07 Jun
07Jun

In continuity with the posts of scientific reports on fuels, which are afflicting the financial life of Brazilians with their constant highs and consequently reflecting on the prices of products, the integrated relationship between fuel prices at resale stations and their sales in cubic meters has now been analyzed. Therefore, this survey aims to answer questions such as: does the increase in the price of a certain type of fuel decrease your sales? or increases of another kind? or do they both diminish? The opposite is also worth analysing: can increases or decreases in sales affect your interrelated prices?

Analyses by Selfregressive Vectors (VAR) were adopted because they are time series, as well as the interrelationship between the variables. The samples were collected in monthly data for the State of Santa Catarina for the following time series:

  • Gasoline sales to Dealer Stations (m3)
  • Ethanol Sales to Retail Stations (m3)
  • Sales of diesel oil to Dealer Stations (m3)
  • Average resale price of gasoline to dealer stations (R$/L)
  • Average ethanol resale price at retail stations (R$/L)
  • Average resale price of diesel to dealer stations (R$/L)

All series were collected and organized for the period between January 2013 and April 2022, transformed into Natural Logarithmic. We analyzed the stationary of the variables, procedures of necessary differentiations, lags, p-values of approval in the equations developed in the autoregressive vectors, granger causality tests and impulse-response function. Data were collected exclusively at the Brazilian Open Data Portal, organized by the National Agency for Petroleum, Natural Gas and Biofuels (ANP, 2022). The series were smoothed by Natural Logarithmic procedures and then passed to the first difference in order to obtain approvals in the parking tests, according to the table of results of the p-values of the ADF test below: 

Parking ADF Test
Time SeriesP-value
VENDA_GASOLINA0.01
VENDA_ETANOL0.01
VENDA_DIESEL0.01
PREÇO_GASOLINA0.01
PREÇO_ETANOL0.01
PREÇO_DIESEL0.02

The results will be exposed in table format to facilitate observations, highlighting only the main impulse-response graphs in sequence of the results of predictive equations and granger causality tests. The autorregressive vectors expressed 1 lag of delay in the monthly analyses. 

Two caveats deserve attention: 

(1) when a variable obtained an approximation to the approval in its p-value of the matrix of the autorregressive vectors (close to 10%), the p-values of granger causality were also approached to confirm its adhering to the analyses or not. Confirming the adherence of the variable, its Impulse-Response Function graph was rotated.

(2) the analysis of granger causality does not run for self-impacts (autoregressive) of Y. Thus, only the p-values of the vector matrix and the FIR were approached for analysis of its relevance in the predictive equation. 

Equation 1: Diesel Price - State of Santa Catarina

preço_dieseldiff = preço_etanoldiff.l1 + preço_dieseldiff.l1 + preço_gasolinadiff.l1 + venda_gasolinadiff.l1 + venda_etanoldiff.l1 + venda_dieseldiff.l1 + const
                                       Estimate Std. Error t value Pr(>|t|)
preço_etanoldiff.l1 -0.124750 0.141679 -0.881 0.38065
preço_dieseldiff.l1 0.457728 0.148737 3.077 0.00268 **
preço_gasolinadiff.l1 -0.042900 0.216251 -0.198 0.84314
venda_gasolinadiff.l1 -0.030867 0.038025 -0.812 0.41883
venda_etanoldiff.l1 0.040005 0.019790 2.022 0.04585 *
venda_dieseldiff.l1 0.046234 0.034931 1.324 0.18860
const 0.007075 0.002702 2.618 0.01019 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05'.' 0.1 ' ' 1
F-statistic: 4,906 on 6 and 102 DF, p-value: 0.0001912


Granger CausalityFIR – Impulse ↑ in X
YXResultadoP-valueLagY Result
Preço_dieselVenda_etanolAprovado0.0029331

Conclusions:

1. An increase in the sale of ethanol increases the price of diesel for up to five months

2. A boost in the price of diesel, increase your own price in up to five months


Equation 2: Gasoline Price - State of Santa Catarina

preço_gasolinadiff = preço_etanoldiff.l1 + preço_dieseldiff.l1 + preço_gasolinadiff.l1 + venda_gasolinadiff.l1 + venda_etanoldiff.l1 + venda_dieseldiff.l1 + const
                                  Estimate Std. Error t value Pr(>|t|)
preço_etanoldiff.l1 -0.288219 0.146140 -1.972 0.0513 .
preço_dieseldiff.l1 0.249168 0.153420 1.624 0.1074
preço_gasolinadiff.l1 0.224879 0.223059 1,008 0.3158
venda_gasolinadiff.l1 0.012678 0.039223 0.323 0.7472
venda_etanoldiff.l1 0.020531 0.020413 1.006 0.3169
venda_dieseldiff.l1 -0.031368 0.036031 -0.871 0.3860
const 0.006582 0.002787 2.361 0.0201 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05'.' 0.1 ' ' 1
F-statistic: 3,646 on 6 and 102 DF, p-value: 0.002549

 

Causalidade de GrangerFIR – Impulso ↑ em X
YXResultadoP-valueLagResultado em Y
Preço_gasolinaPreço_etanolAprovado0.0030161

Causality of Granger
YXResultP-valueLag
Preço_gasolinaPreço_dieselFailed0.14131


Conclusions:

