Growth and instability in crop production in Eastern India

M a n a b e n d u C h a tto p a d h y a y C h ira n iib N e o gi S a n a t K u m a r M a it y Indian Statistical Institute. Calcutta

1. In tr o d u c tio n

T h e states of eastern India (viz. Assam, Bihar, Orissa and W e s t B e n g a l ) are considered by and large the poverty-ridden r egi ons of India. A b o u t 30 per cent of the total rural poor live in these f our states. A m o n g these four states, the incidence of rural p o o r is hi gh in Bi har ( 14 . 9 pet cent) and tow In Assam ( 2 0 per c e n t ) . It may be noted in this connecti on that the introduction of n e w agri cultural t e c h n o l og y in different parts of India since the m i d - s i x t i e s have i mp r o v e d the pr oducti on conditions and conseq­

u en t l y the i nc ome and empl oyment and yet the states of eastern I ndia fall b ehi nd the overall national proformance. For example, t hese states have experi enced drastic falls in per capita production a n d yi el d per hectare d u r i ng the current decades. These states ha ve recor ded drastic decline in gr oss cr opped area per person, but h a ve an i mpressi ve increase in the number of workers per hectare.

T h e s t a t e - w i s e annual g r o w t h rates in the food grain production b e t w e e n 1971 - 8 5 fell behi nd the rate of g r o w t h of population in all these states. H o w e v e r , these states wi t h relatively lower level of

net irrigated area have re la tive ly higher cro p p in g intensities.

W i t h this b a c k g r o u n d of producti on conditions in Eastern I nd i a this p aper attenriDts to examine empirically whether the pattern of g r o w t h in c r o p producti on in some states of eastern Indi a ( v i z . Bihar, Or i ssa and West Ba n g a l ) has recorded any

significant change dur i ng the period 1950 to 1988. In the ra n growth, h o w does one explain the year to year flutuation resul thereby in instability in agricultural producti on ? T hi s is a sub, wh i c h has received w i d e r attention in the countr y in the r e o years in the context of agricultural devel opment strategy across t technologically l aggi ng regi ons in particular.

T h « study is di vi ded into three sections. In Section I, w examine the g r o w t h rates in producti on of foodgrains, its c o n s t i ­ tuents and other mai n cash crop (jute) bet ween different periods along wi t h the cor respondi ng rates of g r o w t h in area and p r o d u c ­ tivity. Section II deals wi t h the measure of instability and the trends in instability for the period 1 9 5 0 - 8 8 . In Secti on III, w e discuss the relationship between g r o w t h and instability for the regions under study and mske some c on c l udi n g observations.

I 2. G r o w t h o f A g r i c u l t u r e 2.1 M e th o d o lo gy

For the purpose of analysis of secular trend to the statistical series of agricultural production, yield and acreage for the period 1950- 51 to 1987- 88, w e have fitted three different curves : nam ely the straight line of the form fx ( t ) a -f- b t where 'a' p o sitiv e and 'b' implies a diminishing rate of gr owt h ; the semi - l og cu rve of the form f j ( t) = a (I + 0* i mplying a constant rate o f g ro w th r ; the

t

Gompertz curve of the form f3 ( t ) = ebc w h i c h al l ows for both a diminishing rate of gr owt h when b < 1 and c < 1 and an in cre a sin g rate of gr owth wh en b > 1 and c > 1. T he goodness of fit of a curve has been measured by the statistic D, where D rep re se n ts the ratio of the Residua! Sum of Squares (i.e. the sum o f sq u a re s of deviations of the observed values from the trend v a lu e s ) to the Total Sum of Squares of deviati ons o f the observed values from the trend values to the Total Sum of Square-

2 ( y t - y t* ) 2

of the observed values, i.e., D — --- , y t =

observed

2 ( y t - y ) 2

value, y t* trend vafue, and y = mean value of the y' s. Thus, out of three alternative forms of cur ves, on e must choose that form w h i c h bears the l owest value of D, i ndicating thereby the good­

ness of fit on the fitting of trend cur ves1.

2 2 S e c u l a r T r e n d 1 9 5 0 - 51 t o 1 9 8 7 - 8 8

Presented in T a b l e 1 are the estimates of the best fitted trend equations for some maj or c o p s obtai ned separately for Bihar, Orissa, We st Bengal and India as a w h o l e over the period 1950-51 to 1 9 8 7 - 8 8 .

