and a useful benchmark will be what happened in the rest of Canada with other native-born Canadians.

We're also interested in the relative position of the different cohorts because that tells us something about changes over time in the actual use of transfer programs by immigrants.

What you can see right here is really the main part of the story that we're going to see for welfare. For most immigrant groups, even up to the immigrant groups with the highest use of social assistance, their use barely reaches the level of that of native-born Canadians.


So just as a first cut, there is no evidence at all in these data that immigrants use social assistance more than native-born Canadians. That's the first important finding, and that's going to survive even when we look at other specifications; we control for age and all kinds of things like that.

The second important finding - and this again is going to carry through - has to do with the assimilation, or the changes over time. As you see, almost every one of these cohorts has an increase in the use of social assistance over the five-year period 1985-90. Now, of course this may be a special period. Who knows? We can never tell. But at least if natives are used as a benchmark, there seems to be increased reliance on social assistance by immigrants.

The numbers are still low, and lower than for native-born, while the newer cohorts...and particularly the cohort that arrived in 1986-90 sticks way up there as having the most reliance on social assistance.

We also looked at the same results for women. I'm not going to spend as much time on this, because otherwise we will not have time to talk about it. Basically the pattern is very similar. The numbers, though, are lower in families than for women alone - not surprisingly.

Well, why might this be the case? The most interesting question is really why the performance of immigrants seems to be deteriorating and why we might be observing this increase in program participation.

To answer that question, let's look at what is going on in the incomes of these households. The first thing is to look at earnings, because family earnings are really the main input, so to speak, if somebody goes to a welfare office in their community to apply for welfare.

This also gives you some idea about the overall picture of how immigrants are performing in the Canadian economy. This is family earnings from self-employment and the labour market, so it doesn't include income transfers and so on. It's just what immigrants are earning.

There are the native-born numbers. You can see most of the other immigrant groups that arrived up to the 1970s are already earning more than the average Canadian. You can also see that immigrants' earnings are going up faster than those of the native-born.

This is what we mean by assimilation of earnings. Immigrant earnings generally grow faster than native-born Canadians'. That's a standard result, a standard finding.

Some of the cohorts' earnings aren't going quite as fast. You can can see some of these cohorts, particularly the more recent ones...they haven't been in Canada as long, so you wouldn't expect them necessarily to have the same level of earnings here. This cohort, as you will see, sometimes doesn't do particularly well. This cohort, surprisingly, based on the other assimilation rates...their earnings grew very rapidly over that last half.

So one might be very pessimistic if one looked at data from before 1985 about how this group of immigrants would do. In fact, Michael Baker and I had a paper where we were somewhat pessimistic. But as it turns out, they actually did fairly well, although they haven't caught up.

You can see the initial position here of this cohort - these are the immigrants who arrived from 1986 to 1990 - in 1990 is a little better than the 1981-85 cohort was in 1985. So their starting position is about the same. It hasn't got any worse.

So the income side doesn't look too bad. Unfortunately, that's not all that goes into determining welfare use or, in some sense, how well off families are going to be, because we haven't made any adjustments for the number of kids people have, the family size, where they live, the cost of living, and so on. So one other measure of how well families are doing is to use some measure of poverty.

One measure of poverty - it's not necessarily a perfect measure - is what's called the low-income cut-off, or LICO, a StatsCan measure. In this data set we have an indication whether or not the family's income is evaluated as being below the cut-off.

You can see that when you look at the more recent cohorts, again, the percentage of those who are actually what we'll call poor is very high, and much higher than for native-born Canadians. It's also the case that for many of these groups, with time in Canada the percentage who are poor goes down, but for this particular cohort people got poorer. This cohort, as we saw, did quite well, and fewer of them were poorer.

But the most recent cohort, even though they didn't do so badly in terms of the average household income, are much poorer. You're talking about 35% to 36% of families that arrived in Canada at that point being classified as being below the poverty line. The native-born are around 18%.


We do a bunch of statistical things to try to control for age differences across cohorts, because people aren't all of the same age. They don't live in the same places, and so on. This is just a very brief summary of the kinds of things we found. In order to save time, I won't go into all the details.

But what did we basically find? In terms of the dynamics, the YSM or assimilation. for families we find positive assimilation. That means that with years in Canada, immigrants become more reliant on social assistance. That is independent of whether you control for even their economic outcomes.

The second thing we find is that there are cohort effects. That's what the little stars mean. This means that each individual cohort is different from the others. In particular, we find a pattern of deteriorating outcomes, if you view collecting welfare as a bad outcome.

Finally, in terms of the differences between immigrants and natives - this is quite important to note, though - despite that grim picture, once you adjust for differences in these groups, the differences are always negative, except for the most recent cohort, IM8690, which showed no difference from natives. Everyone else uses it less than natives do.

So as grim as the picture we can paint for immigrants is, if you use the native benchmark, they don't look so bad in terms of use of welfare programs.

There are a couple of other minor points worth mentioning. The differences from the native-born that we observe are, to some extent, explained by the differences in their income levels. Immigrants generally are poorer than native-born, particularly the most recent groups, and that explains some of the difference between the two groups' take-up rates.

Because the assimilation profile is basically unaffected by controlling for their economic outcomes, it suggests that, whatever it is that's causing increased use of welfare programs, it's hard to explain. You can't explain it by changes in their economic position. We don't know what would explain it.

Of course, one thing that's important is that some of this sort of grim news may be transitory, because we used the 1986 census to look at the question of why we see so much welfare use. Ideally, what we'd like is a data set that had all the classes of immigrants, refugees and so on. We don't have that.

We used the 1986 census and found that immigrants from countries that had a higher fraction of refugees coming to Canada collected welfare more than other sets of immigrants, which may provide a link with the stock of refugees and backlog of refugees that occurred in the late 1980s but has subsequently diminished. That might have been associated with the increased reliance on welfare. So we don't know. With more recent data we could find out what really happens.

Mr. Hanger: Could I interrupt you for a moment?

Prof. Benjamin: Sure.

Mr. Hanger: Why did you include refugees in there? They're in a separate category in the sense that their needs are going to be different from those of the average immigrant that's accepted into the country.

