Paper presented at the CSLS session on the Canada-U.S. Manufacturing Productivity Gap at the annual meeting of the Canadian Economics Association, May 29-31, 1998, University of Ottawa, Ottawa, Ontario.

Introduction

There has been no narrowing of the gap in labour productivity levels in manufacturing between Canada and the United States in recent years. Indeed, just the opposite has occurred, with the gap increasing significantly. According to recently released BLS figures, since 1981, output per hour in manufacturing in the United States has increased, on average, 1.3 percentage points per year faster than in Canada. This trend has obviously important implications for Canada's cost competitiveness and the relative standard of living.

This development in perplexing. For a number of reasons, one might have expected Canada to have enjoyed faster manufacturing productivity growth than the United States over the last decade, with a narrowing of the productivity differential between the countries, as had been the case in the pre-1980s period. These reasons include continued potential for technological catch-up or convergence in Canadian productivity levels toward those in the United States given the lower initial Canadian levels, and structural reforms affecting manufacturing (e.g. replacement of the Manufacturing Sales Tax by the GST, reduction in trade barriers through the FTA and NAFTA, deregulation, etc.)

The Centre for the Study of Living Standards (CSLS) is organizing a project to shed light on the causes of slower manufacturing productivity growth in Canada and the rising Canada-U.S. manufacturing productivity gap. ¹ This purpose of this paper (and this session) is to launch the debate by providing an overview of the issue and suggesting directions for research.

The paper is divided into two major parts. The first provides context by reviewing trends in manufacturing productivity, The second part presents a list of possible explanations of the situation, and briefly comments on the relevance on the explanation. These explanations include slower output growth in Canadian manufacturing, differences in the methodology of measuring real output in manufacturing between U.S and Canadian statistical agencies, a slower rate of innovation in Canadian manufacturing, a relative deterioration in the quality of the Canadian workforce in manufacturing; slower wage growth in Canada, leading to greater substitution of capital for labour; and a slower pace of introduction of productivity-enhancing human-resource practices and organizational changes in the 1990s.

general trends in manufacturing

Combining this level and the relative productivity growth rates produces a fall in Canada’s productivity relative for manufacturing from 87.4 per cent in 1981 to 72.3 per cent in 1996, a precipitous decline (Table 2).

In both the 1980s and the 1990s, Canada’s output per hour growth in manufacturing has lagged that in the U.S. In the peak-to-peak 1981-89 period, output per hour increased at a 1.7 per cent rate in Canada versus 3.2 per cent in the U.S. From the 1989 peak to 1996, (not a business cycle peak, but the most recent year for which data are currently available), output per hour in Canadian manufacturing advanced at a 2.0 per cent average annual rate, compared to 3.1 per cent in the U.S.

The pace of output growth, and the relationship between output and productivity growth changed markedly in Canada between the 1980s and the 1990s, more so than in the U.S. In the 1980s, output growth advanced at a 2.7 per cent average annual rate (not much below the 3.2 per cent registered in the U.S.) and both productivity growth and labour input growth contributed to output, with growth rates of 1.7 per cent and 1.0 per cent respectively. But in the 1990s output has increased at only a 1.0 per cent average annual rate (compared to 2.6 per cent in the U.S.), with all the gains coming from productivity growth (2.0 per cent per year) and total hours falling drastically at a 1.0 per cent annual rate.

The key development in the 1990s in Canadian manufacturing is not any absolute deterioration of productivity growth (in fact output per hour growth actually picked up 0.3 percentage points per year and output per person employed 0.6 per cent), but the collapse of output growth. Since U.S productivity growth continued to outpace that in Canada, the relative deterioration of Canada’s manufacturing productivity performance vis-a-vis the U.S. continued, although at a slightly lower pace (a differential of 1.1 points in output per hour growth compared to 1.5 per cent in the 1980s).

In March of 1998, following revisions by the Bureau of Economic Analysis at the Department of Commerce of its real output measures, the BLS substantively revised its manufacturing output series for the 1992-96 period. Instead of a 12.7 increase in output between 1992 and 1996, output now rose 20.2 per cent, a 7.5 per cent upward revision. The growth rate for the 1989-96 period rose from an average annual rate of 2.2 per cent to 3.1 per cent. It is important to note that without this revision, there would be minimal differences in manufacturing productivity growth between Canada and the U.S. in the 1990s.