W e observe from Tabl e 1 that as far as the criterion of lowest D value is concer ned, trends in the gr owt h of acreage, out­

p u t and yie ld for f o c d cr ops as a wh o l e in the three states of eastern India give t w o distinct patterns for period 1950-51 to 1 9 8 7 - 88 . Thus, in the cases of Bi har and Orissa, linear curve q ive s the bast fit for all the variables ( except yield) under study w h i c h is consistent wi t h the retardation hypothesis. On the other hand, s e mi - l oga r i t hmi c curve provides the best fit for all the vari-’

a b le s in tha case of We st Bengal w h ' c h confirms that the growth of acreage, output and yi el d of f ood crops remained unchsnge dur­

ing the entire per i od of 38 years. Interestingly the results Gbtained for West Bengal confor m wi t h the al l - I ndi a trends in the growth of output and yi el d in particular.

A b o v e observati ons thus sugg est a stagnation in the growth of foodgrains ou t pu t in India mai nly because of the diminishing

1. This m e th o d o lo g y was adopted b y Rudra (1970, 1982) based on earlier data o f agricultural p rod uctio n and was strengthened b y us through som e fresh results ( Chattopadhyay and Bhattacharya 1987)■ To the best o f our know le dge, n o b o d y has till n o w challenged the test based on the use o f ‘O ' statistic.

rate of growth of cropped area. Al t hough We s t Bengal gives the similar picture as in the case of India, the other t wo states of Eastern India (viz. Bihar and Orissa) strongly differ from the all- India trend. T h e performance of agriculture in these t w o states is not satisfactory in terms of acreage, output and yield of food crops in general. T o take a v i ew in this matter a little more, an analysis of data on acreage, output and yield separately for each major crop seems to be very important.

It is found from the same table ( Tabl e 1) that in the case of rice which is the main crop in the states of eastern India, linear trend ( wh i c h implicity assumes a declining rate of g r o w t h ) gives the best fit for all the variables and semi log trend ( w h i c h i m p l i ­ citly assumes a constant rate of a r o w t h ) gives the best fit for Orissa and West Bengal for all the variables except acreage in the case of West Bengal. G r o w i h of acreage in the case of rice for West Bengal seems to be declining as the Gomper t z curve of the type b < 1 and c < 1 gives the best fit. It may be noted that the all-India results for rice are fully consistent wi th the trends of West Bengal and partly of Orissa but not of Bihar A l l - India trend results relating to rice s h o w that the g r owt h of acreage has been declining and the gr owth of pr oducti on and yield r e ma i n ­ ing unchanged during 1950- 51 to 1 9 8 7 - 8 8

Coming n o w to the next most i mportant crop wheat, w a find from the same table that while the g r owt h of acreage and o u t ­ put for whe^t in Bihar and Orissa remained unchanged, the growth of yield rate registered a d e f i n i n g trend. T h e reverse ie true in the case of West Bengal. All - I ndia results, however, largely differ from the results obtained for the states of eastern India. It is seen that at all-India level, growth of wheat producti on and yield remained constant and the gr owt h of acreage of wheat decl ined during this period.

An important commercial crop cultivated in eastern India is jute. Declining rate of gr owt h of yield for jute is a p he n ome n on r ot only for the states of eastern India but for.the country as a wh o l e (Table 1). As a result, although the growth of acreage for jute

r emai ned u nchanged o v e r time, the g r o w t h of output for jute did not register any u p w a r d cha n g e due to the di mini shi ng rate of g r o w t h of yi el d rate.

A b o v e analysi s relati ng to the g r o w t h of agriculture for some maj or crops in India in general and eastern India in particular rule out any transformati on in the pr oduct i on condi ti ons. W e do not f i nd any striking result for any cr op or for a n y region under study w h i c h pr ovi des suffici ent g r oun d for suppor t i ng breakthroughs in agri cultural p r o d u c t i o n. T h i s amou n t s to the rejection of the pro­

posi t i on that there has at all been a ny green revolution over the co u n t r y . In order to exami ne this pr oposi ti on a little more, anal ysi s of data separately for the t w o periods- namely, the pre­

green revoluti on per i od ( 1 9 5 0 - 5 1 to 1 9 6 6 - 6 7 ) and post-green r e vol ut i on peri od ( 1 9 6 7 - 6 8 o n w a r d s ) w o u l d be useful. 3

2 . 3 G r o w t h o f A g r i c u l t u r e B e f o r e a n d A f t e r A d o p t i o n o f N e w T e c h n o l o g y