Prof. Benjamin: Right. We wondered if this was the case. We observed this high use of welfare by immigrants in the 1986 to 1990 cohort. We thought: this seems to be quite a quantum leap from the previous groups.

One thing we know about that group is that there was an increasing backlog of refugees who did not have permission to work. So they would have no choice but to go on welfare. We just thought: well, if that's what's going on, then we can't really infer anything about immigrant behaviour or the labour market from that. We wanted to see whether there was some, at least, indirect evidence that it was restrictions on work or refugee status itself that had something to do with collecting welfare.

Again, without using administrative data, all we can do is say something as vague as that those who came from countries where there was a larger fraction of refugees, where the immigrants from that country were mostly refugees, are more likely to use welfare, which is really not very clean. You wouldn't be able to convict refugee status for causing this pattern.

Mr. Hanger: What I'm asking is this. The needs of the refugee are well-known. They're going to be different from those of the average immigrant population. That's why I'm surprised that you would include it in your data.

Prof. Benjamin: It's because we don't know.

Mr. Hanger: It's going to skew the results in a way.

Prof. Benjamin: Absolutely. We have no idea of what class of immigrant anybody is. We just know when they came to Canada.

Mr. Hanger: All right. That was it.

Prof. Benjamin: The IMDB database, which is the more administrative database, has its own problems. But one of the good things about it is that it does have the class of entry, so you can sort out issues of family versus independent class, which is really the central question.

That's the welfare. The other program, of course, is a little bit more complicated, and that's unemployment insurance.


First of all, we'll look at men. The reason we look at men and women separately is that their labour force patterns are quite different, as are the eligibility rules for UI, in terms of maternity and so on. It's a little bit cleaner in some way, a little easier to look at men.

The graph is much less clean in terms of the nice, simple patterns we observed with the social assistance. You can see the one thing that does emerge is that overall immigrants use unemployment insurance less. That's something that survives any kind of statistical analysis, as you'll see, though we also see increasing reliance in the sense that more recent immigrants use unemployment insurance more.

But the other thing that is disturbing, as was the case in welfare, is that you can see increased use of unemployment insurance by almost all immigrant groups. So the dynamic side is a little bit puzzling, but again, it's not as if it's leading to levels of use higher than natives. It just means that in terms of assimilation and convergence they're becoming more like native-born Canadians. So that's the basic picture for men.

The one thing to remember, though, when you're thinking about unemployment insurance is that, unlike welfare, you can't just go and apply for it; you have to qualify for it. So you have to actually work enough to qualify to collect unemployment insurance. The level you're going to get is going to depend on your earnings level. Your wage rate's going to determine different sorts of things. So it's important to try to adjust for those differences and look after those differences.

The other thing that's important in order to collect unemployment insurance is that you have to actually be unemployed. For women, we have pictures that just look at these two issues. The women's pictures look similar to the men's.

The first thing you can see is that over time in Canada immigrants become more likely to report unemployment. There is the positive assimilation. Again, the 1990 graphs are higher than the 1985 graphs. So if people are more likely to be unemployed with time in Canada, this is kind of strange.

They're also more likely, generally speaking, to become more eligible for UI, because they're working more; they're engaging in the labour market more.

So these two things are going on at the same time as the positive earnings growth and everything else. It makes the picture a little bit more complicated when you recognize that the participation in UI is not just having low income; it means qualifying, working, earning enough, and so on.

To summarize the unemployment insurance results, then, you get very similar results for men and women. First of all, in terms of dynamics you get a very strong assimilation profile for men, meaning that there are very significant increases in use of UI by immigrants with time in Canada. Again, they don't exceed native-born, but they do grow over time.

The second thing is that there are strong cohort effects. That means the initial positions of immigrants or the current positions are quite different, controlling for years in the country again, in that the more recent immigrants are more likely to use UI than previous immigrants. This is controlling for province of residence and all kinds of things.

The final thing is the difference for natives. You can see, again, it's uniformly negative. It's still lower than native-born. If you use the native-born benchmark, the use of UI doesn't really look so bad.

Let's look at some bottom-line differences. For men, the differences between immigrants and natives are largest when you account for differences in their economic outcomes, meaning their eligibility and their wage rates and earnings and so on. What that means is that assimilation into UI, the growth of increased likelihood of using UI, is not explained by immigrants becoming more eligible because they work more; it's something quite different, in some sense unexplained.

For women, the differences are only somewhat reduced by controlling for economic outcomes, which suggests again that whatever is explaining this assimilation, it isn't just that immigrants are accumulating qualification for UI.

The cohort effects for both men and women are consistent with the deteriorating position for immigrants. That's another finding we have. That's the unemployment insurance picture. As I say, it's a little harder to interpret.

On the last program, there are no graphs. I could show you graphs, but it's not worth it. It's in the handout.


Immigrants are more likely to receive rent subsidies. So they're more likely to live in public housing, basically. That's even controlling for where they live. But they also, with time, assimilate out of subsidized housing. There isn't even evidence of cohort effects, which we expected to see.

When you think about rent subsidies versus these other transfer programs, it depends on the market rent. In Toronto, it's not impossible to imagine that the rent subsidy might be in the order of a few hundred dollars a month. When you add it over a year, you're looking at about the average welfare recipient's receipts from welfare, or even UI. So this program, conditional on receiving it, although it's not as widespread as these other programs, is actually still big dollars.

What are the conclusions? Basically, there is no evidence that immigrants make a disproportionate use of transfer payments than the native-born, though some of the most recent cohorts are as likely as the native-born. Generally there is evidence of cohort effects, which suggests that more recent cohorts are more likely to use transfer programs than their predecessors. So there seems to be some deterioration in the position of immigrants in Canada.

There's also some evidence, at least for social assistance, that some of this deteriorating position may be the result of an increased percentage of refugees, or a backlog that occurred just at that point in time.

But the most puzzling result for us is the dynamics, the growth, the change, the general trend towards increased participation in programs, especially UI, with years in Canada. Perhaps we're a little facetious about this, but one way to think about this is that perhaps part of being Canadian is participating in these programs, that when immigrants first get here they are either unaware of the programs or they don't know the parameters of the programs, especially UI. They didn't have UI where they came from, and it's a big surprise to them to find out you can actually participate in this particular manner. So this could be due to increased awareness of program parameters, or some other unobservable Canadian trait that we don't know.