In additional to labour productivity, Canada’s total factor productivity performance in manufacturing was also markedly inferior to that in the United States. The CSLS productivity database (www.csls.ca) shows that total factor productivity in Canadian manufacturing rose at a 1.0 per cent average annual rate from 1984 to 1995, compared to a 1.3 per cent increase in the United States.

Slower productivity growth in Canadian manufacturing industries compared to U.S. manufacturing is a generalized phenomenon. According to data compiled by Statistics Canada (1996), over the 1985-92 period, all industries at the two-digit SIC industry level (15 out of 15) experienced slower total factor productivity growth in Canada (Table 3).

Canada’s manufacturing productivity performance in perspective

It is important to put Canada’s manufacturing performance in perspective by looking at its historical performance and by comparing it to that of the aggregate economy and other sectors. Four observations are particularly noteworthy.

First, the pace of productivity advance in manufacturing has been greater than the economy-wide average. In the 1981-89 period, aggregate output per worker grew at a 1.4 per cent average annual rate, compared to 1.9 per cent in manufacturing. In the 1990s, the differences widened, with manufacturing productivity growth accelerating to 2.0 per cent per year and aggregate economy productivity growth falling off significantly to 0.6 per cent, in part because of the negative productivity growth in certain service industries (finance, insurance and real estate and community, business and personal services (Table 4)

Second, despite the slight pick-up of manufacturing productivity growth to 2.0 per cent in the 1990s, productivity growth in this sector still remains less than one half the 4.5 per cent rate of increase enjoyed in the 1960-73 period (Table 5).

Third, unlike manufacturing, many industries in Canada are not experiencing significantly inferior productivity growth to comparable U.S. industries (Table 4). Indeed, in the 1981-92 period, transportation and communications experienced faster productivity growth in Canada, while the rate of change for community, business and personal services was the same in both countries.

Fourth, since 1981, Canada has had by far the lowest growth in productivity in manufacturing in the G-7. The output per hour increase of 1.9 per cent per year was only 60 per cent of the G-7 average of 3.2 per cent (Table 5). Canada’s performance was markedly inferior to that of all other G-7 countries, with the sixth place country, Germany experienced a rate of advance 0.8 percentage pints higher.

Implications of lagging productivity for international competitiveness

Over the 1981-96 period, the relative deterioration of Canada’s manufacturing performance compared to the U.S. resulted in a deterioration in our competitiveness in U.S. markets, as measured by unit labour cost trends, offset somewhat by the depreciation of the value of the Canadian dollar (Table 6). As the growth of nominal hour labour compensation in 1981-96 was remarkable similar in Canada and the U.S. (4.3 per cent and 4.1 per cent respectively), the 1.3 percentage point slower productivity growth in Canada translated into a comparable differential in domestic currency unit labour cost growth, to Canada’s disadvantage. The 0.9 per cent average annual depreciation in the value of the Canadian dollar between 1981 and 1996 offset much, but not all of the greater increase in domestic currency unit labour costs in Canada (2.4 per cent versus 0.9 per cent) so that in U.S. do llar terms, unit labour costs increased 1.5 per cent per year in Canada, compared to 0.9 per cent in the U.S.

This section of the paper reviews a number of possible explanations for weaker manufacturing productivity growth in Canada relative to the United States since 1981.

One point before proceeding is to note that since total economy labour productivity growth in Canada has not significantly deteriorated relative to the United States since 1981, factors affecting productivity in all sectors may be less important in accounting for Canada’s laggard manufacturing productivity performance than factors specific to manufacturing.

output measurement issues

Table 7 gives the shares of constant ($1992) output for two-digit manufacturing industries in Canada and the United States in 1992 and 1996 and the percentage change in output between these two years based on a crude concordance of industry classifications. Between 1992 and 1996, the total output of the manufacturing sector increased 24.5 per cent in the United States versus 19.5 per cent in Canada. But one industry, electrical and electronic products accounted for 45.7 per cent of the increase in output in the United States, versus only 10.0 per cent in Canada. Output in the electrical and electronic products sector rose an amazing 120.5 per cent in the U.S. in the four-year period (raising the sector’s output share from 7.9 per cent to 16.4 per cent), compared to only 26.4 per cent in Canada.