T r e n d results for the first per i od ( 1 9 5 0 - 5 1 to 1 9 6 6 - 6 7 ) , pre­

sented in T a bl e 2, s h o w that the trends in the g r o w t h of acreage, o u t p u t and yi eld for f o o d cr ops as a w h o l e are more or less con­

sistent wi t h the trends for the entire peri od (i.e., 1950- 51 to 1 9 8 7 - 8 8 ) for ali the states under st udy. T h u s , for food crops the g r o w t h of output and yi el d decl ined dur i ng the pre- green revolution per i od in the cases of Bi har and Or i ssa wher eas the g r o wt h of acreage remai ned c o n s t a n t for these t w o states. Al l - I n d i a trend results are. ho w e v e r , consistent wi t h the hypothesi s of constant rate of g r o w t h of agri cultural output, yi el d and acreage during the 2. We have d e vid e d the p e rio d into tw o sub-periods in terms o f an overall n o tio n o f spread o f n e w agricultural tech­

n o lo g y. In fact, this division m a y n o t be applicable for a ll crops a n d fo r a ll the re gions o f India . We g e t such an im pression w h ile w e attem pt to examine the instability questions o f the variables w h ic h have been discussed in the fo llo w in g section.

pre-green revolution period and the We st Bengal results bear very close resemblance to the all India results.

Comi ng n o w to the results for the green revolution period (i.e., between 1 9 6 8 - 69 and 1 98 7 - 88 ) , w e find that the same retardation hypothesis recording the gr owt h of output and yi el d for food crops can be accepted unambi guousl y not only for Bihar and Orissa but for West Bengal as well. As a matter of fact, in the green revolution period, straight line provides the best fit in most of the cases of gr owt h of acreage, output and yield for food crops for all the three states of eastern India. For all- India, no difference can be made between the p r e - a n d post- green revolution periods in terms of gr owth of acreage, output and yield for fond crops

Thu s it is clear from the above analysis that there is no major changes in terms of growth pattern wi t h the introduction of ne w agricultural technology for India in particular and also in the cases of Bihar and Orissa. T h e break however coul d be located in the case of West Bengal but onlv in the case of producti on and vield of food crops. But this break does not register any u p w a r d chanae of arowth of output and vield of food crops All these amount to acceptinq the proposition that wi t h the introduction of new technoloay in agriculture, there has not been any structural chana* in the composi tion of foodgrains producti on in eastern India as in India as a whole. B^caus^. wi th n e w t echnol ogy either the trend fitted over the entire period wo u l d s h o w an increasing rate of growth or there w o u l d have been a break in the trend calling for t wo separate trend curves to be fitted to the t w o periods.3 N o n e of these possibilities were supported bv the fitted trend' curves for eastern India as well as India as a whol e.

Although the gr owt h of agriculture for f ood crops reveals a smooth trend so as to rule out any transformation in the p r o d u c ­ tion conditions, a crop- specific analysis mi ght have to be a s o m e ­ what different picture. Our crop- specific trend results, presented in the same tables ( Tabl es 2 and 3), once again rule out any

"• '

3. For an extensive discussion of this p o in t see, Rudra (1 9 8 2 ).

t ransformati on t o wa r d s the incraesing rate of growth of agriculture.

In fact, some breaks in the ti me trends coul d be located in the cases of wh ea t and jute for West Be n g al , rice for Orissa and wheat for Bi har. But in no cases, the breaks indicate the positive struc­

tural transformati on in terms of increasing rate of growth. Similar c o n c l us i on can be d r a w n regarding the crop- speci f i c trend results for India as a wh ol e.

T o sum up, agri cultural g r o w t h in the three states of Eastern India and India as a w h o l e , speci ally foodgrains, has been either stagnant or v e ry mar gi nal over the per i od 195 0 -51 to 1987-88.

T h i s is clearly due to the fact that the rate of growth of cropped rre ? has been d e cl in i n g over time. Between the three states, W e s t Bengal s h o w s relati vely hi gher overall rates of growth in 38 years in f o o d - g r a i n s pr oduct i on ( 1 . 7 8 per cent per annum)..

