Gasoline, Diesel Fuel And Jet Fuel Demand In South Africa

Abstract

The paper investigates the price and income elasticity of gasoline (petrol), diesel and jet fuel demand in South Africa using autoregressive distributed lag (ARDL) models. We compare elasticity estimates for 1982Q1-2010Q4 with estimates for 1998Q1-2010Q4. Price and income elasticity estimates for gasoline remain unchanged compared to previous estimates and robust across smaller sub-periods. Similar to recent findings for other developing countries, income is the dominant driver of South African diesel demand even when controlling for the increased number of diesel vehicles. Similarly, income dominates jet fuel demand, a finding that is robust to controls for international tourist departures and is consistent with international findings.

Notes

1 The Main Supply Agreement (MSA) of 1954, concluded among South African petroleum companies, divided the country into two main fuel supply regions: the “in-land” region supplied exclusively by Sasol and the “coastal” region supplied by the coastal refineries (Swart 2010). Although the MSA was terminated in 2003, these definitions continue to be used widely in the industry.

2 Recent supply problems at fuel pump stations (especially in the in-land region) are, at least in part, also the result of insufficient planning – although inertia in energy policymaking was probably the main underlying driver.

3 Discrete choice models have become popular in recent decades, but the data requirements of these models prevent their application in the current context and most fuel demand studies follow our approach.

4 Some studies have considered demand for transport fuels by sector (Dimitropoulos, Hunt and Judge, 2005). We focus on aggregate demand, as our fuel sales data cannot be disaggregated by sector. Also, we do not include sector-specific variables in our aggregate demand functions. For example, we do not include agricultural production as an independent variable: income is likely to be closely related to agricultural production, biasing the aggregate income elasticity estimates.

5 Seasonal adjustment technique is the X12 procedure developed by the U.S. Census Bureau. The literature suggests that alternative techniques perform equally well.