A N E MPIRICAL A NALYSIS OF P ASS -T HROUGH OF O IL P RICES TO I NFLATION : E VIDENCE FROM N IGERIA. * AUWAL, Umar Department of Economics, Ahmadu Bello.

AN EMPIRICAL ANALYSIS OF PASS-THROUGH OF OIL PRICES TO INFLATION: EVIDENCE

FROM NIGERIA.*

AUWAL, UmarDepartment of Economics,

Ahmadu Bello University, Nigeria – West Africa

[email protected] , [email protected]

+234 803 227 4567 , +234(0)705 727 6029

1

ORGANIZATION OF THE WORK

Introduction. Received Knowledge vs. Objective(s). Data Source and Estimation techniques. Models Specification. Unit Root test. Models Estimation, interpretation and

Analysis. Summary and Conclusions.

2

1.0 INTRODUCTION Oil prices have risen sharply over the last year,

leading to concerns that we could see a repeat of the 1970s, when rising oil prices were accompanied by severe recessions and surging inflation.

The oscillation of global oil prices has always been a major concern in market instability. This instability resulted into inflation. Consequently, the price of oil and inflation are often seen as being connected within a cause and effect framework. As oil prices move up or down, inflation follows in the same direction. The reason why this happens may be that oil is a major input in the economy - it is used in critical activities such as fueling transportation or goods made with petroleum products - and if the costs of intermediate input rise, so should the cost of end output (http://www.investopedia.com/ask/answers/06/oilpricesinflation.asp) .

3

INTRODUCTION (CONT…) Crude Oil Prices Period of high price strategy in the oil market Period of substantial decrease in crude oil

prices - It reached a peak of $147 in july,2008 and decrease to $38.6 in December, 2008 and now is below $80 (Abosedra, 2009).

Nigeria : (i)a mono-cultural economy

(ii)recognized as one of the most volatile economies in the world

Volatility: a major constraint on development Causes: planning more problematic and

investment more risky (Ukwu et.al, 2003) 4

2.0 OBJECTIVES OF THE STUDY:

There have been many papers that have examined pass-through of oil price fluctuations to exchange rate as well as some that have examined pass-through to domestic inflation.

Many of the recent studies have concentrated on the relationship between an country’s characteristics and the pass-through of oil price fluctuations in that country.

The objective of this paper is to empirically analyze the pass-through of oil price shock to inflation in Nigeria. Specifically – It examines the

historical relationship between oil price shocks and inflation in light of trend analysis and some recent research, and

Estimate and analyzes the impact of oil price and exchange rates on inflation.

it uses monthly data from 2003:01 to 2012:10.

Received Knowledge Objective(s)

5

3.0 DATA SOURCE AND METHODOLOGY

Monthly data :2003:01 - 2012:10

Type : Crude Oil Prices, Exchange rates and Inflation

Source: Central Bank of Nigeria’s website – Data and Statistics division

OLS, VAR-VECM and Granger Causality model were employed to analyze the data.

Data Source Models employed

6

TREND ANALYSIS

7

8

20

40

60

80

100

120

140

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

OIL_PRICE

9

110

120

130

140

150

160

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

EXCHANGE_RATE

10

0

5

10

15

20

25

30

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

INFLATION

Model Model SpecificationOLS, VAR-VECM This study employs OLS and VAR-VECM for empirical

analysis and only focusses on three chosen variables: Oil Price (Bonny light, $/B), Exchange rates (N/$) and Inflation (All items – year on change).

Where Inft is Inflation rate for period t. OPt is Crude Oil Price for period t, and EXt is exchange rates for period t. To get the best result, the equation must be in log for all variables.

Pre – estimation tests conducted

Lag selection criteria, Johansen test of Co-integration, System equation estimation

4.0 Models Specification

11

),( EXOPfInf tttt EXOPInf

210

tttt EXOPInf lnlnln 210

)(OPfInf

12

ADF – TEST

Variable Order of Integration Critical ValuesComputed Values

Crude Oil Price I(1)-3.4900 (1%) -2.8874(5%)-2.5804 (10%

-4.672726

Exchange Rate I(1)3.4900 (1% )-2.8874 (5% ) -2.5804 (10% )

-4.329422

Inflation Rate I(1)-3.4900 (1%)

-2.8874 (5%)

-2.5804 (10%

-4.267187

5.0 Unit Root Test

6.0 MODEL PRESENTATION, ESTIMATION

& ANALYSIS OF THE RESULTS:

13

ORDINARY LEAST SQUARES OUTPUT

14

Dependent Variable: INF    Method: Least Squares    Date: 04/22/13 Time: 16:38    Sample (adjusted): 2003M01 2012M09  Included observations: 117 after adjustments  

Variable Coefficient Std. Error t-Statistic Prob.                    

