FINA Committee Meeting

STANDING COMMITTEE ON FINANCE

COMITÉ PERMANENT DES FINANCES

EVIDENCE

[Recorded by Electronic Apparatus]

Tuesday, April 27, 1999

• 0934

[English]

The Chairman (Mr. Maurizio Bevilacqua (Vaughan—King—Aurora, Lib.)): I'd like to call the meeting to order.

Today the House of Commons finance committee commences its study of productivity and economic growth and the impact that these variables have on the standard of living of Canadians. The committee has been intimately involved in the process by which the Government of Canada has eliminated annual budgetary deficits and has finally, after three decades, reversed the trend of the growing government debt. This indeed has been a formidable task.

• 0935

Through our pre-budget consultations, we have been able to make Canadians part of the process by which fiscal policy is formulated. But now that the immediate fiscal challenges have been addressed and dealt with, it is time to focus our gaze further into the future. Elimination of the deficit enables us to reflect upon the longer-term concerns and trends and to think about investing in our future and our children's future.

The committee believes we now have an opportunity to do just that, to reflect upon those developments that have made us a strong and prosperous economy, and to take these lessons into the future so that in fact our children and our grandchildren may inherit an even stronger and more prosperous economy and a more vibrant society. Last year's committee report, Facing the Future: Challenges and Choices for a New Era, was a first step in this process, as it recommended a commitment by the government to a productivity covenant. Again, we believe Canadians should play an important part in developing a strategy for promoting economic growth.

We tend to forget that many of the newly industrialized countries that today enjoy a standard of living similar to that which we enjoyed three decades ago were living in abject poverty at that time. For them, economic growth has clearly been a benefit.

Looking at the rest of the industrialized world, we know that higher rates of economic growth are indeed achievable, but how do we match those growth rates of the United States and the rest of the industrialized world? We believe enhanced productivity holds the key, but as policy-makers, we must satisfy Canadians that improving productivity is something that will benefit them directly and concretely. In addition, we must be able to put our belief into practice. Just saying that we need to be more productive is not good enough. We must use the policy tools at our disposal to achieve this goal and to do so in a way that is fair and equitable to all Canadians.

That is why we have begun this important study. With your help, we intend to better understand what is needed to improve the standard of living of Canadians and to find the tools that will allow the prerequisite conditions to be established.

We are here to help Canadians to get more from their efforts, to enable them to look forward to higher real incomes in the future, and to enable them to have access to new and more fruitful investment opportunities. In other words, we want them to get more. That is more than just wishful thinking. It is in fact a necessity.

I would like to welcome the first panel here today. We have faith that your knowledge and expertise will help to shed light upon this important aspect of government policy on this important societal issue. After all, improving the standard of living of Canadians in its many facets is what government is truly all about.

Last fall the committee examined the report of the Task Force on the Future of the Canadian Financial Services Sector. That study held the promise of a profound reform of an important sector of the Canadian economy. This study of productivity holds the promise to be equally profound in shaping the future of the Canadian economy. Canadians are up to the challenge, but they need the assistance and guidance of experts such as you.

On behalf of all members of this committee, I welcome you to our first hearings, and we all look forward to profiting from your wisdom and your guidance. Welcome.

We will now begin with Mr. Stewart Wells, assistant chief statistician, national accounts and analytical studies, Statistics Canada. Welcome.

• 0940

Mr. Stewart Wells (Assistant Chief Statistician, National Accounts and Analytical Studies, Statistics Canada): Thank you, Mr. Chairman.

I'd like to begin by thanking the committee for asking us to appear before it today to discuss productivity measurement. It's certainly an important issue, as you've said, and one very much in the news recently.

As I look around the table at the witnesses who are here today, I'm tempted to say that if this were the movie Casablanca, and you, Mr. Chairman, were Claude Rains, I would say that you've rounded up the usual suspects, not all of them, perhaps, but some of them.

My comments will be of an introductory and fairly elementary nature. My colleague John Baldwin, who is responsible for developing the productivity estimates at Statistics Canada, will provide the committee with more technical but nevertheless interesting, and maybe even more interesting, aspects than I will be touching on.

Productivity is frequently identified as one of the two elements—and you alluded to this indirectly at least, Mr. Chairman—that enable long-term increases in living standards, the other element being increases in the labour supply and the quality of that labour. I have a handwritten overhead that I did this morning, which may be helpful.

Leaving aside issues of income distribution, if the living standard is defined as the GDP, or total real output per capita, the percentage change in that ratio, the percentage change in GDP per capita, can be approximated as the sum of the percentage change in GDP per hour worked, plus the percentage change in the ratio of hours worked to employment, plus the percentage change in the ratio of employment to population.

If those last two, the percentage change of the ratio of hours worked to employment and the ratio of employment to population, declined, for example—and Mr. Baldwin will have some information on this—then the increase in the standard of living will be less than the increase in productivity as it's measured by GDP per hour worked. I think it's a useful relationship to bear in mind as you continue your examination.

For most mainstream economists, productivity includes the effects of technological change and innovation that is either labour saving or capital saving, or both. It also includes the effect of changes in management practices and industrial organization, and it may or may not include improvement in skills and the level of education. It does normally include the effects of cleaner streets, lower crime rates, air pollution or resource depletion, all of which some people would consider to be of great importance to the standard of living. In this context, however, productivity can be defined as how much output or production is to be obtained from a given amount of one or more inputs.

Whatever the measure that is adopted, we can then calculate the level of productivity or the change in the level of productivity. Levels are more useful as a component of cross-country comparisons of living standards at a point in time.

The inherent difficulty in making those cross-country comparisons is the exchange rate at which one chooses to transfer, since it all has to be expressed in a common currency. Rates of change of productivity are more useful for studying how an industry or a country is doing from one year to the next, or one period to the next.

The three most common measures of productivity are labour productivity, which is the amount of real output per unit of labour; output, whether a level or change in level, which is usually measured in constant prices; and labour, as the number of people employed or the hours of work.

Capital productivity is output per unit of capital. Again, output is measured in constant prices, and capital is measured in constant prices.

The more complicated formula that gets some attention, and these days increasingly more attention, is multifactor productivity. Some people call that “total factor productivity.” The distinction is not particularly important.

Multifactor productivity takes account of the amount of labour and of capital that is used, and often includes materials used in the production process. It is usually measured according to the following formula.

The change in multifactor productivity equals labour's share of output, the share of output that goes to labour, multiplied by the change in average labour productivity, plus capital share of output, multiplied by the change in average capital productivity, plus the share of output going to materials used, multiplied by the change in the average productivity of materials inputs. So the weighted average of those things is multifactor productivity.

• 0945

These three formulas are used virtually everywhere—in Canada, in the U.S., and at the OECD, for example—basically because they are relatively easy to compute. When I say virtually everywhere, I have to take note of my good friend Erwin Diewert over here, who may well be using a more complex formula than some of the rest of us.

A noteworthy characteristic is that these formulas measure everything, except what they say they do. Labour productivity reflects the effects of technological change, additional capital, better management processes, etc., but not pure labour. Capital productivity measures the effect of everything but pure capital. Multifactor productivity is the residual between the inputs listed and everything else.

It has been said that the estimate of productivity in the multifactor sense is a measure of our ignorance. That's something worth bearing in mind. What we don't know, and what you can't explain by the other factors, appears residually in there. So it's a combination of many things.

Because everybody uses basically the same formula, differences in the results depend on how output shares are calculated, how capital and labour inputs are defined, how output itself is defined, and the level of aggregation you work from—whether you start at a low level of aggregation at the industry level, or whether you do it at the economy-wide level. The results we get can differ a fair amount because there's much room in there for measurement error, as you'll be hearing about this morning. But they tell us quite a bit about our use of labour and capital—how our use of labour and capital has changed over the years.

I will close with the following observations. Since the determinants of productivity change take place at the factory or on the shop floor, it's not immediately obvious. This is a problem you're going to have to face as you continue to study this. It's not immediately obvious what the levers are that might be used to modify future outcomes.

On that note, I'll ask my colleague John Baldwin to begin.

The Chairman: Thank you, Mr. Wells.

Mr. Baldwin.

Dr. John Baldwin (Director, Micro Economic Studies and Analysis, Statistics Canada): Thank you. I will just try to give you roughly some idea of the complexity of the measurement issues.

Stew has already discussed the general approach that's taken to measuring productivity at Statistics Canada. Then I will talk briefly about the recent release we produced that has had some public attention.

As Stew has indicated, the measurement of productivity basically involves taking a look at changes in output, measuring all of the changes in output, changes in capital, changes in materials input, and changes in labour. Productivity estimates, therefore, are effectively first differences of first differences. We look at the differences in the rates of change in output and subtract from them the changes in the rates of change in outputs, appropriately weighted. Because it's a first difference of first differences, it's an extremely difficult measure to produce. All of the errors that take place in any of the series end up embedded in the final estimate. For that reason we approach the matter very carefully.

The productivity division is effectively at the top of the pyramid, relying upon the expertise of a vast number of other divisions within Statistics Canada that provide us with the inputs to this final formula. The formula itself has been developed over time. This is very much an exercise that has changed and been modified over the last 20-odd years.

At one time, most statistical agencies produced nothing but labour productivity. It was a measure that was relatively easy to comprehend: the rate of change of output relative to the rate of changes in labour. If you do it in growth rates or in absolute terms, you know just how much you're producing per unit of labour in the system. But the academic world insisted we could do better; we could indeed produce something more meaningful in the way of a multifactor or total factor productivity measure.

• 0950

As government research agencies, such as the Economic Council of Canada, experimented with producing those numbers, showing how they could be done and how they contributed to our understanding of what was happening in the economic system, this statistical agency gradually took it upon itself to try to produce these numbers. It did this not because it was the only one that could produce them, but because it required a fair amount of detail and the coordination of a large number of programs—the sort of thing it found it was able to do quite well.

We produce these numbers because we have a great deal of detail at the individual series level—whether it comes to measuring labour or capital—that we can put together. More importantly, we have this at a finer level of disaggregation—at a very fine industry level that is not available to outsiders because of confidentiality provisions that pertain to statistics gathered by Statistics Canada.

We therefore start by paying a fair amount of attention to how we'll measure the labour inputs. At the present time, we measure the labour inputs as the hours worked by both self-employed and employed individuals in the economy. The labour force division provides us with those data, and those data are then manipulated to make sure they're corrected for differences in the number of days workers can work per year because of differences in numbers of holidays and differences in strikes that occur. It's a fairly detailed exercise that eventually gets down to producing a measure of person-hours worked each year, which becomes the input into the multifactor and labour-productivity measures Stew has discussed.

The second input into these series that's important is the capital input. When I was going through undergraduate and graduate work, there were still massive debates about whether we could even measure capital. We've turned from debating whether we can measure it to how we can measure it.

Capital series inputs to these exercises require that we start first by having some idea of how much investment is being made in machinery and equipment and structures in this economy; then those investment series have to be summed to come up with a total capital stock that becomes an input measure to these formulas. That requires a great deal of detail to be developed, because we need information on how long each piece of machinery or equipment lasts, what is its length of life, if we're going to decide over what period we're going to sum those investments to get a capital stock. We also need some idea of how rapidly the capital stock we have depreciates—how rapidly it becomes obsolete and wears down.

The Statistics Canada investment and capital stock division has developed a new set of surveys over the last 15 years that provides Statistics Canada with much more detail, asset-wise, on the nature of depreciation actually taking place than we once had. That division puts together those estimates and comes up with capital stock estimates, which are then used in our program.

At the same time as considerable effort was being placed upon developing measures of labour input and capital input, we also developed, through the input-output division, detailed estimates of the purchases and sales of many industries in this country. That detail is provided at both aggregate and very disaggregate levels.

When the productivity numbers are developed in Canada, we start with a detailed 195-industry level. Stew already alluded to the fact that we believe it's important to start at a disaggregated level for several reasons, one of which is that there is a fair amount of interest in the actual performance of individual industries.

Then the productivity group pays great attention to several things. First is how it weights the individual industry data when it's producing an aggregate. Mr. Wells referred to the fact that we're talking about aggregate productivity, but aggregate productivity comes from hundreds of thousands of commodities in many different industries. Some decision has to be made as to how you weight those industries together when you move to the aggregate level. Therefore, detailed formulas are developed for that purpose.

Secondly, detailed decisions have to be made on how deflation techniques will be used to come up with real changes in output. I know this is a technical issue, but it's worth raising because you will often see different estimates, and statisticians and other people will tell you that different types of deflation techniques have been used and that's what produces those different results.

• 0955

But different types of deflation techniques can have dramatically different effects on rates of growth of what we're measuring. I was talking to Australian statisticians just last week, and they told me that if they used one acceptable price index, the rate of growth of investment in their country over a 10-year period in the mid-1980s had been 86%. If you used a different but completely acceptable way of deflating, it was 18%. The difference between 86% and 18% is huge. Therefore, experts have to decide how you should average these various ways of approaching change in the economy.

Then we put together our estimates that start at a disaggregate level and are then aggregated up by the weighting system, and that effectively gives us a similar rate of productivity growth no matter what level we're dealing with.

Let me turn now to a few of the details that are relevant in trying to understand some of the Statistics Canada productivity estimates that have been recently released. First of all, when you start to measure productivity, you have to make a decision as to whether you're measuring the entire economy or just subsegments of that economy.

We produce productivity numbers for what we call the business sector only, that is, we remove the government sector and sectors within the economy that are mainly government oriented. So federal government departments will be removed from these estimates, but so will health and education in Canada, because health and education are mainly provided by the public sector. Because of the way in which national accounts are actually accumulated, productivity growth in health and education is generally zero. The outputs of these sectors are usually measured by the inputs of these sectors. Therefore, by definition there can be very little difference between rates of growth of output and rates of growth of input.

We believe, therefore, that the only thing we can appropriately measure is the private sector. When you're making comparisons with other countries, you must keep that in mind. The private sector may be much larger and different in other countries.

