WP/NPA -01/22

RESEARCH WORKING PAPER SERIES

Petroleum product subsidies in Ghana: Does the rich benefit more than the poor?

Joseph Wilson1

National Petroleum Authority (NPA)

The findings, interpretations and views expressed in this paper are entirely those of the
Author(s); they do not represent the views of the National Petroleum Authority (NPA).
Research working papers are ongoing research of the Author(s) and are published to elicit
comments and to stimulate discussion.

1
Director, Research Monitoring and Evaluation, [email protected]
Abstract

This study investigates the direct welfare impact of fuel price increases using Ghana living
standards survey data. The study argues that the incidence of increase in fuel prices falls on the
rich more than the poor, therefore, introducing subsidies to mitigate the impact will benefit the rich
more than the poor. The findings suggest that fuel price subsidies are regressive with the rich
benefiting three times more than the poor. The study recommends a policy that ensures a direct
cash transfer to cushion the poor against hikes in fuel prices or social intervention programs that
directly impact on the poor.

Key words: Fossil fuel subsidy, Budget share, Petroleum products Pricing, Price deregulation

1|Page
1.0 Introduction

A survey of existing literature indicates that there are numerous studies on the economic, social
and environmental impact of fossil fuel subsidy (FFS) reforms. Owing to the impact of FFS, many
countries have undertaken subsidy reforms to alleviate the effects of FFS on the economy. Several
studies including Burniaux et al., (2009) and Jensen and Tarr (2002) have highlighted the impact
of FFS reforms on terms of trade, greenhouse gases (GHG) and household’s income. For example,
Burniaux et al., (2009) have noted that global carbon dioxide (CO2) and greenhouse gas (GHG)
emission would reduce by 13 percent and 10 percent respectively if 20 of the Organization for
Economic Co-operation and Development (OECD) countries removed subsidies on fossil fuel and
electricity. Additionally, the removal of FFS reduces pressure on governments’ budgets and allow
revenue to be spent on social priority areas such as health and education (Koplow and Dernbach,
2001).
Though the removal of FFS has been identified to positively impact on the economy and
environment (Saunders and Schneider, 2000; Burniaux et al., 2009), there are concerns that higher
fossil fuel prices are regressive with a higher effect on poor households. Coady, El Said,
Gillingham, Kpodar, Medas and Newhouse (2006) estimated both the direct and indirect effects
of fossil-fuel subsidy reform in Bolivia, Ghana, Jordan, Mali and Sri Lanka. Their findings suggest
that there is a direct effect of hikes in fossil fuel prices on aggregate income and this ranges from
0.2 percent to 0.9 percent, with the effect being distributionally neutral. However, in countries
such as Ghana, Jordan and Sri Lanka, the impact of the fossil fuel hikes was regressive with a
higher effect on the lower income group than the highest income group. Similarly, Freund and
Wallich (2000) examined the effect of energy subsidy reforms in Poland and established that
decline in welfare was higher for the richest quintile (8.2%) than for the poorest quintile (5.9%)
when they assumed a zero elasticity of demand. The impact ranged from 4.6 to 7.6 percent of total
household’s budget indicating that the more elastic the demand, the lower the welfare loss was.
Theoretically, the removal of FFS results in higher fuel prices and reduces the consumption of
petroleum products and consequently limit GHG emissions. However, in poorer countries, the
removal of FFS on cooking fuel such as Liquefied Petroleum Gas (LPG) could have debilitating
effects because it could lead to reliance on biomass for cooking and heating thereby putting
pressure on the forest. Whatever be the case, subsidy reforms would result in economic loss;
therefore, some measures would have to be instituted to compensate the losers while efforts are
put in place by governments to redirect the income saved from subsidies to transfers or social
programs that directly affect the poor (International Energy Agency (IEA), 1999).
In Ghana, before deregulation of the downstream petroleum sector, retail prices of petroleum
products were fixed by the government. Retail prices were consistently lower than the market
prices due to subsidies on the products resulting in a high fiscal cost to the government. To ensure
gradual removal of petroleum products subsidy, the National Petroleum Authority (NPA) was
established in 2005 to, among other things, implement the deregulation policy for the downstream
petroleum sector.