What we do see, though, is this evidence of convergence. Immigrants start out not using UI, and they grow to the levels of native-born Canadians, even though their economic circumstances are improving.

Those are the basic conclusions from our research. If you want to see what relevance they may have, we can talk about that.

The Chairman: Thanks very much. We'll turn the floor over to Mr. Nunez.


Mr. Nunez: Do you have any specific data on the various regions where the immigrants settle? For example, do the immigrants settling in BC use UI or welfare more or less than those who settle in Quebec? Do you have any stats on this?


Prof. Benjamin: I don't have the exact figures, but let me dig out some numbers here. We have specifications where we control for the region of residence for everybody in the same way, which is to adjust for the fact that in different regions there are different rates of use of any of these particular programs. We also have a specification where we allow immigrants to have their own particular propensity in each of these regions.

I have a table I can provide. It may actually be in the Diminishing Returns book. There are also some of these appendix tables, which most sane people would avoid reading because there is just an unbelievable amount of numbers.


What we do find are the overall differences by region. If you want I can rank them. They show the actual total use of programs and are not specific to immigrants. As I say, we control for the fact that in B.C. immigrants may use programs more or less than they do in Ontario, but I don't have those actual numbers here. I just have the overall figures of total use in B.C. relative to total use in Ontario and Quebec and so on. But we do adjust for that. It doesn't make any difference. The overall picture of immigrants stays pretty much the same.


Mr. Nunez: As far as the country or region of origin is concerned, for immigrants and refugees, would you say that Europeans make less use of social assistance, unemployment insurance or rent subsidies than people from Africa, for example? Do you have information on that?


Prof. Benjamin: We don't have information on country of origin in the Survey of Consumer Finances. The only source of data that has the country of origin mixed in with program participation is the census. While we've looked at the role of the source country in terms of earnings assimilation, we haven't looked at the role of the source country for program participation.

The one indirect piece of evidence we have is the result I reported earlier, that there is a correlation between countries that provide more refugees and immigrants using welfare more. I can't remember exactly which countries had the most refugees, but it's not usually Europe. It tends to be Africa and South America, as well as parts of Asia. But that's the only pattern we have. We haven't really studied it broken down by country of origin because the data just really aren't there.

There is another study using administrative data that has a little more evidence on source country. We're only now looking at source country even for earnings in some sense, because it's a riskier area to be looking at.


Mr. Nunez: Today, our immigrants and refugees have an easier access to the Canadian social assistance system, but they still do not use the system as much as the average Canadians. Do you have any recommendations to make on this issue? Do you believe that this trend could stop or even be reversed? For example, could we use the selection process of immigrants and refugees to slow down or stop this trend?


Prof. Benjamin: That's really the $64,000 question. It's hard to imagine that it would continue. It deteriorated quite a bit over the 1980s. One possibility could be that the immigrants themselves are of the same average quality as every other immigrant has been over time. In fact, when we look at earnings we look at anything we can in terms of the observable characteristics of immigrants. It's clearly important to know the average education level of immigrants, since that's an important selection criteria.

Mr. Nunez: The education level of immigrants is higher than the....

Prof. Benjamin: It's higher, and it hasn't really gone down much over time. So at least from the point of view of observable selection criteria, there's no simple answer to how increased selection would affect the outcome of immigrants.

A lot of different hypotheses have been offered as to why we observe these declining outcomes. One could be that the capacity of the Canadian economy has changed.

We have been looking at the role of ethnicity recently and the different performances of different ethnic groups and the first-born or native-born ethnic groups. We can see that among immigrant groups certain occupations used to be very important, at least anecdotally, in terms of the ways people could get jobs if they didn't have skills. Construction is an example. There are very few immigrants in construction from a lot of different countries.


If you take away those low-skill, high-paying jobs, we know that overall in Canada, and particularly in the U.S., the actual earnings of individuals who happen to have low skills are dropping over time relative to those with high skills. So it may be the case - we don't know - that one of the reasons we're observing immigrants doing worse is that if you come to Canada and don't have a lot of skills, you're just going to get hurt a lot more now than you would have if you had arrived in the 1960s when there were good, high-paying, low-skill jobs.

One piece of evidence to support this is something that could be looked at in more detail. If you remember, we looked at the overall earnings of families and saw that they were going up for all immigrants, even the more recent ones, but the poverty rates were going up as well. That suggests that while the average may be doing fine, there is a group of immigrants that is doing particularly badly.


Mr. Nunez: This means that economic categories do exist, that is to say that the gap between the rich and the poor does exist also within the immigrant population. This is true in Canada, in the US and perhaps everywhere else. There is more poverty, but there are also people who are richer than before. Do you see the same trend among immigrants?


Prof. Benjamin: Yes, one way to classify it is the gap between the rich and the poor, but it's highly correlated in terms of the gap between the highly educated and the poorly educated. It isn't just that the rich are getting richer and the poor are getting poorer in some simple way. There are the returns on schooling. If you're highly educated, your earnings are going up, and if you're poorly educated, your earnings are going down or staying flat.

People don't know why that's occurring, but it's one of the explanations of why you're seeing this increase in inequality that's highest in the U.S. It's not as strong in Canada, but there is still this widening of the gap between high and low earners in Canada. When immigrants came to Canada they used to be able to just go along with the low-skilled individuals and grow with everyone else. But if low-skilled individuals in Canada aren't growing much either, the low-skilled immigrant has two strikes against him.

That may explain what's going on here to some extent, which does suggest the possibility that selection criteria might help. You could try to predict who's going to be on the bottom and not admit such people if you wished to do so. But that's just speculation based on these numbers. We don't know whether there would be such an observable characteristic that you could use to predict performance in the Canadian economy. It just appears that differences are emerging.


Mr. Nunez: Are you going to complete your research? For example, are you going to have any stats for 1994? Do you think the situation keeps changing, even today?


Prof. Benjamin: It's hard to imagine that things wouldn't have changed somewhat, because 1990 was not a bad year and the time period from 1985 to 1990 was one of big expansion. So one would imagine that things would have been good before the recession.