We also focus on growth rates, not on levels. We do that primarily because we have a huge infrastructure that develops data on prices that can be used to come up with real changes in quantities and inputs. We do not have as detailed a program that allows us to make correction for cross-industry and cross-country prices. Therefore, we do not provide formal comparisons of productivity levels in this country with those of other countries.

Finally, as I indicated before, we tend to build up our data from a highly disaggregated level. That can often provide us with a difference from other people who provide estimates that start at an aggregate level.

How much time do I have left, five minutes?

The Chairman: Keep going.

Dr. John Baldwin: Let me now say a few words about the recent releases of Statistics Canada.

On March 23 we released historical revisions to our productivity series that resulted from the historical revisions to the national accounts. The historical revisions to the national accounts take place on a regular basis, every five years, and take into account the changing structure of the economy. We regularly get rebased output levels from the system of national accounts that then must be put into our new and revised revisions of productivity. At this point in time capital stock revisions are also rebased, and rates of growth are also changed. This historical revision occurs periodically, every five years, and the changes that occur in productivity estimates are similar in this decade to those that have gone on before in that the productivity estimates subsequent to the historical revisions have been revised slightly upwards.

The revisions we've given you in terms of the multifactor productivity estimates now cover the period up to the middle to late 1990s, up to 1997, and they now show a more or less consistent path of multifactor productivity over the last three decades. We do not show that productivity growth has dramatically increased over these three decades. In fact, what we show is a rather steady growth rate over these three decades of about 0.7% in multifactor productivity, which is slightly higher than for what we call our labour productivity measures.

• 1000

These measures are substantially below the pre-1973 estimates. So our new release shows that there is slightly more growth in the 1990s than we had before we'd made the historical revisions, but there's more or less constant growth subsequent to 1973, and that growth is not very rosy. That growth is relatively low.

Several conundrums were proposed to Statistics Canada after we released our numbers. The first has to do with the conundrum—which many people have noted and which Stew mentioned at the beginning—between the standard of living that appears to have been falling and our numbers that show that productivity growth is relatively constant. This bar chart I'm putting up on the screen shows the rate of growth in GDP per capita in the 1980s and 1990s. The rate of growth in GDP per capita is the blue bar on the left of each of the two panels. You can see that the rate of growth in GDP per capita in that blue bar has fallen from the 1980s to the 1990s, and that accords with most people's perceptions of what has been happening in the 1990s.

The red bar on the far right is the rate of growth in GDP per hour worked. This is not exactly our labour productivity measure, but it's something that's close and akin to GDP per capita. You can see that GDP per hour worked, a labour productivity measure, has been more or less constant over these two decades, which more or less accords with the far more detailed data I referred to earlier.

Why do you have reasonably constant rates of growth of GDP per hour worked or productivity but a decline in the simple standard of living measure? Of course, it comes from the other two aspects Stew was talking about. The rate of growth in the number of jobs or the number of hours worked per person in the economy has fallen very dramatically between the 1980s and the 1990s. We have more or less constant productivity growth, but because of the extent to which we're employing people in the economy, GDP per capita has fallen over these two decades.

The other conundrum people have been confused about is the difference between the estimates of Statistics Canada and the those of another semi-official governmental body, the OECD. The OECD, in its economic outlook last year, indicated that Canadian productivity growth over the last two decades was negative. The bars on the far left of that chart show you the multifactor productivity growth the OECD included in its report as of last year, which it was discussing recently in Canada. The two bars on the right show the productivity growth we estimate for these two decades. The yellow bar is for the 1980s, and the blue bar is for the 1990s. You can see a considerable difference between the far left and the far right.

There is, however, a second estimate produced by the statistics directorate of the OECD. It's in the middle. You can see that there are very little differences between the OECD statistics directorate estimates and ours. We can indeed reconcile those two very easily by looking at the differences in inputs that are used in those formulas. Effectively, the OECD uses employment rather than hours worked. Canadian employment has been going up faster than hours worked because we've been increasing part-time employment. When we use their inputs, employment rather than hours worked, our numbers almost coincide. There's no statistical difference between the two series.

I understand from a discussion with the OECD that the numbers on the far left are being revised and will be reissued in the coming months. They're not official yet, so I cannot discuss them.

That ends my presentation. Thank you.

The Chairman: Thank you very much, Dr. Baldwin.

We'll now hear from Dr. Erwin Diewert, an economist from the University of British Columbia. Welcome.

Dr. Erwin Diewert (Economist, University of British Columbia): Thank you.

I will try to cover some of the same ground Stew and John covered. I don't think it will hurt to repeat some of the material, because it all just becomes a blur after a while.

My instructions were to cover three topics: one, productivity defined; two, the history of Canadian productivity; and three, issues related to the measurement of productivity.

• 1005

Let's look at productivity defined. The simplest thing to do is to start with a very simple economy, one output and one input. You're looking at annual productivity, so you calculate how much output is produced this year, divide by the amount that was produced last year, and it gives you one plus the growth rate of outputs. Then you do the same for inputs. The amount of input used this year divided by the amount of input used last year gives you one plus the growth rate of inputs. Take the difference and you get the growth of output minus the growth of input productivity.

So this is good. The bigger that number is, the more output we're getting per unit input; and that's why we love productivity if it's positive.

The many output, many input world is very similar. We just do this for all the outputs in the economy and all the inputs in the economy, but we have to weight them by something. Different products are more or less important in the economy. So what we do is take an average of the output shares this year and last year for each commodity, and use that as a weight for the specific output we're considering. And then we do the same for the inputs. We look at the value share of an input in total inputs this year, the same value share last year, take an average of the two, and then use that as a weight for the growth rate of this particular input.

So that's really it for total-factor productivity. You have all the output growth rates minus all the input growth rates, weighted by their importance in the economy in the two years under consideration.

What about labour productivity? The output concept is the same there, but we restrict the inputs to either just hours—treat each hour of work as the same thing—or weighted hours, weighted by wages and the share of labour income. Both have their usefulness, but economists tend to like to look at the first comprehensive measure because it's a measure of the free lunch in society. The bigger the multifactor productivity is or the total-factor productivity is, the more we're getting for nothing, and so that is good.

So that's my introduction to what productivity is. I suppose questions can come later, so I'll carry on.

Topic number two, the history of Canadian productivity, is a fascinating area. John Baldwin talked about the new Canadian productivity numbers. By coincidence, a former student of mine, Denis Lawrence and I just wrote a paper called Progress in Measuring the Price and Quantity of Capital, and what we did was look at different capital concepts—again, John and Stew alluded to this—to see what kind of difference it would make in measuring Canada's total-factor productivity. So we have these numbers from 1962 to 1996, which we can compare to the Statistics Canada numbers.

The new Statistics Canada numbers are not published in any great detail. There was a release on March 23 of this year in which they came out in abbreviated form, but René Durand, who's sitting in the audience, was kind enough to send me the detailed output and input numbers that make up this data. So I processed his numbers and compared them with our numbers, and I found that from the period 1963 to 1996 the new Statistics Canada total-factor productivity estimates were 1.2% per year, rather than the 0.7% that you mentioned, John, so perhaps you can explain that to me.

• 1010

In the Diewert-Lawrence estimates, using a comparable methodology, we found total-factor productivity growth over the period of about 0.55% per year—so about half. You can look at the breakdown in different periods, and it's more or less similar. The new Statistics Canada estimates give us productivity numbers about 0.5% a year higher than what we got.

I'll talk about some of the differences in part three, and they are issues related to the measurement of productivity. Can we explain these differences? I'm going to break up my remarks on this topic into five parts.

The first issue is what sectors of the economy are covered. Stew and John talked about this. You could do productivity measures for the entire Canadian economy, but then we have the problem of public administration, which is that it's hard to measure your output. We know you're working hard here, but your output in the national accounts will be measured by your input, rather than the real value that you perform here. So we generally exclude public administration.

Then there's the entire economy less public administration and less schools and hospitals. So the Statistics Canada numbers excluded education and medical services as well. Again, the outputs there are not measured separately, and so if you include them it will drag down your productivity estimates. In the Diewert-Lawrence study we included hospitals and education. So this is one explanatory factor for why our numbers are lower.

So that's something to be kept in mind when you're doing international comparisons too: try to make sure that like is being compared with like.

The second issue related to the measurement of productivity has to do with how we're going to measure the outputs. We can do this in one of two ways: we can look at the outputs of each industry or firm or establishment and work it up from that level, or we can short-circuit all that detailed inter-industry flow and just go to the final demands that are delivered to consumption, investment, and foreign trade.

So the Diewert-Lawrence stuff looked at the final demand approach and the Statistics Canada approach looked at building it up from the industry statistics. From the numbers René gave me, there wasn't much difference between the two methods of compiling the output, about 0.1% per year difference. So that's not enough to explain the big differences. But I want to alert you that there are different ways of measuring aggregate output.

The third measurement issue of course relates to inputs, and Stew Wells covered some of this. What inputs should we include? All productivity studies include labour, so that's labour productivity if you just include that. Then if you add capital, multifactor productivity...except people like me say, what about land, and what about inventories? These should be included as well. There should be natural resource inputs and environmental variables. Stew Wells mentioned this could be done, and in fact Statistics Canada is working on a natural resources module in their suite of statistics. So in future years it may be possible to take this into account.

Just as an aside, in the Diewert-Lawrence computations, the rates I gave you just included labour and capital. But we also had some computations that included land and inventories. Adding these two inputs tended to increase our measured productivity growth from about 0.55% per year to about 0.7% a year. So that's just an aside.

• 1015

The fourth area where we have difficulties is related to the problem of decomposing value flows into their price and quantity components. It's relatively easy to measure values, so when you go to a business and ask it for its revenue or its costs, they can generally give you some idea of what they are. But then when you go further and say, “Of your revenues, can you break that up into price and quantity components for me? Tell me how many units of this good you produced and what the average price was on that good and what the average price was”—and you know the average firm is producing thousands, perhaps hundreds of thousands of products—it's a major job to decompose these value flows into price and quantity components. So that is a general problem.

There are specific problems associated with specific areas in the national accounts, so let's look at labour. Statistics Canada estimates labour income, but then to decompose that into a price and quantity component is not so easy either. The Diewert-Lawrence measure of labour over this period, 1962 to 1996, grew 235%, whereas the Statistics Canada growth was only 184%. There's a 28% difference, so most of our differences in our estimates can be traced to this labour problem.

What about the wage rate growth? Our wage rate grew more slowly, about 15% more slowly; so the Statistics Canada growth rate of wages over the period of 1962 to 1995 was 885%, while Diewert-Lawrence growth was 771%. It doesn't quite add up. If we're too big in our growth of labour by 28%, our wage rate growth should be 28% below. However, we're covering different parts of the economy; there are the hospitals and the educational sector that are different. However, there's still this huge difference, and I don't think that will explain it altogether.

I should say that I don't exactly believe our wage estimate is so good. It's from the OECD. It's a total compensation index, and the Statistics Canada measure should be much better than that. The focus in this paper was the alternative treatment of capital, and we didn't focus so much on the labour side. But that gives you an idea of how these numbers can depend on just the choice of one series.

Another problem area I'd like to mention is intermediate inputs. Firms produce outputs, but they also use materials and other inputs produced by other firms as inputs. We have a bit of a grip on the value flows there by industry, by firm, by establishment, but the decomposition of these value flows into price and quantity components is very sketchy indeed, and so there's a scope for a lot of error in this particular part of the accounts.

Treatment of capital: There are different ways of doing that. I won't go into it. It doesn't make that much difference for productivity purposes, because the share of capital is fairly small in total input.

The final problem in this fourth area has to do with the measurement of outputs, the problems of quality change, which Stew alluded to. When I started out my little simple explanation of productivity, I said, consider an economy with two goods; one output, one input. What do you do in an economy that's producing new goods, because then for the new output this year there's nothing to compare it to in the previous year? This introduces some slippage into the system, and it takes resources to figure out how to deal with this problem.

• 1020

So that is my fifth point: What should be done to improve productivity statistics? I believe you should consider giving Statistics Canada a bigger budget to work on these price measurement problems.

That may seem surprising, because you might have thought I was going to be a bit critical here. But I am a little bit critical in the sense that, you see, the statistical system has evolved following agriculture as the primary producer in the economy, into manufacturing, and now the economy has shifted into services in a big way. But the statistical system is still mired in the structure that was started or worked on in the 1940s and 1950s, and the whole statistical system hasn't moved to better coverage of services.

That is an exaggeration, because we have NAICS, the North American industry classification, which has wonderful detail on services. But to implement it is going to take resources. I believe it is very vital that we do that and that this is an urgent thing. If we're going to follow the new economy, we need to have the statistical tools to measure the new economy.

I conclude there.

The Chairman: Thank you very much, Dr. Diewert.

We'll now hear from Dr. Andrew Sharpe, from the Canadian Centre for the Study of Living Standards. Welcome.

Dr. Andrew Sharpe (Executive Director, Centre for the Study of Living Standards): Thank you very much for the opportunity to address the committee.

[Translation]

In the long term, the only way to improve the standard of living of Canadians is to achieve productivity gains. Therefore, understanding productivity measurements and trends is crucial to the policy development process.

My presentation will be in three parts: in part one, I will attempt to define productivity; in part two, I will review the history of productivity in Canada; and in part three, I will discuss productivity measurements.

[English]

Turning to my first part, on the definition of productivity, I will not repeat what has been said so well by our three previous speakers. I will just supplement the points that have already been made.

The first point I want to make on the definition of productivity is that there is a misconception in the public about what increased labour productivity means. Often people feel that increased labour productivity is caused by worker speed-ups—workers working harder, longer hours, or at greater intensity of work. It is true that one can get, at least in the short term, improved productivity from that source, but increased labour productivity really reflects working smarter, working with better organizational modes or with more human capital, or working with better equipment, much more than it reflects working harder. In that sense, it's important that the public understand that increased labour productivity is not through speed-up but through basically working smarter.

Another point I want to make in the definition area is the fundamental difference between labour productivity and total-factor productivity and how those concepts are used. Basically, total-factor productivity really shows us how efficiently we are using our total resources—human, labour, capital, and raw materials.