2|Page
However, subsidies on petroleum products were continued until July 2015 when subsidies on
petroleum products such as gasoil, gasoline and LPG were completely removed. The removal of
subsidies occurred at a time when international oil prices were as low as $20/barrel.
Critics of petroleum product subsidies argue that real income losses may be substantial, as higher
fuel prices may not only imply higher prices for petroleum products consumed directly by
households, but also higher prices of other goods which use petroleum products as intermediate
goods in the production process (Nwachukwu et at, 2011).
Others believe that petroleum product subsidies compete for limited resources that could otherwise
be used to provide other essential services, widen the scope for rent-seeking and commercial
malpractices (Sharma, 2013), discourage both supply and demand-side efficiency improvement
and promote non-economic consumption of petroleum products (Sarkar and Singh, 2010). Thus,
the government could use the subsidies to fund well-targeted pro-poor programmes that would
directly benefit the vulnerable in society.
Following from the foregoing arguments, this paper addresses three fundamental questions:
1. Are the poor the real beneficiaries of the price subsidies in Ghana?
2. What is the magnitude of distribution of the impact on the various income groups in Ghana?
3. Is the impact of fuel price regressive or progressive in Ghana?
Addressing the above questions is crucial because, in Ghana, anytime fuel prices are adjusted
upwards, there are agitations for government to cushion the poor by either reducing taxes on
petroleum products or by absorbing the price differential by introducing subsidies. Studies on the
removal of FFS subsidies on households in Ghana have not adequately addressed the issue of the
impact of FFS. A study that comes close to this study is the one conducted by Coady et al. (2006),
which investigated the social and fiscal cost of fuel subsidies. They evaluated the magnitude and
distribution of fuel subsidies and how they impact on poor households in Bolivia, Ghana, Jordan,
Mali and Sri Lanka. They carried out budget share analysis using household fuel expenditure
obtained from household expenditure survey of the afore mentioned countries. The study
established that the direct effect of increase in fuel prices ranges from 0.9 percent for Mali to 2
percent for Jordan. However, the direct effect is either distributionally neutral (Bolivia and Mali)
or regressive (Ghana, Jordan and Sri Lanka). This study differs from the previous study because it
uses more recent data on a household survey (Ghana Living Standards Survey rounds six and seven
(GLSS6&7)) to carry out the before- and after- analysis of subsidy removal. Secondly, the study
carries out micro analysis to determine the impact of FFS removal on households in urban and
rural areas as well as the regional level.

The remainder of this paper is organised as follows: Section 2 presents a review of relevant strands
of studies on the subject matter; Section 3 provides the methodology and data employed; Section
4 entails the presentation and discussion of results and Section 5 concludes the study with the
provision of policy recommendations.

3|Page
2.0 Literature Review
2.1 Pricing of petroleum products in Ghana
The pricing of petroleum products in Ghana is guided by three main policy objectives, namely: (i)
full cost recovery of investment based on import parity benchmarking; (ii) revenue generation by
the imposition of taxes and levies on the petroleum products; and (iii) uniform prices nationwide.
The pricing at full cost recovery is meant to assist the players in the industry recoup their cost of
investment. There are two pricing windows and prices are reviewed every two-weeks (first and
sixteenth of every month). Pump prices are computed based on the free-on-board (FOB) prices,
supplier’s premium, taxes, levies and margins. The FOB is the price of petroleum products on the
international market and the supplier’s premium is expected to cater for all costs associated with
importing the petroleum products into the country such as jetty fees, port charges, demurrage cost,
in-plant losses, financing cost, etc. The margins include unified petroleum price fund margin,
primary distribution margin, petroleum products marking scheme margin, LPG compensation and
distribution margin, marketer’s margin, dealer’ margin, etc.

Ghana started the implementation of petroleum price deregulation in July 2015. Prior to the price
deregulation, prices of petroleum products were set by the National Petroleum Authority (NPA).
The introduction of the price deregulation shifted the responsibility of price setting to the Bulk
Distribution Companies (BDCs) and Oil Marketing Companies (OMCs). The BDCs are
responsible for setting-up ex-refinery prices and the OMCs are responsible for setting-up ex-pump
prices. The NPA monitors to ensure that prices conform to the petroleum pricing formula.

2.2 Impact of fossil fuel subsidy reforms

Empirical studies on FFS reforms tend to focus on three key issues such as environmental,
economic and social effects.

2.2.1 Environmental and economic impact of FFS subsidy reforms

Research has identified positive effects of FFS removal on the environment and economic growth.
For example, IEA (1999) examined the impact of subsidy reforms in OECD and non-OECD
countries and established that removal of FFS increased annual GDP by 0.73 percent and further
decreased CO2 emissions by 16 percent due to a decrease in energy use by 13 percent. Similarly,
Saunders and Schneider (2000) investigated the effect of FFS on OECD and non-OECD countries
for the period 1995 to 2010. They found that removal of FFS resulted in increase of GDP by 0.45
percent for non-OECD countries in 2010 and 0.1 percent for OECD countries. Additionally, they
established that in countries where fossil fuel prices increased, GHG emissions fell, but they were
offset by the rise in emission in other countries. Burniaux et al. (2009) examined the impact of the
gradual removal of unilateral and multilateral energy subsidies of OECD and non-OECD countries
using price-gap data from the IEA. They found that multilateral energy subsidy removal resulted
in an increase in GDP and real income of non-OECD countries such as India, China and Brazil,
but declined in some countries, including Russia, the oil-producing countries and the non-EU
4|Page
Eastern European countries. Regarding the impact of fossil subsidy removal on the environment,
they found a reduction in world CO2 emissions by 3.9 percent in 2020 and 13 percent in 2050.
However, GHG emissions were reduced by 3.1 percent in 2020 and are expected to reduce by 10.2
percent in 2050.