During the recession, immigration levels were still high, and we observed this increase in dispersion of returns to education or skill or wages. I would guess in looking at the numbers from 1985, which we'll be able to do very shortly, that things won't have improved.

The only caution I would put on that is we had predicted, based on the 1986 census, that those in the group who arrived from 1981 to 1985 were going to do terribly, but they surprised us. They did very well. So I'm reluctant to make any forecast without any information. My guess would be that I don't think things will have turned around, but they may not have deteriorated more.


Mr. Hanger: You state that there is mounting evidence that the labour market is losing its ability to provide new immigrants with the economic opportunities enjoyed by their predecessors. So what should Canada do to change its immigrant intake?

Prof. Benjamin: That's a hard question. It does appear...this isn't so much a result of the research here, that we feel there's evidence of the declining ability of the Canadian economy to provide all immigrants with good jobs; it's more a study of earnings that shows that.

As I say, except for the surprisingly good performance of the 1981 to 1985 cohort...and we don't know why they did so well. We have a graduate student currently looking to see why they did much better than expected. But we don't really know the source of this decline. If it is something along the lines that the ``skilling'', so to speak, of the Canadian economy, the increasing needs for certain sets of skills, particularly high education levels...if that's what is really the explanation, then clearly if one wants to formulate immigration policy primarily on the basis of the potential success of the immigrants, then it would probably be more important to use assessment criteria like those applied to the independent class.

Mr. Hanger: If you're going to do a study of this nature, I would look for some conclusion in this respect: are we going in the wrong direction? Is the point system we're using right or wrong, or should we use some other gauge?

Prof. Benjamin: George Borjas, who's an economist in the U.S., has done a study comparing Canada with the U.S. on immigration policy, patterns and outcomes of immigrants, and so on. He doesn't feel the point system has made a big difference. It doesn't really apply to a high fraction of the immigrants much any more. He uses the U.S. as a comparison, where family reunification has always -

Mr. Hanger: What do you think?

Prof. Benjamin: What do I think? Well, patterns almost identical to those we have observed in Canada have been observed before in the U.S., where they don't have a point system. My speculation would be, then, that if you want to minimize the chance that people are going to ``fail'' in Canada, or that they're going to have to use these support systems, then it's probably better to apply the point system to more people. That would be my guess.

The only question - and as an economist, that's what I would say - is there are other reasons and other policy concerns. I recognize that in immigration policy family reunification and refugee policy is not really primarily motivated by economic concerns. But I think we do have to recognize - and I think we all have - that the days when people can come to Canada with nothing but the shirt on their back as a poor peasant farmer from somewhere and actually expect to perform in Canada...that just can't occur any more. The economy has probably changed so much that you really do have to have a certain set of skills matched to the Canadian economy - although a lot of entrepreneurial skills and other kinds of skills which may or may not appear on any form that can be applied on the point system would also probably be very welcome in Canada.

I think of all the.... Again, it's purely anecdotal. We've tried to look at self-employment and stuff, but living in Toronto...there are a lot of self-employed immigrants who probably didn't have skills.

Mr. Hanger: You would suggest the point system be maintained still on the evaluation of independent immigrants?

Prof. Benjamin: Yes.

Mr. Hanger: It has some merit.

Prof. Benjamin: I think so. We really don't know a lot...and it's stunning that we don't, but there's never been a data set that has allowed anybody to track the earnings of immigrants by their class of entry. Those kinds of information will be available in the next year or two, because the database has been assembled. But even it's going to be imperfect, because the database just has income, it doesn't have anything to doesn't have a lot of education information. There's no information on native-born and so on.

Mr. Hanger: In view of your findings about immigrants' use of social assistance, and mostly centred around the area of the sponsorship aspect, how do you feel about the proposed policy of a bond, a financial guarantee?

Prof. Benjamin: That would presumably apply only to the sponsored class. I can't really say what I think on that one, in terms of the following. We really don't know whether the family unification class is more likely to use welfare than the independent class, and for the independent class the bond wouldn't apply.


Mr. Hanger: It's my understanding that the most widely accepted area is the sponsorship, where there's a breakdown in the commitment of the sponsor and the subject sponsored of course is on social assistance.

Prof. Benjamin: In which case you're saying that the sponsors have effectively reneged on a commitment.

Mr. Hanger: For one reason or another.

Prof. Benjamin: Right - including their inability to pay, as well as maybe not wanting to. A bond would certainly help to enforce the commitment made by the sponsors and would certainly be there to help to contribute to the backup if the sponsor wasn't able to fulfil their role. So in that sense it's probably a good thing.

Of course, there are probably distributional implications too, in that some people may not be able to come up with the bond, and to the extent that there are non-economic motivations for family unification, then we may be going against that.

Perhaps that's the direction in which we wish to go: to trade off from family concerns and issues to more economic ones, given the potential costs involved.

But we have to recognize that that's what we're doing. It really is a statement that you can't bring your grandparents over because we can't afford, as a country, to risk having to support them.

Mr. Hanger: On a personal note, I've talked to many immigrants who have come here and have taken up citizenship who wish to sponsor their parents on that basis, but there's no provision there at all for them to do it.

Prof. Benjamin: All right.

Mr. Hanger: They would guarantee those services, that they would pay up front.

Prof. Benjamin: So a bond would help them?

Mr. Hanger: Yes. That's good.

Mr. Assadourian (Don Valley North): I received this summary, in the line my colleague used, but I want to use it for different purposes.

It says that the labour market is losing its ability to provide immigrants with the economic opportunities enjoyed by their predecessors. Do you think that this trend, if it continues, is going to have an impact on the social and economic behaviour of the immigrants, and how much damage to our economy is it going to create if we don't face this problem now?

Prof. Benjamin: That raises another issue that economists don't know a lot about. What I've been focusing on, and what all my research has focused on, is what we have actual information on, how the immigrants themselves are doing. There is only indirect evidence of how the economy is able to accommodate them. The actual impact of the immigrants on the economy we know very much less about.

What I can say is that we don't have any evidence that there are adverse effects. The one possible exception to that is if they use transfer payments too much.

I would suggest that there's no real evidence of a crisis. It's not as if there's clearly deterioration of the outcomes of immigrants and we have to change things a lot.

