HOW DO PETROL PRICES RESPOND TO VARIATIONS IN CRUDE OIL AND THE EXCHANGE RATE? EVIDENCE FROM SOUTH AFRICA

Naape B.; Mathebula N. C.

DOI: 10.29141/2658-5081-2022-23-3-2

EDN:XDHURZ

JEL classification: B22, B27, C13

Baneng Naape Ndzalama C. Mathebula

University of the Witwatersrand, Johannesburg, South Africa University of Johannesburg, Johannesburg, South Africa

How do petrol prices respond to variations in crude oil and the exchange rate? Evidence from South Africa

Abstract. Energy resources are a significant production input due to being at the bottom of supply chains. The study investigates the response of petrol prices to variations in the domestic and international components. Methodologically, the research rests on the neoclassical economics. Monthly time series data of the South African Department of Energy and Mineral Resources spanning from 2002 to 2021 is analysed by means of econometric modelling, including unit root analysis, structural vector autoregression, impulse response and variance decomposition. The impulse response function indicates that basic petrol prices respond positively to their own shocks and to shocks in Brent crude oil prices and this response is substantial in size. On the contrary, basic petrol prices respond negatively to shocks in exchange rates, albeit the response is small in size. The variance decomposition reveals that variations in basic petrol prices are largely explained by their own shocks in the short run and by shocks in Brent crude oil prices in the long run. Domestically, petrol pump prices respond positively to shocks in basic petrol prices, inland transport costs, wholesale and retail margins while their response is negative towards shocks in the Road Accident Fund levy and muted towards shocks in fuel taxes. Given the rise in global crude oil prices due to geopolitical woes and their considerable share in the overall domestic petrol pump price as shown in the study, the government should consider implementing price-based policies such as indirect subsidies through the reduction of taxes and levies on petroleum products, and targeted income subsidies to provide the much-needed financial relief to households and businesses. In addition, it can focus on reducing the reliance on oil through energy efficiency improvement and diversification into non-petroleum sources of energy.

Keywords: petrol prices; exchange rates; crude oil; pricing; South Africa.

For citation: Naape B., Mathebula N. C. (2022). How do petrol prices respond to variations in crude oil and the exchange rate? Evidence from South Africa. Journal of New Economy, vol. 23, no. 3, pp. 23-42. DOI: 10.29141/2658-5081-2022-23-3-2. EDN: XDHURZ. Article info: received April 11, 2022; received in revised form May 19, 2022; accepted June 20,

2022

Баненг Наапе Университет Витватерсранда, Йоханнесбург, ЮАР

Ндзалама К. Матебула Университет Йоханнесбурга, Йоханнесбург, ЮАР

Влияние обменного курса и стоимости нефти на цену бензина в Южной Африке

Аннотация. Энергоресурсы играют решающую роль в производстве и реализации товаров и услуг. Исследование направлено на изучение изменения цен на бензин под влиянием внешних и внутренних факторов. Методология работы опирается на неоклассическую экономическую теорию. Методы включали расчет единичного корня, структурную векторную авторегрессию, импульсные отклики и декомпозицию дисперсии. Информационную базу составили ежемесячные данные временных рядов за период 2002-2021 гг., размещенные на сайте Министерства энергетики и минеральных ресурсов Южной Африки. Функция импульсного отклика показывает, что базовые цены на бензин положительно и значимо реагируют на собственные шоки и шоки цен на сырую нефть марки Brent и демонстрируют негативную, хотя и не очень значимую реакцию на шоки обменных курсов. Декомпозиция дисперсии свидетельствует о том, что колебания указанных цен во многом объясняются их собственными шоками в краткосрочный период и шо-ками цен на сырую нефть марки Brent в долгосрочной перспективе. Розничные цены на бензин положительно коррелируют со скачками базовых цен на бензин, издержками на внутреннюю транспортировку, оптовыми и розничными наценками, в то время как их ответная реакция на изменения размера сбора в Фонд дорожно-транспортных происшествий является отрицательной, а на шоки налогов на топливо - умеренной. С учетом роста мировых цен на сырую нефть из-за геополитических проблем и их значительной доли в общей цене на бензин внутри страны правительству следует рассмотреть возможность проведения ценовой политики. Например, можно ввести косвенные субсидии путем снижения налогов и сборов на нефтепродукты и целевые субсидии для оказания финансовой помощи домохозяйствам и предприятиям. Кроме того, целесообразно снизить зависимость экономики от нефти посредством повышения энергоэффективности и диверсификации за счет альтернативных источников энергии.

Ключевые слова: цена на бензин; обменный курс; сырая нефть; ценообразование; ЮАР.

Для цитирования: Naape B., Mathebula N. C. (2022). How do petrol prices respond to variations in crude oil and the exchange rate? Evidence from South Africa. Journal of New Economy, vol. 23, no. 3, pp. 23-42. DOI: 10.29141/2658-5081-2022-23-3-2. EDN: XDHURZ. Информация о статье: поступила 11 апреля 2022 г.; доработана 19 мая 2022 г.; одобрена 20 июня 2022 г.

Introduction

Globally, energy plays a crucial role in the production and facilitation of goods and services. The availability of electricity and fuel has long been recognised as an economic driver. Specifically, fuel has generated revenue and covered numerous government fiscal deficits while enabling mobility for motorists and businesses globally [Alvarez et al., 2011]. Each country's fuel pricing structure makes up the overall price of fuel. The long structure of fuel pricing in

South Africa has raised the overall price of fuel over time, which has negatively affected both businesses and households, including through higher food prices, increased public transport fares and electricity generation, to name a few [Ashley, Tsang, 2013]. In South Africa, the key variables that impact on the price of fuel are the exchange rate, fuel levies and taxes, refinery costs and the price of crude oil. In addition, there are other factors that influence the price of fuel indirectly. They include the inflation rate, interest rate, political stability and the level of economic development. These factors are often overlooked in the economic literature. The significance of this study lies in its relevance to the current economic status quo in South Africa. Fuel prices and levies have increased firmly in the last two decades regardless of the domestic economic climate. To some extent, this increased the cost of living in the country which prominently affects those in low-income and middle-income households. The fuel pricing structure in South Africa comprises several components such as the basic fuel price, fuel levies, transport costs and profit margins [Sun, Hong, Xu, 2013]. It is worth noting that the basic fuel price is largely determined by the international crude oil price and the USD to ZAR exchange rate. As such, an increase in the international oil price or depreciation of the ZAR exchange rate would translate into an increase in the basic fuel price while a decline in the international oil price or an appreciation of the exchange rate would translate into a decline in the basic fuel price [Lescaroux and Mignon, 2008]. Of concern however, Sabelo [2018] finds that decreases in international oil prices or an appreciation of the exchange rate do not always translate into decreases in the basic fuel price in South Africa.

