Price discovery analysis of green equity indices using robust asymmetric vector autoregression

- Green equity indices considered spanning global, regional and sectoral.
- Use of robust asymmetric vector autoregression model
- Evidence of wide range of interactions within and across markets
- Broad equity indices and NBP natural gas identified as main causal influences
- Recognition of multiple comparisons bias and application of correction techniques

Covering the first commitment period of the Kyoto Protocol (2008-2012), we perform a price discovery analysis to determine Granger causality relationships for a range of prominent green equity indices with the broader equity and commodity markets. Three pivotal contributions are made. Firstly, an expanded database is used that gives greater depth to the price discovery analysis relative to previous literature. Prominent global, regional and sectoral green equity indices are considered, as well as a broader set of commodities including crude oil, natural gas and emissions. The inclusion of natural gas recognises its role as the transition fossil fuel to a low carbon economy. In addition to the main European Union Allowance traded under the EU Emissions Trading Scheme, Certified Emissions Reduction (CER) prices are also included in the emissions database to capture activities under the global Clean Development Mechanism. Secondly, a problem with conventional symmetric vector autoregression is that its implementation commonly leads to large occurrences of insignificant parameters. Therefore, as a first layer of robustness, we utilise an asymmetric vector autoregression model to perform the Granger causality testing, which addresses this limitation by means of allowing different lag specifications among the system variables. Thirdly, explicit recognition is made in our study of the multiple comparisons bias inherent in our high-dimensional testing framework, which is the non-negligible likelihood of identifying statistically significant results by pure chance alone. As a second layer of robustness, we utilise a generalised Holm correction method to control this source of bias. At conventional statistical significance levels, we find that the FTSE 100 and FTSE Global Small Cap equity indices have a causal effect on all of the green equity indices, with limited evidence of causality in the opposite direction. Within the green equity markets, we find evidence that the chosen sectoral index has a Granger causal effect on one of the two global indices considered and also the regional index. This price transmission provides modest evidence that the global green economy is becoming ever more integrated. NBP gas is shown to have a causal effect on all of the green equity indices, whereas we find no such evidence for Brent oil. The former observation may reflect the increasing role of gas as the transition fuel to a low carbon economy, playing a key role in decisions on power generation mix and associated capital investment. Finally, we find no evidence that EUA or CER prices have a causal effect on green stocks, consistent with previous findings and likely reflecting the excessively low prices being commanded for compliance permits in the European emissions markets.