2.2.2 Social effects of FFS subsidy reform

Other studies have looked at the social effect of removal of FFS with a particular emphasis on
incidence of FFS removal on rich and poor households. Jensen and Tarr (2002) investigated the
effect of FFS reforms on the poor in Iran and established that subsidy reform would benefit poor
rural households more than poor urban households with gains of 200 percent to rural poor
compared to 100 percent in an urban household if the money was directly transferred to the poor
in equal amounts. They concluded that welfare impact was high even without cash transfers. The
findings were confirmed by a study conducted by Saboohi (2000) on subsidy reforms in Iran and
found that increase in energy prices increased rural household cost of living by 33.7% compared
to 28.7% of urban households. Similarly, Andriamihaja and Vecchi (2007) investigated the impact
of the increase in petroleum prices on living standards in Madagascar and established that the
incidence of increase in price was higher for low-income households (2.1%) than high income
households (1.5%). They concluded that the subsidies benefited the rich more than the poor. These
findings were confirmed by a similar study carried out in Mali by Kpodar (2006). He established
that increasing prices of gasoil and gasoline affected the rich households; however, increasing
prices of kerosene affected the poor households. Rao (2012) confirmed the findings of the study
carried out in Mali when he conducted a similar work in India and concluded that subsidies on
kerosene were regressive since they did not provide any financial benefit to the poor. He
recommended that subsidies should be replaced with government transfers. In contrast, Kebede
(2006) established that subsidies on kerosene prices did not significantly change the overall cost
of living of households.

3.0 Methodology

3.1 Model Specification

Following Coady et al. (2006), the study specified petroleum products price subsidy as the
difference between the full pass-through price and the actual price of petroleum products sold at
the pump. The petroleum products price subsidy was specified as:

𝑃𝑆 = 𝐹𝑃 − 𝐴𝑃 (1)

where 𝑃𝑆 is price subsidy, 𝐹𝑃 is full pass-through price and 𝐴𝑃 is actual pump price.

The welfare impact of petroleum products subsidy on the households could be direct or indirect.
The direct impact on the households is because of higher prices for fuels consumed for cooking
and transportation. The indirect impact is through higher prices for other goods consumed by the

5|Page
households is because of higher fuel costs. The study focused on the direct welfare impact of
petroleum products subsidy because of insufficient data to estimate the indirect welfare impact of
petroleum products subsidy.

Household budget share is specified as:

𝐻𝐹𝐸
(2)
𝐻𝑇𝐸

Direct effect of fuel price increase is specified as:

𝐻𝐹𝐸
𝐷𝐸 = × 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 𝑖𝑛𝑐𝑟𝑒𝑎𝑠𝑒 𝑖𝑛 𝑓𝑢𝑒𝑙 𝑝𝑟𝑖𝑐𝑒 (3)
𝐻𝑇𝐸

where 𝐷𝐸 is direct effect of fuel price increase, 𝐻𝐹𝐸 is the household’s expenditure on petroleum
products and public transport, 𝐻𝑇𝐸 is the household’s total expenditure.

The magnitude of the impact of fuel price increase depends on the importance of cooking and
private transport cost in the total household consumption, as well as on the fuel-intensity of other
goods and services. The distribution of the impacts across different income groups also depends
on the relative importance of these factors across income groups. For instance, if the consumption
baskets of higher-income groups are relatively more fuel-intensive than for lower-income groups,
then the impact on the former will be greater than on the latter.

3.2 Data

The study used Ghana Living Standards Survey rounds six and seven (GLSS6&7) data from the
Ghana Statistical Service (GSS). The GLSS data is a nationwide household survey designed to
generate information on living conditions of households in the country. The GLSS6 survey was
conducted from 18th October 2012 to 17th October 2013 and that of GLSS7 was conducted from
3rd October 2016 to 3rd October 2017 (GSS 2019). The survey captured information on the
demographic characteristics of households, education, health, employment, migration, tourism,
housing conditions, household agriculture, household expenditure, etc.

This study used data on the expenditure of households in the country to estimate the welfare impact
of the fuel price increase and benefits of fuel subsidies. Data on total household expenditure
(excluding rent) and household expenditure on LPG, kerosene, petrol/diesel and transport were
used to estimate the budget share of the petroleum products. The mean household expenditure for
each quintile is shown in Table 1. Data on fuel price subsidies for 2014 were computed from data
on full pass-through price and actual pump prices obtained from NPA. This period was selected
because of availability of data.

