Skip to content

The Cost of Rigidity: The Case of the South African Labor Market

Literature › Works

Johannes Fedderke (2012). The Cost of Rigidity: The Case of the South African Labor Market. ERSA Working Paper 290. Economic Research Southern Africa (ERSA).
Type working paper
Scope Post-1994 South African economy
Method synthesis review
Period studied 1970–2010
Regions South Africa
Themes labour markets wages, inequality poverty unemployment, manufacturing industrial policy, education human capital, methods data sources, mining migrant labour, comparative theory
Citation key fedderke2012

Argument

The paper argues that high, persistent unemployment in South Africa stems from rigidities in both labor and output markets. Real labor costs keep rising despite mass unemployment, and small output shocks trigger large job losses. A simple Cobb-Douglas growth and labor-market model shows that the wage, supply and output elasticities are interdependent, which lets the author organise a selective review of existing studies. Output-market pricing power (high mark-ups, rising concentration, heavy regulation) lowers productivity growth and raises labor-market rigidity. Evidence on demand elasticities suggests labor is mispriced relative to productivity. The paper concludes that policy must pursue either stronger growth or lower real labor costs, and that current policy pursues neither.

Evidence and method

Secondary literature (mostly peer-reviewed econometric studies, many by the author) on labor supply and demand elasticities, mark-ups and concentration. Data: SARB output, employment and unit labor cost series; Banerjee et al (2008) unemployment rates; manufacturing censuses and large sample surveys; World Development Indicators; OECD R&D and regulation indicators; the author's own capital-labor ratio and CR10 concentration calculations for 66 countries.

Findings

  • The 2008-09 output fall was mild but the employment fall was far larger, indicating strong labor-market rigidity. (p. 2)

    The decline in real aggregate output in South Africa was of the order of 1.5%, while aggregate employment fell by 6.7%.

  • Real labor costs rose even as unemployment roughly doubled between 1995 and 2005. (p. 2)

    the real cost of labor not only showed no sign of falling, but in fact rose dramatically instead.

  • Employment growth lagged population growth over 1970-2005, so rising unemployment followed of necessity and the demand side is critical. (p. 11)

    From 1970 - 2005 the population grew at 2.7% per annum. Over the same period, employment grew at 1.3 % per annum.

  • Manufacturing mark-ups are far higher than in the US, indicating pricing power in output markets. (p. 7)

    Empirical results indicate the presence of a mark-up of price over the marginal cost of production in the South African manufacturing sector that is two to three times higher

  • Higher mark-ups reduce productivity growth, which indirectly depresses employment. (p. 8)

    a ten percent increase in mark-ups would lower productivity growth in South Africa by 1.6% to 2.5% per year

  • In gold and uranium mining, falling real labor costs coincided with employment growth, and rising costs with labor shedding. (p. 15)

    Periods of labor cost moderation, or declines in real labor costs over this period by contrast have seen growth in employment in the sector.

  • The conventional -0.7 wage elasticity is an average that hides much higher values in the biggest employing sectors. (p. 16)

    with some sectors reported with considerably lower elasticities (mining at -0.15), but also sectors with considerably higher elasticities (services at -0.95 and manufacturing at -1.0).

  • South African production is highly capital-intensive relative to other middle-income countries, favouring a high labor demand elasticity reading. (p. 18)

    As is clear from the data, capital intensity in South Africa is amongst the highest in the sample.

  • Real wages rose about twice as fast as demand-side factors warranted, implying labor mispricing, while trade liberalization was not the culprit. (p. 19)

    is found to have mandated an increase in real wages of on average 0.5% per annum over the last three decades of the twentieth century. Since in fact real wages have increased by 1% per annum

  • Labor is priced too high relative to productivity for full employment to be feasible. (p. 21)

    In short: the price of labor is simply too high at prevailing productivity levels for full employment to be feasible in South Africa.

  • Recent experience supports the Class 1 reading, in which growth policy is the main lever for employment. (p. 22)

    the fact that employment responded positively to the period of growth recovery during the 2000s, as well as strongly to the economic downturn post 2008

  • The demand-side data base is deficient: manufacturing censuses ended in 1996 and are not comparable with later surveys. (p. 11)

    South Africa conducted manufacturing censuses on either a biannual or triannual frequency over the full 1917 - 1996 period.