Wage Differentials among Ownership Groups and Worker Quality in Vietnamese Manufacturing

Abstract

Wholly foreign multinational enterprises (WFs), joint-venture multinationals (JVs), state-owned enterprises (SOEs) pay higher wages than domestic private firms in Vietnamese manufacturing. In large samples of medium–large (20+ employees) firms, conditional differentials accounting for worker education and occupation, as well as capital intensity, size, and shares of female workers, were substantially smaller, but positive and significant. Wage levels and differentials varied substantially among industries. Conditional differentials remained positive and significant for WFs and JVs in most of the 11 industries examined, but estimates of SOE-private differentials were insignificant in most industries. Robustness checks using 2007 data yielded similar results.

Key Words:

• Multinational enterprises
• manufacturing
• wage differentials
• state-owned enterprises
• ownership

Jel Classification:

• F23
• J31
• L60
• O53

Acknowledgements

This paper is one output of the research project “Multinationals, Wages, and Human Resources in Asia’s Large Developing Economies”, which was funded by the Asian Growth Research Institute (formerly the International Centre for the Study of East Asian Development) in fiscal 2014 (ending March 2015). We thank AGI and the University of Danang, School of Economics for financial and logistic assistance. Valuable comments were also received at an AGI Seminar on 25 February 2015. Responsibility for all opinions expressed and any remaining errors or omissions are the authors’ alone.

Notes

1. In the combined sample of all manufacturing plants, intercept dummies are used to capture industry-specific effects. The industry-level regressions are more general in that they allow intercepts and all slope coefficients to differ among industries

2. These are region 1 (Hanoi and Ho Chi Minh City) and region 2 (Hai Phong, Dong Nai, Binh Duong, Ba Ria Vung Tau, Quang Ninh, Da Nang, and Can) in the minimum wage classification.

3. Output is converted using a manufacturing output deflator at the two-digit level of Vietnam’s Standard Industrial Classification (VSIC). Capital is converted using the deflator is for fixed-capital formation from the national accounts (General Statistics Office, various years a, various years b).

4. In addition, only limited information is collected from very small local firms with 10 or fewer employees (Jammal et al., 2006).

5. For example, tertiary shares increased or decreased by more than 6 percentage points for SOEs and JVs in the metals group, general machinery, and electronic machinery, JVs only in wood and furniture, and SOEs only in transportation machinery. Although these large changes are not impossible and there were large economic changes in 2007–2009, variables like shares of workers by educational background do not usually change much in a short period of time.

6. One referee suggested that we also control for the effects of firm profitability as a proxy for demand-side effects. However, we prefer to exclude this variable because firm profitability reflects both supply- and demand-side effects, firm profitability (and productivity) is likely to be correlated with capital intensity and size, and we want to retain comparability with previous studies.

7. The regions are Hanoi, the Red River Delta, the North Mountainous Area, the Central Coast and Central Highland Area, the South East Area, Ho Chi Minh City, and the Mekong Delta (used as the base dummy). Industries are defined at the three-digit level for industries having at least two firms of each ownership type in them; three-digit categories are combined or two-digit categories used in industries with fewer firms.

8. Females tend to earn less than males because they tend to be less educated and have less experience in high paying jobs, and because they are discriminated against in the workplace and when educational resources are allocated.

9. Perhaps surprisingly, the omission of SP affected estimates of coefficients on SH by much less (change of +26% to +27%) than estimates of coefficients on SM (change of +67% to +70%). Omission of SP also had relatively large effects on the estimates of coefficients on KI (change of +25% in the lagged specification and +77% in contemporaneous specification), but coefficients on KI remained insignificant or only weakly significant at the 10% level in all estimates. Most importantly, all major qualitative results (coefficient signs and significance levels) were similar with or without SP included.

10. Because dependent and independent variables are in natural logs, conditional differentials are calculated as the exponential value of the relevant coefficients (a₇, a₈, a₉) from estimates of Equation (1).

11. Data downloaded from www.gso.gov.vn on 22 January 2014.

12. To conserve space, Table 7 only provides wage differential coefficients and results of testing the null hypothesis that JV-private and WF-private differentials were equal. Other details (e.g., other slope coefficients and the equation information provided in Table 6) are available from the authors upon request.