3.3 Estimation Technique

6|Page
The study carried out budget share analysis to estimate the welfare impact of petroleum products
price increase and subsidy. The budget share was estimated for petroleum products such as LPG,
kerosene, transportation fuel (petrol/diesel) and public transport (cost of fares) for both rural and
urban households. The budget shares were multiplied by the percentage price increase to estimate
the impact of the fuel price increase. The estimation of the magnitude of impact across the different
groups was carried out by grouping the households into various consumption quintiles2 and
estimated the direct welfare impact of fuel price increase on the various quintile groups. To
analyze the impact of fuel price increases on the welfare of households, we carried out various
scenarios using actual fuel price adjustments.

Table 1: Mean annual household expenditure

Mean annual household Mean annual household
Quintile group expenditure (GHS) GLSS VI expenditure (GHS) GLSS VII
First (bottom 20%) 3,924 4,114
Second 5,833 7,538
Third 7,444 9,849
Fourth 9,238 11,837
Fifth (top 20%) 14,665 17,813
Source: Authors own computation using GLSS VI &VII data

4.0 Discussion of Results

4.1 Results based on GLSS VI data

This section presents the result and discussion of budget share of household’s fuel consumption
as well as impact of fuel price increases and direct benefit of petroleum products subsidies using
GLSSVI data.

4.1.1 Budget shares of household’s fuel consumption at the national level

Table 2 presents the results of household’s budget shares on LPG, transportation fuel
(diesel/petrol), kerosene and public transport (fares). The results indicate that the households spent
0.2% of their total expenditure on LPG for cooking and 1.7% on transportation fuel (diesel/petrol)
per year. Regarding kerosene, households spent 0.1% on the product per year. On the other hand,
households spent 3.3% of their annual expenditure on public transport.

Table 2: Household’s budget share of petroleum products and public transport

Products Budget Share
LPG 0.002

2
Divided the total households into five according to ascending consumption level

7|Page
 Transport (diesel/petrol) 0.017
Kerosene 0.001
Public transport 0.033
Source: Authors own computation using GLSS VI

4.1.2 Direct impact of fuel prices increase by consumption quintile

The results of the analysis of the distribution of the impact of fuel price increase across different
income groups are presented in Table 3. The impact of fuel price increase on the various income
groups depends on the relative importance of the petroleum products across income groups.

Table 3: Budget share across different income groups

Product Bottom 20% Second Third Fourth Top 20%
LPG 0.0001 0.0007 0.0011 0.0027 0.0035
Transport fuel
(petrol/diesel) 0.0090 0.0091 0.0091 0.0109 0.0278
Kerosene 0.0012 0.0014 0.0013 0.0010 0.0006
Public transport 0.0168 0.0254 0.0318 0.0350 0.0383
National Total 0.0272 0.0366 0.0433 0.0496 0.0701
Source: Authors own computation using GLSS VI

The results revealed that households in the lower-income group allocated a smaller consumption
budget to LPG, petrol/diesel and public transport. The higher-income group allocated a higher
consumption budget to LPG, petrol/diesel and public transport. While the lower-income group
allocated 0.01% of their income to LPG, the higher-income group allocated 0.35% of their income
to LPG. Concerning transportation fuel, the lower-income group allocated 0.9% of their income
to diesel/petrol while the higher-income group allocated 2.8% of their income to diesel/petrol.
Similarly, the percentage of income spent by the higher-income group is almost twice that of the
lower-income group. The lower-income group was found to spend higher (0.12%) on kerosene
than the higher-income group (0.06%).

The above results indicate that the impact of the increase in fuel prices was progressive across the
different income groups except for kerosene which was regressive (see Table 4). For instance, the
impact of a 10% rise in fuel prices is 2.5 times higher on the top 20% income group than the bottom
income group.

This is expected because the budget share of the households in the higher-income group is higher
than that of those in the lower-income group. Also, it is expected that those in the lower-income
group will spend more on kerosene compared to those in the higher income group. These findings

8|Page
are contrary to the findings by Coady el al. (2006) and Andriamihaja and Vecchi (2007) who
established that the impact of fuel price increase was higher on the poor than the rich.

Table 4: Direct Impact of Fuel Price Increase across different income groups
Product Bottom 20% Second Third Fourth Top 20%
LPG 0.0001 0.0008 0.0012 0.0029 0.0038
Transport fuel
(petrol/diesel) 0.0099 0.0100 0.0100 0.0120 0.0305
Kerosene 0.0014 0.0015 0.0014 0.0011 0.0006

Public transport 0.0185 0.0279 0.0350 0.0385 0.0421

National Total 0.0299 0.0402 0.0476 0.0546 0.0771
Source: Authors own computation using GLSS VI

4.1.3 Distribution of subsidy benefit by consumption quintile

Table 5 presents the results of the share of the total benefits from subsidized fuel prices of each
income group. The estimation of the total subsidy benefits was carried out using the actual price
subsidy of the petroleum products for 2014. In 2014, the total price subsidies on LPG, petrol/diesel
(combined), kerosene were GHS0.90, GHS4.42 and GHS2.99 respectively.