Labour productivity is really more how much output, basically, we can get for an hour of labour or a person working over the course of the year, and ultimately it is the labour productivity, not total-factor productivity, that's behind the increase in our living standards, measured as, say, GDP per capita.

I'll give you an example. Before the Asian crisis, in the 1970s and 1980s, east Asia had a massive increase in its standard of living through increases in labour productivity. The amount of output per worker in Korea, Singapore, and other tiger countries was going up by 4%, 5% and 6% a year. On the other hand, their total-factor productivity was not doing particularly well, because they were accumulating capital at a very rapid pace. So there were very large increases in capital input, and if you factor that into the total-factor productivity equation, it didn't show very well at all. But ultimately they were very successful in increasing the standard of living of their population.

• 1025

From my perspective, I feel it's labour productivity that should really be the key concern of the committee. Of course, in addition, there are very severe measurement problems in total capital, in measuring total-factor or multifactor productivity, which we don't have as much in labour productivity.

Another key definitional issue is the difference between levels and growth rates, and this was mentioned before. But in the productivity discussion, we also hear that Canada's productivity is doing poorly. What does that mean? Does that mean our growth rate in productivity has been slower than that of our competitors, or does it mean that our level productivity is lower than that of our competitors? Of course it's our level that determines our standard of living. Often people are confused between those two concepts, and it's very important when you speak of productivity whether you're referring to growth rates or productivity levels.

Turning to my second part, on the history of productivity in Canada, there are some key points here. The golden age of productivity in Canada was from 1945 to 1973. We had 3% productivity growth a year. If you look at the rule of 72, at 3% a year, that means we doubled our productivity, and hence our living standards, in 22 years.

Since 1973, the period known as the great productivity slowdown, our labour productivity averaged around 1% a year. Again, with the rule of 72, it would take 72 years to double our standard of living at a 1% productivity growth. So the overall slowdown in living standards in the last 25 years is to a large degree related to our slower productivity growth.

Of course, the causes of that productivity slowdown are still poorly understood by economists, and that's really a topic for another day. You could have a whole session on just the causes of slower productivity growth in the last 25 years.

Many people thought that in the 1990s we would see an improvement in productivity growth because of the information technologies. It hasn't happened yet, and this is known as the productivity paradox, why information technologies seem not to have led to an improvement in productivity. It may have happened in the United States, because the American productivity in the last three years has gone up enormously. Some people think that just reflects a strong economy; other people think it's a shift in trend in productivity growth. In Canada, we haven't seen as great an improvement in the last three years.

The important point is that in the 1990s we have not seen a deterioration in our productivity growth vis-à-vis the 1980s, and the 1970s as well. We have not done worse than the United States, and I guess that's another day's discussion when you look at international comparisons.

That's in terms of our levels. Basically our growth rates have been the same in the 1990s as in the 1980s, and the same as the United States. In terms of our levels, until 1973 we were improving our relative level vis-à-vis the U.S. because we were growing at a more rapid rate. Since 1973 we have basically been about the same as the U.S., so we're basically stalled at 80% of the GDP per worker employed that the United States does. So there's still a major potential to improve our standard of living if we can catch up to the American level. From that perspective, there's still a lot of potential in Canada for improved productivity growth.

Turning briefly to productivity measurement issues, again, Erwin went over these very well. I would like to reiterate that productivity data requires data on current dollar output, data on prices, data on capital stock and intermediate goods. The data requirements to develop productivity statistics are great, and all the problems in the statistical system of estimating various quantities come home to roost, so to speak, in the productivity data. We have all the problems that were mentioned in terms of the price indexes, the problem with prices of quality adjustment, the problem of measurement in sectors like the government sector, where we don't have a measure of output independent of labour input. Therefore, it creates a lot of work for economists in terms of productivity measurement.

The bottom line, though, I would say, is that the errors in productivity measurement would have a major impact on our productivity series. Often the errors are offsetting, and they've always been with us, so they're not a new phenomenon. I don't think the overall productivity measurement errors are going to have a major effect on productivity growth, particularly labour productivity growth, although of course there's lots of room for honest disagreement among experts in these areas.

• 1030

Now, to conclude, I think the committee in its hearings and deliberations should concentrate on what we know and not get bogged down too much in what we don't know, which is certainly a big area.

To summarize, we do know three things. The first is that in the 1990s the deterioration in our standard of living did not reflect a decline in productivity growth. The deterioration in our standard of living more reflected the poor performance of the labour market and the falling employment-population ratio. However, over longer periods of time, certainly since 1973 and into the future, our standard of living is going to reflect more and more the trends in productivity.

The second key point is that despite what's going on with the growth rates of productivity, we still have a productivity problem because of our much lower level. We're only 80% of the U.S. level of productivity, and we have catch-up potential there in going to maybe 90% to 95% of the U.S. level. That means there is still a major challenge facing Canadians in improving our productivity performance in terms of levels relative to the United States.

Thank you very much. I'll stop there.

The Chairman: Thank you very much, Mr. Sharpe.

Now we'll hear from Dr. David Slater. Welcome.

Dr. David Slater (Individual Presentation): Greetings, Mr. Chairman.

Being somewhat older than my colleagues on the panel, also a little out of the loop on the current productivity statistics, I'm going to offer some historical perspectives that may be helpful to you. I have made available to you a copy of a little article I wrote a couple of years ago called Setting the Scene: The Post-WWII Canadian Economy. This was an introduction to A History of Business Economics in Canada, which I wrote. I'm also going to refer at times to this little book on war finance that I did a few years ago.

You should have both my speaking notes and a copy of this post-WWII Canadian economy setting, which you can take away.

In this paper I begin with four summary points from Norrie and Owram's recent economic history of Canada. That's the best recent Canadian economic history. It's well written and a useful thing for the committee to have a look at.

The four summary features that they point out are as follows:

First, taking the whole of this experience, is significant growth and structural change for Canada.

Second, Canada remains a small, open society in commerce, ideas, and ideologies.

Third, in terms of the overall macroeconomic performance, there have been two distinct periods. The first was from the end of the war to the end of the early 1970s, which was a period of growth and prosperity as Andrew has just mentioned. In the second, which was after 1973, economic growth fell off sharply, inflation and unemployment rates rose, and government deficits soared. Even if the inflation and government deficits have receded from the 1980s perspective, the continued poor performance in productivity and unemployment add force to the fundamental Norrie and Owram question. Was the “golden era” of Canadian economic performance during the first three post-war decades the norm and the subsequent two decades an aberration, or the reverse?

Fourth, they point out that the economic difficulties after 1973 seriously challenged the faith in economic management.

Now, when looking at the golden era of Canadian post-war economic growth, I argue, first, that it should be examined against the background of the Great Depression and the enormous accomplishment of Canada's wartime economic performance. You get a different view of the thing if you go back to 1939 and you take 1945 or 1960 as your starting point for comparison. I'm arguing that to get a real perspective on Canadian performance, you really have to go back to 1939.

Out of a labour force of not much more than 5 million people, at the peak of the war effort about three-quarters of a million were in the armed forces, and another three-quarters of a million were devoted to war production. Yet in a few years, Canada's real gross national product increased by 60%, and its output per employed member of the civilian labour force by about 30%. The difference in output and productivity between a fully employed and an underemployed economy was demonstrated, I believe, for all time for Canadians. I've stuck on the back of these reading notes a little summary of growth indicators, including real GDP, civilian persons with jobs, and a kind of “productivity index”. That's the first point.

• 1035

The second point I would make is that the wartime productivity experience was attributed by some analysts to special wartime conditions. We concentrated production on long runs of a small number of goods and activities, in contrast with the pre-war Canadian industry, which produced small amounts of nearly everything behind high tariff walls. We had exceptional export opportunities during the war due to wartime financial arrangements and the cut-offs of competing sources of supply. We had exceptional work commitments by Canadian men and women.

Now, there is something to this story about these special conditions, but I believe the most interesting feature is that the wartime productivity improvement did not disappear after the end of the war, when presumably those special conditions disappeared. Instead, with very little pause, Canada's output and productivity continued to increase strongly in the quarter century after 1945.

Third, the conventional economic historians' theories of Canadian economic growth turned out, I believe, to be inadequate and incomplete. These were, first, the Mackintosh-Innis theory of export staple development; second, a heavy emphasis on government and government-assisted megaprojects; and third, the inflow of American capital into resource exploitation.

They are all elements of the successful Canadian performance after the end of the war. But I argue that too little weight has been given to the strong indigenous factors that led to a remarkably large and effective development of the Canadian domestic economy, rooted in Canada's wartime experience or its private as well as public efforts at post-war reconstruction and development.

I sketch in this paper some of the fundamental changes to the Canadian economy between the beginning and the end of the war. The economy became more productive and diversified. We had a more confident skilled labour force. New industries that were on the cutting edge of advanced technology were introduced. There was the development of supplier networks that didn't exist before. For the first time in Canadian history, there was a fully developed internal migration network, and there was the lifting of the yoke of foreign debt, which had been a curse for us, particularly through the 1930s.

I add, looking to the post-war period, the latent potentialities of the development in the early post years, such as the world's accumulation of scientific and technological knowledge that had been underexploited during the Great Depression and during the shrinkage of world trade between the wars; the educational opportunities that could be given strong support at the end of the war; and the potential sources of high-quality skilled immigrants from Europe. More Canadians than ever before, at the end of the war, had significant financial assets from their accumulated wartime savings. Canadians as well as their governments were determined not to blow their opportunities on an inflationary spree at the end of the war. They wanted to achieve something for the huge wartime efforts they had made, something of lasting value, and that coloured private and governmental activities during this period.

In my paper, I pick out a few markers of the golden age performance. I'll leave to your reading the comments about the post-war Canadian agricultural revolution, which combined enormous increases in productivity with massive reductions in the agricultural labour force.

I'll leave to you also the post-war triumph of the car and truck and everyday transportation for Canadians. The triumph of the car and truck was post-war, not pre-war, in Canada. There was the replacement of networks of gravel and mud roads with paving—and any of you who have driven Saskatchewan mud roads in the spring or gravel roads in the middle of the summer will know that they were not exactly a joy.

I do draw your attention, however, to two things I make note of in the paper. One of the greatest scarcities at the end of the war was housing. In fact, the nation's housing stock was smaller in 1945 than it had been in 1929. In the first 15 years following the war, 1.6 billion new housing units were built, an increase of roughly 50% in the total housing stock. Moreover, most of the new housing met the minimum standards of the National Building Code, which was of enormous importance.

• 1040

This was a tremendous accomplishment of the domestic economy. It took place despite the residential mortgage market being a shambles at the end of the war, except for persons with substantial assets. I remember that Queen's University lost most of its endowment on Saskatchewan farm mortgages in the 1930s.

Clifford Clark, David Mansur, and the president of North American Life, who was on loan to Clark, devised the multi-faceted mixed public and private programs and institutions that led to more and better housing, city design and infrastructure that was needed, all under the leadership of the Central Mortgage and Housing Corporation. By the 1960s, that corporation was able to bow out of the mortgage lending business, although it continued in mortgage insurance and financing of social housing.

Capital investment from domestic sources was also an important factor in post-war economic growth. I refer to a paper of Malcolm Urquhart from 1988, which is the classic paper on that subject. While there was significant net foreign capital inflows in the middle to late 1950s and the late 1970s, the dependence on such inflows for the high amount of investment activity that occurred in Canada in those years was much less than during other periods of high investment. To put the matter differently, the bulk of the post-war investment was financed by domestic Canadian saving.

In the paper I run through a whole series of the factors, policies and so on that contributed to the success of this golden era. I'm going to leave those to you.

They lead me to three generalizations that may be relevant to your current productivity concerns. First, the Canadian growth experience between 1939 or 1945 and the early 1970s was sui generis by historic standards and in part reflected a conjuncture of transitory forces. The untapped potential of scientific and technological knowledge was unusually large, and likely to be a less powerful force after 25 years of exploitation than at the start.

Eventually the most pressing needs for housing, sewers, sewage treatment plants and roads would be caught up. Eventually machines would relieve some of the opportunity for well-paid manual labour in the mines, forests, fisheries as well as agriculture. The best and most easily exploited resources had been brought under economic development. Excess labour in the agricultural sector had been moved out. Opportunities for increased productivity continued to exist, but they were less rich and provided benefits on a less even-handed scale across the population.

Second, by the early 1960s—as evidenced by the first reports of the Economic Council of Canada—continued productivity growth at the post-war pace was pretty much taken for granted. Policy attention turned to completing the social and tax programs that had been envisaged in the documents of the Dominion-Provincial Conference on Reconstruction at the end of the war, but not executed.

The federal-provincial fiscal transfer arrangements were strengthened, public health insurance was introduced, the Canada Pension Plan and the Quebec Pension Plan were introduced, and the GIS program was strengthened. I'm not arguing against any of those things. Those are the things that caught the attention, rather than improving productivity.

It was assumed that the country could meet the additional costs of these new or large programs from the largesse of continued growth in output and productivity. It's interesting to go back and read the first two or three reports John Deutsch turned out at the Economic Council of Canada. There was never a whisper in those reports that prospects for output and productivity would be any less in the period following the middle 1960s than they had been since the end of the war.

My third observation is that the mismanagement of fiscal and monetary policy from the 1970s on made the performance of productivity, employment and unemployment much worse than their potential. Post-1973 potential was probably below that of the golden era, but not all that bad by historic standards. But we fell far below the potential most of the time.

• 1045

Thank you, Mr. Chair.

The Chairman: Thank you very much, Dr. Slater.

We'll now hear from Dr. Dale Orr. Welcome.

Dr. Dale Orr (Economist, WEFA Canada Inc.): Thank you, Mr. Chairman.

I've circulated a copy of this paper and I'll be referring to some charts you'll find in it. This paper is an update of an earlier paper I wrote in December. I have a couple of comments on the conclusions from that earlier paper.