In view of the above, the study aims to statistically examine the response rate of basic fuel prices to shocks in international oil prices and the exchange rate in South Africa. The study is going to contribute to the existing body of literature by providing a statistical analysis of the drivers of fuel prices in South Africa as well as the response of basic fuel prices to changes in the domestic and international markets. Also, to the best of the authors' knowledge, such a research has not been conducted in the South Africa, a gap which this study intends to fill. The study is particularly crucial and relevant in a time wherein fuel prices have reached record highs amid variations in the global oil price and the exchange rate. Also, higher fuel prices have devastating effects on households, businesses, and the economy. The next section provides a review of theoretical and empirical literature on the influence of international and domestic factors on the overall fuel price.

Literature review

The availability of fuel remains a fundamental factor in the supply chain of every economy [Brown, Yucel, 2002]. Moreover, the impact of its pricing and cost cannot be overlooked since they provoke a chain reaction in aspects outside the economic space. Fuel prices are one amongst numerous things affected by the pandemic and at the centre of causal relationships behind the economic growth. Thus, they are key in understanding the economic condition of households and businesses. In South Africa, the price of fuel is determined by numerous factors which are subject to economic shocks and political occurrences. The price of fuel is South Africa encompasses five elements: the general fuel levy which amounts to 31 %, Road Accident Fund levy (18 %), basic fuel price (freight and insurance costs, cargo dues, storage, and financing), wholesale and retail margins as well as distribution and transport costs1. The prices are based on 93 octane petrol (inland) and 95 octane petrol (coastal) pricing structure. The basic fuel price in particular, is largely determined by the exchange rate and the price of Brent crude oil. Apart from the above-mentioned elements, there are other macroeconomic factors that indirectly

1 Automobile Association. (2021). Five things movement must do to reduce fuel prices in South Africa. Press release. https://aa.co.za/five-things-government-must-do-to-reduce-fuel-prices-in-sa-aa.

influence the price of petrol such as the strength of monetary and fiscal policy. For example, changes in the interest rate and political instability affect the exchange rate and consequently, the price of petrol. In the same line, the government may shift the burden of increased expenditure and higher interest payments to taxpayers through increases in several tax rates including fuel levies. Figure 1 illustrates the composition of fuel prices in South Africa. The South African Department of Energy and Mineral Resources (hereinafter referred to as Department of Energy) provides a complete list of all the components that make up the fuel pricing structure and this includes international and domestic factors.

Petrol pump price

X

International factors

Basic petrol price

Brent crude oil

Exchange rate

Freight & insurance

X

Inland transport cost

I_

Domestic factors

X

Government levies

Fuel levy

Customs & excise

Road Accident Fund

X

Profit margins

Wholesale margins

Retail margins

X

Other levies

Pipeline levy

DSML

Slate levy

Equalization levy

Fig. 1. South African petrol pricing structure1 Рис. 1. Факторы ценообразования на рынке бензина в ЮАР

International factors pertain to the basic fuel price, which is largely determined by the Brent crude oil, the South African exchange rate to the US dollar as well as freight and insurance costs. According to the Department of Energy, freight cost refers to the transport costs of refined petroleum from the export refinery centres to the South African ports. The government has no control over the above-mentioned costs but has an influence on selected components such as the exchange rate. Domestic components include inland transport costs, government levies, customs and excise levies collected in agreement with the Southern African Customs Union. Road Accident Fund (RAF) is a government levy collected to compensate third party victims of road accidents. Other components include wholesale and retail margins which are set and fixed by the Department of Energy based on the actual costs borne by the service station operator while selling petrol and marketing petroleum products. Lastly, other levies comprise the pipeline levy, slate levy, equalisation levy and demand side management levy (DSML). It is worth noting that DSML only applies to 95 unleaded fuels and is aimed at curbing the demand for 95 octane petrol to avoid unnecessary octane waste given its dire economic consequences. The slate levy is imposed to balance the slate account provided the daily calculation of fuel is lower than the basic fuel price reflected in the fuel price structure at that time [Cheung, Thomson, 2004].

1 Source: own compilation based on South African Department of Energy and Mineral Resources. http://www. energy.gov.za/files/esources/petroleum/petroleum_pricestructure.html.

With the recent hike in fuel prices, it has been noted that their determinants can explain the price increase. This relates to the instable nature of these determinants with specific reference to the exchange rate and the price of Brent crude oil. The Automobile Association states that it is due to a weak rand, economic policy, and poor management of the country's fiscal affairs that motorists are paying such high fuel prices. Though the fuel prices are not controlled by the state, it is worth bearing in mind that it does have an influence on the exchange rate, fuel levies and fiscal policy [Metcalfe, Dollan, 2012], all of which directly and indirectly determine the price of fuel. Figure 2 illustrates the trends in Brent crude oil prices and the exchange rate as international factors behind the basic fuel price. Herein, the two factors display an inverse relationship. That is, when the rand is strong against the US dollar and the price of crude oil is low, the price of basic fuel becomes relatively cheaper whereas a weaker rand and higher crude oil prices would result in relatively higher basic fuel prices. Nonetheless, the price of crude oil averaged out at 62.1 US dollars per barrel in March 2007, before accelerating to an all-time record high of 132.7 US dollars per barrel in August 2008 amid the emergence of the 2008 global financial crisis and subsequent disruptions in global supply chains. As the 2008 global financial crisis intensified, the average price of crude oil fell sharply to 40 US dollars in December 2008. During this time, the rand was trading at 9.95 rands to the US dollar while the domestic price of basic fuel was 3.65 rands per litre. This evidences that the combination of a stronger rand and lower Brent crude oil price results in lower basic fuel prices.

