0305-750x/X6 $3.00 + 0.00
World ~evelq~mrr~t, Vol. 13, No. h. pp. 757-762. 19X6.
Pergamon Journals Ltd.
Printed in Great Britain.
Impact of Agricultural Research and Extension Crop Productivity in Pakistan: A Production Function Approach MAHMOOD
HASAN
KHAN:‘: and ATHER
Simon Fraser University,
Rurnahy,
HUSSAIN
on
AKBARI
British Columbia
Summary. -This study analyzes the contribution that agricultural research and cxtcnsion made to crop productivity in Pakistan from lY.55 to 19x1. Using a production function approach. it &mates the (marginal) internal rate of return to these activitica. The cstimatcd rate of around 36% is in the rang reported for other countries. Considering this high rate of return and the low in Pakistan. the obvious policy public spending on agricultural rcacarch and extension recommendation is to allocate more :md higher quality rcsourccs t(l these activities to facilitate a rapid increase in agricultural productivity.
1. INTRODUCTION Several studies have shown that investment in agricultural research has a high payoff.’ However, in many underdeveloped countries, public and private investment in research on agriculture is accorded a low priority. This reflects the general neglect of Research and Development (R and D) as a “soft” area. There is thus a tendency to allocate minimal resources to those activities, such as research, education and extension, that generate new knowledge and transfer it to farmers in hopes of improving productivity and increasing profits. Many of these countries are dependent on borrowed technology from the more developed countries. This has been particularly true for several grain and some fibre crops. A few internationally-funded institutions are also involved in applied research on selected crops. Most underdeveloped countries. including Pakistan, are engaged, at best, in adaptive research. This study focuses on the contribution of agricultural research to crop productivity in Pakistan from the mid-1950s to the early 1980s. There are at least two good reasons for this study. First, while crop production in Pakistan has increased quite impressively, public funding for agricultural research has remained low. although there has been visible improvement in the funding picture in recent years. The second reason is that this issue has not been systematically analyzed with respect to Pakistan. Studies by Pray (1978) and Nagy (1983). based on the index
number approach, use data from the Pakistani Punjab for a few specific crops. This paper is divided into four sections. The following section gives a background on changes in crop production and public investment in agricultural research and extension since the mid-1950s. The third section discusses the methodology for analyzing the contribution of agricultural research to crop production in the period 1955-56 to 198&81. Sources and limitations of the data are also presented. The final section includes the results and their interpretation, with some comments on policy implications. 2. CHANGES IN AGRICULTURAL PRODUCTION AND INVESTMENT AGRICULTURAL RESEARCH
IN
Agricultural growth in Pakistan has been neither stable nor balanced over the last 35 years. Aggregate growth rates clearly indicate that agriculture stagnated in the 1950s. but experienced a somewhat robust growth in the 1960s. There was then a visible slowdown in the *The authors are grateful to the Pakistan Agricultural Research Council for data on expenditures for agricultural research and extension. They thank Nuzhat Ahmed, Faiz Mohammad and Mahmood Iqbal for their help in preparing an earlier version of the paper. Comments by two anonymous referees have also been very valuable. Needless to add. the authors alone are responsible for the contents.
7%
WORLD
early to mid-1970s. recovering That
in
with
the
agriculture
Iargcly
in
on weather
last two cotton
to three
years,
Punjab
vinces. particular
still
quently.
these
resources. been
Edible
Iargcly for
poultry,
LJneven of
the
is
growth of
more
serious
aspect
levels.
Khan
The
for
recent
years.
but
the
Punjab
rice
in
rainfed
yields
1Y5Os.
dramatically in
has
exposed the
kept
increased
III Sind.
the IY7Os.
or falling
in wcathcr
nology.
Cotton
Mexico)
pines)
have
undergone
It is well rice,
institutions
and
This ment
growth
was aided, of
in
us2 of fertilizer. and subsidies
in the absence sugarcane
of
Story
of
Icvcl
in
state
crops
have
Revolution (from
the
support
less
then
Domcatic one-quarter
mainly
although
in
has
research
in
through Council
of the gross
from
about
period.
other
IOSI
Pakistani
countries
amounted Rs.20
to
((IS
agricultural
country
of
3.
2.
one-tenth of
1%
put
it
Rs. IO0
(US
in
$7) peg-
hectare.
These
of public
and
ot
differ-
activitie
per
In of
on agricul-
these
allocations
Third
funds
cxtcnsion
in
;I
World.’
METIIODOLOGY
The
contribution in
changes
AND
productivity
can
that
these
limitations.’
and
at
DATA
tion
services.