Excluding this sector, manufacturing output only advanced 14.6 per cent in the United States versus 19.0 per cent in Canada. This suggests that much of the greater productivity gains in the United States in the 1990s are due to this massive increase in the output of the electronic and electrical equipment sector, commonly known as the high-tech sector. A key question is whether this increase is real or rather represents a statistical artifact associated with the price indices for high-technology output such as computers. This question is part of the larger issue of whether Canadian and American statistical agencies use different methods in the estimation of real value-added, with obvious implications for the reliability of international productivity comparisons.

output growth

The weak pace of output growth in Canadian manufacturing in the 1990s was mentioned earlier. This situation could explain, at least in part, why productivity growth was not faster given the many favourable structural influences. Rapid demand and output growth boosts productivity through economics of scale associated with long production runs and the spreading of overhead costs, and by learning by doing. This relationship is known as Verdoorn’s Law. Domestic demand growth was very weak in the 1990s, due to tight monetary policy and the ensuing high interest rate in the early part of the decade and fiscal retrenchment in response to the deterioration of government finances in mid-decade. While foreign demand was strong, it was insufficient to offset the weakness of domestic demand and bring the economy back to potential output. Had manufacturing output growth in the 1990s been at a rate comparable to that of the 1980s or at the pace the U.S enjoyed in the 1990s, it is very likely the productivity growth would have been higher.

innovation

As technical change is the driving force behind productivity growth, an obvious explanation for slower manufacturing productivity growth in Canada is a slower pace of technological change in Canada, due to less innovation. Two pieces of information suggest that Canada does not have a growing innovation gap with the United States, although there still may be a significant gap.

First, innovation surveys of Canadian and U.S. manufacturing have found a decreasing gap in the use of advanced technologies between Canada and the United States. Baldwin and Sabourin (1996) found that in 1989, 58 per cent of establishments in Canada used at least one advanced technology, compared to 74 per cent in the United States. By 1993, this 16 percentage point gap had been cut in half to 8 per cent as now 73 per cent of establishments in Canada used at least one advanced technology, compared to 81 per cent in the United States.

It is interesting to note that this innovation gap is concentrated among the medium and small-sized firms. In 1989, there was no innovation gap for establishments employing 500 or more workers, and in 1993 little gap (Table 8).

Second, the relative gap in innovation effort between Canada and the United States has declined in recent years. Canada has been making more effort in terms of the share of GDP devoted to R&D, a crude indicator of the intensity of a country’s innovation effort, while the United States is making less in the last 15 years, although the absolute level of innovation remains greater in the U.S. Canada’s share of GDP devoted to R&D increased from 1.24 per cent in 1981 (and 1.05 per cent in 1976) to 1.59 per cent in 1995, while the share in the U.S. fell from 2.91 per cent in 1986 to 2.45 per cent in 1995 (reduced military R&D accounts for some of his decline).

It is often noted that Canada generates only a very small proportion of the new technology it introduces, drawing on the rest of the world and the U.S. in particular for its stock of innovations. A key mechanism for the transfer of technology has been foreign direct investment in Canada as foreign-owned firms have easy access to the best-practice technology in their home country. Changing patterns of FDI in Canada in recent years may have affected the transfer of technology, to the detriment of innovation in Canada.

human capital investment

A key long-run determinant of productivity is the quality of the labour force, in large part determined by the quantity and quality of investment in human capital. Canada’s slower manufacturing productivity growth may reflect slower growth in the quantity and quality of the workforce in Canada relative to the United States.

There does not appear to be any evidence that Canada’s human capital is markedly inferior to that of the United States, or that it has deteriorated relative to the United States in recent years, as the following data illustrate

Recent research suggests that innovative human resource practices and organization changes can have a significant positive impact on productivity growth. Canada’s lagging manufacturing productivity growth may reflect a relative failure to have implemented these practices relative to the United States. Again evidence is limited as there are few studies than conduct comparable workplace surveys in the two countries that would allow one to assess this hypothesis. Anecdotal evidence suggests that this is not a key factor in explaining Canada’s slower manufacturing productivity growth.

real wage growth

Labour productivity is affected by the relative price of labour. High wage levels or rates of growth of wages relative to other factors of production provide an incentive for firms to substitute capital for labour, while low wages or wage growth have the opposite effect. Slower manufacturing productivity growth in Canada could reflect slower growth of real wages or low wage levels relative to the United States.