On the other hand, there could be some very localized outcomes that we don't understand that might actually be very serious. It could be the case that in cities like Toronto, where most immigrants go, and to some extent Vancouver, outcomes are even worse than they are if you average in the other areas. We control for the region, but we don't look specifically at these potentially localized effects, where the strain on welfare provision of a particular locality may be much more serious than it appears to be from just looking at the overall picture.

Of course we can't tell people that they have to live here or there, because it's not enforceable and it's probably not a good idea. Given that people are going to locate and concentrate in certain cities, as are most Canadians, that's something you have to take into account even in letting in a total number. But that is something we don't know much about.

We don't know much about the hypothesis I alluded to, that's it's just the de-skilling or that they're outside the skilling of the Canadian economy, in which case potentially you could fix that simply by adjusting the admission standards. We don't really know the answer to that at all.

I wouldn't say that there's evidence that there's a sudden crisis.

Mr. Assadourian: Do you see any relationship between the income of individuals or families and their assimilation in Canada? If someone is rich, will they assimilate faster than someone who is poor, and likewise for education?


Prof. Benjamin: What you end up finding when you do that is that because assimilation is measuring a rate of convergence, the rate tends to be higher if you're further away. The people with the highest assimilation rates are those who start off lower because if they're going to converge, which they on average do, then they have to grow faster.

So it becomes hard to look at the exact question you want because it's hard to separate this out; it's almost an identity that has to occur. If you started out lower and you're going to catch up, then you had to grow faster. It's hard to actually separate out what the actual effect of education is.

To be more precise, way back when we first started this we were looking at immigration by source country from the 1970s and 1980s, and we would find the very lowest assimilation rates for immigrants from the U.K. That's because they basically started out at about the same as where Canadians were, so they didn't assimilate. Assimilation rates were actually higher for immigrants from less similar countries to Canada than the U.K., but they started out lower as well.

It's really hard to get at. I know the exact question you're looking at. Certainly the overall performance in terms of future income is conditional on where you start. That's the most important thing. If you start higher you'll end up higher, there's no question.

Mr. Assadourian: On my final point, if you don't feel you should answer you don't have to answer; it's a free country.

You mentioned in your presentation many times that you find it alarming or disturbing that more immigrants are using UI or more immigrants are going to welfare, although these levels are lower than with natives. Why do you describe that as alarming? That's a little part of assimilation, being the same. Why do you blame me when I'm the same as you? You're native-born and I'm not native-born.

Prof. Benjamin: I don't find the immigrant behaviour alarming; I find Canadian behaviour more alarming. Presumably immigrants come here and they're not aware of the program or they're aware of it and they don't use it. Perhaps they don't feel they need it. The longer they're here they think what a good idea, so they use it. I think it raises more questions about the UI system than it does about immigrant behaviour. That's why I find it alarming.

As I said, when you observe people who may be no different from Canadians in some ways, except that their earnings started out lower and they may have slightly different information when they first arrive, their overall standard of living is rising, their weeks worked are rising, their wages are rising, earnings are rising, but they're also using UI more. It's a bit of a puzzle to me. What they're doing is just becoming more like Canadians. That's what I find alarming. But that is the more facetious interpretation.

Mr. Assadourian: I have one question. Have you ever done a study to find out how Canadian-born people would react if they became immigrants to another country?

Prof. Benjamin: The only study I know on that has been done by this other guy I mentioned, George Borjas, who looked at Canada-U.S. What people have done is compared how Canadians do in the U.S. and how the Americans do in Canada. It's not an ideal comparison because a lot of people work in Canada for a few years and vice versa, so they're not permanent immigrants in the way we usually think.

In those studies the Canadians do better in the U.S. than vice versa. The argument given for that is that the American economy is just a lot bigger, and the people who go to the U.S. are primarily at the very top, so they're people who are going to earn a lot. In the U.S. that's not necessarily the case. It's just a more unequal economy in the U.S. so if you're going to go you're going to win and you're going to win well. In the U.S. your income may be above average, but not a lot. There are not a lot of Canadians going to other countries for us to see how they would perform, at least on that dimension.

That's one of the problems. If you're looking at comparing immigrants to natives in any dimension, immigrants are by and large different in some ways. They're people who are willing to leave their countries and leave everything behind to start a new life somewhere. Their drive and everything is not going to be equal to that of the average Canadian; it's usually going to be greater. That's what we accept. That's one of the reasons we like immigration.

My guess would be that if the average Canadian went somewhere else they wouldn't do quite as well as the average immigrant here, but those who do go do well because they tend to be those who are willing to move.

Mr. Assadourian: The reason I ask is that we assume that most Canadians are as well or better educated than people in Third World countries. We make assumptions.

Prof. Benjamin: Sure, that's a reasonable assumption.


Mr. Assadourian: If education is a factor in being successful, then the average Canadian with education who goes to Third World country A,B, or C will do just as well as a new immigrant coming to this country with an education.

Prof. Benjamin: Yes, they'd do just as well within that country. They wouldn't do as well as if they stayed. That's the big difference. If you've got a high level of education....

I do a fair bit of research and development economics, so I know about the Indonesian case. The returns to education in Indonesia if you're Indonesian are much higher than they are in Canada. If you have education in Indonesia, then you make a lot more than if you don't. If I went to Indonesia and wanted to be a professor, then I could be in the very top of the income distribution, but I would make less money than I do here.

So it's true that the education returns within the country would be higher there, and an educated Canadian would do very well elsewhere, but they wouldn't necessarily do as well as if they stayed. That's why people don't leave.


M. Nunez: You say that Canadians do not immigrate much and that there are no studies about those who do. However, there are some who go to the U.S. professionals for instance, and I would like to know how they perform there? Do they catch up with the American average? Are they over or under the average? Do you have any data on this?


Prof. Benjamin: They start out much higher than average and they continue to grow faster than average. The main reason for that is, as I say, you're looking at a very select group of people who leave. People usually make a choice to leave because they think they can do well. So on average they tend to be more highly educated, certainly than the average immigrant to the U.S., but more highly educated than the average American, and their earnings grow faster. So they tend to do very well in the U.S.