The
that
of cotton
to evaluate search
in
in
number
uses
two
ways:
and
approach.
the second
reasons.
Second, parts
First.
final
method it
allows
allow
other
countrics
the
method
contribution The
comparisons and
also
to an
yielded
the production total
and perhaps this
Pakistan.
would
return
such as research
is that the
of
to
approach. Each of advantages and
to the mwrrgc
assign
developniiijor
estimated
study
research
single-enterprise)
of the rrlctrgirltrl rate of return
sources.
reason
(or
index
three
approach.
can
agricultural
function has its
as opposed
the first
tant
be
or
This
least
measurement input.
of
aggregate
production methods
complen~entary
$1 .50)
research
the
in Table
To
on
As
in the sonic
output
to about
the lowest
1%
one-half
extension.
spending
IYtiOs.
it incrcascd
about
over
<;ross
to about
for conipai-isons
shown
spent
just
and
public and
to
of agricultural
research
are among
1%
arc
and
value
in the early
figures
society
I’%, of its GDP the gross
the
1050s
of all crops,
of
increased
I ‘I/;, of
in the
interesting
with
Pakistan
value
research
I, ha
of
(
one-third
More
ently.
agricultural
in Table
of 1% of GDP
share
different
for
channclcd
Rcsuarch
one-tenth
Product
(b) the
(tubewcll)
varieties of
from
crops.
It arms
and
public
government
on
shown
had
conditions
prices
federal
expenditure
surplus
rt3earch
two
arc
on
government\.
in agricultural
(a) the
a key role in the
key inputs.
ettects
Public
for
these
entirely
Agricultural
tcch-
Philip-
seeds of wheat
agents)
in
due to of
Its role
Pakistan
lY5Os.
fluctuation\
by the rapid
of high-yi,elding the
or
cotton
the
the
education
of extension
funds
and extension.
has been either
poor
of
the
farm
poor
The
Pakistani
yields
for
fell
yield
(canal) and ground
water.
the
by international
of course.
surface
but
Pakistan.
to
creased
tural in
resistance
Wide
have played of
crops, and
level
Green
adapted
by local researchers. sustained
in
in
and rice
mainly
the
static
lY6Os.
Its
known that hybrid
produced
in
again bccamc static
the
in the
(i‘rom
Orr.o,r~ri
have been mainly
and
in
NWFP.
production
and sugarcane
if any share
for
alnlo\t
the years.
wheat
mainrice
wheat
disease
and
in the yield of sugarcane changes little
been
I_ack of impI-ovcment
the yield
over
rice
and
IRRI
and
stagnant
IY605.
of the Several
way since the
of weather.
remained
in the
have
local
is no less disturbing.
Punjab
static
the
to the vagaries
sugarcane
in the
particularly
of seed and limited
control.
The
(lY82)
has
year\
pesticide.
(PARC‘).
changes
in the
10705.
cx-
Public
and
depend
The
except for
research.
the provincial
recent
one side
Wheat
Punjab
technology.
have
IYHO.
rem;;ined
somewhat
in the
cotton
quality
the
and
Pakistan.
and
for
funding. through
of growth
trend
wheat
crops:
sector.
it i4 le\s apparent
‘of
services
C’onse-
institutions.
fertilizer
extension
in
and its
are carried
in these arcas.
agricultural
sector
is only
bc noted.
of
on the
of technology
state-funded
distribution
Therefore.
effect
crops. in Pakistan
does little
the training
predicament.
to
the
by
sector
(including
(lentils)
Siddiclui
mainly
private
national
attention
in
1950
of cotton
rose
most
3 similar
and
pro-
major
in ;i su5lainecl
dwarf
region\
The the
the
the
four
pulses
upward
tained
regions
sugarcane.
and source5
fact\ niav
IY6Os.
that
livestock
from
and
Frontier
is of uneven
have increased
Iate
in the rainfed
and
and
crops
interesting yields
provinces
The in
the trends
major
out
agriculture
yield four
of the
of crop output
problem
analyzed
by the failure
get the
oils
two
to farmers
on
ignored.
have had little
these
and adaptation
and
crops
of
Creation
and the agricultural
cotton.
levels
dissemination
to note
depend
rice.
dependent
incentive\
and
yield
in the
North-West
in general
wheat,
still
inputs
IYXOs.’