First, Canada's standard of living grew more slowly in the period from 1990 to 1997 than over the 1980s. Second, Canada's standard of living grew more slowly over the period from 1990 to 1997 than that of the U.S.

I have a couple of comments on definition. For purposes of what I'm doing today, it's very important to keep at the front of your mind that our standard of living can be decomposed into two general terms—a productivity set of terms and an employment-labour market performance set of terms. That's been outlined by Stew Wells. In each of those productivity terms there's a couple of components, and there's also a couple of components on the labour market side.

Just to tie what I'm going to say in with an important point Andrew Sharpe made, I will also talk about rates of growth of productivity in the 1990s compared to the 1980s, and rates of growth of productivity in Canada relative to the U.S. I will focus entirely on the economy overall. I don't have comments to make on individual sectors.

When I looked at the determinants of our standard of living over the 1990 to 1997 period, I concluded first that Canada's weak productivity performance was not the most important reason for our weak performance with respect to standard of living. The most important reasons for Canada's weak standard of living performance were related to weak employment performance.

The decline in our labour force participation rate accounted for 74% of the decline in the improvements in our standard of living, when we compare the 1990s to the 1980s. Weaker productivity growth only accounted for 12% of that weaker standard of living performance. This general conclusion is consistent with what you heard from John Baldwin and Andrew Sharpe.

Our failure to match the Americans' rate of growth in living standards over the 1990 to 1997 period was due about equally to relatively weaker productivity, relatively weaker reduction in the unemployment rate, and relatively weaker labour force participation.

To put that in colloquial terms, I would say the weak performance in our standard of living is not so much due to the lack of productivity of those who are employed; it is due to the fact that a lower fraction of our population is employed. In turn, the fact that a lower fraction of our population is employed is not so much due to people willing and able to work but not being able to find work, as it is to the fact that a growing portion of our population was not actively seeking work over that 1990 to 1997 period.

There are several reasons, in turn, why the labour force participation rate fell over the 1990s. Some of these are a policy problem and some of them may not be. Of course, there are the discouraged workers, but there's also the fact that a higher fraction of people was taking early retirement and a higher fraction was going to university.

I turn from that to put, as a background for my comments, the OECD's warning they set up with their report in December. The OECD report warned that Canada was likely to suffer a major setback in its standard of living unless our productivity performance, as well as our participation in the labour force, improved significantly.

What I bring to you today is an update of the economic forecast for Canada over the next five years, as well as an update of the economic forecast for the U.S. over the next five years.

• 1050

My company, WEFA Incorporated, is, I think, the world's largest economic consulting company. You can see on the back page that we have about 15 offices in 90 countries of the world.

These are the forecasts for Canada and the U.S. that were made just this month, which I'll be commenting on.

Can we expect our economic performance to improve? If you look at the table on page 5, you'll see that over the next five years we expect economic growth to be a little bit stronger, at about 2.5% a year, relative to the 2% average over the 1990 to 1998 period. Employment growth is also forecast to be considerably stronger, at just over 2%. That's quite a bit higher than the 1% we averaged over the earlier part of the 1990s.

Productivity, if we measure it as GDP per employee, for example, is forecast to barely grow at all. So the good news is higher output growth. You might think higher employment growth is also good news, and from many perspectives it is, but when you get very high employment growth and you're measuring productivity as GDP per employee, productivity will not be strong over the next five years.

Will productivity growth in the future keep up with the Americans? The table on page 6 shows that the current forecast for the U.S. economy over the 1999 to 2004 period calls for economic growth to average 2.4%. With employment growth at 1%, this leaves productivity growth at 1.4%. In the future, therefore, WEFA expects the U.S. economy to grow at almost the same rate as the Canadian economy. The more rapid employment growth in the Canadian economy will result in significantly lower productivity growth than in the U.S. economy. I should note here that where I'm using GDP per employee for productivity, I'm also assuming that the hours worked per employee over the next five years will be at about the level it was in 1998. Nobody has really good forecasts for changes in hours worked per employee.

Turning from that to forecasts of our standard of living, will we do better than in the past? If you look at the table on page 7, you'll see that over the 1990 to 1998 period our standard of living, as measured by GDP per population, grew at an average rate of just under 1% per year. The good news is that over the next six years, our standard of living is forecast to grow at a much faster pace, about 1.4% a year.

In terms of interpreting these forecasts, the key point is that economic growth is forecast to pick up, and employment growth is expected to pick up even more sharply. We're expecting a healthy increase in our standard of living measured as GDP per population, even though our productivity growth is forecast to slow down. Over the next few years improvements in our standard of living will be driven by improved labour market performance rather than by productivity performance.

Will we close the standard of living gap in terms of our improvements? Will we improve our standard of living over the next five years as much as the Americans will improve their standard of living? The table on page 9 indicates that over the medium term the Americans are forecast to have about the same rate of GDP growth but slightly weaker population growth. This combination will lead to an average increase in standard of living of about 1.5%, which is just marginally above our 1.4%.

There are a couple of things to conclude from this work. The most important is that given these most recent economic forecasts of the Canadian and American economies, our standard of living is expected to increase at a much better pace over the 1999 to 2004 period than it did over the earlier years of the 1990s. In fact, it's 1.4% versus 0.7%.

We're forecast to almost match the pace of the Americans over the 1999 to 2004 period. Canada's improved standard of living over the forecast period will be driven by a fall in the unemployment rate and an increase in labour force participation. This will permit a stronger pace of growth in the standard of living in spite of falling productivity growth.

• 1055

Improved productivity is the only route to improvements in our standard of living over the longer term. However, our standard of living is determined by several labour market performance factors as well as by productivity. It so happens that over the 1990s our weak standard of living performance was caused primarily by weak labour market performance rather than weak productivity. Over the medium term our standard of living performance is expected to strengthen as a result of stronger labour market performance in spite of worsening productivity performance. The recent economic forecast data indicate that with regard to improvements in our standard of living, the path we are on is not too bad, but we can and should do better.

The key issue for this committee and the government in choosing the theme for the next budget is whether you want to focus on productivity. If you do focus on productivity as we've defined it, you will miss the important labour market factors that have been so critical in determining changes in our standard of living in the past decade and are forecast to be most critical over the medium term. However, I think you have a dilemma. If you focus more broadly than productivity, to focus on policies that will improve our standard of living, can you then provide a meaningful focus for the budget? Has the government ever spent any money on anything that couldn't be justified as increasing our standard of living?

Thank you.

The Chairman: Thank you very much, Dr. Orr.

We'll now hear from Dr. Rick Harris, an economist from Simon Fraser University. Welcome.

Dr. Rick Harris (Economist, Simon Fraser University): Thank you, Mr. Chairman. I'm glad to be here today, and I wish you well in your endeavours.

After teaching productivity for a number of years, I must say that I hope you do better than my average student, whom I apparently managed to confuse thoroughly on this issue. Having read the media, I think we tend to be doing the same thing here in Canada.

So what is productivity? I think we all know what it is in terms of our own life. It's a measure of how efficiently we produce whatever it is we produce, and then going up to the economy-wide level, it's a measure of how efficiently the economy uses its resources to produce goods and services.

I think maybe a way of putting it in context is to think about what productivity is not. Productivity is not consumption. Consumption and production are quite different things, particularly in a very small, open economy and in an economy in which a lot of services are provided at the household level. Productivity is not incomes, productivity is not wages, it's not profits, it's not employment. We heard from many witnesses today that of course these are very important issues, and Canada shows quite a different performance with regard to some of these measures than it does in productivity. It's obviously related to productivity, but it's not the same thing.

Productivity is not asset values. Productivity is not the value of the NASDAQ or the S&P 500 or the Dow Jones. Many people argue that these are closely related. The so-called new economy view is that high rates of productivity growth are driving these asset valuations, but when economists discuss these issues, this is not what we mean.

Nevertheless, at the intuitive level I think the most important relationship is productivity to long-run standards of living or income. Many economists have written about this historically, beginning with Adam Smith, and this is probably the largest question in economics: what drives the wealth of nations? It would be far beyond me to discuss all of those things today, except to argue that this is a very big issue and, quite frankly, the productivity issue confronts virtually all points of view on what it is that determines the wealth of nations, including the very important issue of the role of markets, government, and individual incentives.

One thing I do want to note, particularly in the Canadian case, is the difference between income and production. Let me give you some examples. Take the oil industry. Syncrude has remarkable productivity levels by any international standard, but it has very low prices for what it's producing at the moment. Fortunately, things have got a little better recently for those of us who own oil stocks. Prices are a very significant input to our standard of living, but productivity does not measure prices. It measures the efficiency with which we produce the goods that are then sold in the marketplace.

Take the example of Microsoft. How can you become rich? Have a world-class monopoly. In fact, have lots of them if you can. Monopolists have high revenues. They don't need to be productive. The point is that they have a very secure and protected market.

• 1100

Canada could become a nation of Céline Dions, artists. Artists have no productivity growth at all, by definition, but their value, what they produce, is highly valued in the international marketplace. So we have to distinguish what's valued versus how productive we are at doing it. As a nation, it's very important that we do both of these things, so it's important to extend our discussion beyond the simple question of productivity to the question of producing value-added.

We heard a lot today about the measurement to date. I'm not going to say anything more about that, but I do want to turn to two broader questions. One is, what do professional economists looking at the broader historical evidence know about the determinants of productivity? Assuming we could measure it, which of course, as we talked about here, is a big “if”, there's a very substantial part of our profession that is engaged in the business of trying to figure what it is that drives productivity growth. I'm sure you're going to be confronted with this question in your deliberations.

I have read hundreds and hundreds of those studies, and I can tell you that there are literally only three broad determinants for which there is any compelling evidence.

The first one is investment. Investment is overwhelmingly the single largest determinant of productivity growth in the post-war period. No matter how you cut the data, high investment-to-GDP ratio is the single biggest number.

The second one is human capital and skills. You don't need to be a rocket scientist to notice that people who have a lot of skills are being paid a lot of money. There's a reason for that: they are very productive. This comes out in virtually all the evidence, and the evidence gets stronger every day. This is creating a lot of problems for income distribution in countries like Canada, but it's the reality.

The third one is trade, especially for smaller countries. Countries that are open to international trade consistently have higher productivity growth rates than those that are closed to trade.

Beyond that, there are many, many worthy candidates of productivity change, but the evidence is all substantially less decisive—things like R and D, population growth, taxes, inflation performance. You can go on and on. There are literally hundreds of studies. But quite frankly, when you use a broad variety of international comparative evidence in the post-war period, it's difficult to make those cases, however virtuous or compelling they may be at the intuitive level.

However, I think it's important to emphasize that even after we have controlled for what we think determines productivity growth, the remaining uncertainty is absolutely tremendous. If you take a typical cross-country growth regression that controls for all of these variables that we think are important, you can move a country by a factor of two, in terms of its level, in less than 15 years, just accounting for what you don't know. So the point about our ignorance is a very compelling one, and I think the basic issue of the measurement is that it would be nice to reduce the ignorance, but it's going to be difficult.

My last point is about policy levers. This is where the rubber hits the pavement on this subject. Ultimately, one is going to have to go from productivity to making some statement about policy. Measurement is important in this, although it's important to remember that measurement is backward-looking, not forward-looking, and that is a significant problem.

Are they a useful guide to policy? I can think of two important examples.

A very important historical example in Canada was the Auto Pact. Prior to the Auto Pact, industry economists had done detailed industry studies of Canadian automobile plants and U.S. automobile plants, and the productivity numbers were very compelling. There was a very substantial gap of about 40% to 33% at the time. Ultimately, that led to the Canada-U.S. Auto Pact, which was the precursor to now, probably our most successful export industry and an example of a spectacular consequence of a combination of good data with good policy.

So there's an example where you can use the case that good measurement exercises lead ultimately to good policy. However, that has not always been true. If you look at studies, sometimes things don't work out. If we look at our resource industries, Canada has had a long history of very successful, capital-intensive technological development in resources, but right now, resources are not proving to be much in terms of an income generator. If you look at long-term forecasts for commodity prices, no matter how productive we get at taking the stuff out of the ground, it's not going to get us rich. At least that's one forecast. That might turn out to be wrong. But the point, again, is that productivity studies, in and of themselves, do not necessarily lead to a forecast of those future sectors that might do well and will ultimately lead to jobs and income.

• 1105

In your deliberations, you are going to hear from a variety of interest groups. I can just imagine the list of things that are going to come up—taxes, tax reform, subsidies to innovation, more R and D, improved foreign access to the Canadian markets, do something about the lazy small firms, do something about the dynamic small firms, brain drain, better education, manpower training, reduced trade barriers, and so on.

The problem you're going to have is that the numbers that are talked about here today by the measurement community bear a rather distant relationship to those policies in most instances, even though we all have our opinions and we're willing to sort the scientific evidence that exists about their impact. This is the nature of the beast.

Productivity change is inherently long-term change. It involves very long forward-looking policies, which in some cases involves a combination of good luck and good policy. That's just the nature of the problem.

Think about human resource development. We now have compelling evidence being produced by child psychologists and other people that very early infant childhood development is extremely important in terms of subsequent human capital acquisition. You literally have to have a 30- or 40-year horizon when you talk about developing those kinds of resources.

Secondly, there's what's called the technological wave, or the general purpose technological development view of history, which is that technological change, which is the big driver of income in the economy, which simply drives what we ultimately measure as productivity, is characterized by these big waves. The last wave was the Fordism wave, electricity, industrialization, which hit the economy in the 1920s and then came to fruition in the 1950s and the 1960s.

Now, of course, the argument by these advocates of the new economy view is that information technology is a new wave, and if we take that view seriously and we look at what historians have told us about this, it may well be the case that one has to wait 20 to 30 years before productivity statistics ultimately reveal the actual improvements in the real quality of goods and services, real consumption incomes that we care about.

So I think you have a very difficult job, and I wish you the best in your deliberations.

The Chairman: Thank you.

Now we'll get to the question and answer session. Mr. Epp, we'll begin with you. I'll give you approximately ten minutes for your questions.