The other interesting kinds of comparisons one can make, which it's very important to do in some dimensions.... If you want to factor out the mobility of an individual, the fact that people who move are different from those who don't move, then you can compare not the average Canadian to the average immigrant, but the average mobile Canadian even within Canada.

People migrate within Canada all the time. In some sense we treat that as not moving, but of course it is. In some ways, to move from Quebec to Ontario is moving less than if you move from New Jersey to Ontario.

Those kinds of comparisons we don't really make. They're much more specialized kinds of comparisons and address only very simple kinds of questions about trying to adjust for this mobility factor that we think immigrants, and migrants of any kind, have.

People have done it in the U.S. in other contexts, particularly looking at black-white earnings differentials.

Within Canada, no such studies have been done. But it is generally the case that people who move do so because they think they will make more money. So they usually do pretty well, compared to the average person who didn't move.


M. Nunez: Do you know how many Canadians have settled in the U.S.?


Prof. Benjamin: I've heard differing numbers. I've heard that there are something like 600,000 Canadians living in Los Angeles alone, but I don't know how true that is. It depends on what you call a Canadian.

In the U.S. census you can identify people whose parents are Canadian or Canadian-born, but very few identify Canadian ancestry as some kind of ethnicity. So except for the immigrants, it's hard to know who's Canadian. The numbers might be in the order of a million or something over time. That has accumulated over the last 50 years. That's just a guess.


Mr. Nunez: You have talked about assimilation, which is not a word that immigrants in general don't usually like. In Quebec, for example, we almost always talk about integration. Do you think the results of your research would be different if the concept was different? What do you mean by assimilation?


Prof. Benjamin: The term ``assimilation'' in our study is very precise. I recognize that there are other interpretations of the word ``assimilation'', much as when we talk about cohort quality there are a lot of value-laden possible interpretations. Our interpretation of ``assimilation'' is very specific.

It is just looking at the growth or change over time of the outcomes, whether they be in participation in a program, earnings, or whatever.


Usually when people have talked of assimilation, and even economic assimilation, they were thinking in terms of immigrants starting below natives and rising towards natives, catching up. That's why it was called assimilation. But in Canada - and to some extent in the U.S., but more in Canada - quite often the immigrants actually pass. So really it is just a matter of growing earnings. It's not converging in some sense or integrating with the rest of Canadians, because they actually pass them.

So assimilation in terms of thinking of it as people becoming like Canadians in their income and so on is not quite the right way to think about it. It really is just how, through time, their economic outcomes become the same.

The other interesting thing about assimilation is just trying to figure out how that occurs, what the mechanism is by which somebody comes in earning less than everybody else but their earnings grow over time. Certainly if you go to your employer and say you're an immigrant, they're not going to start giving you bigger raises just because of that. So the question really is how do their earnings grow? Why do their earnings grow faster than those of comparable Canadians?

There are a number of hypotheses about that. One of them has to do with the investment on the part of immigrants in their skills once they get there. They're spending more time investing in acquiring skills that are useful in the Canadian labour market and they invest at a higher rate than the average Canadian, so their earnings grow faster. That's the explanation that's often provided. Of course it's very hard to measure something like that.

Mr. Dromisky (Thunder Bay - Atikokan): Your work is very fascinating. I can see you're totally immersed in it and very enthused and dedicated to your task. I think you find it very exciting, at least as being presented here.

However, there are a lot of variables you did not consider in your studies here and I'm sure you will be considering them in the years to come, as you continue in this field; and I hope you will. The great danger at this stage of development is that we make a lot of attributes about groups of people. That's quite common in committees of Parliament, where we think we know the answers and we come to all kinds of conclusions based on half-truths and wishy-washy statements we've heard about, gossip in our communities, and so forth. So we have to be very careful about the kinds of attributes we make about these people and how we interpret the information that's in your study.

We have to be really careful when we look at the immigrant and his introduction into the workforce and in essence what the market holds for him. Twenty-five or thirty-five years ago, immigrants, once they became involved in the marketplace - I'm talking about the labour force - more or less would have a job guaranteed for life, if they wanted. But that situation has changed drastically today.

I think that's a very important area that has to be studied. Attitudes have to be studied too. But we can't jump to conclusions and say, well, they learn the Canadian way, and if they don't want to work they can get unemployment insurance. I don't think that's the answer. I think the answer is more in what is available for them.

Most immigrants will work very hard, at two or three jobs if they can find them, in order to be an example of success to their relatives and their friends back in the country from which they came. That's part of the story, as you indicated already.

Do you have any plans to do any research in this area, a study of the evidence that's available at present in this area?

Prof. Benjamin: Yes. As we suggest, there's the puzzle about why assimilation in UI, for example, would be positive, why people become more likely to use it. As I say, when you try to control for, or adjust for, the changed economic circumstances of immigrants, which include the probability of being unemployed, and whether or not they are eligible, their earnings and wage rate and everything else, what we're saying is that even when you look at that and control for the fact that immigrants may be more likely to be unemployed, they're still less likely than native-born to use unemployment. But even that propensity goes away over time. That's all we're saying. That's our interpretation.


That's why we try to control for the fact that there may be different economic outcomes. That's why we argue, at least in terms of interpretation, that you can't disentangle the performance or outcomes in the labour market from looking at transfer programs, though from the dollar standpoint the taxpayers may not care why. They just want to know how much it costs. Nonetheless, the real puzzle is why the initial conditions of the immigrants are declining in the first place.

And you're right, the concept of a lifetime job, that ability to just go and get a good union construction job, as I'm sure people were able to do in Toronto in the 1960s or 1970s, doesn't exist any more. The nature of the workplace itself and the nature of the labour market are changing for everybody.

As we said before, it may be the case that this change in the overall economy is what immigrants are being caught in. If they don't have the particular set of skills that is a good match for that type of economy, then they may be less likely to be successful. But we don't know. That's just a possibility.

I was at a conference recently where people were looking for evidence of overall increases in part-time work or the loss of full-time jobs in the economy. That turns out to be a really hard question to answer, and there's no clear answer that it's really where we're going. So it's a really tough one. I don't think any data are ever going to allow us to give a definitive answer as to what really is going on. Inevitably, you're grasping and saying this group does something kind of differently compared to another group.