demonstrated
and Sind
of wheat and
It is significant
economy
is
first
momentum early
and
Pakistan
crop in the Punjab
the
cept
1070s
was well
then by the failure of
the growth
late
DEVELOPMENT
return
to
and extension
the most
impor-
has not been tried of
results
agricultural of
of research help
in
func-
draw
re-
this
\tud\
payoff
with
policy
in-
CROP
PRODUCTIVITY
759
IN PAKISTAN
Table 1 Grossdomestic producl (GDP), gross vulue of crops (GVC) cmd e,rpendirure.s on ngriculiurul resew& und exrrnsion in Pukisrurl, 1955-X 1 (in million rupees)
GDP GVC Expenditure research
First Five-Year Plan I YSS-60
Second Five-Year Plan 196&6S
Third Fivc-Year Plan 1965-70
79.703 23,083
101,131 26.97X
140.476 36.538
305.306 74.233
Fifth Five-Year Plan 197X-81 152,x2 I 33,017
on 3x.3
as “% of GDP as ‘% of GVC Expenditure extension
Non-Plan Period lY7&7X
’
X6.6
126.2
0.05 0.17
0.12 0.47
220.7
0.06 0.24
0.07 0.30
104.5 0.07 0.32
on 55.7
as ‘% of GDP as % of GVC
177.6
0.07 0.24
241.9
0.18 0.66
776.5
0.17 0.66
Source: Data for GDP and GVC are from Pakistan Economic expenditures on research and extension are from PARC. *All figures are in constant prices of 1959-60.
23x.2
0.25 1.05
0.16 0.72
Survey fY84-85;
data for
Table 2. Gross domeslicproduct
(GDP), gross value of agriculrural output (GVAO), number of farms, cultivared urea, expenditures on agricultural research and extension in Pakistan, 198041
Gross domestic product Gross value of agricultural Number of farms Cultivated area Expenditure on agricultural as as per per
output
research
per cent of GDP per cent of GVAO farm cultivated hectare
Expenditure
Rs. Rs.
on agricultural
as per cent as per cent
extension
billion billion million m. hectares million
0.05 0.17 30.18 6.05
Rs. 287.66
of GDP of GVAO
million
0.11 0.40
per farm per cultivated
Rs. 250.97 Rs. 70.35 4.06 20.25 Rs. 122.53
Rs. Rs.
hectare
70.85 14.21
Source: Data for GDP and GVAO are from Pakistan Economic Survey 198445 and Agricultural Statistics of Pakistan 1982.
plications for public spending on agricultural research and extension. The basic production relation can be expressed in the following form:
Q = A f(X. e),
(1)
where Q is the value of crop output, A is technical progress or the shift factor, X is the value of all physical inputs, and e is the error term. The underlying properties of Equation (1) are assumed to be of a linearly homogeneous production function, with constant returns to
WORLD
760 scale
and
time.
As
input.
Hicks we
neutral arc
Equation
technical
using (I)
progress
indice\
can
of
over
output
cxptascd
be
IlEVEI~OI’ML:N’I‘
and
procedure, without The
as:
we
dummy
“good”
and
IYSS-S6
to
xsumcd
We can then measure technical pi-ogre5s as a function of expenditure on agricultural research md
extension.
data.
from
Since
IYSS-S6
the effect
we to
of weather
are
using
time-series
IYX(~Xl.
w’c incorporate
on output
in the estimating
lies ‘‘It ‘3 5% line.
seems
i’r agricultural
sion,
and II,
“bad”)
effects
ycnrs.
of
The
(0,
research
and L>: arc: dummy
the
depict
pr~~ductivity
on agricultural
ix,).
over for
(3).
tivity.
11,. is estimated
total
value of cl-op output inputs. crops
output
is
Physical labor
inputs
include4 tubewell\ and were
not
aggregate IYSS-56 input
input
indices
provides
input
of have
in
and IYXt~Xl, land.
labor,
IYX2,
IYXS).
The ment
4.
expenditures been
data
provided
of
the
level
(\et‘
though
than
using
\iicIi
of weather
;I single
rainfall.
ax
prox)
It is
L’L’C’II
index
;I composite
rainfall, Khan
admittcrl-
the effect
humidity
and
Siddiqui.
of and
IYE).
AND
IN-I‘ERPRETATION
labor, into
=
100.
R.
is the total
not
PARC.
develop-
agricultural
by all lcvcls
values arc
for
on
but
of into
research they
of coefficients
bv the OLS
in Equation
method.
high :ILltocorrc‘l~,tic,n was made by using
We also wanted
to analyre
(3)
In view of the
(D.W.