Mr. Ken Epp (Elk Island, Ref.): Thank you, Mr. Chairman, and thank you to all of you who presented us with what I believe must have been lucid presentations, even though some of it went over my head.

I have a whole bunch of questions with respect to the measurement of productivity. It seems to me that the general population, myself included, do mix thoughts of productivity with standard of living. We all have this sense that we're not as well off as we were 20 years ago. That certainly is true of my family.

When I first came out of university, that was before the days of the technological on-off switch on having babies, and so ours came right away. We paid our rent, we paid our utilities, and with the money left over we went on a vacation. Twenty-five years later, we pay our rent, we pay our utilities, and we wonder whether we're going to meet our monthly payments on anything we borrowed, and we have very little left. So our standard of living seems to have gone down.

I have my own theories about that, but I've been told by you not to mention it, and that is that the tax bite has nothing to do with productivity. But I have an idea that it does greatly affect our standard of living. So we have a bit of a problem here in that I think we need to really discipline ourselves to talk about productivity and to not confuse it with standard of living. Am I right, and do any of you have anything to add to that comment?

Dr. Andrew Sharpe: That's a very good point. For example, we've been talking here today about GDP per capita, which has been up slightly in the 1990s, and using that as a measure of standard of living. But there are other measures of standard of living related to that, for example, personal income per capita or disposable personal income per capita.

If you look at personal income per capita, there haven't been any gains in the 1990s. If you look at 1989 as the peak year and then look at 1997, there's basically no change on a per capita basis in real terms.

If you look at disposable personal income per capita, we're actually down 5% in 1997 vis-à-vis the peak year in 1989, because of increased taxes. So from that perspective, one is worse off.

• 1110

Of course, one has to realize that the increased taxes have gone to serve Canadians in many ways, such as paying down the debt, increased transfer payments and so on. But I think the public perception is related more to real disposable income per capita, as opposed to GDP per capita. That definitely has done worse than GDP per capita.

Mr. Ken Epp: But you were talking about the standard of living right now.

Dr. Andrew Sharpe: Right.

Mr. Ken Epp: I have a question for Dr. Baldwin. He mentioned something about the measure of productivity being the first difference. I have a meagre understanding of this since I used to teach math at a technical institute, so I know a little about the first derivative. But the phrase was used here that this is the first difference of first differences. I don't really understand that.

Are you saying that if the productivity in a country remains unchanged, we say that productivity is zero?

Dr. John Baldwin: No. Let me try to give you a simple analogy. To work on something Erwin was talking about—the manna from heaven phenomenon that many economists associate with this productivity—if we get a 6% increase in output and we've increased all of our inputs, however measured and aggregated, by 5%, then those two are differences. We have a rate of change of 6% and a rate of change of 5%. We measure the growth in productivity as the difference between those. If we take 6% minus 5%, we have 1% growth from somewhere that we didn't plan on, so to speak. That's what I meant by a first difference of first differences.

Mr. Ken Epp: Okay. So that's the growth in productivity. But how do you measure the productivity itself?

Dr. John Baldwin: That reverts to the issue Dr. Wells referred to earlier, which was the issue of levels. How much do we have in terms of output per person or output per unit of capital? That's simply the measure of output, however measured, in 1992 dollars or otherwise, divided by the measure of inputs or number of people, as Erwin has used, or number of hours worked, as we have used.

Mr. Ken Epp: Would you all agree that the best measure is probably gross domestic product over population? Is that the best single measure?

Mr. Stewart Wells: The best measure of what?

Mr. Ken Epp: Of productivity.

Mr. Stewart Wells: No.

A witness: No.

Mr. Ken Epp: Good. There's unanimity there.

Mr. Stewart Wells: It's a measure of the gross standard of living, without paying attention to some non-monetary factors and the distribution of income.

Mr. Ken Epp: Okay, so I made the error I accused myself of—mixing up the standard of living with this.

Mr. Stewart Wells: Yes.

Labour productivity at the global level is GDP divided by some measure of employment. There's a difference. You can move from there to the standard of living measure, taking into account changes in the rate, population and hours worked. But the GDP per capita is a standard of living issue, not a measure.

Now that I'm speaking, I'd like to introduce something that is, I suppose, on a personal level. I've listened to all of the stuff about taxes too, and I agree with Andrew's responses. But he kind of slid over or glided past this notion that, well, the taxes might be spent on something. Personally, I think that's something Canadians undervalue to a very great extent.

We're in a time when a lot of people seem to think taxes are being sucked into a black hole, never to appear again, and they just put a drain on the system. Yet when you talk to people and ask them to compare their state of living here with what it would be like in other countries they've visited, almost inevitably they feel very good about this country. A lot of that is because taxes are spent on things we want. When we're talking about the standard of living, we should not forget that. That's a personal view of the world.

Mr. Ken Epp: So you agree with United Church Minister Phipps, who says we should just have a lot of fun and enjoy paying our taxes.

Mr. Stewart Wells: I do. I have a relative who got quite cross at me when I said I didn't mind paying taxes. But yes, I basically agree with that.

Mr. Ken Epp: By the way, I grew up on a farm in Saskatchewan, and my dad used to say “Don't ever complain about paying taxes, because it proves you have an income.” So I suppose there's some truth to that.

• 1115

I'd like to ask about the population effect. It seems to me that there are several factors here that affect productivity. One is simply the gross number in our population, but secondly, there is also the degree to which they can be productively employed, and this is dependent on education. I think of this very greatly because I worked in education in a technical institute for a great number of years, and I thought I was making a positive contribution to the productivity of the country because we were training students to do things they would not have been able to do without that training.

So I think education is a very important input factor into our productivity, but you're saying you don't measure education in terms of the productivity itself, which is understandable; it's an indirect input cost. But I don't think you're suggesting by this that we should reduce our emphasis on educating our young people. So would you please tie that together?

The Chairman: Mr. Harris.

Dr. Rick Harris: You're absolutely correct. The output of the educational sector is what results in the stock of skills and human resources we produce. As I said, amongst the determinants of productivity growth that we can identify clearly, it is the second most important factor. So you're absolutely correct. It's important to remember that these measures you're talking about of productivity are largely accurate with respect to the private sector. We're basically talking about private sector business productivity.

Most people have given up trying to measure, because of the problems we talked about, the output of the public sector—in particular, say, the educational sector—but we do know that the productivity of the private sector is clearly related to the skills and education of their workers. And the skills and educations of their workers are in turn related to the activities of the education sector, so you're absolutely correct.

Dr. David Slater: If I may, I would add two points on this. First, one of the phenomena that were significant in the 1950s and 1960s was a very large inflow of well-trained migrants. The skills in working with stone and brick and so on that came with the Italian migrants were just fantastic. We in fact in this country relied for our highly skilled artisans, including tool and die makers and so on, very largely on the importation rather than generating these people ourselves.

I think that one of the really important neglects, despite several serious tries, has been the neglect of the development of highly skilled artisans in this country, both the development of them and the respect of their role and contribution. It's not just a matter of banging people through university degrees.

The Chairman: Okay.

Mr. Stewart Wells: Could I add to that, Mr. Epp, that in Statistics Canada, as John explained, we don't include public education, but it's not because we don't believe Rick and Dave aren't both absolutely right on this issue, it's because we don't know how to measure it directly. We think it feeds in clearly indirectly, but we don't know how to measure it and we tend to measure it then by the number of people working in the education system, or something. That means you're defining away productivity growth, and it's why we restrict ourselves to the private sector.

But you're absolutely right that education is an extremely important...I guess Rick says the second most important contributor to productivity growth.

Mr. Ken Epp: Having worked in that field, I think I'm very cognizant of the danger of using statistics to measure this, because in the education environment one of the measures often used is what percentage of your students passed? I can get them all to pass; just give them a passing mark and they'll all pass. On the other hand, if two-thirds of the students for some reason or other did not grasp the material, whether it was the fault of their instructor or any other factors, and I failed two-thirds of them, I was considered a failure.

So the measurement is very important, and I think the statisticians and others who measure productivity have a very large burden of responsibility here to make sure we're measuring something real and not something that's just a phantom.

The Chairman: Thank you.

Dr. Baldwin.

• 1120

Dr. John Baldwin: I want to reiterate a point that Rick has made. I emphasized earlier that the numbers we build up at Statistics Canada are built up at the industry level; so we can look at the differences across industries and what they're correlated with and it doesn't take much time before the issues that Rick raised become perfectly evident. The rates of growth at industry level are much higher in industries where measures of the education of the labour force are higher. These productivity growth rates are definitely higher in those areas.

They're also much higher in areas where you have much greater capital investment taking place at the present time. Even after we've taken into account our capital stocks in estimating productivity, we find these very strong relationships. They're there and they're easy to see, and whether one calls them determinants or correlates in order to understand the system, it's quite evident that the education system feeds into the estimates we have. So even though we miss the direct productivity performance of the education system, we're catching it across the entire system.

The Chairman: Thank you. Mr. Epp, your time is up, you know that.

Mr. Ken Epp: I don't have time for one more quick one? Okay, maybe you'll let me get back.

The Chairman: I'm going to actually follow up on one of the comments you made, Mr. Harris, when you referred to investment, human capital and skills, and trade as the three major determinants of productivity growth. Then Mr. Wells said taxes really don't play into it. But taxes have a direct impact on investment. Can I have comments from both of you on that issue?

Dr. Rick Harris: Obviously taxes affect investment, but the point is we look at determinants of productivity growth, and once you've controlled for the level of investment, and then we look for an additional impact of taxes, that appears to be small. But no one is saying that taxes are not an important determinant of investment. So I think we have to talk about the immediate approximate correlates or determinants of productivity growth versus other.... Infrastructure, legal system, taxes, all these things can be incredibly important, but we don't have—this is what I'm talking about—the direct connection.

The Chairman: Mr. Brison.

Mr. Scott Brison (Kings—Hants, PC): Thank you, Mr. Chairman. Thank you all for your erudite interventions. It's been very helpful.

To tag onto the chairman's point relative to taxation and investment levels, as you suggested, Dr. Harris, it's difficult sometimes to develop the policies for where the rubber hits the road, I think you were saying.

Government cannot directly increase investment in the economy, but it is through policies like tax policy that government could if there was direct government investment. Some would argue that's a good thing, but we don't have quite the flexibility we once may have had to do this sort of thing. If you look at Ireland's example of where tax policy has had a very direct effect on foreign direct investment in Ireland, I would argue that this capital—and I would appreciate your feedback on this—has enhanced productivity and growth in productivity, particularly in high-tech industries because of the nature of the tax relief provided.

Other structural impediments to productivity that may exist, and on which I believe we could effect change, would be the regulatory burden that may exist in Canada relative to other jurisdictions. Again, I would appreciate feedback from any of you on that and whether or not the regulatory burden that exists is something we should take a hard look at. These are things like interprovincial trade barriers, and you mentioned that as one issue.

That would be one point. I'm going to mention a couple of things here, and perhaps that would be more productive ultimately.

The other area would be that the Statistics Canada information or analysis does not take into account health care and education. Those are publicly delivered by and large in Canada. The growth in these areas, particularly in the U.S., has been on the private sector side. There is strong movement in the U.S. towards private delivery in education, and in fact there has been in health care as well, but there are arguments to be made that private delivery of education and other public services can in fact improve the degree to which those public services meet the needs of private individuals.

• 1125

Isn't it a little bit behind the times to not include the delivery of education and health care in the statistics? If our trading partners are delivering those privately, and if we're comparing our productivity numbers to our trading partners'.... Perhaps we're lagging behind in those areas, but they should still be accounted for, because there may be better ways.

Lastly, in regard to labour market flexibility, to what extent should we be looking at labour market policies—everything from minimum wage to employment insurance policies, etc.—that may create disincentives in the economy for what would be productivity enhancement behaviours individually or collectively, whether it's by our corporate communities or individual Canadians? Labour market flexibilities would be one area for which I would appreciate your feedback on policies we should be looking at.

Mr. Stewart Wells: I'll address the treatment of education.

Let me just repeat that the basic reason we don't include it is that we're not happy with the way we measure output in that area, and if Erwin Diewert gets his way, we'll get more money to spend on such measurement. I'm being a bit facetious there; we could always use more resources. But we are exploring better measurement of education in that area, and if were to do a better job of that, we'd be more inclined to talk about productivity for the total economy than we are currently.

Now, you're quite right, I think, that if in the U.S. they put education in, and more importantly get it in the business sector.... I don't believe they measure any better than we do; therefore, it might lower the U.S. estimates of productivity growth a little compared to ours. That's possible.

Our data would suggest that's not going to make serious differences in our comparison of the growth rates, and on the total business sector we've done a little better that the States in recent years. So I don't think that's really distorting our comparisons at the present time. But I can agree that we would like to get better output measures of hospitals, schools, and certain other information-related services. We do a lot of services fairly well, but there are some we don't do as well and we have to work toward improving them. You're quite right.

Dr. David Slater: Can I add something on the tax side?

I have two or three points. You mentioned Ireland. Well, we all know that individual jurisdictions can play the game of offering tax concessions to suck things away from other jurisdictions. If everybody plays that game, nobody wins, and the record on that is pretty clear. The states in the United States have played this game on and on. I had occasion to look at this a few years ago, and the net pay-offs in terms of real gains to the nation are very minimal out of that.

The second point I would make is that of course a good regime of business and personal taxation is important. On the business taxation side, you do have in hand a splendid piece of work that was done by the Mintz committee, and Rick's wife contributed a wonderful chapter on resource taxation—if I'm allowed to give her a pat on the back.

The third point is that I think the evidence is really pretty clear that when you factor the differences in the cost of how health care and education are financed, what taxes are, etc., Canada and the U.S. make fair comparisons.

• 1130

Bob Brown, who is adviser to the Department of Finance, has done very careful work, and he's not a guy who gets up and screeches about anything.

It is pretty clear that at the moment, in the Canadian tax and transfer system at the personal level, people in the upper middle range are taxed more heavily than they are in the United States for many, many comparable groups of people. Probably the single biggest factor, of course, is that the top marginal rates in Canada apply at such a low level and the top marginal rates in the U.S. apply at three times that level. It's a very, very big difference.