0 0

ifSco-^LnvotvooovO'—i

OOOOOOO*—I»—I»—I»—I»—I»—I»—I»—I»—I»—ICNCN

ooooooooooooooooooo

(N(N(N(N(N(N(N(N(N(N(N(N(N(N(N(N(N(N(N

— Average Brent crude oil price — Exchange rate (USD to ZAR)

Fig. 2. Trends in average crude oil price and ZAR exchage rate, 2003-20211

Рис. 2. Динамика средних цен на нефть и обменного курса южноафриканского ранда, 2003-2021

It is worth noting also that a stronger rand has the potential to partially offset the effects of higher crude oil prices on the basic fuel price. For example, in March 2012, the average price of crude oil stood at 125.5 US dollars per barrel while the corresponding USD to ZAR exchange rate was 8.01. During this period, the basic fuel price amounted to 12 rands per litre although one could have expected it to be relatively higher as a result of higher crude oil prices. The Covid-19 pandemic likewise had devastating effects on the average price of crude oil and the strength of the rand against the US dollar. During March 2020, the price of Brent crude oil reached an all-time record low and stood at 18.3 US dollars per barrel while the rand depreciated against the US dollar to reach 15. However, this provided the much-needed financial relief

1 Source: own calculations using the data from South African Petroleum Industry Association. (2021). https:// www.sapia.org.za/Overview/Old-fuel-prices.

to households and businesses as the basic fuel price stood at 4.79 rands per litre. The total fuel pump price however, stood at 15.84 rands given the number of levies that feed into it. Recently, there has been a rising concern over the ominous increase in fuel levies that have been growing every year over the last decade [Cavalcanti, Jalles, 2013]. According to the Organisation Undoing Tax Abuse (OUTA), fuel levies and taxes make up almost 65 % of the fuel price (see Figure 3).1 In 2004, the fuel price was below 5 rands per litre, however, after this period fuel prices have been relatively high given the gradual increase in fuel levies and taxes. In 2021, fuel taxes, RAF levy and other levies constituted nearly 40 % of the total fuel pump price.

25

20

и

>ч

О) Он О

у 'С

Он

15

10

co^invor^oooNO^cNcO'stimvor^ooaNO^ OOOOOOOi—I 11 I—I I—Ii—Ii—Ii—Ii—II—I I—I {N <N ooooooooooooooooooo

OJ<4C4<N<N(4(4<N<N<4iN<NC4iN<N(4(4<N<4

Basic petrol price (95 octane) I Inland transport cost

I Fuel tax

Wholesale and retail margins

Road Accident Fund I Other levies

Fig. 3. Trends in petrol price breakdown, 2003-20212 Рис. 3. Структура цены на бензин, 2003-2021

Furthermore, Mabugu, Chitiga and Amusa [2009] echo that while increasing fuel levies and taxes the government has been overlooking the tax base for which the taxes are collected and a clear rationale for the increment of taxes and levies on fuel prices. These irrational hikes in fuel tax have been given policy significance since August, 2006 [Mabugu, Chitiga, Amusa, 2009]. To reiterate, Ncanywa and Mgwangqa [2018] report that the general growth of fuel taxes and levies brings about a negative economic relationship: the production costs of petroleum products increase thereby decreasing priority expenditure and pushing up unemployment over time. One rationale justifying excise fuel tax applies to the promotion of using public transport and lift clubs to mitigate the effects of air contamination and achieve cost saving strategies [Ncube, Shimeles, Verdier-Chouchene, 2012]. Yet, it turns out to be unsound, because the country has less taxpayers than grant recipients. Next, Newbery [2005] highlights that most fuel taxes are excise taxes that burden producers while the standard argument suggests that levies should be concentrated on the final consumption. Then, the most illogical justification of these taxes is to fill fiscal deficits where the tax system has shortfalls. Newbery [2005] further emphasises that to address them the country should overcome political unwillingness to hold an effective policy reform on fuel taxes.

1 Organisation Undoing Tax Abuse. (2021). Governments unsustainable fuel related levies and taxes. https:// www.outa.co.za/blog/newsroom-1/post/opinion-governments-unsustainable-fuel-related-levies-and-taxes-1037.

2 Source: own calculations using the data from South African Petroleum Industry Association. (2021). https:// www.sapia.org.za/Overview/Old-fuel-prices.

As fiscal policy has equal importance to the fuel prices, we considered a study by Mabugu, Ribichaud, Masonnave and Chitiga [2013]. It found that there is the relationship between the 2008 economic crisis and an unsustainable fiscal policy, though afterwards the country saw a significant fiscal consolidation. Burger and Calitz [2021] point to the same situation with the height of the pandemic. After the February 2020, government expenditure was adjusted and became larger than the estimated budget presented by former minister Tito Titus Mboweni. In real terms, state expenditure amounted to 145 billion rands vs the 36 billion rands planned leaving a difference of 109 billions of unintended state expenditure [Burger, Calitz, 2021]. Though in times of the national state of disaster the latter suffices, poor fiscal policy produces an unintended impact in the long run [Mabugu, Ribichaud, Masonnave, Chitiga, 2013]. According to Keynes economic theorisation, in times of economic difficulties the state should adjust its taxes and spending and plunge the finances back into the economy to achieve and maintain good fiscal and monetary policy [LeBlanc, Chinn, 2004]. Consequently, such an inverse relationship between deficits and surplus without any sign of fiscal consolidation especially with a shrinking tax base in the country yields a poor economic outlook [Edelstein, Lutz, 2009]. It can be argued that, though the 2008 fiscal woes were consolidated within a short period of time, at present a quick swift fiscal consolidation is rather unlikely, because the economic growth in preceding years was weak. To maintain revenue collection there has been a general increase in fuel taxes despite an economic contraction in the past decade.