=
0.50).
the GLS
method.
the separate
and joint
of research
and exteri\ion on agricultural productivity by estimating the coefficients of these variables In two variants of Equation (3). In the
first
variant,
included research to
be
them
as ;I single
highly
extension
The
variant variable. Since
(joint) we
were
are
of the second
were
second
expenditures
correlated,
results
and
variables.
and extension
estimation tion
research
as separate
found
reporting
variant
the
of Equa-
(3):
A, = 106.254
+ 0.0064XR
(9.067)
+ X.SXlfI,
(3.626)
-
(2.7XX)
have
In the estimation
= 1.9 I; figures
R’ = 0.60; D.W. are
‘t’
I 1.741111 (3.377)
in parentheses
statistics.
All coefficients in the equation are significant at the YY’% level and their signs are as expected. The coefficients of good and bad ycarb clearly indicate that growth in agricultural productivity has been significantly affected by weather in Pakistan. The Durbin-Watson statistic is close to 2.0 or within a reasonable range. The
marginal
value product
and extension
each year.
converted
Data
published.
output,
(Pakistan.
on
incurred
and
1Y7Y-X0
data about capital
IYSY-60.
who
output
years,
tried
For and
Wizarat’s
the
estimation
impact
an
estimation
and
lYX(~XI.
by the
(19X1).
The w’as first
data
and
additional
current of
are
the
followed
expenditure
to
rupees
and
dw,elling4
land.
of
constructing
and extension
IYSS-Sh
constant
Farm
1YSY-60
then
variable.
The
stock
tractors
by Wizarat
fertilizer,
government.
Capital
arc convert4
the
two
current
time
RESULTS
entered
area).
because
values
with
v;~Iuc
IYSY-60. in agricu-
etc.).
based on official
research and
research from
for
con-
WC have used the output
We
indice\
stock.
stock
constructed
methodology
of
(cultivated
public).
index
IOO. The
population
The
details
methods.’
land
and
to lY78-7Y
is
=
are excluded
and capital
index
rupees
(cattle.
available.
fertilizer.
of
IYSY-60
active
(private
index
value of all majoi
animals
implements
the
output
and capital
draft
produc-
the total
include
fertilizer.
;I
The
of the interval
reasonable.
to construct
tcmpcraturc
and
by the index of calut2 of
constant
(economically
ture).
.$qricultural
The
with
in
output.
which
(“good”
by dlvlding
by concerting
and minor
crop
on each side
interval
including
observed
of
of
ill-c
below
dummy
Equation
physical
Ijne
wcatber
difficult
correction
all
during
yeaIF
abo\,c and
way of capturing
output
variables
structed
represent
bad
and
the choice
to bc ;I more
variable\. to be explained shortly. provide an indirect method of isolating the cffcct of wcathcr on crop output. We should now explain the specification of in
I).
output
and exten-
variables
abnorn;al
of K but
arbitrary. it i5 similar to the con\eIifor confidence It’vcI\ in statistics. It
I\; indirect.
/i is
tlenti
While
wcathcr. il,
Iving
trend
usctl
anti crop
be those the
is sonicwhat
I),
(kXKi
b3nd ‘
more
values
for
IYXO-XI.
from
on
expenditui-e
years
hand
variable
where
variables.
“bad”
to
tion
equation:
the cumulative
LISC
a lap.
marginal search.” long-run return
is Rs.37.54.
return The
one
estimated
marginal from
to
rupee MVP
product,
one rupee
(MVP)
which
of research
reflects invested
approximates
or the expected
invested
a high in
rethe total
in research over time. To cstimatc the (marginal) internal rate of return (IRK) to agricultural research in Pakistan.
CROP
PRODUCTIVITY
a lag structure of an inverted “V” type is assumed.’ The estimated marginal product would then be an approximation of the area under the inverted “V” lag. The calculation of IRR requires that the future returns be discounted by a rate of interest that makes the present and future returns equal. Assuming a lag structure of IO years. with the high point of the inverted “V” in the fifth year, the marginal internal rate of return is about 36% for agricultural research and extension in Pakistan. Lag structures of longer duration for crop research have been assumed in the United States! The shorter lag assumed in this study seems reasonable because (a) almost all agricultural research in Pakistan is basically adaptive and (b) it is consistent with other studies. The estimated (marginal) real rate of 36% on investment in research and extension in Pakistan compares favorably with the payoff reported for other countries.” The policy implications of our study are clear: that more investment should be devoted to activities like agricultural research, education and extension. This is narticulnrlv important for those crops of whi& the yielii levels are still very low and an be increased by
IN PAKISTAN
761
introducing new seeds or varieties to farmers. Significant yield improvements have been experienced only by wheat. rice (IRRI) and corn. A second and related aspect of increased investment in research and extension is of qualitative improvement in delivering research output to farmers. Recent studies have revealed serious problems of fragmented and largely uncoordinated research and extension services in the country.“’ A major reorganization of the agricu!tural research system. at least at the federal level. and the new experiment of extcnsion by training and visit (T and V) in selected areas of the Punjab and Sind provinces are two of the policy changes introduced since IWO. There are still major problems of inadequate allocation of resources to research in the provinces, lack of coordination of research and extension services. and poor quality of extension agents used on the front line. It is Important to note that the much expanded activities of PARC at the federal level are seen with great skepticism in the provinces, where most of the regional agricultural institutes wait for funds and resources to filter down to their level.