You're right, you have to very, very careful. That's why I prefer to stick to what the data can tell me, but even within that one has to be very careful, because data are just data.

Mr. Dromisky: I like your answer.

The Chairman: From one PhD to another.

Mr. Hanger.

Mr. Hanger: I have a question about one of your graphs. It's figure 2, receipt of social assistance. Looking at the comparison there, the native-born maintained the same level of propensity to receive social assistance. Is that how -

Prof. Benjamin: Exactly. This is for the women. This is about 6.3% and it didn't change over the -

Mr. Hanger: I assume that level was maintained right across.

Prof. Benjamin: Yes. I don't know what the numbers are now, but actually I think it's fairly comparable. If you look at the family, this is one of the bigger figures I have -

Mr. Hanger: I was more particularly interested in this graph in the sense that it maintained a constant level for the native-born, yet the 1985 and 1990 comparisons for the immigrant women in some of those years changed dramatically. Why was there such a change?

Prof. Benjamin: That's what we're identifying as the assimilation. This is the positive assimilation into social assistance. That's what we call assimilation, the change in the propensity to use the program relative to the native benchmark.

We're saying we don't know. It is not explained by the changing income levels. It's not explained by changing poverty rates because the poverty rate doesn't really change very much.

The next picture on the very last page is the below low-income cut-off, in figure 6. This is something you would think would be a key explanatory variable, a key thing that would explain the use of welfare programs. It's the percentage of poor. You can see that the percentage of native-born women who were poor went from 15% down to something below that over the time period.

For many of the immigrant groups up to 1975, the percentage of poor went down as well. From 1976 to 1980 it went up and from 1981 to 1985 it went down. So the overall picture is that except for that one cohort, the percentage of poor went down but the percentage collecting welfare went up.

That's why we say we don't know what it is. We can't really explain it with anything an economist might like to have to explain a variable, which would be some change in outcomes that would cause the use of transfer programs.


It's the same with the UI. We can't explain it. When you look at changing economic circumstances, it doesn't line up with the changing use of the programs. That's the puzzle.

Mr. Loney (Edmonton North): Professor Benjamin, in your research, did you find any evidence that certain immigrant communities were making more or less use of social assistance than others?

Prof. Benjamin: As I say, when we were looking at the Survey of Consumer Finances, which is a pretty good data set to look at these things, it does not have.... By ``communities'', you mean by ethnic group or by city?

Mr. Loney: Ethnic breakdown.

Prof. Benjamin: We don't have an ethnic breakdown of use of transfer programs. The sample sizes are far too small in those data. They don't have the information, but even if they did, it wouldn't be big enough.

The census is the best way to look at that specific question, and we haven't looked at it. We have looked at earnings by ethnic groups, the performance of ethnic groups in the economy on that basis, and there are differences. But we can't possibly explain them.

To the extent that those are correlated with transfer programs, that would be interesting. But as I said at the beginning, our primary focus is really looking at the labour market outcomes, and there are obviously consequences for transfer programs that may follow.

But the interesting thing about ethnic groups is that if you look at just immigrant earnings by ethnic group, there are not that many big differences across ethnic groups. There are bigger groups in the next generation of immigrants than there are in the actual immigrants themselves. That may feed even more into welfare programs. But that's just speculative.

The Chairman: Does anybody object if I ask the researcher if he has any questions?

Mr. Kevin Kerr (Committee Researcher): Thank you.

In your essay in Diminishing Returns, you note that probably the most important policy implication of your work is the setting of levels. But that subject isn't broached again in the conclusion. Do you have anything to say about what the current levels should be?

Prof. Benjamin: We suggest at the beginning that the percentage of immigrants times the level will give you the biggest impact on the treasury, so to speak, and that's where the policy implication comes in. We don't say, and it doesn't really follow from our results, the level should go up or down. If you did raise the level, the level of welfare payments would go up over what it was before, but again, it's the idea of this native-born benchmark, that the marginal person you let in would still have a lower propensity to use welfare than the existing population. So I don't think there are any obvious implications there for levels.

The bigger implication - and again it doesn't follow from the numbers; it's more by putting together the other bits of information - is more on the selection criteria, as to whether you might tilt the balance, as it is now, towards a more independent class if one believes economic outcomes explain the use of welfare programs. The only caution to that is that, as we pointed out, a lot of the changes in behaviour that we observed in terms of the propensity to use programs are not explained by the economic outcomes.

So while the actual level of use of any group will be primarily explained by their economic outcomes, the growth in Canada, the change over time, is not. But I still think the bottom line would be that the selection criteria have a bigger impact than the numbers.

Mr. Kerr: You don't seem to want to come down one way or the other on whether you think the composition should be rebalanced.


Prof. Benjamin: Whether it should or not, as an economist I would say that if one is primarily interested in economic outcomes, it seems to be the case that there is some weak evidence. I wouldn't say it's overwhelming, but there is very suggestive evidence that skills are important and that if you can tailor immigrants to the Canadian economy, they might on average do better. I don't think we can predict perfectly whether anybody is going to succeed or not, but on average, if that was the benchmark we wanted to use for setting immigration policy, the argument would be towards more independent and less family.

On the other hand, we don't admit refugees on the basis of economic criteria, because that's not the motivation for the immigration. I've clamped on that one. Clearly there are other important considerations for determining the number of family immigrants and the number of refugees, independently of their performance in the Canadian economy.

Another point that I didn't make that I think is very important is if there's no evidence that immigrants impose a cost in excess of anybody else in the economy, and there's no evidence of any ``damage'', then the immigrant is better off here than they were before, or else they wouldn't have come here. So it's not clear that you could open the doors and say that everybody who wants to come in can come in. They'll be better off, but unless there's evidence that they're going to hurt other people, who cares?

That's an extreme case of an argument that says: in choosing the numbers in some particular way, according to some economic policy, unless there's evidence that the immigrants are actually harming the economy, you could let in as many as you want. But what we're seeing here is there is some suggestion that....

The evidence we have is that the immigrants themselves aren't doing as well, but, in some sense, who cares? If the immigrants are happy, if they want to come here, even if they're doing worse than their predecessors, why should that matter, unless there's some negative effect on the rest of the economy?