In line with the Keynesian taxation theory developed in the 1940s, no industry should be left to function on a free market basis, but rather, the government should play an active role in regulating and influencing the market [Carruth, Hooker, Oswald, 1998]. The latter assumes that economic growth is achieved through an effective fiscal and monetary policy where excise taxes and levies should be collected on the basis that it does good to the employment in the state1. In addition, Keynes argued that fiscal policy should play a major role in times of economic crisis by introducing tax rebates [Keynes, 1936] while the state expenditure should be covered by the surplus collected during an economic boom [Odhiambo, 2010]. The logic suggests that when people pay lower taxes during an economic crisis, they have more money to spend and invest which, in turn, yields demand; accordingly, economic expansions and employment are known as a positive feedback loop [Lardic, Mignon, 2006]. However, the illustrations in the case of South Africa show the direct opposite. Fuel levies have been increasing steadily despite the unfavourable economic climate.

Of equal importance to the variations in fuel prices remains the exchange rate. Bala and Chin [2018] revealed that the vulnerability of the exchange rate is due to its intricate link to politics. This talks to factors of political stability, sovereign debt, trade terms, political upheavals, and elections. Due to the uncertain nature of the above factors this leaves the exchange rate at a highly vulnerable position of dependence. This way, researchers points to a specific time in the political space of South Africa where the administrative capacity of the government resulted in policy uncertainty [Lalbahadur, 2017; Ngwakwe, Sebola, 2019]. There was a sharp fall in the value of 16 rands to the dollar due to an unforeseen cabinet reshuffle [Krabodia, Soni, 2016]. Moreover, the cabinet reshuffles of 2017 did not only weaken the rand but downgraded the sovereign credit rating of the country. These reshuffles can be attributed to policy uncertainty which makes the rand unstable.

The fuel prices' response was 13.40 (93 octane) and 13.54 (95 octane) rands per litre [Ros-souw, 2017]. The worst was seen during the 2021 July unrest and lootings which were triggered by the incarceration of the former president Jacob Zuma. The rand responded by 17.13 rands to

1 IMF. (2014). What is Keynesian economics? https://www.imf.org/external/pubs/ft/fandd/2014/09/basics.htm Accessed 18 March 2022.

the US dollar while petrol prices spiked to 17.30 (93 octane) and 17.39 (95 octane) rands per litre. Though the lootings signalled an entrenched socioeconomic and inequality woe, the trigger to these remained political factors1. Due to the negative relationship between the exchange rate and the fuel price, the response has been adverse especially with the current economic state of South Africa. Despite all discussed, the sole aim is to find effective solutions to mitigate the impact of the economic and political factors that affect fuel prices. The Automobile Association has proposed practical ways in which the government can reduce the price of fuel. The first step is to scrutinise the current pricing structure while auditing and recalculating the existing elements and clarifying their relevance. The reduction of RAF levy by improving its management and governance and maximising road safety and semi-privatisation of RAF should be considered.

Materials and methods

This section describes the research materials and methods and the approach executed in order to estimate the response of petrol prices to domestic and international components. The study employed monthly time series data from 2002 to 2021. The data was obtained from the Department of Energy archives. The method was guided by earlier studies [cf. Aucott, Hall, 2014; Berument, Sahin, Sahin, 2014; Kargbo, 2018].

The model for basic petrol prices can be specified mathematically as:

yt = (30 + (1 BCOt + ft EXRt + et , (1)

where yt is the dependent variable represented by the basic petrol price including both 95 octane and 93 octane fuel prices; BCO is the average Brent crude oil price; EXR is the exchange rate used to import petroleum products; et is the idiosyncratic error term.

It is sufficient to note that the basic petrol price is largely determined by the Brent crude oil price and the prevailing market exchange rate. The second model, which encompasses domestic factors, can be expressed mathematically as:

yt = a0 + a1 BPPt + a2 Xt + et, (2)

where yt is the dependent variable represented by the petrol pump price including both 95 octane and 93 octane fuel prices; BPP is the basic petrol price consisting of the Brent crude oil price and the exchange rate; Xt is a vector of domestic factors that make up the petrol pump price including fuel taxes, customs duties, Road Accident Fund levy, inland transport costs, wholesale margins, retail margins, pipeline taxes, equalisation fund, slate and pipeline levies; et is the idiosyncratic error term.

The study employed a multiple regression analysis. Several pre-estimation tests were conducted prior to the regression analysis. This included an informal overview of the variables and unit root analysis. The variables were examined for unit root by means of the Augmented Dickey - Fuller unit root test [Dickey, Fuller, 1979]. Following this, the study employed the structural vector autoregression (SVAR) given the structural breaks in the data. The optimal lag length was identified by means of a standard VAR. Given that the objective of the study was not to estimate the impact of petrol prices but rather the response of petrol prices to shocks in the energy market, the impulse response function and variance decomposition were executed by means of SVAR. The last point of analysis was examining the estimated residuals for autocorrelation and heteroskedasticity.

1 International Federation of Red Cross and Red Crescent Societies. (2021). Emergency plan of action South Africa: Urban violence. Situation report. https://reliefweb.int/report/south-africa/south-africa-urban-violence-emer-gency-plan-action-epoa-dref-operation-n-mdrza010.

iНе можете найти то, что вам нужно? Попробуйте сервис подбора литературы.

Results and discussion

The findings of the study are discussed in this section. This includes an overview of the data, stationarity analysis, SVAR estimation, impulse responses and variance decomposition. The section begins through an overview of the data in Figure 4. The y-axis measures the price of fuel in South African cents while the x-axis measures the period.