BLS (1997) produces data on hourly wages for production workers in manufacturing, expressed in U.S. dollars, for industrial countries (Table 9). In 1996, hourly wages were 6.2 per cent higher in the United States compared to Canada( $17.70 versus $16.66). a reversal of the situation of the early 1990s. In terms of growth rates of wages in domestic currency, Table 6 shows that both nominal and real hourly compensation growth was virtually identical in the two countries over the 1981-96 period. This suggests differences in wage growth have not contributed significantly to productivity growth differences.

other factors

In addition to the factors outlined above, other possible explanations of slower manufacturing productivity growth in Canada include:

- weaker capital investment and capital-labor growth in Canadian manufacturing;

- a less favourable regulatory environment for productivity growth in Canada; and

- slower adjustment by Canadian manufacturing firms and workers to shocks because of rigidities.

This paper has discussed in a very preliminary manner some of the possible cause of slower labour productivity growth in Canadian manufacturing relative to that in the United States since 1981. As noted earlier, the major recent revision in the US manufacturing constant dollar output series has transformed a situation of comparable manufacturing productivity growth in the two countries in the 1990s to one of significantly faster growth in the United States. The factors behind this revision need investigation to ensure that comparisons of official productivity statistics are meaningful, particularly for high-tech industries. The existence of lagging manufacturing productivity in Canada in the 1990s relative to the United States does not appear particularly robust, depending on statistical revisions in the U.S. real output series, which could be further revised in the future, and the behaviour of output in one sector, that is electrical and electronic products.

The second key factor that can account for slower relative productivity growth in Canadian manufacturing in the 1990s (but less in the 1980s) is the less favourable macroeconomic environment. With output only advancing at a 1.0 per cent average annual rate since 1989, Canadian manufacturers have less opportunity to exploit economies of scale and longer production runs that U.S. manufacturers where demand was growing at a 2.6 per cent rate.

Despite the above tentative conclusions, our understanding of Canada’s deteriorating manufacturing productivity relative to the United States is poor. It is hoped that the CSLS project on this topic will be able to provide a definitive answer to this important question.

Allard-Saulnier, Marie (1993) "Productivity of Manufacturing Industries in Canada and the United States," Canadian Economic Observer, Statistics Canada, July, pp. 4.1-4.23.

Baldwin, J. and D. Sabourin (1996) "Technology and Competitiveness in Canadian Manufacturing Establishments," Canadian Economic Observer, Statistics Canada, May. pp.3.1-3.15.

Beede N. and Kan H. Young (1998) "Patterns of Advanced Technology Adoption and Manufacturing Performance, Business Economics, Volume XXXIII, Number 2, January, pp. 43-48.

Conference Board (1997) "Perspectives on a Global Economy: Technology, Productivity, and Growth: U.S. and German Issues," Report Number 1206-97-RR, Winter.

de Jong, Gjalt (1996) "Canada’s Post-war Manufacturing Performance: A Comparison with the United States," Research Memorandum GD-32, Groningen Growth and Development Centre, University of Groningen, December.

Denny, Mike, Jeff Bernstein, Mel Fuss, S. Nakamura, and L. Waverman (1992) "Productivity in Manufacturing Industries, Canada, Japan, and the U.S., 1953-86: was the productivity slowdown reversed?" Canadian Journal of Economics, XXV, No, 3, August.

Magun, Sunder (1998) " Productivity and Cost Performance: Canada Versus the United States," paper presented at the CSLS session on the Canada-U.S. Manufacturing Productivity Gap at the annual meeting of the Canadian Economics Association, May 29-31, University of Ottawa, Ottawa, Ontario.

Pilat, Derek (1996) "Labour Productivity Levels in OECD Countries: Estimates for Manufacturing and Selected Service Industries," OECD Working Papers No. 169.

Rao, Someshwar and Tony Lempriere (1992) Canada’s Productivity Performance (Ottawa: Economic Council of Canada).