1. An increase in the price of ethanol, increases the price of gasoline for up to two months


Equation 3: Ethanol Price - State of Santa Catarina

preço_etanoldiff = preço_etanoldiff.l1 + preço_dieseldiff.l1 + preço_gasolinadiff.l1 + venda_gasolinadiff.l1 + venda_etanoldiff.l1 + venda_dieseldiff.l1 + const
                       Estimate Std. Error t value Pr(>|t|)
preço_etanoldiff.l1 0.034789 0.151782 0.229 0.819
preço_dieseldiff.l1 0.228756 0.159343 1,436 0.154
preço_gasolinadiff.l1 0.288392 0.231671 1,245 0.216
venda_gasolinadiff.l1 0.013101 0.040737 0.322 0.748
venda_etanoldiff.l1 -0.008122 0.021201 -0.383 0.702
venda_dieseldiff.l1 0.008521 0.037422 0.228 0.820
const 0.003719 0.002895 1,285 0.202
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05'.' 0.1 ' ' 1
F-statistic: 4,657 on 6 and 102 DF, p-value: 0.0003181

 

Causality of GrangerFIR - Impulse ↑ in X
YXResultP-valueLagY-bottom result
Preço_etanolPreço_dieselApproved0.0019891

Conclusion:

1. An increase in the price of diesel, increases the price of ethanol for up to six months


Equation 4: Diesel Sale - State of Santa Catarina

venda_dieseldiff = preço_etanoldiff.l1 + preço_dieseldiff.l1 + preço_gasolinadiff.l1 + venda_gasolinadiff.l1 + venda_etanoldiff.l1 + venda_dieseldiff.l1 + const
                                      Estimate Std. Error t value Pr(>|t|)
preço_etanoldiff.l1 -0.4151218 0.4123432 -1,007 0.316443
preço_dieseldiff.l1 -0.0029791 0.4328838 -0.007 0.994522
preço_gasolinadiff.l1 0.9957006 0.6293754 1,582 0.116735
venda_gasolinadiff.l1 -0.1040774 0.1106691 -0.940 0.349215
venda_etanoldiff.l1 -0.0535512 0.0575961 -0.930 0.354684
venda_dieseldiff.l1 -0.3880267 0.1016643 -3.817 0.000232 ***
const -0.0001491 0.0078640 -0.019 0.984915
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05'.' 0.1 ' ' 1
F-statistic: 6,656 on 6 and 102 DF, p-value: 5.787e-06

 

Causality of GrangerFIR - Impulse ↑ in X
YXResultP-valueLagY-bottom result
Venda_dieselPreço_gasolinaApproved0.043851

Conclusions:

1. An increase in the price of gasoline, slightly increases the sale of diesel for up to three months

2. A boost in diesel sales, strongly increases your own sale for up to a month


Equation 5: Sale of Gasoline - State of Santa Catarina

venda_gasolinadiff = preço_etanoldiff.l1 + preço_dieseldiff.l1 + preço_gasolinadiff.l1 + venda_gasolinadiff.l1 + venda_etanoldiff.l1 + venda_dieseldiff.l1 + const
                       Estimate Std. Error t value Pr(>|t|)
preço_etanoldiff.l1 -0.105382 0.442312 -0.238 0.8122
preço_dieseldiff.l1 -0.007173 0.464346 -0.015 0.9877
preço_gasolinadiff.l1 -0.087595 0.675118 -0.130 0.8970
venda_gasolinadiff.l1 -0.292986 0.118713 -2.468 0.0152 *
venda_etanoldiff.l1 0.005935 0.061782 0.096 0.9237
venda_dieseldiff.l1 -0.051120 0.109053 -0.469 0.6402
const 0.005182 0.008436 0.614 0.5404
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05'.' 0.1 ' ' 1
F-statistic: 2,024 on 6 and 102 DF, p-value: 0.06918

Conclusion:

1. An increase in gasoline sales, strongly increases its own sales for up to a month

Equation 6: Ethanol Sales - State of Santa Catarina

venda_etanoldiff = preço_etanoldiff.l1 + preço_dieseldiff.l1 + preço_gasolinadiff.l1 + venda_gasolinadiff.l1 + venda_etanoldiff.l1 + venda_dieseldiff.l1 + const
                       Estimate Std. Error t value Pr(>|t|)
preço_etanoldiff.l1 -3.643958 1.016909 -3,583 0.000522 ***
preço_dieseldiff.l1 0.147747 1.067566 0.138 0.890200
preço_gasolinadiff.l1 2.489217 1.552148 1,604 0.111867
venda_gasolinadiff.l1 0.284777 0.272929 1,043 0.299225
venda_etanoldiff.l1 -0.125733 0.142042 -0.885 0.378139
venda_dieseldiff.l1 -0.493544 0.250722 -1.968 0.051725 .
const 0.004581 0.019394 0.236 0.813728
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05'.' 0.1 ' ' 1
F-statistic: 3,615 on 6 and 102 DF, p-value: 0.002722

 

Causality of GrangerFIR - Impulse ↑ in X
YXResultP-valueLagY-bottom result
Venda_etanolVenda_dieselApproved0.016391

Causality of GrangerFIR - Impulse ↑ in X
YXResultP-valueLagY-bottom result
Venda_etanolPreço_etanolApproved0.0023761

Causality of Granger
YXResultP-valueLag
Venda_etanolPreço_gasolinaFailed0.39581


Conclusions:

1. An increase in diesel sales decreases ethanol sales for up to three months

2. A decrease in the price of ethanol decreases the sale of ethanol for up to five months


References:

ANP. Preços de Revenda e de Distribuição de Combustíveis. Disponível em: https://www.gov.br/anp/pt-br/assuntos/precos-e-defesa-da-concorrencia/precos/precos-revenda-e-de-distribuicao-combustiveis/. Acesso em 02 maio 2022

ANP. Vendas de derivados de petróleo e biocombustíveis em: https://dados.gov.br/dataset/vendas-de-derivados-de-petroleo-e-biocombustiveis acessado 10 maio 2022 

Comments
* The email will not be published on the website.