Basic petrol price (95)

Basic petrol price (93)

Average crude oil

Exchange rate

10 15 20

Petrol pump price (95)

0 5 10 15 20 Petrol pump price (93)

5 10 15 20 Fuel tax

0 5 10 15 20 Transport cost

0 5 10 15 20 Wholesale margin

0 5 10 15 20 Retail margin

0 5 10 15 20 Customs & excise

10 15 20

Fig. 4. An overview of the variables, 2000-2021 Рис. 4. Динамика переменных модели, 2000-2021

The basic fuel prices, including 95 octane, 93 octane fuel and diesel (0.005 sulphur) increased sharply between 2000M1 and 2021M12. Also, the prices appeared to move in the same direction as the average crude oil price. On the contrary, the exchange rate was constant with the exception of a large deviation in 2001M6. Domestic factors that feed into the fuel pump price, including fuel taxes, transport costs, wholesale margins, retail margins and customs and excise, likewise steadily went up except for custom and excise which remained fixed at 0.40 rands

throughout the entire period. Due to this, the variable dropped as it resulted in substantial serial correlation. The next point of analysis involved examining the variables for unit root by means of the Augmented Dickey - Fuller unit root test. The results are presented in Table 1.

Table 1. Unit root analysis Таблица 1. Анализ единичного корня

Variable Level 1st Diff 2nd Diff Outcome

Basic petrol price (95) -2.29 -12.19* - D(1)

Basic petrol price (93) -2.29 -12.23* D(1)

Crude oil price -2.57 -10.90* D(1)

Exchange rate -15.22* -12.24* D(0)

Petrol pump price (93) -0.55 -11.97* D(1)

Petrol pump price (95) -0.81 -11.94 D(1)

Fuel tax 0.91 -1.96 -31.99* D(2)

Transport cost 1.20 -18.38* - D(1)

Wholesale margin -2.11 -15.92* D(1)

Retail margin 1.30 -16.82* D(1)

The basic petrol price, average crude oil price, petrol pump price, transport cost, wholesale and retail margins were found to be stationary after first differencing while the exchange rate, on the contrary, was found to be stationary at level. Fuel taxes were neither stationary at level nor first difference but rather after second differencing. Following this, the SVAR technique was carried out to estimate the response of petrol prices to changes in their components. Further, international and domestic components are considered separately.

International components. International components that make up the basic petrol price include the Brent crude oil and the exchange rate. Other components, although relatively small in size, include freight and insurance costs. The response of basic petrol prices (95 octane) (shock 1) to shocks in Brent crude oil (shock 2) and the exchange rate (shock 3) is illustrated in Figure 5. The y-axis measures the response to a one standard deviation while the x-axis measures the period.

Accumulated response of basic petrol price (95) to Shock 1

Accumulated response of basic petrol price (95) to Shock 2

Accumulated response of basic petrol price (95) to Shock 3

~1-1-1-1-1-1-1-г

123456789 10

8 9 10

123456789 10

Fig. 5. Accumulated response of basic petrol price (95 octane)1 Рис. 5. Накопленный отклик базовой цены бензина (с октановым числом 95)

The impulse response function reveals that basic petrol prices (95 octane) respond positively to their own shocks in the short, medium and long run. Similarly, basic petrol prices responded

1 Note: accumulated response to SVAR innovations ±2 standard errors.

positively to shocks in crude oil prices throughout the entire period. Hartley and Medlock [2014] likewise found that basic fuel prices and crude oil prices move in the same direction at least in the short run. It is worth noting also, that the response of basic petrol prices to their own shocks and shocks in crude oil prices was found to be substantial in size. In contrast, the response of basic petrol prices to shocks in the exchange rate appear to be neutral in the short and medium term albeit the response becomes negative in the long run.

The response of basic petrol prices (93 octane) (shock 1) to shocks in Brent crude oil (shock 2) and the exchange rate (shock 3) is illustrated in Figure 6. The y-axis measures the response to

Accumulated response of basic petrol price (93) to Shock 3

.8 .6 .4 .2

______________

.o "■=":::::;;;________

123456789 10

Fig. 6. Accumulated response of basic petrol price (93 octane)1 Рис. 6. Накопленный отклик базовой цены бензина (с октановым числом 93)

a one standard deviation while the x-axis measures the period.

Accumulated response of basic petrol price (93) to Shock 1

Accumulated response of basic petrol price (93) to Shock 2

-1-1-1-1-1-1-1-Г

123456789 10

1-1-1-1-1-1-1-1-Г"

123456789 10

Basic petrol prices (93 octane) behave similarly to basic petrol prices (95 octane). For example, basic petrol prices (93 octane) respond positively to their own shocks and to shocks in the average Brent crude oil price in the short and long run. On the contrary, the response of basic fuel prices to shocks in the exchange rate is found to be muted in the short run and negative in the long run, albeit the magnitude is relatively small in size. These findings are in line with the studies of Gagnon, Hall, Brinker [2009] and Berument, Sahin, Sahin [2014] who found that the response of basic fuel prices to changes in Brent crude oil is relatively larger than the response of basic fuel prices to changes in the exchange rate. The variance decomposition was employed to estimate the forecast errors of each variable in relation to its own shock. The findings are summarised in Table 2. In the short run (1-3 years), shocks in basic petrol prices (95 octane) are largely explained by their own shocks. In the medium term (4-6 years), shocks in basic petrol prices (95 octane) are explained by their own shocks (42 %) and shocks in the Brent crude oil price (58 %). In the long run (7-10 years), shocks in the basic petrol price (95 octane) are explained by their own shocks (38 %) and shocks in the Brent crude oil price (62 %). To a small extent, shocks in the basic petrol price (95 octane) are attributable to shocks in the exchange rate (0.15 %), at least in the long run.

The variance decomposition was likewise carried out for basic petrol prices (93 octane) and the results were not far different from those obtained for basic petrol price (95 octane). As can be seen from Table 3, in the short run, shocks in basic petrol prices (93 octane) are explained by their own shocks (54 %), shocks in Brent crude oil prices (46 %) and shocks in the exchange rate (0.8 %). In long run however, shocks in basic petrol prices (93 octane) are explained by their own shocks (43 %), shocks in Brent crude oil prices (57 %) and shocks in the exchange rate (0.8 %).

