THE MOST MERCIFUL 

 In the recent years, food price inflation has risen very 
  sharply  at  global  level.  According  to  Commodity 
  Research  Bureau  (2009),  the  overall  and  food 
  inflation  rates  at  global  level  stand  at  16.5  and  30.2 
  percent  respectively  by  November  06,  2007.  This 
  high food inflation persists in most of the countries in 
  the world. 

 Reduced  level  of  poverty,  increase  in  per  capita 
  income, urbanization and change in dietary habits are 
  the  main  reasons  of  sharp  increase  in  demand  and 
  prices of some basic food items. 

 Because  of  higher  food  inflation  households  have  to 
 make reductions in some areas of food consumption 
 leading to malnutrition. 

 Malnutrition  results  in productivity  losses  of  up to  10 
 percent  of  lifetime  earnings  and  GDP  losses  of  2-3 
 percent.                                                (Alderman, 2005)

 High  inflation  erodes  the  benefits  of  growth  and 
 leaves the poor worse off. 
                                         (Esterly and Ficsher, 2001) 
 It  hurts  the  poor  more,  since  more  than  half  of  the 
  budget of low wage earners goes toward food. 

 It redistributes income from fixed income groups to 
  the  owners  of  assets  and  businessmen  and 
  increases the gap between rich and poor.  
                                         (Khan et al, 2007)
 Pakistan has also experienced high food inflation of 
  17.5 percent and 26.6 percent in 2007-08 and 2008-
  09  respectively.  Moreover,  food  inflation  remained 
  more  than  10  percent  on  average  from  1972  to 
  2009, in the whole history of (West) Pakistan. 
     Historical Inflationary Trends 1971-72 to 2008-09
         (Annual percentage change, period average)
     Years               Overall CPI            Food CPI 
       70’s                 13.3                   13.8
       80’s                  7.2                   7.9
       90’s                  9.7                   10.1
    2000-01                  4.4                   3.6
    2001-02                  3.5                   2.5
    2002-03                  3.1                   2.9
    2003-04                  4.6                   6.0
    2004-05                  9.3                   12.5
    2005-06                  7.9                   6.9
    2006-07                  7.8                   10.4
    2007-08                  12                    17.5
    2008-09                 22.4                   26.6
2009-10 (Jul-April)         11.5                    12
                      LITERATURE REVIEW

 Two Schools of Thoughts (sources of inflation)
￿    Monetarist
     Friedman (1968, 1970 and 1971), Schwartz (1973)
     Demand-side factors (money supply, real money balances)

2.   Structuralist
     Sunkel (1958), Streeten (1962), Olivera (1964), Baumol (1967) and
     Maynard and Rijckeghem (1976).
     Supply-side factors (food prices, administered prices, cost of
     production, wages and import prices )
   Bhattacharia and Lodh (1990)
    Supported the strctulists model of inflation for India

   Balkrishnan (1992, 1994)
    Prices of food grains were determined by per capita output,
    per capita income in agriculture sector and government
    procurement of food grains.

   Khan and Qasim (1996)
    Money supply and wheat support prices showed positive
    relation with food price inflation and agriculture output was
    negatively cointegrated with food price inflation.
 Khan and Schimimelpfenning (2006)
 Found that broad money growth and private sector credit growth
 were the key variables of inflation in Pakistan. Support prices
 influenced inflation only in short run.

 Hasan et al. (2005)
 concluded that supply shocks (production of agricultural goods)
 have negative impact on food price inflation. Impacts of support
 prices of wheat and expectations were positive and highly
 significant on food price inflation. Money supply showed an
 insignificant impact on agriculture food prices
 Lorie and Khan (2006)
 Concluded that there is only a weak evidence of the existence of
 long run co integration between domestic prices, international
 prices and support prices for key agricultural goods in Pakistan.

 Dorosh and Salam (2006)
 There is little effect of increasing procurement prices from
 government on overall prices. In recent years, production short
 falls, particularly in 2004, and hoarding are the major reasons for
 price increases.
                  MODEL SPECIFICATION

 Economic literature on inflation provides some inflation models
  that incorporate the demand and supply side factors (Hassan et
  al., 1995; Khan and Qasim, 1996; Callen and Chang, 1999;
  Bokil and Schimmelfennig, 2005 and Khan and Schimmelfennig,

 Following Khan and Schimmelfennig (2006), the stylized hybrid
  monetarists-structulists model given below is formulated to
  capture the effect of certain demand and supply side factors of
  food price inflation in Pakistan.
The above quation can be rewritten for estimation purposes as follows:

t= 1, 2, 3, …., 37.
(time period ranging from 1972-2008)
FPIt= Food Price Inflation (CPI food as proxy of Food Price Inflation)
   in time t
FPIt-1= one year lag of FPIt(as proxy of inflation expectations)
M2GRt= Growth Rate of Money Supply (M2) in time t
PGDPt= Per Capita GDP(in Pak rupees) in time t
ASPt= Agriculture Support Price (rupees/40kg of wheat) in time t
FXt= Food Export (as percentage of merchandise export) in time t
FMt= Food Import (as percentage of merchandise imports) in time t.
 Stationarity and Non-stationarity
 A stationary series is time invariant and fulfills the properties of
  ‘constancy doctrine’ i.e. constant mean and constant variance
  and co-variance. In contrast, a non-stationary series violates one
  or more properties of ‘constancy doctrine’.

