Nonlinear Cointegration and Relevant Market Definition: A Study of Grocery Markets


Principal Investigator
 
 
Co-Investigator(s)
 
Grant Type
Research Grants Council - Faculty Development Scheme
 
Project Code
UGC/FDS15/B06/14
 
Amount awarded
$385,940
 
Funding Year
2015
 
Duration of the Project
24 months
 
Status
Complete
 
Start date
01/2015
 
Completion Date
12/2016
 
Abstract
The grocery industry in Hong Kong has increased its concentration over the last three decades. High concentration may give rise to market power, which in turn causes anti-competitive behaviour and high grocery prices that lower the standard of living among consumers, especially those in low-income households. However, based on the theory of contestable markets, high concentration may not necessarily result in excessive market power. On the contrary, its positive impact can allow the dominating few large supermarket chains to take advantage of the economy of scale and transfer the efficiency gains to benefit the community.

If there is fit and proper competition in the grocery industry, the positive benefits should out-weigh the negative costs to the consumers. To assess the degree of competition in the grocery industry, the common approach in the competition policy literature is to first define the relevant market, which identifies all substitutes that are competing in the market. Fewer substitutes reflect a lack of competition where sub-optimal productive and allocative efficiency would likely occur as a result. Competition authorities in foreign countries have recognised the usefulness of econometric cointegration tests in defining the relevant market and have used the test results to support policy and court decisions.

However, the cointegration tests previously adopted by the anti-competition research studies do not consider transaction costs. The existence of these transaction costs are however important because if the potential gain from price differentials of similar products does not outweigh the transaction costs, the arbitrage process to equalise the two product prices will not take place. Hence, when transaction costs are ignored in the estimation, though these two similar products are competing fiercely in a fit and proper competitive market and their prices synchronise closely, the testing results may still mistakenly conclude that these two products are not competitive substitutes.

This research project overcomes this weakness of the previous studies by embedding the transaction costs into nonlinear price-based cointegration tests that specifies different types of regime-switching and adjustment processes. We adopt these econometric methods to examine the price dynamics in the grocery industry in Hong Kong. If possible, we also apply the same methodology to some other economies of interest.
 
Keyword(s)
Grocery shopping
Supermarket shopping