1 Note: accumulated response to SVAR innovations ±2 standard errors.

Table 2. Variance decomposition of basic petrol price (95 octane)

Таблица 2. Дисперсия базовой цены бензина (с октановым числом 95)

Period Standard error Shock 1 Shock 2 Shock 3

1 0.08 100.00 0.00 0.00

2 0.13 75.52 24.39 0.08

3 0.18 54.69 45.23 0.08

4 0.21 48.51 51.42 0.07

5 0.23 47.31 52.63 0.07

6 0.24 46.17 53.77 0.07

7 0.25 44.14 55.79 0.07

8 0.26 41.87 58.05 0.08

9 0.27 39.91 59.98 0.11

10 0.28 38.24 61.61 0.15

Note: structural factorization is used.

Table 3. Variance decomposition of basic petrol price (93 octane)

Таблица 3. Дисперсия базовой цены бензина (с октановым числом 93)

Period Standard error Shock 1 Shock 2 Shock 3

1 0.08 100.00 0.00 0.00

2 0.14 75.23 24.68 0.08

3 0.18 54.16 45.76 0.08

4 0.21 47.82 52.09 0.08

5 0.23 46.42 53.50 0.07

6 0.24 45.15 54.78 0.07

7 0.25 43.12 56.81 0.07

8 0.26 40.89 59.03 0.08

9 0.27 38.96 60.92 0.10

10 0.28 37.33 62.52 0.14

Note: structural factorization is used.

The estimated SVAR model was analysed for residual diagnostics. This includes the autocorrelation and heteroskedasticity tests. The results are presented in Table 4.

Table 4. Residual analysis

Таблица 4. Анализ остатков

Dependent variable Autocorrelation White heteroskedasticity

Basic petrol price (95) 0.94 0.06

Basic petrol price (93) 0.91 0.07

The autocorrelation test revealed that residuals in both models (93 and 95 octane) do not suffer from autocorrelation. This is evidenced by the probability values which are above the 5 % significance level. The probability value for the basic petrol price (95 octane) was found to be 94 % while the probability value for the basic petrol price (93 octane) was found to be 91 %.

Furthermore, the White heteroskedasticity test likewise revealed that the models do not suffer from heteroskedasticity given that the probability values are above the 5 % significance level. Further we will detail the findings on the domestic components.

Domestic components. This part presents findings on the response of the fuel pump prices (shock 1) to variations in domestic components including basic petrol price (shock 2), fuel tax (shock 3), Road Accident Fund levy (shock 4), transport costs (shock 5), wholesale margins (shock 6) and retail margins (shock 7). The response of fuel pump prices (95 octane) to these shocks is illustrated in Figure 7. The y-axis measures the response to a one standard deviation while the x-axis measures the period.

Accumulated response of petrol pump price (95) to Shock 1

iНе можете найти то, что вам нужно? Попробуйте сервис подбора литературы.

Accumulated response of petrol pump price (95) to Shock 2

Accumulated response of petrol pump price (95) to Shock 3

123456789 10

Accumulated response of petrol pump price (95) to Shock 4

123456789 10

Accumulated response of petrol pump price (95) to Shock 5

123456789 10

Accumulated response of petrol pump price (95) to Shock 6

123456789 10

Accumulated response of petrol pump price (95) to Shock 7

123456789 10

6 7 8 9 10

Fig. 7. Accumulated response of fuel pump price (95 octane)1 Рис. 7. Накопленный отклик розничной цены бензина (с октановым числом 95)

The response of the fuel pump price (95 octane) to its own shocks was found to be positive and significant throughout the period. Fuel pump prices generate a positive respose to shocks in basic petrol prices, transport costs, wholesale and retail margins in the long run, albeit it is neutral in the short and medium run. The response of fuel pump prices (95 octane) to shocks in fuel taxes was found to be muted throughout the entire period whereas their response to the RAF levy was found to be negative in the short and long run.

1 Note: accumulated response to SVAR innovations ±2 standard errors.

A summary of findings on the response of fuel pump prices (93 octane) to domestic components is provided in Figure 8. The y-axis measures the response to a one standard deviation while the x-axis measures the period. Similarly, the response of fuel pump prices (93 octane) to internal components is no different from the response of fuel pump prices (95 octane). This is particularly due to the fact the prices of both 95 octane and 93 octanes differ by a relatively small margin. Nonetheless, fuel pump prices (93 octane) respond positively to shocks in the basic petrol price, transport costs, wholesale and retail margins but respond negatively to shocks in the Road Accident Fund levy. In addition, the response of fuel pump prices (93 octane) to shocks in fuel taxes is somewhat neutral throughout the entire period. On the downside, we fail to compare our findings concerning the influence of domestic components with earlier studies given that research in this field is scant in South Africa.

Accumulated response of petrol pump price (93) to Shock 1

Accumulated response of petrol pump price (93) to Shock 2

Accumulated response of petrol pump price (93) to Shock 3

123456789 10

Accumulated response of petrol pump price (93) to Shock 4

123456789 10 123456789 10

Accumulated response of petrol pump price (93) to Shock 5

Accumulated response of petrol pump price (93) to Shock 6

8 9 10

123456789 10

Accumulated response of petrol pump price (93) to Shock 7

123456789 10

Fig. 8. Accumulated response of fuel pump price (93 octane)1 Рис. 8. Накопленный отклик розничной цены бензина (с октановым числом 93)

The variance decomposition analysis, which measures the forecast error of a one standard deviation, was performed for fuel pump prices (95 octane). The results are provided in Table 5.

1 Note: accumulated response to SVAR innovations ±2 standard errors.

Table 5. Variance decomposition of fuel pump price (95 octane)

Таблица 5. Дисперсия розничной цены бензина (с октановым числом 95)

Period Standard error Shock 1 Shock 2 Shock 3 Shock 4 Shock 5 Shock 6 Shock 7

1 0.05 100.00 0.00 0.00 0.00 0.00 0.00 0.00

2 0.09 98.88 0.63 0.03 0.01 0.00 0.44 0.01

3 0.10 98.56 0.46 0.02 0.24 0.22 0.44 0.07

4 0.11 97.99 0.39 0.02 0.38 0.78 0.37 0.06

5 0.12 97.15 0.39 0.03 0.51 1.37 0.41 0.14

6 0.12 95.99 0.53 0.03 0.62 1.87 0.63 0.33

7 0.13 94.72 0.66 0.03 0.69 2.33 0.99 0.58

8 0.13 93.47 0.71 0.04 0.76 2.73 1.46 0.84

9 0.13 92.29 0.71 0.05 0.81 3.05 2.03 1.06

10 0.14 91.22 0.69 0.05 0.85 3.28 2.63 1.26

Note: structural factorization is used.

