Some empirics on the convergence of carbon dioxide emissions intensity across US states

ABSTRACT

This study examines the convergence of carbon dioxide emissions intensity across the 50 US states including the District of Columbia over the period 1997–2013 using a battery of tests. The results from cross-sectional tests of β- and σ-convergence lend support for overall convergence in carbon dioxide emissions intensity. On the other hand, panel unit root tests with allowance for endogenous structural breaks reveal that relative carbon dioxide emissions intensity contains a unit root, hence the absence of stochastic convergence.

KEYWORDS:

• Carbon dioxide emissions intensity
• convergence
• US states

JEL Classification:

• C33
• Q47
• Q48
• Q54

Notes

¹ Energy Policy and Conservation Act of 1975, National Energy Act of 1978, Energy Security Act of 1980, Energy Policy Acts of 1992 and 2005, and the Energy Independence and Security Act of 2007.

² Going forward, we refer “states” to include the District of Columbia.

³ In their analysis of regional income convergence, Carlino and Mills (1993) suggest testing for stochastic convergence with the use of the relative measure given by Eq. (3).

⁴ The CD test follows asymptotically the standard normal distribution for stationary dynamic and unit root heterogeneous panels with a relatively short time period, T, and a large number of states, N.

⁵ Bootstrap blocks of m = 20 are used in our analysis. Results are not sensitive to the size of the bootstrap blocks.