Table 5: Distribution of subsidy benefits by consumption quintile (%)

Product Bottom 20% Second Third Fourth Top 20%
LPG 0.9 6.9 11.1 26.5 34.4
Transport fuel
(petrol/diesel) 439.1 443.7 440.8 532.1 1,349.8
Kerosene 40.9 45.9 43.3 31.7 18.2
National Total 480.9 496.4 495.2 590.3 1,402.4
National
(Average) 160.3 165.5 165.1 196.8 467.5
Authors own computation using GLSS VI

The results reveal that, on average, the top income group received about three (3) times more of
the subsidies than the bottom income group. Likewise, the top income group received about three
(3) times more of the subsidies than the bottom 40% of the income group. The situation is even
worse at the individual product level. For instance, almost all subsidies on LPG went to the top
income group with the lowest income group receiving just 1%. The top income group received
more than thirty (30) times the subsidies received by the lower income group. Similarly, the top
income group received about three (3) times the subsidies on petrol/diesel than the lower-income

9|Page
group. However, regarding kerosene, the bottom income group received more subsidies than the
top income group.

The above findings lend support to the assertion that there is a spillover of fuel subsidies to the
rich. Thus, the rich benefit more than the poor when there are fuel subsidies. The analysis indicates
that there is substantial leakage of subsidy benefits to the higher-income group which defeats the
rationale for petroleum products subsidy.

The findings support the findings of studies conducted by Kpodar (2006), Vecchi and
Andriamihaja (2007) and Rao (2012). For example, Kpodar’s (2006) study in Mali established that
increasing the price of petrol and diesel affected the rich more than the poor. Therefore, the rich
would benefit more than the poor when fuel was subsidized. Also, Vecchi and Andriamihaja and
(2007) found that, in Madagascar, increase in prices of petroleum product was higher on the poor
than the rich and concluded that subsidies benefited the rich more than the poor.

4.1.4 Budget share of rural and urban households

Table 6 presents the budget share of rural and urban households on LPG, Kerosene, transportation
fuel (diesel/petrol) and public transport. The results reveal that the budget share of rural households
on all the products is lower than that of households in the urban areas except kerosene. The higher
budget share on kerosene recorded in the rural areas is because the households in rural areas
consume more kerosene than those in urban areas.

Table 6: Budget share of households in the rural and urban areas

Transport Public
Location LPG (diesel/petrol) Kerosene transport Total Average
Rural 0.0007 0.0144 0.0013 0.0270 0.0434 0.0108
Urban 0.0036 0.0196 0.0006 0.0380 0.0618 0.0154
Source: Authors own computation using GLSS VI

The budget shares of rural households on LPG, kerosene, petrol/diesel and transport were 0.0007,
0.0013, 0.0144 and 0.0270 respectively, suggesting that households in the rural areas allocated
0.07% of their total consumption budget to LPG, 0.13% to kerosene, 1.44% of petrol/diesel and
2.70% of public transport.

On the other hand, the budget shares of households in urban areas on LPG, petrol/diesel and public
transport were higher than those of households in rural areas except for kerosene. For example,
households in the urban areas allocated 0.4% of their consumption budget to LPG compared to
0.07% of the households in rural areas. Similarly, households in the urban areas allocated 1.96%
more of their consumption budget to petrol/diesel than households in rural areas. Regarding public
transport, urban households allocated 3.8% of their consumption budget to public transport.

10 | P a g e
Overall, households in rural areas in Ghana spent less consumption budget (1.1%) on petroleum
products and public transport compared to their urban counterparts. This suggests that the impact
of fuel price increase is higher among the urban dwellers than the rural dwellers because the rural
household’s consumption basket is less fuel intensive.

4.2 Results based on GLSS VII

This section discusses the result of the budget share analysis of household’s fuel consumption as
well as impact of fuel price increases and direct benefit of petroleum products subsidies using
GLSSVII data4.2.1 Budget shares of a household’s fuel consumption at the National Level

Table 7 exhibits the results of the household’s budget shares on LPG, transportation fuel
(diesel/petrol), kerosene and public transport at the national level. The results show that households
spent 0.4% of their total consumption budget on LPG for cooking and 1.7% of transportation fuel
(diesel/petrol) per year. On the other hand, households spent 0.02% of their consumption budget
on kerosene per year. Households spent 4.8% of their annual consumption budget on public
transport.