OP -0.061926 0.015108 -4.098788 0.0001EX 0.111950 0.035382 3.164009 0.0020C 1.327876 4.604224 0.288404 0.7736                  

R-squared 0.150036    Mean dependent var 12.19744Adjusted R-squared 0.135124    S.D. dependent var 4.902558S.E. of regression 4.559316    Akaike info criterion 5.897529Sum squared resid 2369.759    Schwarz criterion 5.968354Log likelihood -342.0054    Hannan-Quinn criter. 5.926283F-statistic 10.06164    Durbin-Watson stat 0.244726Prob(F-statistic) 0.000095      

GRANGER CAUSALITY TEST

Pairwise Granger Causality TestsDate: 04/22/13 Time: 17:31Sample: 2003M01 2012M12Lags: 2    

 Null Hypothesis: Obs F-Statistic Prob.               

 EX does not Granger Cause OP  115  1.70448 0.1866 OP does not Granger Cause EX  2.61122 0.0780

 INF does not Granger Cause OP  115  0.70705 0.4953 OP does not Granger Cause INF  0.52597 0.5925

 INF does not Granger Cause EX  115  0.91917 0.4019 EX does not Granger Cause INF  0.47069 0.6258

               15

CO-INTEGRATION – OIL PRICE TO INFLATION

16

Date: 04/22/13 Time: 16:30    

Sample (adjusted): 2003M04 2012M09    

Included observations: 114 after adjustments  

Trend assumption: Linear deterministic trend  

Series: INF OP       

Lags interval (in first differences): 1 to 2  

Unrestricted Cointegration Rank Test (Trace)  

Hypothesized   Trace 0.05  

No. of CE(s) Eigenvalue Statistic Critical Value Prob.**

None *  0.107121  16.32007  15.49471  0.0375

At most 1  0.029413  3.403341  3.841466  0.0651

 Trace test indicates 1 cointegrating eqn(s) at the 0.05 level

 * denotes rejection of the hypothesis at the 0.05 level

 **MacKinnon-Haug-Michelis (1999) p-values  

VECM ESTIMATESError Correction: D(INF) D(OP)

CointEq1 -0.136008 -0.215053   (0.04255)  (0.12746)  [-3.19644] [-1.68723]     

D(INF(-1))  0.196687 -0.022662   (0.09350)  (0.28008)  [ 2.10365] [-0.08091]     

D(INF(-2))  0.083269  0.018721   (0.09499)  (0.28454)  [ 0.87665] [ 0.06579]     

D(OP(-1))  0.007182  0.364237   (0.03169)  (0.09491)  [ 0.22667] [ 3.83752]     

D(OP(-2)) -0.005914  0.075168   (0.03196)  (0.09573)  [-0.18506] [ 0.78523]     C  0.036832  0.424848   (0.19651)  (0.58866)  [ 0.18743] [ 0.72172]

17

VECM OUTPUT – SYSTEM EQUATION

18

Dependent Variable: D(INF)    Method: Least Squares    Date: 04/22/13 Time: 16:32    Sample (adjusted): 2003M04 2012M09  Included observations: 114 after adjustments  D(INF) = C(1)*( INF(-1) + 0.0961409050262*OP(-1) - 19.2942120304 ) +

        C(2)*D(INF(-1)) + C(3)*D(INF(-2)) + C(4)*D(OP(-1)) + C(5)*D(OP(-2)) +        C(6)      

                    Coefficient Std. Error t-Statistic Prob.                    

C(1) -0.136008 0.042550 -3.196438 0.0018C(2) 0.196687 0.093498 2.103650 0.0377C(3) 0.083269 0.094986 0.876645 0.3826C(4) 0.007182 0.031685 0.226674 0.8211C(5) -0.005914 0.031956 -0.185056 0.8535C(6) 0.036832 0.196511 0.187432 0.8517

R-squared 0.106455    Mean dependent var 0.047368Adjusted R-squared 0.065087    S.D. dependent var 2.154033S.E. of regression 2.082755    Akaike info criterion 4.356456Sum squared resid 468.4896    Schwarz criterion 4.500466Log likelihood -242.3180    Hannan-Quinn criter. 4.414901F-statistic 2.573369    Durbin-Watson stat 2.008545

COEFFICIENT TEST – WALD TEST APPROACH

Wald Test:    Equation: EQN    

Test Statistic Value df Probability              

F-statistic  2.573369 (5, 108)  0.0306Chi-square  12.86684  5  0.0247

Null Hypothesis: C(1)=C(2)=C(3)=C(4)=C(5)=0Null Hypothesis Summary:  

Normalized Restriction (= 0) Value Std. Err.              

C(1) -0.136008  0.042550C(2)  0.196687  0.093498C(3)  0.083269  0.094986C(4)  0.007182  0.031685C(5) -0.005914  0.031956

Restrictions are linear in coefficients.

19

7.0 SUMMARY AND CONCLUSION The co-integration between oil price and inflation

variable exist at 5% significant level in the long run. For granger causality test, we found that the inflation

does not granger cause to the exchange rate but it does granger cause to the oil price.

The oil price does granger cause to the inflation but it does not granger cause to the exchange rate.

The exchange rate does not granger cause to both of the variables (Inflation and Oil Price).

So, the oil crude price can give an effect on inflation. If the rate of crude oil price changes, the inflation also changes.

The finding will contribute to Nigerian government in making policy towards crude oil price to avoid from the inflation.

20

THANK YOU FOR YOUR ATTENTION

21