NOTES 1. See Evenson (IYX4). (lYX2). pp. 242-243. 2.
See Pakistan
3.
See Daniels
(1985). and
pp
358~359
statistical
Nestel
and
Ruttan
appendix,
p. 22.
(1981).
4. Two excellent reviews are given in Norton Davis (1981) and Schuh and Tolini (1979).
and
5 Wizarat has not included fertilizer as an input. She gives no reason for this omission. WC have also reconstructed the value of capital stock for the whole period, because the livestock population reported by Wizarat does not match with numbers given in various census reports and other published documents. It seems that the numbers of draft animals in Pakistan have fallen because of the rapid spread of tractors in recent years.
6. Marginal Value Product (MVP) of rcscarch extension is estimated by the following formula:
and
u, = d(QlX)ld(R) d, = 1/X dQldR rr,.X = dQldR. 7. Evenson’s (1967) work is still a standard on lag structures. 8.
See Bredahl
9.
See Ruttan
and
Peterson
reference
(IY7h).
(19X2). pp. 242-243.
10. Studies by the World Bank (1981) most relevant here.
arc
perhaps
the
REFERENCES Bredahl. M., and W. Peterson. “The productivity and allocation of research: U.S. agricultural experiment stations,” Amrricm Jortmul of Apkdmrrrl Ecortomic.s, Vol. 5X. No. 4 (1976): pp. h8&hY2. Daniels. D., and B. Nestcl. Reseurch AI/m&m 10 Agriculturul Research (Ottawa: International Development Research Ccntre. lY81). Evenson. R. E.. “The contribution of agricultural
research to production,” Journcrl OJ Furm E‘cot~ omics, Vol. 49, No. 5 (lY67), pp. 1415-142.5. Evenson, R. E.. “Benefits and obstacles in dcvcloping appropriate agricultural technology.” in C. Either and J. Staatz (Eds.). A~riculrural Drrvlormwrrr in /he Third World (Baltimore: The Johns Hopkins University Press. lY84). pp. 34X-361. Khan, M. H.. and A. S. Siddiqui, “Growth and
762
Nagy, J. C.. high-yielding
WORLD
DEVELOPMEN’I
“Estimating the yield advantage ol wheat and maize: The use of Paki\t;m’s
Norton. Cr. W.. and J. S. Davis. .‘Ev;d;atjt;g returns to ag,ricultural rescarch: A rwiew.” Amcv+c~~r Jorrn~rrl o/ A~ridrurul Ecor~otmc~.~. Vol. 63. No. 4 ( IYXI ) pp. 6X5-6”)!, Ministry of Finance. f’crlti.srtr,~ Ewrrornic Pakistan, S~rrwv fY868.5 (Islamahad: IYXS).
Pakistan. Ministry’ of Food and Agriculture, /2gricd /uru/ .Qurivdc.s of Pukisrurz /‘IN2 (Islam;rh;d: IYX3).
Pray, C. E.. “The economics 01 agricultural rcscarch in British Puninh and Pakistan Punish. lYOSp7S.” Ph.D. Dissertation (Philadelphia: . University 01 Pcnnsylwnia. lY7X). Ruttan, V. W.. Agricrtlrwrrl K~w~wchPolic~y(Minneapolis: University 01 Minnesota Press. IYX?). Schuh. G. G.. and H. Tolini. “Costs and benefit\ 01 agricultural rescarch: The state of the arts,” World Hrrr~k S/off’ Workirrg fcrper. No. 3hO (Washington. D.C.: The World Bank. 197’)). Wiarat, S.. “Technological change in Pakistan‘s agriculture: IYS3-54 to lY7X-7Y.” ffrki.\/crrl Ucvdoprw~~~ R~vicw’. Vol. 20. No. 4 (IYXI), pp. 427-445. World Rank. I’dktan Stuff Appr~kd Rqmrl of 111c~ A,qricrrlrrwd Rtwcrrcl~ Projcc~ (Washington. D.C.: IYXi).