Table 7: Households budget shares on petroleum products/public transport (GLSSVII)

Products Budget Shares
LPG 0.0035
Transport (diesel/petrol) 0.0169
Kerosene 0.0002

Public transport 0.0480

Source: Authors own computation using GLSS VII

The results, based on GLSSVII, indicate that the household’s consumption budget was higher for
all the petroleum products and public transport compared to GLSSVI data. This could be attributed
to an increase in fuel price because of the implementation of the price deregulation policy
introduced in July 2015.

Further analysis of budget shares based on rural and urban households reveals that households in
rural areas allocated less consumption budget to LPG compared to households in urban areas (see
Table 8). Additionally, households in rural areas spent less of their consumption budget on
transportation fuel (diesel/petrol) and public transport compared to urban areas. The only product
that households in rural areas spent on more than their counterparts in urban areas is kerosene.

Table 8: Budget share of rural and urban households (GLSSVII)

Products Rural Urban
LPG 0.0015 0.0050
Transport (diesel/petrol) 0.0145 0.0187

11 | P a g e
 Kerosene 0.0002 0.0001
Public transport 0.0389 0.0549
National Total 0.0551 0.0787
National Average 0.0138 0.0197
Source: Authors own computation using GLSS VII

4.2.2 Direct impact of fuel price increase by consumption quintile

The results of the incidence of the impact of fuel price increases across different income groups
are presented in Table 9. The results reveal that the households in the lower-income group allocated
a smaller consumption budget to LPG, petrol/diesel and public transport. The higher-income group
allocated a higher consumption budget to LPG, petrol/diesel and public transport. While the lower-
income group (bottom 20%) allocated 0.01% of their income to LPG, the higher-income group
allocated 0.55% of their income to LPG. Concerning transportation fuel, the lower-income group
allocated 0.1% of their income on diesel/petrol while the higher-income group allocated 2.8% of
their income to diesel/petrol. Similarly, the percentage of income spent by the higher-income group
on public transport is almost twice that of the lower-income group. The lower-income group was
found to spend higher (0.02%) on kerosene than the higher-income group (0.01%).

Table 9: Budget shares of various household’s consumption quintile (GLSSVII)

Bottom
Products 20% Second Third Fourth Top 20%
LPG 0.00007 0.0008 0.0022 0.0039 0.0055
Transport fuel
(petrol/diesel) 0.0113 0.0100 0.0083 0.0092 0.0277
Kerosene 0.0002 0.0002 0.0002 0.0002 0.0001
Public transport 0.0258 0.0312 0.0388 0.0502 0.0607
National Total 0.03739 0.0422 0.0495 0.0635 0.0939
National Average 0.00935 0.01055 0.01237 0.01587 0.02348
Source: Authors own computation using GLSS VII

Further analysis was carried to ascertain the impact of the recent increase in prices of petrol, diesel
and LPG due to the imposition of energy recovery levy, and sanitation and pollution levy. The
government imposed an energy sector recovery levy of GH₵ 0.20p on petrol and diesel and GH₵
0.18p on LPG. Also, a sanitation and pollution levy of GH₵ 0.10p was imposed on petrol and
diesel. This translated into an increase of about 10% for petrol, diesel and LPG. The results of the
impact of recent fuel price increases on the household’s consumption budget are presented in Table
10. The results reveal that the impact is marginal on the various consumption quintiles of
households which consume LPG. However, the impact is higher for the higher-income group than
the lower-income group since the higher-income group allocated a higher consumption budget to
petroleum products.

12 | P a g e
Table 10: Impact of fuel price increase on household’s consumption quintile based on GLSSVII
(% of total consumption)
Product Bottom 20% Second Third Fourth Top 20%
LPG 0.01 0.09 0.24 0.43 0.60
Transport fuel
(petrol/diesel) 1.24 1.09 0.91 1.01 3.05
National Total 1.25 1.18 1.15 1.44 3.65
Source: Authors own computation using GLSS VII

4.2.3 Regional household budget share

Tables 11 and 12 show the results of the regional household budget shares based on GLSSVI&VII
data. The results based on GLSSVI indicated that households in the Greater Accra region had the
highest budget shares on LPG followed by Central and Western regions. This could be attributed
to the fact that in the Greater Accra region, most households use LPG as cooking fuel, and,
therefore, their budget share on LPG tends to be higher. Even though the consumption of LPG in
Central and Western regions in 2013 was low compared to Ashanti region, the budget shares of
households in the former regions were higher. This is due to the fact that the total household
consumption budgets of Western (GHS16 million) and Central (GHS12 million) are small, thus,
resulting in higher budget shares. Regarding transport fuels, households in the Northern region had
the highest budget shares (0.0292) followed by Upper West (0.0291) and Greater Accra (0.0224)
respectively. This could be attributed to the fact that the total consumption budgets of households
in Northern and Upper West regions are low, thus, resulting in higher budget shares. Households
in the Greater Accra region had the highest budget shares (0.0548) on public transport followed
by Eastern (0.0359) and Western (0.0350) respectively.