There are things that can be done, but again I would be inclined to go along with Rick's observation. I don't think that if you were to improve the tax system significantly it would be your primary answer to improving your productivity levels or rates of growth.

The Chairman: Dr. Harris, what do you think of that?

Dr. Rick Harris: Now we're getting into an area where people can have reasonable disagreements. I'm much more enamoured of the Ireland example than I think David is. It seems to me that Canada, as a small country, like it or lump it, cannot change what other countries are doing. And the fact of the matter is that we do not have economic activity in Canada. We do not have the investment. And you can have as high a productivity level as you like; it's completely irrelevant. If you read the Mintz report carefully, I think you'll see that's essentially what they're saying. So yes, taxes are an important determinant on investment and on work effort, and maybe they're affecting the migration of skilled workers out of Canada at the moment. So these are important things.

Coming back to my basic point, these are basically what economists think of as traditional resource allocation issues. It's very difficult to make a strong case that you're going to drive a very significant long-term increase in your growth rate, but it may help a great deal to lower the gap between Canada and the United States that Andrew talked about. So I think those are two separate questions, and it would be helpful if you could keep them clear.

The Chairman: Dr. Diewert.

Dr. Erwin Diewert: Yes, I'd like to respond to some of your points as well, and I agree with Rick on the Irish example. If we reduce taxes on capital, then economic theories suggest we'll have more capital accumulation. This may not increase productivity, except for the correlation of investment with increased TFP, but that's tenuous. We don't understand exactly why that occurs. But this does play into what Rick finished up with and what Andrew is saying. It should increase the amount of capital we have available per worker and bring our standard of living up closer to the U.S. level. It won't have an effect beyond that, but at least we could get up there.

The other two points—interprovincial barriers and labour market flexibility—are important points, I believe, that are very difficult to put into our productivity statistics, and I think you're right to remind us that our measurement techniques are somewhat limited and we can't really address all of these issues.

The Chairman: Mr. Sharpe.

Dr. Andrew Sharpe: Yes, I just wanted to comment on the issue of the regulatory burden. It's true that if we got rid of a lot of regulations, we probably could improve our productivity. But productivity is not the be-all and end-all of economic life. Basically, the goal is to improve the quality of life of Canadians, and that's a much broader concept than, say, GDP per worker or GDP per capita—much, much broader.

So there are reasons we put in regulations, for example, to improve the environment or to cut CO² emissions. There are all sorts of good reasons that we have these regulations. So we always have to have a balance between the optimal regulatory burden in terms of improving the quality of life of Canadians and any negative effects on productivity growth.

It's not clear to me that we've gone too far in increasing the regulatory burden on business in Canada in terms of reducing the productivity growth, and I would argue that there is not a lot of evidence overall that the regulatory burden has had a major negative effect on productivity growth in Canada. So I just want to stress this balance. There are reasons for regulations. They can improve the quality of life for Canadians, even if they have some possible negative effects on productivity.

The Chairman: Thank you, Mr. Brison, and thank you, Dr. Sharpe.

We'll move to Mr. Szabo, followed by Dr. Bennett.

• 1135

Mr. Paul Szabo (Mississauga South, Lib.): Thank you very much, gentlemen. It's a very challenging subject to deal with, and we're going to have to have a lot of stuff repeated to us. There were some things that stuck in my mind, though, and I want to share them with you so that you get an idea of maybe what the public reaction is to this debate on productivity.

There was a remark made by John Baldwin, I think, with regard to measuring labour output and looking at employed and self-employed people, and he said that after we gather this data, then we manipulate it. That word “manipulate” tended to stick in my mind. Then Mr. Diewert started to provide us with a list of all of the reasons your data could be soft or lead you astray, and there were at least five in the last segment of your presentation. There also was this whole idea that we're only measuring the private sector and we can leave out all the government umbrella stuff. Another remark was that on balance all our mistakes balance each other out, so we don't have to worry about them.

Dr. Erwin Diewert: I didn't say that.

A voice: I think “offset” was the word.

Mr. Paul Szabo: They offset each other, and therefore we don't have to worry about things.

I think it certainly does paint a picture about this issue of productivity, about the understanding and probably about the credibility index with regard to how the discussion about productivity will be viewed. So we're going to have to work very hard to get it down to some basic nuts and bolts.

The issue of the private sector thing makes me a little concerned. It would say to me that if governments were to be inefficient, to overemploy, to divest high productivity areas, and basically to be a fat government, that would improve productivity, because it would increase the participation rate in the private sector. It would basically take all the bads out of the private sector and harbour them under the government umbrella.

The point about the comparability of the private sector between countries to me sounds as if, in fashioning themselves, governments could in fact compensate for their inability to improve productivity in the private sector. So it really looks as if, over the longer term, governments have a lot to do with their productivity areas.

But I did have one question. This isn't nuts and bolts to me. The comparison for us, obviously, will be with the U.S. I read in the material that their productivity rate is three times higher than Canada's is, but I don't see it visibly. There's something there that I'm missing, as to what are the indicators.

The question on my mind is about the economies of scale implications. It seems to me that when things such as the GST came in, three doughnuts was a snack but more than three doughnuts was food. So we have these conventions that tend to make things a little different. The cost of three doughnuts was more per unit than buying a dozen. To me, economies of scale tend to mean that larger countries probably have an opportunity to be more productive simply on the basis of economies of scale, because they can average down on a number of scales.

I don't know if I've told you how frustrated I am right now, but I'm working hard to understand if it's realistic to simply compare a productivity rate in Canada with a productivity rate in the U.S. without understanding the differences between us and whether some of those are permanent differences or so long term that they're effectively permanent, and which are the temporary or variable differences we can do something about.

The Chairman: Thank you, Mr. Szabo.

I'll go to Mr. Wells. I think you wanted to clarify the second last point.

Mr. Stewart Wells: Thank you. What you saw me react to was the suggestion that the U.S. productivity rate was three times the Canadian rate. I don't know where that comes from. I don't think anyone here has said that.

• 1140

From what I've seen of other calculations, if you talk about levels of productivity, the productivity rate has been something like—I think Andrew referred to it—80% of the U.S. rate for 20 years, and probably for 40 or 50 years. I think your point that we may have difficulty doing anything about it is worth paying attention to. It has been in place a long time, and it's not going to be easily gotten rid of. It's something you would want to take into account in your deliberations. But I don't know whether that's three times—

Mr. Paul Szabo: I'm sorry, I can clarify that. It isn't the productivity rate in the U.S., it's the growth rate.

Mr. Stewart Wells: Do you mean the growth rate of real output?

Mr. Paul Szabo: We'll find the actual reference.

The Chairman: Dr. Diewert.

Dr. Erwin Diewert: Perhaps I could just comment briefly on the economies of scale issue. In theory, our free trade agreement should allow us to achieve the economies of scale the Americans are achieving, but my colleague Helliwell tells me that borders matter, and hence the amount of integrated trade between Canada and the U.S. is not as great as the free trade agreement would lead us to believe.

So, yes, I think the U.S. will have some advantages because of the large scale of their economy, plus the willingness of workers to migrate. I find it astounding that in this last expansion in the U.S. unemployment rates all over the U.S. have more or less equalized as workers have moved from place to place. Obviously, we don't have that situation here in Canada. Maritimers are reluctant to leave, Quebeckers are reluctant to migrate, etc.

The Chairman: Are there any further comments?

Professor Harris.

Dr. Rick Harris: I disagree on the scale issue. There's virtually no evidence I can see that growth rates are a function of scale. In fact, some of the most spectacular growth rates have been achievable by the small countries, including high-income small countries. The recent examples are the Netherlands, Ireland, and Finland, all extremely small countries with unbelievable growth rates. I just don't buy this scale argument as a serious restraint on Canada's growth. Now, I'm not suggesting there aren't problems with the trade agreement or structural problems in the Canadian market, but I don't think we can point to evidence that says that scale is going to be a long-run drag on Canadian economic growth. I don't buy it.

Mr. Paul Szabo: I just want to follow up on that. I'm not sure I would disagree that you could have various growth rates depending on economies of scale. It could be all over the map. It depends on whence you started. This is the reason that sometimes you want to know what absolute growth is as opposed to the percentage growth. That's the difference.

But in terms of economies of scale, my point is that if I have a higher base on which to average down and, say, to negotiate more favourable trade arrangements, lower unit pricing, etc., this has to be reflected in the cost of inputs, but I would still get the same value of outputs as those who would have to pay more for inputs on a per-unit basis.

For example, there could be economies of scale on labour. It could be that I'm McDonald's and I can buy in bulk, and I can demand lower wage settlements because I don't have any unions, and I can control that, etc. The power of size tends to be to the advantage of the producer. Therefore, my outputs, which can compete with everybody else's even though I've got a lot lower....

It effectively gets the profits, and all of you told us that profits have nothing to do with productivity. The economies of scale I'm talking about are directly related to the cost of inputs in the measurement of productivity.

Dr. David Slater: I'd like to get into the economies of scale discussion. I think you really have to look at this a lot more carefully, because a lot of the changes that have taken place have made some of the traditional economies of scale less important than they used to be.

If you look at the successful Canadian knitting industry in Winnipeg, which generates wonderful multicoloured sweaters and things of that sort, you'll see that the knitting industry is run off of computer-driven programming and so on. If you look at an automobile factory now, it doesn't turn out 1,000 of this and 1,500 of that; it turns out one with pink this and something else that, and so on. The ability now to get efficient production on smaller scale, and changing mix and changing output, is really an enormous change. I think on the scale side you have to look at it very carefully.

• 1145

Again, if you look at a traditional example of economies of scale, look at what happened to big steel in the U.S. compared with small steel. It's the new, smaller, electric ovens-based steel industry that just licked the pants off big steel in the United States.

The Chairman: Thank you, Mr. Szabo.

Mr. Orr, and then we'll go to Dr. Bennett.

Dr. Dale Orr: Yes, the paper I'd completed here made the comment that our standard of living performance over the 1990s was very weak. It was weak relative to our own historic standards in the 1980s and it was weak relative to the U.S. in the 1990s. And now that we're getting into a discussion of why might that be, what are the levers to change that, I would make two points.

I think the reason we observed that weak standard of living performance in the 1990s was twofold. One was the aggregate demand in the economy. We didn't, over the 1990s, stimulate the economy to ever really pick up from the recession of 1990-91. There still remains an output gap in the economy of 1% to 2%, so we had very weak aggregate demand. Improving that would have improved our standard of living performance.

The other is that there's a very weak incentive, relatively speaking, relative to our history and relative to the Americans, to participate in the labour force. I think this deserves a lot of attention, because the point I was making is that our weak standard of living performance is not due to weak productivity as we define productivity; it's due to weak performance in the labour force. We have fewer people participating in the labour force. Yes, we've reduced our unemployment rate, but it's still almost 8%, and the Americans are at 4.2%. Those are serious problems, and that's what's eating into our standard of living.

Now, why is it that we didn't reduce taxes in the 1990s to stimulate aggregate demand? Why is it that we didn't reduce taxes in the 1990s so we could give people more incentive to join the labour force, to stay in the labour force, and to work hard? We couldn't afford to. Why? Because contrary to what was said earlier over here, 30% of our tax money really is going right off the top into what almost everybody considers to be a big black hole. And that's the interest charges on the debt. The reason we couldn't afford to stimulate the economy and reduce taxes to provide the incentive we need is that we all send a dollar to Ottawa and 30¢ goes right off the top to pay debt charges. The debt-to-GDP ratio is simply too high to afford the tax reductions we needed. Now fortunately we're moving into a little bit of a different era.

Thanks.

Dr. David Slater: Chairman, I'm part of the big black hole.

The Chairman: If I can ask you a question vis-à-vis unemployment rates, do the Americans measure unemployment rates the same way as we do?

Dr. Dale Orr: Very close but not precisely. So ours is 7.8%, there's is 4.2%. In talking to someone last week who I think is a real expert on it, I gathered that maybe half a point or something in that order of magnitude might be dismissed as measurement error. So we have a good healthy three-point difference, I think, even if you account for some of that measurement error. Certainly it would be a huge mistake for us to be too complacent and say ours is 7.8% and there's is 4.2%, and it's measurement error. It's a very small part of that difference.

But we have Stew Wells here. Stew, maybe you want to comment on that.

Mr. Stewart Wells: On the measurement error on the employment?

Dr. Dale Orr: On the difference in the unemployment rate between Canada and the U.S.

Mr. Stewart Wells: I certainly don't think it's any worse than you suggested.

• 1150

The Chairman: Okay.

Dr. Bennett.

Ms. Carolyn Bennett (St. Paul's, Lib.): I have three questions.

The first is that there has been a lot of speculation that it's the low Canadian dollar that is making us look like our productivity is not so bad; it's really worse if some companies are perhaps doing better just because of the low Canadian dollar.

The second question is the one that everybody expected me to ask, which is about the free lunch. When Canada does spend its dollars on the education and health care systems and somehow it ends up a null set in terms of productivity, it doesn't count anywhere....

I have two questions. One, are we able to properly compare? I know we can't compare taxes, because I think my patients who had to go and pay $10,000 a year for their health insurance when they moved south of the border know that somehow it was there when they were here. But how do we deal with productivity in those areas like health care? If you guys aren't bull terriers at their heels.... When I first went on the board of the OMA, only 40% of doctors had a fax machine. How do we push productivity in those areas that you're not measuring? How can we make sure it's a priority?

The third question comes from that. If some of the things we're proudest of aren't being measured, in a productivity debate is it possible to leave out the progress indicator kinds of things? Obviously I feel badly that we think huge revenues from tobacco companies and a huge industry around lung cancer somehow shouldn't be on the positive side of productivity, or big oil spills that seem to add up as a positive thing on the balance sheets. Should we always be also doing the progress indicator as we look at GDP or any of these other indicators?

The Chairman: Mr. Harris, you may want to jump in.