Sabourin, D. and J. Baldwin (1997) "Factors Affecting Technology Adoption: A Comparison of Canada and the United States, Canadian Economic Observer, Statistics Canada, August, pp.3.1-3.17.

Sharpe, Andrew (1998) Productivity: Key To Economic Success, report prepared by the Centre for the Study of Living Standards for the Atlantic Canada Opportunities Agency.

Spiro, Peter (1997) "A Macroeconomic Analysis of International Differences in Manufacturing Productivity Growth," Ontario Ministry of Finance, unpublished paper.

Statistics Canada (1996) Aggregate Productivity Measures,1994 cat. 15-204.

West, E.C. (1971) Canada-United States Price and Productivity Differences in Manufacturing Industries (Ottawa: Economic Council of Canada).

World Economic Forum

Table 1
Output, Output Per Hour, and Labour Input in Manufacturing in Canada and the United States
(average annual rate of change)

Year	real output		output per hour		total hours		employment		output per employed person
	Can.	U.S.	Can.	U.S.	Can.	U.S.	Can.	U.S.	Can.	U.S
1981-89	2.7	3.2	1.7	3.2	1	0	0.8	-0.5	1.9	3.7
1989-96	1	2.6	2	3.1	-1	-0.5	-1.4	-0.7	2.5	3.3
1981-96	1.9	2.9	1.9	3.2	0.1	-0.2	0.2	-0.6	2.2	3.5

Source: International Comparisons of Manufacturing Productivity and Unit Labour Costs, Updated Trends for 1996, News, Bureau of Labor Statistics, U.S. Department of Labor, March 27, 1998.

Table 2
Relative Level of Output per Hour in Canadian Manufacturing
(U.S. = 100 in all years)
 

1981	87.4
1982	80.9
1983	83.1
1984	86.6
1985	85.9
1986	84.1
1987	79.4
1988	76.9
1989	77.8
1990	78.6
1991	78.6
1992	80.1
1993	79.8
1994	78.7
1995	74.8
1996	72.3

Source: Based on an estimate of 79.4 for the 1987 benchmark year by de Jong (1996) and growth rates from International Comparisons of Manufacturing Productivity and Unit Labour Costs, Updated Trends for 1996, News, Bureau of Labor Statistics, U.S. Department of Labor, March 27, 1998.

Table 3
Total Factor Productivity in Manufacturing Industries in Canada and the United States, 1985-92
(average annual rate of change)
 

	Canada	United States
Total Manufacturing Industries	-0.6	1.0
Food and beverage	-0.5	0.4
Plastic and rubber	-0.5	0.4
Leather	-0.7	1.2
Textile	-0.3	1.2
Clothing	-0.2	1.0
Wood	0.1	0.9
Furniture	-1.4	-0.2
Paper	-1.5	0.6
Printing	-3.0	-1.4
Primary metals	0.1	0.4
Machinery, electrical and electronic products	0.3	3.5
Transportation Equipment	-0.6	-0.2
Non-metallic mineral prod.	-1.3	0.8
Refined Petroleum	0.0	0.5
Chemicals	0.0	0.8

Source: Aggregate Productivity Measures, 1994 cat 15-204, Statistics Canada

Table 4
Output Per Hour Trends by Sector in Canada and the United States, 1981-92
(average annual rate of change)
 

	Canada	United States
Manufacturing	1.8	2.6
Transport	1.8	0.8
Communication	5.3	4.4
FIRE	-1.0	0.5
Trade	1.9	2.7
retail	1.4	1.5
wholesale	1.8	4.3
CBPS	-0.5	-0.5

Source: Statistics Canada and Federal Reserve Board
 

Table 5
Output Per Hour Growth in Manufacturing in Industrial Countries
(average annual rate of change)
 

	1960-73	1973-81	1981-89	1989-96	1981-96
Canada	4.5	1.7	1.7	2	1.9
U.S.	NA	NA	3.2	3.1	3.2
Japan	10	4	3.9	3.6	3.7
France	6.8	3.8	3.7	2.7	3.2
U.K	4.2	1.3	5.3	3	4.2
Italy	6.4	5.1	4	3.4	3.7
Germany	5.8	3.2	2.3	3.1	2.7
G-7 average	NA	NA	3.4	3	3.2
Belgium	6.9	6.2	4.3	1.9	3.2
Denmark	6.4	4	0.7	2.7*	1.4*
Netherlands	7.3	4.9	3.9	3.4	3.7
Norway	4.8	1.6	2.7	1.3	2
Sweden	6.4	2.1	3.1	4.2	3.6