Table 11: Regional Households Budget Shares (based on GLSS6 data)

Transport
Region LPG (diesel/petrol) Kerosene Public transport
Western 0.0029 0.0151 0.0003 0.0350
Central 0.0031 0.0098 0.0019 0.0346
Greater Accra 0.0041 0.0224 0.0010 0.0548
Volta 0.0018 0.0151 0.0026 0.0310
Eastern 0.0020 0.0066 0.0013 0.0359
Ashanti 0.0028 0.0155 0.0002 0.0342

13 | P a g e
 Brong Ahafo 0.0011 0.0154 0.0001 0.0290
Northern 0.0003 0.0292 0.0011 0.0171
Upper East 0.0008 0.0169 0.0008 0.0118
Upper West 0.0008 0.0291 0.0002 0.0105
Source: Authors own computation using GLSS VI

Table 12: Regional Households Budget Shares (based on GLSS7 data)

Transport
Region LPG (diesel/petrol) Kerosene Public transport
Western 0.0035 0.0051 0.0001 0.0383
Central 0.0035 0.0030 0.0002 0.0451
Greater Accra 0.0062 0.0257 0.0001 0.0825
Volta 0.0044 0.0113 0.0010 0.0470
Eastern 0.0031 0.0048 0.0002 0.0429
Ashanti 0.0038 0.0168 0.0000 0.0555
Brong Ahafo 0.0020 0.0161 0.0000 0.0366
Northern 0.0007 0.0371 0.0000 0.0239
Upper East 0.0019 0.0142 0.0000 0.0166
Upper West 0.0012 0.0404 0.0000 0.0342
Source: Authors own computation using GLSS VII

The results reveal that an increase in the price of LPG tends to affect households in the Greater
Accra more than households in the other regions. Similarly, an increase in the price of diesel/petrol
affects households in the Northern and Upper West regions more than the rest of the regions.
However, increasing prices of kerosene affects households in the Volta and Central regions more
than any other region. Households in the three Northern regions are the least affected whenever
there is an increase in LPG prices because the households in these regions spend less consumption
budget on LPG. Households in Central and Eastern regions spent less consumption on petrol and
diesel compared to the other regions. Regarding public transport, households in the Upper East
and Upper West spent less consumption budget on public transport than any other region.

The results based on GLSSVII indicate that the three Northern regions had the lowest budget
shares on LPG. However, they had the highest budget shares on petrol and diesel compared to
most of the regions. The lowest budget share on LPG is due to the fact the three Northern regions
have lower LPG consumption. Additionally, because the three regions have lower total
consumption budgets, their budget shares on petrol and diesel tend to be higher than other regions.

14 | P a g e
unlike the GLSSVI results, the budget share on LPG is higher for Greater Accra followed by Volta
and Ashanti regions respectively. The results obtained from GLSS6 data and GLSS7 data in terms
of the trend of household’s budget shares on petroleum products and public transport are not
different.

The comparison of regional total budget shares on petroleum products and public transport for
both GLSS6 and GLSS7 is presented in Figure 1. All the regions recorded higher budget shares in
2017 except the Western region. The increased budget shares of households in 2017 could be
attributed to the fuel price increase due to the implementation of the price deregulation.

14.00

12.00

10.00
Budget Shares

8.00

6.00

4.00

2.00

0.00
Greater Eastern Western Central Ashanti Volta Brong Northern Upper Upper
Accra Ahafo East West
Regions

Budget Shares2013 Budget Shares2017

Figure 1: Comparison of total regional budget shares on petroleum products and public transport,
2013 and 2017 (Source: Authors own computation using GLSS VII)

5.0 Conclusion and Recommendation

5.1 Conclusion

This paper investigates the direct impact of petroleum price increases on the welfare of consumers
and benefits of petroleum products subsidy using the information on household’s expenditure data
obtained from Ghana Living Standards Survey rounds six and seven (GLSS6&7).

The study has established that households in the rural areas are less affected whenever there is an
increase in the price of petroleum products because they allocate a smaller amount of their
consumption budget on petroleum products per year. For instance, on average, the households in
rural allocated 1.1% of their total consumption budget to petroleum products. This increases to
1.4% in the case of GLSSVII data. However, households in urban areas are mostly affected
whenever the prices of petroleum products go up. This is as a result of the fact that households in
15 | P a g e
urban areas allocated higher consumption budgets to petroleum products. The households in urban
areas allocated 1.5% of their consumption budgets to petroleum products. This increases to 2% in
the case of GLSSVII data.

