Row over productivity data
But other economists were quick to criticise Adcorp's findings, arguing that the employment services company's conclusions were not sufficiently backed up by concrete information.
Adcorp's employment index numbers, which it publishes monthly, have previously been hotly contested, with critics arguing that its methodology is incorrect and its data sources are insufficiently transparent. But Adcorp's chief economist, Loane Sharpe, has consistently stuck by the firm's research, arguing that its detractors, which include trade unions, are politically motivated.
This week's Adcorp employment index included its analysis of labour productivity, essentially a way to measure economic growth by looking at the amount of goods and service produced by workers.
Using data from Statistics South Africa and the South African Reserve Bank, Adcorp calculated that labour productivity, as measured by output per worker per unit of capital in South Africa, had fallen by 32.5% since 1967. Since its peak in 1993, labour productivity had fallen 41.2%.
This measure of labour productivity was a better way to examine the trend because it included the contributions of other factors of production in the economy such as land, capital and entrepreneurship, according to the report.
Another measure of productivity, more commonly used and defined as output per worker suffered from serious limitations, the report noted.
Using this lens, labour productivity was shown to rise consistently since the 1960s when official records began.
According to this measure, the average worker produced R63162 of real output in a year, whereas in 2012 the average worker produced R143 412 of real output per year, an increase of 127.1%.
According to Adcorp, "the problem with the output-per-worker definition of labour productivity is that output is not only produced by workers" and it failed to account for the contribution of factors such as capital.
It also did not explain why the South African economy's labour intensity was falling; in other words, why returns on capital were rising. Nor did it explain why labour's share of national income was at an all-time low, argued Adcorp.
When labour productivity was measured in a way that included the contribution of other factors such as capital, a decline in productivity became apparent and correlated with these other trends.
Two factors were significant in driving down labour productivity, according to the report: labour legislation and bargaining councils.
"First, dismissal protection contained in the Labour Relations Act (1995) have made managing poor worker performance significantly harder," it said.
"Second, bargaining councils – also created by the Labour Relations Act (1995) – have led wage increases to become decoupled from productivity gains, with the result that wages are no longer an incentive to improved worker productivity."
Miriam Altman, a distinguished research fellow at the Human Sciences Research Council, said there were a number of problems with Adcorp's report, beginning with a lack of information on how it calculated labour productivity.
Adcorp had been unwilling to reveal its methodology, she said. "No information or data is provided to see how it calculates productivity. Measuring productivity well is notoriously difficult and quite neglected in South Africa. This report does nothing to remedy that," she said.
Altman did agree that labour productivity measured as output per worker was not a good measure of the productiveness of the economy, nor did it say much about how productive labour was. It could simply be a reflection of growing capital intensity. Instead, "total factor productivity" that combined trends in both capital and labour productivity was a better measure, she said. Although Adcorp was essentially proposing to use a measure of total factor productivity, which was welcome, its method and use of data was not made clear.
In addition, an economy-wide measure, over such a long period, did not reveal useful insights, Altman said. Since 1970 there had been major structural shifts towards services, significant job losses, capital intensification in agriculture and substantial capital intensification in manufacturing.
"A more interesting exercise would involve sector analysis, looking at capital, labour and total factor productivity, with an understanding of the sector dynamics in regard to capital investment or capacity utilisation," she argued.
"One would find a very different result if analysing goods producing sectors such as agriculture, mining or manufacturing."
Insofar as there had been a decline in productivity, as measured as output per worker per unit of capital, it could be thanks to the greater role services had played in the economy over the years, Altman said. "Over the period in question, about three-quarters of all net new jobs were created in services," she said.
"Therefore, the changing composition of the economy could explain why output per unit of labour per unit of capital might fall over time, since services are inherently less capital intensive."
Adcorp's assertion that the decline in productivity was linked to weak incentives to perform and to wages becoming delinked from productivity was ideological and not based on evidence, she said.
"One cannot attribute economy-wide phenomena to psychological motivation such as worker laziness." Each major sector had its own dynamics, such as the public sector, different sections of manufacturing, mining and agriculture and services.
The connection made to labour regulation was "tenuous at best", Altman said.
Adcorp's measurement and analysis of labour productivity decline did not necessarily explain the falling labour intensity of the economy or the decline in labour's share of national income, in effect a measure of workers' share of gross domestic product.
"Falling labour intensity over this period was caused by different factors such as negative real interest rates in the 1970s, reform of the agricultural sector and mass labour shedding in the 1980s, capital intensification in mining [owing to the increasing depth at which mines operated] and the generally capital intensive orientation of manufacturing," said Altman.
The fall in labour's share of national income was partly caused by rising economy-wide productivity and the poor distribution of income through wage earnings, she said.
Haroon Bhorat, the director of the Development Policy Research Institute at the University of Cape Town, said that the conclusion that bargaining councils and labour legislation were responsible for declining productivity was "a leap which requires firmer evidence than is currently available".
"It may be true, but it has not been carefully estimated and proven econometrically," Bhorat said.
Adcorp economist Sharp stressed that his research showed that labour productivity had increased from 1970 to 1993. "I'm not sure how the services sector's low capital intensity could have a bearing on our finding," he argued.
"In contrast, labour productivity had been falling consistently since 1993, probably because methods of using labour had not matched the competitive requirements of the post-1994 internationally integrated South African economy."
There was merit in looking at labour productivity over a long period of time to gain a sense of where we have been and "possibly where we are heading", Sharp said.
According to Sharp, a number of similar factors helped to explain the fall in both labour intensity and labour's share of national income. These were higher returns on investment in capital, including mechanisation, automation and other labour-saving technologies; low returns on labour as a result of high wage increases that were out of step with productivity; increasingly disruptive labour conditions because of strikes and work stoppages; and high wages relative to our trading partners.