The Long-run Causal Relationship Between Military Expenditure and Economic Growth in China: Revisited

Abstract

This paper re-examines the long-run causal relationship between military expenditure and economic growth in China over the period 1952–2010. An empirical econometric analysis based on a Barro-style growth model is conducted. By employing the Bartlett corrected trace test, which provides better approximations of the finite sample distribution to determine the rank of cointegration, the results support the existence of a single long-run equilibrium relationship between the variables. Furthermore, it is confirmed that the cumulated shocks of military expenditure primarily originate from different components of shocks that relate to economic development rather than the other way round.

Keywords:

• Bartlett corrected trace test
• China
• Cointegration
• Economic growth
• Long-run
• Military expenditure

JEL Codes:

• H5
• O41
• O47

Acknowledgements

We would like to thank the editor and an anonymous referee for the valuable suggestions and comments on an earlier version of this paper. Special thanks go to Prof. Ron Smith for his insights and constructive comments on an earlier draft. Thanks to the CRS and CEA UK conference attendees for their recommendations and comments. Of course, the responsibility for any remaining errors is ours alone.

Notes

¹When the sample is small (as in our case), asymptotic distributions generally provide poor approximations of the true distributions (Juselius, 2006).

²Sources: SIPRI Military Expenditure Database: http://www.sipri.org/databases/milex/; and IMF, World Economic Outlook database, Sep. 2011: http://www.imf.org/external/pubs/ft/weo/2011/02/weodata/index.aspx).

³The reform programme mainly built on the promotion of industry and foreign trade. Thus, significant differences were observed in China after the reforms (e.g. Chinese GDP before 1978 was 5.9 and 10.5% after), even though the growth rate increased in both periods (see Tables I and II).

⁴‘… China in recent decades has developed a consumer-driven economy in which the ability of the state to direct resources towards the military has been severely curtailed’ (Nolt, 2002, 322).

⁵The policies that China adopted after 1949 were based on principles such as non-interference in the internal affairs of other countries known as peaceful coexistence (Khrushchev, 1959/1960).

⁶This involvement is also known as international non-combat operations. For more details, see Watson (2009).

⁷The official Chinese defence budget for 2010 was 532 billion Yuan ($78 billion), but according to SIPRI’s estimates, China’s total military expenditure was 809 billion Yuan ($119 billion). This shows an increase in military spending of 189% between 2001 and 2010 (an average annual increase of 12.5%) (SIPRI 2011).

⁸For a review of the theoretical literature, see Dunne, Smith, and Willenbockel (2005).

⁹The rest of the 14 countries were integrated of different orders.

¹⁰For a thorough discussion of the Feder–Ram and Solow models, see Dunne Smith, and Willenbockel (2005).

¹¹The Barro-style specification was taken up owing to the following recommendations by Devarajan, Swaroop, and Heng-Fu (1996), Stroup and Heckelman (2001), and D’Agostino, Dunne, and Pieroni (2010) who extended the Barro (1990) model to include military spending. Subsequently, Aizenman and Glick (2006) also reformulated the Barro-style model to examine a non-linear relationship between defence spending and economic growth where the effects of military expenditure were augmented by further interaction variables.

¹²Unlike Chen (1993), Wolde-Rufael (2001) and Bo and Xing (2011), we use the GDP in level instead of economic growth as the latter is likely to be stationary.

¹³Population growth is used because population in level is found to be an I(2) variable.

¹⁴Perron (1989) proposes a unit root test that allows for a known or exogenous structural break; a procedure that requires the visual inspection of the data prior to the test.

¹⁵We thank the anonymous referee for the suggestion of using an array of unit root tests to check the time series properties of the observed data.

¹⁶If a residual is larger than |3.5σ| and corresponds to a known event, we include a dummy variable (see Juselius and MacDonald, 2004).

¹⁷This was Mao’s attempt to modernise the Chinese economy.

¹⁸Dividing the sample into two sub-samples, pre-1978 and post-1978, in order to check the robustness of the results to the policy reforms in 1978 is not a feasible way, as the test is likely to lose its power because of the relatively large system and a small number of observations.

¹⁹Some recent high-level official statements support the fact that economic growth in China is a driver of defence spending. Speaking on the anniversary of the founding of the People’s Liberation Army, President Hu Jintao noted that ‘… [China] will gradually increase input in national defence as the economy grows, and continue to modernize national defence and the armed forces in a way that serves the interests of our national security and development’ (quoted in Wood, 2010, 3).