Further analysis based on the various consumption quintiles (income groups) revealed that the
lower-income group allocated a small fraction of their income to petroleum products such as LPG
and diesel/petrol. The results indicated that the impact of higher fuel price increases was
progressive across the different income groups except for kerosene. This is because the budget
share of the households in the higher-income group on fuel products is higher than those in the
lower-income group.

Regarding the distribution of subsidy benefits by consumption quintile, the study found that the
top-income group received about three (3) times more of the subsidies than the bottom-income
group suggesting that FFS is regressive. This suggests that identifying cost-effective social
intervention programs to protect the poor will be more appropriate than subsidizing petroleum
products. These findings justify the complete removal of subsidies on prices of petroleum products
since there is a substantial spillover of benefits of petroleum product subsidies to the rich.

5.2 Recommendation

We recommend that government should implement targeted policies that will directly cushion the
poor instead of subsidizing price of petroleum products since subsidy on petroleum products is
regressive. Thus, the government could provide a direct cash transfer to cushion the poor through
the Livelihood Empower Against Poverty (LEAP) program to absorb the impact associated with
fuel prices increase to avoid leakages of subsidies to the rich. This could be done by using part of
the funds realized from the price stabilization and recovery levy (PSRL) as well as funds that could
have been used to subsidize the product to support the poor through the LEAP programme.

Additionally, government could cushion poor households by introducing mass transportation to

lessen the cost of transportation. While the budget share of bottom 40% of household’s
consumption quintile on transport is 0.057, that of the top 20% is 0.061, which is almost the same.
Therefore, provision of mass transport will help ameliorate the effect of fuel price increase on the
poor.

Acknowledgements

We thank Prof. Kwabena Sarfo Sarfo-Kantankah, Department of English, University of Cape

Coast, and Dr. Kwaku Ohene-Asare, University of Ghana Business School for proofreading the
paper for us.

16 | P a g e
References

1. Burniaux, J. M., Chateau, J., Dellink, R., Duval, R., & Jamet, S. (2009). The economics of
climate change mitigation: How to build the necessary global action in a cost-effective manner.

2. Coady, D. P., El Said, M., Gillingham, R., Kpodar, K., Medas, P. A., & Newhouse, D. L.
(2006). The magnitude and distribution of fuel subsidies: evidence from Bolivia, Ghana,
Jordan, Mali, and Sri Lanka.

3. Freund, C. L., & Wallich, C. (1995). Raising household energy prices in Poland: who gains?
who loses? (No. 1495). World Bank Publications.

4. Ghana Statistical Service (GSS) (2019). Ghana Living Standard Survey Round 6 Report.
Accra. www.statsghana.gov.gh/gsspublications.

5. International Energy Agency (IEA). (1999). World Energy Outlook 1999: Looking at Energy
Subsidies –Getting the Prices Right. International Energy Agency: Paris.

6. Jensen, J. B. (2002). Trade, Foreign exchange, and energy policies in the Islamic Republic of
Iran (No. 2768). World Bank Publications.

7. Koplow, D., & Dernbach, J. (2001). Federal fossil fuel subsidies and greenhouse gas emissions:
a case study of increasing transparency for fiscal policy. Annual review of Energy and the
Environment, 26.

8. Kpodar, K. (2006). Distributional effects of oil price changes on household expenditures:

evidence from Mali.
9. National Petroleum Authority (2015). Historical Prices of Petroleum Products in Ghana.
www.npa.gov.gh/download-media/industry-data/downloads.

10. Nwachukwu, M. U., & Chike, H. (2011). Fuel subsidy in Nigeria: Fact or
fallacy. Energy, 36(5), 2796-2801.

11. Rao, N. D. (2012). Kerosene subsidies in India: When energy policy fails as social policy.
Energy for sustainable development, 16(1), 35-43.

17 | P a g e
12. Saboohi, Y. (2001). An evaluation of the impact of reducing energy subsidies on living
expenses of households. Energy Policy, 29(3), 245-252.

13. Sarkar, A., & Singh, J. (2010). Financing energy efficiency in developing countries—lessons
learned and remaining challenges. Energy Policy, 38(10), 5560-5571.

14. Saunders, M., & Schneider, K. (2000). Removing energy subsidies in developing and transition
economies (p. 14). ABARE.

15. Sharma, S. V. (2013). Implications of Cash Transfers of Subsidies in the Energy Sector in
India. ENERGY MARKET INTEGRATION IN EAST ASIA: RENEWABLE ENERGY AND ITS
DEPLOYMENT INTO THE POWER SYSTEM, 267.

16. Vecchi, G., & Andriamihaja, N. A. (2007). An evaluation of the welfare impact of higher
energy prices in Madagascar.

18 | P a g e