The findings with respect to the fuel pump price were quite interesting in that shocks in the fuel pump price are largely explained by their own shocks both in the short run and long run. For example, in the short run, 98 % of the variations in the petrol pump prices (95 octane) are seemingly explained by their own shocks. In the same vein, in the long run, shocks in the petrol pump prices are explained by their own shocks (94 %), shocks in the basic petrol prices (0.7 %), fuel tax (0.01 %), Road Accident Fund levy (0.9 %), transport costs (3.28 %), wholesale margins (2.63 %), and retail margins (1.26 %). The results regarding the petrol pump price (93 octane) are presented in Table 6.

Table 6. Variance decomposition of fuel pump price (93 octane)

Таблица 6. Дисперсия розничной цены бензина (с октановым числом 93)

Period Standard error Shock 1 Shock 2 Shock 3 Shock 4 Shock 5 Shock 6 Shock 7

1 0.056 100.00 0.00 0.00 0.00 0.00 0.00 0.00

2 0.085 98.97 0.57 0.02 0.02 0.00 0.41 0.00

3 0.101 98.63 0.41 0.01 0.24 0.23 0.41 0.06

4 0.111 98.02 0.36 0.02 0.38 0.82 0.34 0.06

5 0.118 97.15 0.35 0.02 0.50 1.42 0.39 0.14

6 0.123 95.96 0.49 0.03 0.61 1.93 0.63 0.33

7 0.127 94.66 0.62 0.03 0.68 2.38 1.02 0.59

8 0.131 93.39 0.67 0.04 0.74 2.78 1.52 0.86

9 0.133 92.21 0.68 0.04 0.79 3.09 2.09 1.09

10 0.136 91.11 0.67 0.05 0.82 3.33 2.73 1.29

Note: structural factorization is used.

In the short run, 98 % of the shocks in the petrol pump prices (93 octane) are explained by their own shocks. In the long run, variations in the petrol pump prices are less explained by their own shocks (92 %), they are attributable to shocks in the basic petrol price (0.7 %), fuel taxes (0.1 %), Road Accident Fund levy (0.8 %), to transport costs (3.3 %), wholesale margins (2.7 %) and retail margins (1.3 %). The last point of analysis involved examining the residuals. The results are summarised in Table 7.

Table 7. Residual analysis

Таблица 7. Анализ остатков

Dependent variable Autocorrelation Heteroskedasticity

Petrol pump price (95) 0.99 0.62

Petrol pump price (93) 0.99 0.69

The autocorrelation and heteroskedasticity tests confirmed that the estimated residuals do not suffer from autocorrelation and heteroskedasticity. This was proved by the probability values that were above the 5 % significance level in both models. As a result, the null hypotheses of autocorrelation and heteroskedasticity were rejected against the alternative hypotheses.

Conclusion

The primary goal of the study was to investigate the response of petrol prices to variations in domestic and international components. Monthly time series data spanning from 2002 to 2021 was analysed by means of econometric modelling including the structural vector autoregression. The first point of analysis involved examining the variables for stationarity by means of the Augmented Dickey - Fuller unit root test. The basic fuel price, average crude oil price, fuel pump price, transport costs, wholesale and retail margins were found to be stationary after first differencing while the exchange rate, on the contrary, was found to be stationary at level. Fuel taxes were neither stationary at level nor first difference but rather after second differencing. Following this, the structural vector autoregression was carried out to estimate the response of fuel prices to changes in their components. The structural impulse response function indicated that basic petrol prices respond positively to their own shocks and to shocks in Brent crude oil prices and this response was found to be substantial in size. By contrast, basic petrol prices respond negatively to shocks in exchange rates, albeit the response is small in size. The variance decomposition revealed that variations in basic petrol prices are largely explained by their own shocks in the short run and by shocks in Brent crude oil prices in the long run. Domestically, petrol pump prices respond positively to shocks in basic petrol prices, inland transport costs, wholesale and retail profit-margins while on the contrary, the response of petrol pump prices was found to be negative towards shocks in the Road Accident Fund levy and muted towards shocks in fuel taxes. Given the rise in crude oil prices, the government can consider reviewing the current fuel structure and suspending several fuel levies indefinitely to provide the much-needed financial relief to households and businesses. In addition, it can attempt to reduce the reliance on oil by improving energy efficiency and diversifying into non-petroleum sources of energy. The most common alternatives are natural gas and renewable sources of energy such as hydro, wind, solar and geothermal power.

References

Alvarez L. J., Hurtado S., Sanchez I., Thomas C. (2011). The impact of oil price changes on Spanish and Euro area consumer price inflation. Economic Modelling, vol. 28, issue 1, pp. 422-431. https://doi. org/10.1016/j.econmod.2010.08.006.

Aucott M., Hall C. (2014). Does a change in price of fuel affect GDP growth? An examination of the U.S. data from 1950-2013. Energies, vol. 7, pp. 6558-6570. D0I:10.3390/en7106558.

Ashley R., Tsang K. P. (2013). The oil price-real output relationship: Does persistence matter? (Virginia Technical Working Papers). https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.488.4366&rep=r ep1&type=pdf.

Bala U., Chin L. (2018). Asymmetric impacts of oil price on inflation: An empirical study of African OPEC member countries. Energies, vol. 11, no. 11, 3017. https://doi.org/10.3390/en11113017.

Berument M. H., Sahin A., Sahin S. (2014). The relative effects of crude oil price and exchange rate on petroleum product prices: Evidence from a set of Northern Mediterranean countries. Economic Modelling, vol. 42, pp. 243-249. DOI: 10.1016/j.econmod.2014.07.019.

Brown S. P. A., Yucel M. K. (2002). Energy prices and aggregate economic activity: An interpretative survey. Quarterly Review of Economics and Finance, vol. 42, issue 1, pp. 193-208.