Source: International Comparisons of Manufacturing Productivity and Unit Labor Cost Trends, Updated Data for 1996, Bureau of Labor Statistics, March 28, 1998
Note: NA indicates data are not available.* 1989-93
 

Table 6
Trends in Compensation, Unit Labour Costs in Manufacturing in Canada and the United States
(average annual rate of change)
 

	Nominal hourly labour compensation		Real hourly labour comp. (CPI basis)
	Can.	U.S.	Can.	U.S.
1981-89	5.2	4.4	-0.1	0.5
1989-96	3.3	3.7	0.8	0.3
1981-96	4.3	4.1	0.3	0.4

.

	Unit Labour Cost (ULC) in domestic currency		Exchange Rate ULC in U.S. dollars
	Can.	U.S.	Can.	Can.	U.S.
1981-89	3.4	1.2	0.2	3.6	1.2
1989-96	1.2	0.6	-2	-0.8	0.6
1981-96	2.4	0.9	-0.9	1.5	0.9

Source: International Comparisons of Manufacturing Productivity and Unit Labor Cost Trends, Updated Data for 1996, Bureau of Labor Statistics, March 28, 1998.

Table 7
Manufacturing Output by Sector in Canada and the United States, 1992-96
 

	Output Shares (1992$)				Percentage Change 1992-96
	1992		1996
Total Manufacturing	Can.	U.S.	Can.	U.S.	Can.	U.S.
Total Manufacturing	100	100	100	100	19.5	24.5
Wood	5.08	3.01	4.93	2.54	16	5
Furniture	1.71	1.52	1.9	1.42	33	16.1
Primary Metals	5.5	3.67	5.68	3.54	23.2	20
Fabricated Metal Prod.	6.68	6.59	6.92	7.1	23.9	34.1
Machinery	3.95	10.21	5.16	14.06	56.2	71.4
Electric & Electronic Prod.	7.43	9.27	7.85	16.42	26.4	120.5
Transport Equip.	14.24	10.28	16.25	9.11	36.4	10.3
Other Mfg.	3.06	6.99	3.02	4.65	18	-17.1
Food & Beverage	16.24	9.6	14.67	8.53	8	10.6
Tobacco	1.02	1.73	0.85	1.81	-0.1	29.9
Textiles	2.14	2.39	2.17	2.01	20.7	4.7
Clothing	2.67	2.56	2.2	2.03	-1.6	-1.1
Paper	6.01	4.31	5.74	3.57	14	3.3
Chemicals	8.59	11.33	8.39	10.74	16.7	18
Petroleum Prod.	1.01	2.65	0.97	2.55	14	19.9
Rubber and Plastics	3.98	3.58	4.53	3.85	36.2	33.6
Leather	0.41	0.45	0.29	0.36	-15.8	0
Non-metallic	10.27	9.85	8.46	7.81	-1.6	-1.3

Source: Gross Domestic Product by Industry, cat 15-001, Statistics Canada, October, 1997; Survey of Current Business, Bureau of Economic Analysis, 1998.

Table 8
Innovation in Canadian and U.S. Manufacturing
(% of firms using at least one technology)
 

Establishment Size	1989		1993
	Canada	U.S.	Canada	U.S.
20-99	50	67	70	75
100-499	81	89	85	94
500 plus	98	98	94	97
total	58	74	73	81

Source: Baldwin and Sabourin (1996).
 

Table 9
Hourly Compensation in U.S. Dollars for Production Workers in Manufacturing
 

Year	Canada	United States
1975	$5.96	$6.36
1980	8.67	9.87
1985	10.94	13.01
1990	15.84	14.91
1992	17.03	16.09
1993	16.44	16.51
1994	15.85	16.87
1995	16.04	17.19
1996	16.66	17.7

Source: International Comparisons of Hourly Compensation Costs for Production Workers in Manufacturing Updated for 1996, Bureau of Labor Statistics, February 9, 1998