Dr. Rick Harris: First of all, on what we think of as education and health care, I think Andrew actually organized a conference here a couple of years ago, or maybe it was a year ago. There are people who work in this area, and there are people who spend most of their professional career examining the productivity of the educational sector, or examining the productivity of the health care sector. And detailed country-by-country comparisons are available. They do not show up, though, in these aggregate statistics.

So it would be helpful for the debate if that evidence was brought to bear on this issue. And that is of course quite separate from the issue of how one values it.

Mr. Brison asked about the privatization issues. We can have health care delivered through a variety of mechanisms in the OECD countries, but we can still in principle at least do benchmarking or best practice about what is good practice across countries and see where Canada is on that score.

On the valuation of the output I think it's much more difficult, because inherently it reflects social judgments, political judgments, about the relative merits of these activities, particularly when they're provided through a non-market system. So I think that while we may enter them into a comparison of our personal evaluation of living standards, I see no resolution between us and the Americans as to how you're going to put a number on that ultimately.

Ms. Carolyn Bennett: Even in terms of best practices, that productivity may be doing way too many operations, may be giving antibiotics for colds, may be giving all kinds of things, and really in the sustainability of a system we want those things stopped so we can save the $7 billion a year we know we could if we moved to best practices. Again, it is in the free lunch department, so actually moving to best practice is a good thing for the economy if you're actually paying for it out of the public purse.

Dr. Rick Harris: We could disagree about the details, but in principle the point of view you're articulating is correct.

Dr. Dale Orr: I have a comment on health care. I think there are at least two very fundamental problems with the way we operate our health care system in Canada.

The first is that there's a zero perceived marginal cost to the consumer. So when the person going into the hospital or to the doctor sees no incremental cost to them, of course demand is going to exceed supply, of course we're going to have waiting lists. That's a fundamental economic problem. If you set a zero or negligible perceived marginal cost, you're going to have demand exceeding supply all the time for everything.

• 1155

Secondly—

Ms. Carolyn Bennett: But education can help with that. Who would want an operation they didn't really need? I think the dollar part is really getting us into trouble, because who can afford the six weeks off work for a piece of surgery they didn't really need, if they'd had the proper educated choice? So just the dollars about going into the hospital is very different from six weeks off work.

Dr. Dale Orr: Yes, of course, but an awful lot of people are using the health care system for things other than surgery, going for all sorts of things for which, if there were a price they had to pay, the usage of the system would be reduced.

Let me make my second point.

Ms. Carolyn Bennett: This is something they're told to do.

Dr. Dale Orr: Yes. My second point is that the suppliers of the system are really in an unusual position of being able to create the demand for their services. That is, the doctors, in particular, are the main source of information for most people as to whether they should demand health care services or not. It's a little bit difficult to get around that, but that's a fundamental economic problem.

I think the efforts that are now being made to give us better measurement in the health care system are a very important step in the right direction. But I just want to say on the margin that there are two fundamental economic difficulties here in the way we're delivering health care that I think should be addressed, and I think they would both contribute to increased productivity in the health care system.

The Chairman: Thank you.

Dr. Sharpe.

Dr. Andrew Sharpe: I have a couple of responses. One is about the relationship between the low value of our dollar and productivity.

First, it doesn't affect the level comparisons, because when we do level comparisons we don't use the actual value of the dollar; we use the purchasing power parity value of the dollar. So, in principle, it shouldn't have any effect on the level comparisons.

Whether it explains our poor productivity performance, the argument is that if there's a low value of the dollar, we're very cost competitive abroad, our manufacturers get lazy, and there's deterioration of our ultimate performance. I don't think there's a lot of evidence of that. Some people on the panel may disagree.

The argument there is just the converse: if we have a very high dollar, we're going to have much better productivity growth because there's going to be this massive incentive for firms to become more productive. If you look at the early 1990s when we had a very high dollar, up at 88¢, we didn't see any improvement in productivity. In fact, we had a relatively poor performance at that time.

Also, if you take that argument to its logical conclusion, any additional cost on firms would result in productivity gains because it would force them to make it more productive. So if we raised wages, it would become more productive; if we imposed taxes, it would become more productive. Maybe there's a certain element of truth in that, but I don't think we want to take that to a large conclusion and start imposing taxes to improve the productivity of Canadian firms.

On your point about progress indicators, I agree completely with you. In fact, last fall, the Centre for the Study of Living Standards released an index of the economic well-being of Canadians. We find our index of economic well-being increases at a much slower rate than GDP per capita. So the overall economic well-being of Canadians is not increasing as quickly as GDP per capita over, say, the last 25 years.

The reason for that is that there has been a large increase in the economic insecurity of Canadians due to higher unemployment and cuts in our social safety net, particularly in the employment insurance program. That's picked up in polls of economic security. So it's important to realize that productivity is not the same thing as economic well-being.

The Chairman: Mr. Wells.

Mr. Stewart Wells: I don't want to get into this issue of whether there are surgical freaks or operation groupies or something, but I did want to support something Dr. Bennett said.

The productivity measures and indexes we've been talking about here today are good indicators in a broad general sense. What you're talking about, which is micro-productivity estimates, or examining the detail of productivity in a specific industry or occupation or something, is really the way one has to go if one wants to influence productivity, in my judgment. We do some of that work.

I strongly support your views in that sense. When you want to find out something about the medical business or the hospital business, you should focus on that and do a detailed or really good examination of the nature of the industry. But we haven't been talking about that, and we don't pretend that our indexes will really tell you much about that.

• 1200

Dr. David Slater: Can I add one point for Dr. Bennett, a little bit counter to the point Dale has made? He was talking about classroom demand and supply curves, and so on.

One of the things that you get in the American medical system is 20% of the costs going to accountants and lawyers. I'm not sure that's a very smart way to run a medical system.

The Chairman: Thank you, Ms. Bennett.

We'll hear from Dr. Diewert, and then we'll go to Ms. Redman.

Dr. Erwin Diewert: This is still on the medical problem.

There is a group in the United States, the National Bureau of Economic Research, that is devoted to measuring outcomes and introducing a new approach to the measurement of health care benefits. It's very costly, very time-consuming, and ties into what Stew was saying, that this is kind of the way you would go. But it also ties into what I was saying. Detailed analysis, then, would produce detailed price indexes.

If we ask Stew, of all the price indexes that Statistics Canada has, what percentage of them are devoted to measuring service outputs, it would be something like 10%. So we cover agriculture and commodities wonderfully well, but as far as I know, there are two people working on service price indexes. So that why I say there is a need for this kind of detailed research.

The Chairman: Ms. Redman.

Mrs. Karen Redman (Kitchener Centre, Lib.): Thank you, Mr. Chairman.

I represent Kitchener Centre. We're twin cities, with the University of Waterloo, so I've listened very intently. I wonder why the high technology isn't showing up in Canada, because certainly the impact has been felt in my community.

Further to that, are we lagging behind the United States and other economies in wiring our country? I wonder if anybody has any input they'd like to give us on the legal and regulatory system that should be created to facilitate this.

Dr. Dale Orr: Having spent about a decade in the telecommunications business, I can offer this: we do not lag behind the Americans in wiring. For years, a much higher percentage of Canadian families than Americans have had telephones in their homes. I think right now almost 99% of Canadians and 95% of Americans have telephones in the home.

As well, Canada has for a long time been more wired for cable—to quite a significant extent. As well, the efforts we've had in the last couple of years in terms of bringing the information highway to schools and libraries, as well as to remote communities, have been phenomenal. I'm not sure it's possible to make a really accurate direct comparison with the Americans, but we're doing really well, and I suspect probably better than the Americans in that sense as well.

So that's not where the problem lies. The problem lies in the overall growth in production within their information technology sector, the size of that sector, the growth of that sector, the incomes that are generated within that sector, and how they're able to export information technology, software and hardware all around the world to a much greater extent than we are. That's where we're lagging behind the Americans. It's not in wiring our homes and schools.

The Chairman: Mr. Wells.

Mr. Stewart Wells: On why we don't see the effect of technology in the economy more than we apparently do, I'm going to cite a very well-known American economist, Robert Solow, who has studied productivity for many, many years. He at one time said that defective computers are everywhere except in the productivity statistics. It has been a conundrum that has not been resolved very well, about which there is much argument.

If you'll forgive me, I'll read a paragraph from an article by Robert Solow, a review of a new book on the productivity debate, in one of my favourite magazines. It's The Challenge magazine for January-February 1998, and he's reviewing a study by a Daniel Sichel, I believe.

• 1205

He says:

So why has “the computer revolution” had so little noticeable effect on productivity? [...] Yes, investment in computer and peripheral equipment comprises a large proportion of current business fixed investment—17.8 percent in 1991 and undoubtedly higher now. But productive services provided by the stock of computers (in industry) is rather small, for two reasons: first, the stock of long-lived capital in other forms is very large; second, the depreciation-obsolescence rate for computers is very high, usually thought to be about 25 percent a year. Sichel estimates that computers and peripherals totalled less than 5 percent of the net stock of fixed capital in 1993.

That doesn't answer all questions and Solow doesn't even support it, though everything is found in the book. But I think that is probably a contributing factor.

The Chairman: Yes. Dr. Baldwin.

Dr. John Baldwin: It should also be added that it's recognized by most statisticians that have examined this issue that the sectors that make major usage of computers are those sectors for which the statistical agencies have the most difficulty in accurately measuring prices that Erwin has captures.

Mr. Stewart Wells: [Inaudible—Editor]

Dr. John Baldwin: We're perfectly honest about our deficiencies. If you look at the business services, insurance, or banking sectors in the United States, those are the sectors for which the American statistical system actually shows negative growth rates over time of productivity, primarily, I believe, because of measurement problems.

Erwin is also correct in pointing out that some progress is being made in these areas. In the Canadian statistical system, the adoption of the 1993 UN system of national accounts has now meant that the financial remediation function in the banking sector is producing positive productivity growth where it was once not present. So there very well are measurement issues associated with this particular problem.

But to answer, Ms. Redman, the question on high-tech growth more precisely.... I was going to avoid that, because it's an issue that has to do with the Canada/U.S. comparisons, and I understood that was for another day.

In the release of March 23, we do compare productivity growth for 17 industries in the manufacturing sector for Canada and the United States. We showed that the incredible performance of the American manufacturing sector is essentially located in the two high-tech growth industries where either computers are manufactured or the parts are manufactured. And it is an incredible performance—it's way above their long-term averages. They've been supermen over the last decade relative to what they has done before.

Those are two industries that are dominating their manufacturing sector. That's why they're doing so well in their manufacturing sector. In those two sectors we do better on average than in other sectors, but we're still very far behind the Americans. So it's very much a sectoral problem when we come to measuring differences for manufacturing.

Thank you.

The Chairman: Thank you, Ms. Redman.

We'll go to Mr. Pillitteri.

Mr. Gary Pillitteri (Niagara Falls, Lib.): Thank you very much, Mr. Chairman. I just wonder if my question is in order, but I will try it.

There was some discussion on economies of scale, but being a businessman, if I were thinking of economies of scale in the business I'm in, I'd say I should have stayed at home sleeping rather than trying to start up my business. But I like your analogy, Mr. Harris, on Microsoft and how you compare it to profitability and productivity.

I just want to pick on one sector—and I think you touched on it—the auto sector, specifically the auto sector in Canada prior to the 1960s, the first Auto Pact that came into Canada. We were very vulnerable, so we thought we needed protection.

I do recall that we were in deficit to the United States for some $10 billion—this is prior to the 1960s. Yet, with the protection of the auto sector that came in with the Auto Pact between Canada and the United States, we were still in deficit, even though it did give some protection. They had higher wages than Canada. Canada picked up comparably afterwards—in the 1970s—to try to get people wages, which it did. Yet we were still in deficit. All of this, and we were trying to catch up to the States.

• 1210

Yet today, that is one of the free trade agreement benefits that came into Canada in the auto sector, where we have the surpluses in this country—over $20 billion surplus. Tell me, was it productivity by itself in the traditional manufacturing sector or was it the niche marketing in the parts sector that really turned that around? And today we have higher wages than the States, by the way.

Dr. Rick Harris: I would draw two inferences from your observations.

One is that productivity change in some broader sense of the term is very long-lived, and the process is subject to considerable lags. After the 1965 agreement, there were some initial changes, but it takes a long time to reorient management, labour, and the whole host of industrial practices and to integrate ourselves in this North American-based manufacturing system. But it happened, and it was an amazing success.

Now, your second question goes again to the heart of what we mean by measuring productivity. The gentleman sitting beside you has worked on this problem extensively.

When we think about productivity at the level of the plant, we typically think about all the things you just talked about. That would involve changing the reorganization of the plant, going from having many product lines to one product line, changing the nature of the machinery that's used, etc.

But the broader question, I guess, is what kinds of policy interventions lead to these types of productivity-enhancing reorganizations? Of course, there was in the 1970s and 1980s a tremendous discussion about whether we needed to have sectoral free trade or broader free trade with the United States. It turned out sectoral free trade wasn't an option, as you know. Broader free trade did work.

Nevertheless, there are many advocates of selective industrial policy to achieve the kinds of gains you're talking about. I'm sure you're going to hear from some of them in this room. My general view, personally, is that on average they haven't worked, and given our trade regime, they're not an option for Canada. So that's just a personal note on that particular theme.

The Chairman: Mr. Pillitteri.

Mr. Gary Pillitteri: What I'm getting at, though, is wasn't it a factor also that there was the parts industry rather than the traditional manufacturing? What I'm saying is the Magna.... You're not getting my question fully.

Dr. Rick Harris: Do you want to grab a crack at that?

Dr. David Slater: I can't speak to that.

It is clear that we've had a very successful parts industry develop, that's right. But I am told that among the things that led us to get bigger shares of the production is that our guys didn't appear on the assembly lines stoned on Monday morning and didn't quit, half of them, on Thursday night; and that the defects in cars from poor assembly and so on were so much smaller in Canada than they were in the typical American plant. There were exceptions, but the typical story was really one that had all of those work habit elements, the attitudes of the unions, and all of those sorts of things in play. I believe they have been a part of the success of the Canadian auto industry.

