Skip to content

The effects of strikes in the South African gold mining industry on shareholder value

Literature › Works

Aayesha Seedat (2013). The effects of strikes in the South African gold mining industry on shareholder value. Research report, MCom (Accounting), Faculty of Commerce, Law and Management, University of the Witwatersrand. University of the Witwatersrand.
Type thesis
Scope Post-1994 South African economy
Method econometric
Period studied 2007–2012
Regions South Africa, Witwatersrand gold mining, JSE
Themes unions strikes, mining migrant labour, labour markets wages, business and state, methods data sources
Citation key seedat2013

Argument

The study uses an event-study design to test whether strikes at JSE-listed gold miners reduce shareholder value between January 2007 and December 2012. It finds negative cumulative abnormal returns after strike announcements, concentrated in roughly the first 10 days. The negative effect is clearer for protected than unprotected strikes, and shorter strikes (under 15 days) produce larger negative CARs than longer ones, although the difference between the two groups is not statistically significant. The author reads this as markets pricing in expected wage and cost increases, and she treats the result as relevant to the surge of unprotected strikes after Marikana in 2012.

Evidence and method

Department of Labour list of protected strikes; SENS announcements for unprotected strike dates (McGregor BFA database); daily closing share prices and ALSI index data from McGregor BFA; company reports and press statements (Gold Fields, AngloGold Ashanti, Harmony); 30 strike events (16 protected, 14 unprotected), analysed in Excel and SPSS.

Findings

  • Sample comprises 30 strike events, 16 protected and 14 unprotected. (p. 26)

    A total of thirty strike events were included in the sample. Of the thirty strikes sixteen were protected and fourteen were unprotected.

  • Protected strikes show significant negative CAARs in the first 10 days after commencement, suggesting a gradual market adjustment. (p. 29)

    The fact that cumulative abnormal returns are most significant after 10 days suggests that markets do not immediately incorporate the new information into the share price but seem to take about two weeks to fully adjust.

  • Protected strike commencement CAAR over days 1 to 10 is -7.52% (t=-3.418, significant at 1%) and -9.43% over days 1 to 30 (not significant). (p. 29)

    1 to 10 -7.52 -3.418** -7.05 -3.478** 1 to 30 -9.43 -1.726 -7.11 -1.509

  • Protected strike settlement is followed by further significant share-price falls, contrary to the hypothesis of a positive settlement effect. (p. 30)

    One to 10 days after settlement, t value =-.3.478 and the p value < 0.001 this suggests that 1-10 days after settlement there is a negative reaction attributed to the investor realising that the settlement will add cost

  • Unprotected strike CAARs are not significantly different from zero, though the cumulative trend slopes downward. (p. 33)

    Table 2 above provides a summary of the results for the different intervals during the unprotected strike action. The returns are statistically insignificantly different from zero.

  • Combined protected and unprotected strikes show significant negative CAARs after commencement, including -5.95% over days 1 to 30. (p. 35)

    1 to 10 -3.92 -2.801** -3.60 -2.527* 1 to 30 -5.95 -2.026* -5.21 -1.995*

  • Protected strikes produced a larger fall in CAAR than unprotected strikes over days 1-30 after commencement (-9.43 vs -2.23). (p. 37)

    1 to 30 -9.43 -2.23 -7.11 -3.02

  • Short strikes (<15 days) have a significant negative announcement CAR; strikes over 15 days do not. (p. 39)

    The t test indicates a significant result for the shorter duration, whereas for strikes which occur for more than 15 days, there is no significant result. This result is inconsistent with hypothesis 4.

  • The duration difference is not statistically significant, so the finding rests on a small sample. (p. 41)

    Table 9 and 10 indicate the mean difference of -.05616489 is not significantly different from zero.

  • Trading volumes show no consistent reaction to strike commencement or settlement. (p. 44)

    The graphs in the Appendix B, indicates that the commencement of strikes do not cause any significant reaction from the markets.

  • Author's explanation for why shorter strikes hurt more is speculative: they are seen as more threatening and more violent, and tend to recur. (p. 46)

    A possible reason for the opposite effect is that the shorter strikes are seen as more threatening and are also generally more violent here in South Africa.