Tuesday, October 22, 2019

Effect of Trade and Environmental Treaty Networks on Trade in Environmental Goods & Services




BACKGROUND:
In the light of the failure of the Doha Development Round of the World Trade Organisation (WTO), preferential trade agreements (PTAs) have proliferated in the last 25 years. From around 50 PTAs in force in 1991, this number has grown to 481 in 2019. The PTAs now form a complex network as shown below:

PTAs have evolved from “shallow” agreements that cover tariff and non-tariff barriers to “deep” agreements that cover a host of other issues like labour, environment and intellectual property rights. [1]

Many PTAs now cover environmental aspects such as trade in environmental goods and services, and compliance with multilateral environmental agreements to which the signatories of the PTA are already a party. Bilateral and Multilateral Environmental Agreements (BEAs and MEAs) also form a complex and overlapping network and may include trade in environmental goods and services.
However, traditionally growth in trade and environmental protection have been seen as opposing forces rather than complementary goals. Are the most central actors in the trade network also providing leadership in trade in environmental goods?

NETWORK QUESTIONs:
1)     Which network has greater effect in increasing trade in environmental goods and services – PTAs or BEAs?
2)     Are the most central actors in trade networks also providing leadership in trade of environmental goods and services? Or can we identify new actors in this space?

RESEARCH METHODOLOGY:
Data will be used from the following sites:
2)     RTA Exchange: https://rtaexchange.org/
3)     Database of MEAs: https://iea.uoregon.edu/
4)     World Bank Database of World Trade statistics: https://wits.worldbank.org/countrystats.aspx?lang=en
5)     WTO Environmental Database: https://edb.wto.org/

NETWORK MEASURES:
The network of PTAs as well as the network of MEAs are two-mode, nondirectional networks, which may be given a value as deep or shallow agreement. This will be correlated with the network of actual flows of environmental goods and services across borders, which is a one-mode, directional network, valued by amount of trade (low =1, medium =2, high = 3).

Centrality
Centrality measures will help us see whether the same actors are leaders in signing both PTAs and MEAs and whether this translates into actual leadership in trade in environmental goods. Here, degree centrality is not of as much use as signing of treaties is no guarantee of actual increase in trade. A country could sign a large number of treaties with actors that are peripheral in the international trades network. Hence, Eigenvector centrality will be given more importance. A large betweenness score could indicate that the country acts as an important trade centre in connecting two otherwise loosely connected economies.

Density & Centralisation
Density will help us understand the potential linkages while Centralisation will identify hubs in the trade network.

Subgroup Analysis
It will be useful to identify cliques in the trade of environmental goods and whether flows are predominantly one of the four:
1)     North-North
2)     North-South
3)     South-North
4)     South-South

My hypothesis is the first two types of flows dominate when it comes to trade in environmental goods and services. Subgroup analysis will help us identify the hubs and spokes of the system or identify a core-periphery structure if it exists, in trade in environmental goods and services.

WHY SNA?
Social Network Analysis is a useful tool here besides traditional economic analysis because it can help identify a core-periphery structure in the trade of environmental goods through centrality, whole network and subgroup analysis. It can also help highlight potential linkages in the network and identify new opportunities and potential leaders/hegemons in trading environmental goods.

LIMITATIONS:
If I take the second module, I will only be restricting my analysis to Eurasia due to shortage of time.


REFERENCES:

1 comment:

Christopher Tunnard said...

Sapna, this is a very good idea that you have thought more about since we initially discussed it. You've incorporated net measurements well. I'm assuming that the data is readily available and that you can come up with a sensible approach to coding it. You might get some help from Alex Kersten on that. If the visualizations are complex, you might also consider using Gephi, which does a much better job on complex maps than NetDraw.

Look forward to having you participate in the second module.