That's a question we didn't raise, or we raised at the beginning, but it's important. If we're worried about the welfare of immigrants, in a sense they're free to come and go. If they prefer to stay here, and they're not collecting UI but they're earning less than they used to, why is that a problem?

Mr. Kerr: Focusing on the group of immigrants, foreign-born, who are economically assessed, do you think there is any utility to streamlining or rationalizing the point system? Are there certain factors that you feel should be emphasized more than they are under the current structure?

Prof. Benjamin: I don't know a lot about the details of administration of the point system. I know something about it, and somebody from the Department of Immigration probably knows better.

My guess would be that within the point system it's probably just a matter of fine-tuning it. It might not make a big difference. Whether or not the points are applied at all probably makes the bigger difference.

The point system gives points on the basis of education. We already know that on average immigrants are more educated than the native-born, and that might apply even within the family class to some extent. So I don't know what the evidence is on that. Partially it depends on perceived occupational opportunities.

As an economist, I do not make forecasts. I would imagine that no matter how many economists you put in trying to predict occupational demand over the lifetime of immigrants, it's not going to be perfect, or even very good. So I don't think you should put any more weight on that.

I think it is really more just a question of asking if you are going to put any economic assessment at all, versus none. That probably has a bigger impact than the fine-tuning within it. Given that there are so many things you can't possibly measure that are going to affect an individual's earnings and stuff, you can't possibly have them fill in a form that's going to fail them. No one is going to say that they expect to fail.

So I guess I would be reluctant to give any suggestions on what would improve the point system. I think it's just applying it or not applying it that makes the biggest difference.

Mr. Assadourian: Last year the minister had consultations throughout the country. We had four regional conferences regarding immigration policy for the next 10 years, the levels and what have you.

Did you or your colleague ever participate or have any input in that consultation?

Prof. Benjamin: We didn't.

Mr. Assadourian: You mentioned earlier, who cares how many immigrants come as long as they're happy here and they aren't a burden to society here? The total is now set at between 192,000 to 215,000 per year. Are you comfortable with this figure, or would you increase or decrease it? If you increased it, why would you do that? Would be your suggestion for a fair number of immigrants coming to the country? Would there be a formula in your mind that you can relate to?


Prof. Benjamin: This does not come from my research - there is nothing that says what the number should be - but my own guess would be that it's not easy to make a case that Canada is worse off with more economic immigrants from the independent class. I think it would be pretty hard to make a case that Canada is worse off with more immigrants in the independent class. I think, on average, that we would probably gain more than we would lose.

I don't know about the family class. I think that one's open. But I would doubt, again, that you could make a strong case that the family class is hurting us.

So in that sense I would say that 250,000 was probably fine. But the counter-argument to that is a political one: are immigrants perhaps worse off, as a group, if there's political heat? There have always been episodes in history in which immigrants were perceived as being a problem or that there were too many. It may then be politically sensible to cut back the numbers to reduce the heat, so to speak, on the 200,000 that do come in, rather than have people grumbling and so on. From a social standpoint, it may make sense to keep the numbers lower.

I guess what I'm saying is that I think that cutting down the numbers from 250,000 to 200,000 was politically motivated. It was probably sensibly politically motivated, but I don't think the economic case is there.

Mr. Assadourian: For the sake of argument, say the government comes up with a plan. I'll use the term ``Marshall Plan'', not in terms of economics but in terms of immigrants. Say that in the next five years we are going to bring in 2 million immigrants. These will be the criteria. The purpose of this will be to stimulate the economy. That's because you told us that everybody who came here so far was a plus to immigration rather than a minus.

How would you react? Would you react by saying that this is a sensible thing to do, as it's planned and controlled? Would that positively impact the economy? Never mind the political -

Prof. Benjamin: It would raise the price of a house, which would be a good thing.

I think it would be good. I think Canada can always use more hard-working people. It's not that we're not hard-working here, but I think the more there are, the better. To not accept immigrants, even a lot of immigrants, I think would be a mistake.

First, there will be people who want to come anyway, hopefully. If things get too grim, maybe more people won't want that, but by and large I don't expect that to happen. I think people will always want to come here.

I think it is a positive thing to keep the numbers high. As I say, some of the other evidence in the book suggests that on balance we definitely do better by having more of that. It adds to our economy.

Our population growth alone is not going to be enough to pay the bills later on. So in terms of just little things like smoothing out the age distribution and bringing in working-age people, it's probably a good idea. Somebody is going to have to pay the bills when the baby boomers start retiring.

I think you can look at the big picture in terms of the reasons why we want to bring in highly motivated immigrants. Again, given that other changes occur in the economy, that skill level is going to be important.

The other issue that I'm not the least bit qualified to talk about are the other dimensions to immigration besides economics. There are other pressures on institutions that have to be adjusted to, and it causes other pressures. Those things have to be taken into account.

Mr. Assadourian: So in other words, say it's controlled and planned exactly according to the plan. As you said, say you bring in 2 million. That's the figure we used for the next five years. You'll have to build more schools and hospitals. You'll have to produce more cars. If everything is in place, then it's basically going to be acceptable to you to bring in large numbers like that. It will be beneficial to the overall economy.

Prof. Benjamin: Yes, overall. I don't think anyone has done studies on the impact on the infrastructure, and so on, of immigrants. Some 100,000 of the 200,000 tend to live around Toronto. It is clearly a more crowded city.


So things like that occur that we're not talking about here, and you do get pressures on schools, and so on. There's language training and a whole bunch of other issues that have to be addressed. The ethnic composition of Canada changes as well, and that puts its own social pressures, and so on, that have to be addressed.

Those are the kinds of pressures that I think are going to be more important than the economic ones, and those are things we know nothing about.

The Chairman: Are there any other questions?

Thanks very much for your time. We appreciate your coming down. Your presentation has been very good.

Prof. Benjamin: Thank you. I enjoyed the questions.

The Chairman: You were recommended to us by the department.

Mr. Assadourian: So the department knows what they're doing, then.

The Chairman: They didn't recommend everyone in the book. You were recommended - for what that's worth.

Prof. Benjamin: I'm sure Don DeVoretz gave you a good show, though.

The Chairman: The meeting is officially over. Thank you.