Or i ssa s h o w s relatively poorest rate of g r owt h ( 0.55 per cent per a n n u m ) S o m e w h a t different pictur e is observed in the case of rice pr odu c t i on in these three states. Thu s, West Bengal registers hi ghest rate of g r o w t h ( 1. 7 7 per cent per annum) whereas Bihar registers l owest rate of g r o w t h ( 0 . 6 7 per cent per annum) in rice p r o d u c t i o n in 3 8 vears. So far as n o n - f o o d - g r a i ns crop jute is concer ned, this s h o w s a sustained decrease in the rate of growth in area and yi el d in r espect of all the states. On the whole, it is clear from the analysi s that the n e w t echnol ogy had hardly any i mpact in the aqrarian sector of eastern India. Thi s is mainly due to the lack of assur ed irrigation t h r o u g h o u t the year in these rainfed

a n d d r o u g h t pr one states. In fact, this g r o wt h of agriculture in t hese states is suspected to have been accompani ed by ari increase in the i nstabi lity in agri cultural producti on wh i c h has been b r o u g h t about by the uncertain natural conditi ons particularly in v i e w o f the absence of wi d e s p r e a d use of new agricultural t e c h n o l o g y .

In the light of the pr obl em of agriculture in eastern India d i s c u s s e d above, in the f o l l o w i ng section, w e attempt to measure the extent of i nstabi lity in cr op p r o d u c t i on jn these states.

I I

3. In s tab ility in A g ric u ltu ra l P ro d uctio n 3.1 M e th o d o lo g y

A good number of studies have been made on the time trends in the gr owth of Indian agriculture and on inter-regional variations in the growth of crop production by numerous researchers. H o w ­ ever, not much wor k has been done so far to analyse the g r owt h of agriculture along wi th the fluctuations of producti on, area and yield of different crops at the regional level. Some of the recent studies whi ch are notable in this area are Mehra ( 1 9 8 1 ) . Hazell ( 1982) . Ray ( 1 98 7 ) , Dh a wa n ( 1987) and Boyce ( 1 9 8 7 ) . Most of these studies are at the all-India level and suffer from some serious methodologi cal problems.

The measure w h i c h most of the researchers used in a n a l y­

sing instability in crop producti on was the coefficient of variati ons ( c v ) whi ch is defined as the standard deviation of output from its trend divided by the mean level of output ( M e h r a 1 98 1 ) . T h e standard errors of the simple ( di scon t i n uou s) g r o wt h rates we r e sometimes used as a measure of instability. A n alternative methodology of assessing changes in instability over time suggested by S.R. Sen ( 1 9 6 7 ) is the 'trough' and 'peak' method.

A crucial dr awback of meansuring instability by "estimating the variance or coefficient of variation of the variable ar ound the time trend for t wo separate periods is that it fails to capture the large and frequent fluctuations of the variable concerned. Agai n, this measure shows o n l y the extent of instability of the variable, and does not give any impression about the trend of instabi lity That means, the question as to whet her the fluctuations are s t a b i l i ­ sing over the period or the system becoming^ more unstable, can not be revealed from this kind of measure.

EstimatesofBest FittedTrend Equationfor Different Crops Obtainedfor Different Regions: 1950-51to1987-88

4-*a)

3 CO

“ )

(AQ.

O o

~o LO

o CA

LLo a o <

w o

**

CD0) o'—s _C

§

ao CO DC

■aCD (A CO % E (D

4—' E CM

CA CO UJ

-C CO a .

4—■

§ c _a>

o JD

*0> CO

o 5

cc >

CM CM

05 f-

CM CM CD CM 00

^ I

CM

O

LO CD CO

COCM

oo

CO * -

a>

COo o>

lO

G i

h s CM CM <3 *

o o CO

CO CM

O rt*

0 0 T -

CO

^ CO CM CM CO

w T O.

CO 1-’

CM T— *—

CO O00 LO t -

r* cm CM oCO

q q co co

CM

to

00 00

00 co

LO

co -Q o co -Q o

.

<0 S a 5 ca <

co

+»o 3

T5O 2

<1?

615.4918343.5849502.9783792.8884 10.471433.427515.05519.7844

8.28351.26484128.73713.7281 1.00321.090777.54251.0022

CM CD 03 o in

o> «- 00 o ID r-

CM CM

LO in

r~

03 CD CM t-

O)

CO CM O 03

0)03 0300

CM

00 t- 00 CM 0 0 T—

CO o

co

00 00 00 JN 00 00

w

00 00 CO

w oo in00

<oin

CM

t— o CN 03

o CM

CO 05 00 CD

COin

o

a>

<8 a to eg

v> 0 O <

co

«•«u

■a3

o ²

£

Contd.

West Bengal Acreage O w CD£

00

CM LO r“

00 o cm N.