Burger P., Calitz E. (2021). Covid-19, economic growth and South African fiscal policy. South African Journal of Economics, vol. 89, issue 1, pp. 3-24. https://doi.org/10.1111/saje.12270.

Carruth A., Hooker M. A., Oswald A. (1998). Unemployment equilibrium and input prices: Theory and evidence from the United States. The Review of Economics and Statistics, vol. 80, issue 4, pp. 621-628.

Cavalcanti T., Jalles J. T. (2013). Macroeconomic effects of oil price shocks in Brazil and in the United States. Applied Energy, vol. 104, pp. 475-486. DOI: 10.1016/j.apenergy.2012.10.039.

Cheung K., Thomson E. (2004). The demand for gasoline in China: A cointegration analysis. Journal of Applied Statistics, vol. 31, issue 5, pp. 533-544. https://doi.org/10.1080/02664760410001681837.

Dickey D. A., Fuller W. A. (1979). Distribution of estimators for autoregressive time series with a unit root. Journal of American Statistical Association, vol. 74, no. 366, pp. 427-431. D0I:10.2307/2286348.

Edelstein P., Lutz K. (2009). How sensitive are consumer expenditures to retail energy prices? Journal of Monetary Economics, vol. 56, no. 6, pp. 766-779.

Gagnon N., Hall C., Brinker L. (2009). A preliminary investigation of energy return on energy investment for global oil and gas production. Energies, vol. 2, issue 3, pp. 490-503. https://doi.org/10.3390/ en20300490.

Hartley P., Medlock K. (2014). The relationship between crude oil and natural gas prices: The role of the exchange rate. The Energy Journal, vol. 35, no. 2, pp. 25-44.

Kargbo B. I. B. (2018). Exchange rate, petroleum price and price determination in Sierra Leone (AERC Research Paper no. 353). Nairobi: African Economic Research Consortium. 46 p.

Keynes J. M. (1936). The general theory of employment, interest, and money. New York: MacMil-lan. 472 p.

Krabodia A. M., Soni P. (2016). President Jacob Zuma and South Africa's financial crisis: A Machiavellian debacle. International Business Research, vol. 9, no. 7, pp. 24-54. D0I:10.5539/ibr.v9n7p24.

Lalbahadur A. (2017). The impact of the cabinet reshuffle on South Africa's standing as an emerging power. The South African Institute of International Affairs (SAIIA). https://saiia.org.za/research/is-south-africa-still-an-emerging-african-power/.

Lardic S., Mignon V. (2006). The impact of oil prices on GDP in European countries: an empirical investigation based on asymmetric cointegration. Energy Policy, vol. 34, issue 18, pp. 3910-3925.

LeBlanc M., Chinn M. (2004). Do high oil prices presage inflation? The evidence from g-5 countries (SCCIE Working Paper no. 04-04). https://ssrn.com/abstract=509262.

Lescaroux F., Mignon V. (2008). On the influence of oil prices on economic activity and other macroeconomic and financial variables. OPEC Energy Review, vol. 32, no. 4, pp. 343-380. https://doi. org/10.1111/j.1753-0237.2009.00157.x.

Mabugu R., Chitaga M., Amusa H. (2009). The economic consequences of fuel levy reform in South Africa. South African Journal of Economic and Management Sciences, vol. 12, no. 3, pp. 280-296.

Mabugu R., Robichaud V., Maisonnave H., Chitiga M. (2013). Impact of fiscal policy in an intertemporal CGE model for South Africa. Economic Modelling, vol. 31, pp. 775-782. DOI: 10.1016/j.econ-mod.2013.01.019.

Metcalfe R., Dolan P. (2012). Behavioural economics and its implications for transport. Journal of Transport Geography, vol. 24, pp. 503-511. DOI: 10.1016/j.jtrangeo.2012.01.019.

Ncanywa T., Mgwangqa N. (2018). Impact of a fuel levy on economic growth in South Africa. Journal of Energy in Southern Africa, vol. 29, no. 1, pp. 41-49. http://dx.doi.org/10.17159/2413-3051/2018/ v29i1a2775.

Ncube M., Shimeles A., Verdier-Chouchene A. (2012). South Africa's quest for inclusive development (Working Paper Series no. 150). Tunis: African Development Bank, Tunisia. 28 p.

Newbery D. (2005). Why tax energy? Towards a more rational policy. Energy Journal, vol. 26, no. 3, pp. 1-40. DOI: 10.5547/ISSN0195-6574-EJ-Vol26-No3-1.

Ngwakwe C. C., Sebola M. P. (2019). Differential effects of politically driven policy uncertainty of exchange rate performance. African Journal of Public Affairs, vol. 11, no. 1, pp. 34-47.

iНе можете найти то, что вам нужно? Попробуйте сервис подбора литературы.

Odhiambo M. N. (2010). Oil prices and economic growth in South Africa: an ARDL testing approach. Energy Policy, vol. 38, pp. 2463-2469.

Rossouw J. (2017). Firing of South Africa's financial minister puts the public purse in Zuma's hands. The Conversation, March 31. https://theconversation.com/firing-of-south-africas-finance-minister-puts-the-public-purse-in-zumas-hands-75525.

Sabelo M. (2018). The drivers behind fuel price hikes. Mail & Guardian, July 6. https://mg.co.za/ article/2018-07-06-00-the-drivers-behind-fuel-price-hikes/.

Sun Z., Hong J., Xu X. (2013). Price effect of domestic oil tax under vertically related market structure: Evidence from the United States, EU and Japan. OPEC Energy Review, vol. 37, issue 1, pp. 81-104. https://doi.org/10.1111/j.1753-0237.2012.00228.x.

Information about the authors

Baneng Naape, PhD Candidate (Economics), Research Scholar at the Department of Economics. University of the Witwatersrand, Johannesburg, South Africa. E-mail: banengnaape@gmail.com Ndzalama Cleopatra Mathebula, MA in Politics and International Relations, Academic Tutor at the Department of Politics and International Relations. University of Johannesburg, Johannesburg, South Africa. E-mail: ndzali57@gmail.com

■ ■ ■

Источники