Oil shocks and the US terms of trade: gauging the role of the trade channel

Abstract

Recent theoretical literature claims that demand-driven transmission mechanisms are the key to understand how oil shocks affect the economy. Following this literature, we measure the economic strength of one of these demand-driven channels, the trade channel, in the transmission of oil shocks to the US economy. We use Kilian's (2009) decomposition of oil price shocks to identify three possible sources of oil shocks: oil supply, oil-market specific demand and global demand shocks. We then estimate the impact of each shock on the US terms of trade controlling for nonlinear effects in the sign and the size of the shocks. All oil shocks have persistent and statistically significant effects on the US terms of trade. However, we find that only oil supply shocks have an impact on the terms of trade that is nonlinear in the size of the shock. This last result is in accordance with the theoretical findings in Maravalle (forthcoming).

Keywords:

• oil shocks
• propagation mechanism
• asymmetric effects
• terms of trade

JEL Classification:

• Q43
• E30
• F40

Acknowledgements

The author acknowledges financial support from the Departamento de Educación Universidades e Investigación del Gobierno Vasco (IT-313-07).

Notes

¹ Oil price shocks, through terms of trade movements, transmit to the economy modifying the level and the composition of the aggregate demand and the trade balance. For an empirical analysis on the external adjustment to oil shocks, see Kilian et al. (2009).

² http://www-personal.umich.edu/~lkilian/reaupdate.txt.

³ Oil-market specific demand shocks are interpreted as unpredictable changes in the precautionary demand for oil.

⁴ A standard result in the literature is that the impact of oil shocks reaches a peak after 3–4 quarters. Moreover, results do not change when using 16 or 24 lags.

⁵ GIR take into account variability in both the initial values of the regressors and future values of the structural shocks. Standard Impulse Responses (IRs), instead, do not condition on initial values of the regressors and set to 0 future values of the structural shocks.

⁶ Kilian and Vigfusson (2009) find a similar result for energy price shocks.

⁷ For positive innovations, the size σ increases for σ→infinity. For negative innovations, the size σ increases for σ→0.