 Augmented Dickey-Fuller test was proposed by Dickey and
  Fuller (1979, 1981). It is widely used in economic literature to
  investigate the stationarity of a time series data. Dickey and
  Fuller (1979, 1981) have tabulated critical values for tδ which
  are called ‘τ (tau) statistics’. Dickey and Fuller unit root test
  can be applied under following two steps.
       The Augmented Dickey-Fuller (ADF) Test

 Step 1, OLS is regressed on the following equation and save
 the usual tδ values.

 Step 2

   The existence of unit root is decided on the basis of following
   H0 : for non-stationary if tδ≥ τ
   Ha : for stationarity if tδ < τ
               Johansen Co-integration Test

 Engle and Granger (1987) method finds out only one co-
 integrating vector through two step estimation approach.

 While on the other hand, number of vectors can be found using
 maximum likelihood testing procedure suggested by Johansen
 (1988) and Johansen and Juselius (1990) in the Vector
 Autoregressive (VAR) representation.
                    DATA SOURCES
Annual data from 1972 to 2008
Variables                         Sources
CPI food (FPI)                     Various  issues  of 
                                   Pakistan  Economic 
Agricultural support prices (ASP)  Survey

Per capita gross domestic 
product (PGDP)                    World Development 
Growth rate of money supply       Indicators (WDI) 
(M2G),                            online database by 
Food exports (FX)                 World Bank (2009).

Food imports (FM)
 Augmented Dickey-Fuller (ADF) Test at 1st Difference

  Variables             Trend & Intercept          Prob. Values 
       FPIt                    -4.0928*               0.0156 

      M2Gt                    -7.8567**               0.0000 

     PGDPt                     -3.4095*               0.0173 

      ASPt                     -3.7743*               0.0302 

       FXt                    -8.2416**               0.0000 

       FMt                    -6.0840**               0.0000 

 Note: * represents significant level at 1%. 
           ** represent significant level at 5%.
           VAR Lag Order Selection Criteria

  Lag         AIC             SC                  HQ 
   0        53.61601        53.88264        53.70805 
   1       44.30009*       46.16651*       44.94438* 
   2        44.46346        47.92966        45.65999 

* Indicates lag order selected by the criterion
 AIC: Akaike information criterion
 SC: Schwarz information criterion
 HQ: Hannan-Quinn information criterion 

 Same order of integration one I(1)
 Johansen co-integration
 Maximum Eigen Statistics
 Trace Statistics
            Unrestricted Co-integration Rank Test (Trace) 
   H 0       H1         Trace Statistics          0.05 Critical    Prob. 
 r = 0*     r ≥ 1            141.9786               95.75366        0.0000 

 r ≤ 1*     r ≥ 2            82.89489                69.81889       0.0032 
  r ≤ 2     r ≥ 3            45.08015                47.85613       0.0891 
  r ≤ 3     r ≥ 4            18.41380               29.79707        0.5356 

  Unrestricted Co-integration Rank Test (Maximum Eigen value) 
   H 0       H1            Max-Eigen              0.05 Critical    Prob. 
                           Statistics                Value 
 r = 0*     r ≥ 1            59.08367                40.07757       0.0001 
 r ≤ 1*     r ≥ 2            37.81475                33.87687       0.0161 

  r ≤ 2     r ≥ 3            26.66635                27.58434       0.0652 

  r ≤ 3     r ≥ 4            12.69969                21.13162       0.4803 
  * Denotes rejection of the null hypothesis at the 0.05 level
                      Long Run Relationships
Dependent Variable = FPIt 
    Variable         Coefficient       T-Statistic      Prob-Value 