The Chairman: Thank you, Mr. Pillitteri.

Mr. Valeri.

Mr. Tony Valeri (Stoney Creek, Lib.): Thank you, Mr. Chairman. I know the time is late. I'll try to be as brief as possible.

I believe it was Mr. Slater who, in response to a question, made reference to the smaller steel plants in Canada taking the pants off the larger American steel plants. The point of my question is to focus a bit more on the micro side of this debate. We really appreciate the general discussion about measurement, and Dr. Bennett started the health care question. I just want to follow up with the small business side of it.

I believe Mr. Baldwin has done a fair bit of work with respect to research on small business in the manufacturing sector and its impact on productivity. Donald Daly from York University has also done some work.

• 1215

To quote Mr. Daly's study, he indicated that value-added for a production worker in a plant size of less than 50 is only half that of plants employing greater than 400. In Canada, we've obviously had a much stronger growth in the small business sector than has the United States. Our manufacturing production, in terms of output in small business, is much higher compared to that of the United States, and yet it has been said, according to studies, that productivity in small business is lagging far behind that of firms of over 400.

So I'd like some reaction to this, that given that our employment growth has really taken place in the small business sector, we're now faced with greater employment in a certain sector that is being deemed to be less productive than the larger manufacturing companies. I'm wondering if you could respond to that. Perhaps Mr. Baldwin would start.

Dr. John Baldwin: In terms of facts, the small business sector has increased its share of employment in the manufacturing sector, and thus average productivity measures, whether they be labour productivity or MFP measures—and we don't actually calculate the latter, but let's presume we had some—would show decline or the rate of growth would slow down as a result of all of this, that we've simply shifted part of our industrial population towards smaller firms.

But while we've done that, it would be incorrect to say they've been growing more rapidly generally, using other measures of importance. If you actually used measures of importance such as shipments, small firms haven't actually increased their share of shipments very much over this period of time. If anything, they've fallen.

So basically what has happened is that the large firms have become far more productive in a simple measure in terms of output per worker than small firms have, and as we have downsized our manufacturing sector—remember, overall now we've been losing employment from manufacturing during most of the last 15 years—we've lost most of our employment out of the larger plants. So larger plant productivity growth, in a simple way, output per worker, has been more rapid than in smaller firms.

At that I stop, because, one, I don't have a good explanation as to why this has been happening. Don Daly has some, and if you've read his article you have those, which have to do with management capabilities. Secondly, I'm not sure of the policy implications for this myself. But it does provide us with a conundrum as we worry about productivity growth in this country relative to others.

One has to be careful about presuming that this process has necessarily slowed our productivity growth relative to other countries, because I've just described a phenomenon that may indeed be present elsewhere, and we have an ongoing study that's trying to examine the extent to which this phenomenon is also taking place in the United States at the same time.

And we don't have final numbers on this study. There is some indication that a similar phenomenon has been taking place in the U.S. as well. If that's the case, it has a rather dramatic influence on what we might argue would be policy handles here. If this is something that is indeed endemic to the North American economy, it's hardly likely that we're going to change it ourselves. If it's something that's particular and peculiar to Canada, then one can begin to look at made-in-Canada solutions.

But we don't have answers to that last question I just gave you.

Mr. Tony Valeri: Mr. Daly points out that 64.2% of manufacturing shipments in Canada were made by firms with less than 500 employees, while only 17.7% of U.S. manufacturing shipments were made by the same size of firms in the United States.

So when I read this, I find that.... We in Canada seem to be trying to promote small business, and the rationale is that in this changing environment small business is much more adaptable to change and taking on new technology, and it can compete because of, in most instances, lower wages, less unionization and a whole bunch of other inputs. At the same time, we hear studies that say small business is not as productive as larger business is, yet in Canada the majority of our manufacturing shipments are made by small businesses when we compare that to the United States.

• 1220

So I'm a bit worried that we find ourselves in this situation, and I'm looking for some type of direction as to whether we should be seriously concerned about this or whether we need to just do more research to see whether we have a problem here, or is this another one of those measurement problems we have to deal with?

Dr. John Baldwin: You've raised two issues; first of all, whether or not we generally have more small businesses. I was talking in my response a minute ago as to whether there was a trend to creating relatively more small businesses. We've always had an industrial structure with fewer large plants than the Americans have.

The research you referred to, my research, showed that we have been trending towards putting even more of our employment in these small businesses in the last 20 years, and the reference I just made was to the fact that the Americans appear to be trending in the same direction. So if that's the case, we have a general change going on in industrial structure.

The other issue you raised was whether or not we should worry about the fact that we have inefficient small business, or should we be pleased we have small business because they're in effect the Microsofts of tomorrow. Clearly, differences of opinion can sit on the same desk in this particular instance, because you have both phenomena taking place. You have some small businesses that are dying at any one point in time because they're small and inefficient and simply can't manage it. And you have others that are growing and are going to be the Microsofts of tomorrow.

It's a very dynamic process; you see at any one time only a snapshot. There are some small businesses and there are some large businesses. On average, those small businesses are not as productive as the large businesses, but there's a dynamic set that is growing and becoming just as productive and there is another set that is in the process of dying. It's essentially sorting out the relative importance of these two groups that is the difficult measurement issue, and what we've actually focused on is trying to understand productivity change over time.

Mr. Tony Valeri: When we compare the actual productivity of our small businesses to those of the United States, how do our small businesses compare in terms of productivity?

Dr. John Baldwin: I actually don't have the numbers from the study yet to draw final conclusions.

Mr. Tony Valeri: Thank you.

The Chairman: Thank you. Are there any further questions?

Mr. Sharpe, you said labour productivity, not total-factor productivity, is the determining factor for our standard of living. Is that correct?

Dr. Andrew Sharpe: Right. I'm saying in terms of the link between our GDP per capita, which is a measure of standard of living, it's labour productivity. Now, that doesn't mean total-factor productivity isn't important. The key, though, is that capital accumulation investment, as Rick mentioned earlier, is a crucial variable in terms of our labour productivity, but when we have more capital stock, that doesn't show as increased total-factor productivity because capital is an input.

So there is much stronger correlation between our standard of living and trends in labour productivity. Labour productivity also has a higher rate of growth than total-factor productivity because of the increase in the size of the capital stock.

So in terms of your discussions, I think it's very important to focus on labour productivity. Economists like total-factor productivity. It's certainly more sophisticated, it's a better measure of the efficiency and the use of resources, but from the general public point of view, I think labour productivity is really the most important variable.

The Chairman: Any comments? Do we all agree, or is that what that means?

Dr. Rick Harris: No. I want to make it clear that when I'm talking about productivity I'm talking about what drives total-factor productivity. Investment is the principle determinant in total-factor productivity. One of Canada's best young economists, Jeremy Greenwood, who left about three years ago and is now at Rochester, has done a lot of work. He looks at the post-war data for the U.S., and he attributes 60% total TFP to improvements in the quality of capital goods. And that's delivered through investment. So investment is important for its...if you hold quality constant and you increase the quantity of capital without correcting for quality, workers have more capital to work with and average productivity goes up.

But for the dynamic gains, the growth in total-factor productivity, an important intermediating mechanism is investment. And that's largely through better improvements in the quality of capital goods, which again is another horrendous quality measurement problem that we can't deal with.

Mr. Stewart Wells: For what it's worth, I'd be inclined to agree with Rick that labour productivity is important, but one ought not to focus exclusively on that.

• 1225

The Chairman: Dr. Diewert.

Dr. Erwin Diewert: I think Andrew's right that labour productivity is closer to the bottom-line concern of this committee, the welfare of Canadians. I think maybe some of us are thinking it's useful to factor labour productivity into two parts, the total-factor productivity part and then what happens with capital. So if it grows a lot, labour productivity is going to improve. But you can kind of decompose the problem into the two parts, and there are different policies for the two parts. You can influence capital accumulation by policy levers through tax policy, but it's very difficult to influence total-factor productivity.

The Chairman: We spend a lot of time comparing things. We compare ourselves with the Americans, we compare 1970 with 1990, we compare 1960 with 1974, and so on and so forth, but the bottom line is the following. Whether the rates are going up or down and whether productivity is going up or down are important from a historical perspective, but what we in this committee really want to do is improve our productivity. We're looking at some of the levers we have at our disposal as legislators and members of the House of Commons. In the final analysis the private sector has to come up with perhaps more efficient ways of delivering services, producing products, etc.

When you look at the issues of trade, taxes, investment in R and D and education, and human resources development, and also taking into consideration what Professor Harris said about investment in trade and human resources development, if you were a member of the finance committee, what would your priority be? What would be the first thing you'd recommend?

Dr. Erwin Diewert: Reduce income taxes.

Mr. Stewart Wells: I think in the long run the best place for you and other people to focus their attention is on the growth of human capital, skills, and education. In that sense I'd agree with Andrew. We need a healthy population. It seems to me that is the best role for government to play in this.

As I said at the outset, and I think others have also said it this morning, I have great difficulty finding the levers you're going to push. Some will vote for reduced taxes.

The Chairman: Don't you think that would help?

Mr. Stewart Wells: I don't think it's going to help much, but there's room for disagreement on that.

The Chairman: Don't you think high taxes are disincentives to work or to invest?

Mr. Stewart Wells: There is an argument as to whether our corporate taxes really are higher. Are you talking about personal taxes or corporate taxes?

The Chairman: I'm talking about both.

Mr. Stewart Wells: Corporate taxes as a share of the government's take have risen over the last few years. But there has been a long decline in that, and we're not back to where we were. I have trouble seeing taxes as a particular problem at the moment.

Dr. David Slater: I'd like to add one thing. In the period when you've been fighting deficits, the idea of public investment has been a sort of no-no. I think the net result is that we have really run down our public capital. My old university had to place water buckets around when it rained because they didn't have enough money to repair the roof. When money is tight, you cut down on maintenance. The Ontario road system is just now beginning to be resuscitated, but it really was allowed to run down very considerably. If you look at a lot of our facilities, it's true that we have too much in some things and not enough in others. But I think you really ought to look pretty carefully at the state of the public capital of this country and what's needed to get it into shape in order to make effective contributions.

• 1230

The Chairman: Dr. Sharpe.

Dr. Andrew Sharpe: One factor that hasn't been mentioned is the role of a favourable macroeconomic environment, which I think is very important for strong productivity growth. It is true that our productivity growth in the 1990s has been worse than that of the U.S. and we've had a very poor macroeconomic environment. But because of that catch-up potential with the U.S., if we had had a more favourable macroeconomic environment in the 1990s, I'm sure we would have seen significantly stronger productivity growth than in the United States.

Again, we have to think of the demand side and a robust demand condition as a pre-condition for strong productivity growth through learning by doing, economies of scale, and increasing returns to scale. I think that's just one of many factors, but it's important not to lose sight of this strong macroeconomic environment.

The Chairman: Dr. Sharpe, would you consider the macro conditions to be quite good now?

Dr. Andrew Sharpe: They're certainly better than they were earlier, because interest rates have gone down. Probably there's still some scope for lowering nominal interest rates. Given that inflation is very low right now, our real rates are still not that low from an historical perspective.

There is also the issue of the deficit, how much of that is used for tax cuts, which are stimulatory, plus spending on goods and services, which is again stimulatory, as opposed to, say, just paying down the debt.

it's certainly much more positive than it was in the earlier 1990s, when we had a very restrictive monetary policy, and in the mid-1990s, when we had a very restrictive fiscal policy. But it probably could still be more stimulatory.

The Chairman: Mr. Epp.

Mr. Ken Epp: In Canada we're in the habit of redistributing earnings. We take money from individuals who earned it and give it to those who didn't earn it. It's the same thing with businesses. There are a lot of businesses in Canada that are simply the recipients of government largesse, which is again transferring money from those who earned it to those who didn't earn it. Is that a significant factor affecting productivity?

Dr. Erwin Diewert: It would be if it were really big, but my understanding is that business subsidies are not that big as a percentage of our GDP.

Mr. Ken Epp: They amount to $5 billion a year.

Dr. Erwin Diewert: That could fund a nice tax cut, but it's not going to be the answer to everything.

The Chairman: Are there any further comments? Mr. Orr.

Dr. Dale Orr: When one looks at all of this and asks what the government can do, I think you'll find that tax reductions are pretty high on the list. At the same time, I wouldn't want to exaggerate what would be accomplished by tax reductions, but I think there are tax reductions that can start us moving in the right direction. They can work on all of the levers. They can work on stimulating investment. There are some taxes that can give more incentive to people to get more education and training, and there are also some taxes that can increase the flow of trade by making us more competitive. So there is scope there.

Another one that hasn't been mentioned, which I think deserves mentioning, is that not enough attention is paid in Canada to what we can do to make sure that all of those new inventions, patents, and whatever that are occurring around the world get installed in Canadian plants as soon as possible, that we adopt those best practice techniques, and that they're working efficiently. Not nearly enough attention is paid to that flow of technology and to taking best advantage of what's out there.

There is relatively far more attention paid by people in the government toward trying to stimulate R and D and to create jobs in their riding by having people do R and D. We could double or triple the amount of R and D that's done in Canada and it would still be a very small fraction of all of the technology that's available to Canadian plants and workers. I think we need a lot more attention paid to that technology transfer process.

The Chairman: Thank you, Dr. Orr.

On behalf of the committee, I'd like to thank you very much. This was the first panel. You've certainly helped us with our challenge of defining productivity. You have given us an historical context of this issue as it relates to Canada and brought to our attention the key issues that are relevant to this particular challenge we face.

Tomorrow we will be looking at the changing nature of work; the impact on productivity; the nature of productivity in, for example, the service sector; and the impact of technology, which are some of the issues we've already discussed.

• 1235

This has been an excellent start. We are fully cognizant of the fact that we are challenged by this issue, but I think it's an issue that merits the attention of this committee and of the Canadian public in general.

Thank you.

The meeting is adjourned.