    Constant          -44.90991        -4.941833          0.0000 

     FPIt-1           0.735522          15.78609          0.0000 

      M2Gt            0.073152          1.499076          0.1447 

     PGDPt            0.001740          5.343473          0.0000 

      ASPt            0.055197          4.131034          0.0003 

      FXt             0.479935          3.675908          0.0010 

      FMt             0.272316          2.384839          0.0238 

  R2= 0.9986            F-Statistic= 3656.589         Durbin-Watson 
Adj-R2 = 0.9984        Prob(F-statistic)= 0.0000         = 2.1329 
                    Short Run Relationships
Dependent Variable = FPIt
     Variable       Coefficient         T-Statistic     Prob-Value 
     Constant       -0.163500           -0.219200         0.8284 
FPIt-1               0.800563           3.808210          0.0009 
ΔM2Gt                0.059029           1.530154          0.1396 
ΔPGDPt               0.001114           1.477726          0.1530 
ΔPGDPt-1             0.000463           0.536597          0.5967 
ΔASPt                0.058287           4.446252          0.0002 
ΔASPt-1             -0.006688           -0.219333         0.8283 
ΔFXt                 0.354770           2.831904          0.0094 
ΔFXt-1               0.134124           1.919740          0.0674 
ΔFMt                 0.275443           1.794954          0.0858 
ΔFMt-1               0.009114           0.071144          0.9439 
ECTt-1              -0.991143           -3.614136         0.0015
  R2= 0.915113          F-Statistic= 22.54085          Durbin-Watson
  Adj-R2 = 0.874       Prob(F-statistic)= 0.0000           = 2.092 
                            Diagnostic Tests
Normality Test(Jarque-
                             Jarque-Bera Statistics 
Bera Statistics)                                     Probability = 0.4721 
                             = 1.5011 

Serial Correlation
(Breush-Godfrey Serial       F-statistics = 0.1859    Probability = 0.6696 
Correlation LM Test) 
(Autoregressive              F-statistics = 0.0147    Probability = 0.9044
Heteroskedasticity Test) 
Heteroskedasticity Test
(White Heteroskedasticity  F-statistics = 1.4383      Probability = 0.3075 
Model Specification Test
(Ramsey RESET Test)          F-statistics = 1.4383    Probability = 0.3744 
            Plot of Cumulative Sum of Recursive Residuals

The straight lines represent critical bounds at 5 percent significance level.
  Plot of Cumulative Sum of Squares of Recursive Residuals






                 86   88   90   92   94   96   98   00   02   04   06   08

The straight lines represent critical bounds at 5 percent significance level.

 In Pakistan, food inflation remained 9.9 % on average 
 during  the  study  period  (1972-2008),  some  time  as 
 high as 34.7 % in 1974 and 26.6 % in 2008-09. 

 First of all, stationarity of time series was checked by 
 using  Augmented  Dickey-Fuller  (ADF)  unit  root  test. 
 Results of ADF proved that all the variables were non-
 stationary at level and became stationary at their first 
 differences at 5% level of significance.

 As  the  variables  had  same  order  of  integration, 
 therefore  Johansen  co-integration  was  applied  to 
 find  the  long-run  relationship.  Both  statistics 
 (Maximum Eigen    statistics  and  Trace  statistics  ) 
 confirmed the existence of co-integration and same 
 number (two) of co-integrating vectors. 

 Long  run  coefficients  showed  that  the  impact  of  all 
 independent  variables  on  food  price  inflation  was 
 positive  and  statistically  significant  except  money 
 supply  growth.  All  the  coefficients  had  expected 
 positive signs. 

 On  the  basis  of  empirical  results  we  may 
 conclude  that  food  price  inflation  is  not  a 
 monetary  phenomenon  in  Pakistan.  While  the 
 supply  side  factors  or  structural  factors  have 
 dominant role in determining the food prices.

 In  the  short  run,  only  inflation  expectations, 
 support prices and food exports affected the food 
 price inflation. The negative value of coefficient of 
 ECTt-1,  which  is  (-0.9),  indicated  the  very  high 
 speed of convergence towards equilibrium. 
                     POLICY IMPLECATIONS
   Inflation expectations
      Continuity  and  consistency  in  government’s  economic 
       Strong  policy  statements  and  actions  will  help  to  dampen 
        inflationary expectations

   Support prices
     Government  should  pursue  a  moderate  policy  in  raising 
      support prices to slow down the inflationary  pressures and 
      maintain the reasonable production level of food grains
       Government  may  provide  subsidies  on  inputs  as  on 
        fertilizers, pesticides, diesel and electricity 

       Government should also encourage and support farmers to 
        adopt  modern  technology  for  higher  production  with  lower 
        production cost.  
 Economic growth

    Proper policy for agriculture sector to fill the output gap

    Credit facilities should be provided through formal and informal

    Improve infrastructure, agriculture markets and land ownership system

    Modern technology should be introduced to improve the production of
     food grains, meat, poultry and dairy products
 Growth in Money Supply

    Government should encourage the expansion in private sector credit,
     especially towards the agricultural and its related sectors

    There should be the availability and easy access of loans for all farmers
     for all types of their needs such as expenditure on the use of modern
     technology, inputs, marketing and storage facilities

    Increase in public expenditures on the provision of infrastructure for
     rural areas will also be helpful for optimal utilization of the potential of
     agriculture sector
 Imports of food

     We need to exploit our unrealized yield potential in production of food
      items as God has gifted us with all necessary resources.

     Sound agriculture development strategies and result oriented agricultural
      policies should be adopted by the government to produce foods in the
 Exports of food

     Government should ban the exports of food items until they are over
      and above the domestic needs.

     For price stability in the country, buffer stocks of essential food items
      like wheat, sugar and pulses should be maintained.

     There should be maximum control on smuggling of wheat, rice and
      live stock to neighboring countries through the coordination between
      all the stake holders

To top