CO CN o

CO CO CN CO o

oo CN

LO CN CM 00

CO o

T_

—

CN LO o ! CN 00 o

CO oo'

CM in CM 00 CO r - rs o CO * -

CD

CO r— CN r^* <x> CO

00 O) CN 0) o

/— S

CN ,

CO CO

CN o 1

CN 1^ CO T“

CO r- w

w'

r— o o

00

o CO CO

s. o to <J) LO

CO o CO

'— ' o CO O) CN CM o o s-" 00

LO CM § w CO CD

CO

CDio

c o o=3

“D C

2a>

oo ir>

___ o CD N O ) O

U, «t o

CD ^

Oo co

CM CD

^ °. <R

CO CO O W (£>

^ CO _ CO CO

■2. S

00

CO

CN

O OCO

£ © m

T3 o>

C co

— ©

I I

CO O O )

CO CM CO

o CM CN

0) ©

o CM

T—

00 o 00 ^O)

o CO O)

CO 00 CO to

(O ® 11 - s CO o lo’ o

1

10 CN CO00

00 CN LO

00 00 CO

CO CN *t CN

°? 1 ^{o o}

^ 1 o T—

o00

c

.o 4->

u3

•D O

o>

CM

CN 00 O 00 LO CD 0 CN

LO

00 o

CO CM O fM CO * -

£* CO in cm 00 ooo s r ® -

CM 05

T -CO o

CO - T“

CD JD O

203

<D • _ C *fc- +-* CO

2 3 .1 A

TJ E

C 0

— CO Ui

o

“O c CO

© CM —

£ ~ A

~ ©

-Q c -Q

c — sz

il O

3 , ’5 O ) *-

— "►-»

CO

<D 3

CO ZJ o UrnCO

>

■O c

00 ^ TJ w

c (0 ^{. .} CM CO

(1)

S-g ■ ?

15 O ^

CD ^ . I— {/) * - w c

? .2

W -O

c

<o

© -O

CO o

(0 C D

CO TJ

c

©c

© -O u> C ■“

JD © _C

=> £ t- «n V o

CO©

>

5o

V (0 .9-u c c CL

<o

*o § c o © N n 5 4-*■*—* +-* 4-J H— © o

© 4—* Q. 3 -C4-1

o CO© -Q ^Eo

TJO k.

■o © O Q_

cCO-C_+-*

*4— s T3 C

o CD

IE eh»

o CQ

+-J E ©

c oH- ^-C <

©o w 3 O c/>

CM

a« ow r*

C_> co c ®I

£. o>

a> *-

£ o 5 ^

w in o I

>*- o e 10 I ?

**(0 ••

3 <0 CT C

Ul UJ O - J -o a>

m c o

< £ DC -

"O ®

*>® £ O

~ Q V) u ai ® oa

o a w c

® 5 2 -Q E o +j

w U l

TJ »- at

® ®

* ® E ICO 4^ »—

(A CO UJ CL

5 .

CO S ’ -

© a CC >

CO

in

co

CM

00 *— CO

o o t o eg

C4 o | CM IO o

>mK>

in r - CO r*

^ S

^

n r • O) *-

CM rs co

o CD «N CM °

CO •“

O 00 cm a>

co in lO O) a i co r> •

*—■ io

O CM CO CM co CO CO If) CO T“

CO CO O i CO

N “ > 9

w U)r

o **

as o 05 to co oo

oo oo

CMO)

A a

,

n S

£ o ta <

co u3

■oo

ooo o>

CO CO CM CO ai ui f** T—

I''

£ °>

CO oo

r: co

<*? O) S “>

CD r - 't

V*>

c0 o

o> O) 00 r-

00 f"

oo

00 *-CM

CMin o 00 co 00 CO I to ^ r> *-

lO

-Q O ^{S i o}

2*5

>

Orissa Acreage

#-,s co

<N o

LO Oo CO r -

CM

O CM CD f - r - CN CO O LO

CD00

CD CNCM

<D CM O00 00 c6

CM CO cd in

^ CO T - CM CO o

^ 03

COCO o0>

oCD

oo

oCD CO LO

00 00 K

r^*

CD CO o CM CM

<* CM

r - 00

^ r - CM CO O

r - S2CD

,-s 00

cm O

oCNI O )o CMK

LO

LO O )CO o5

00 T t

^-v LO 00 CN CM O w CM O 0 0 r -

O00 ^CN 00 00 o w oo oo CO r-

CN ot - O) o> oo r~ O

<6

(D .Q _{CD XI}

O

c

o o 3

“D

O 2

©

£

West Bengal Acreage

CO CM

OCO co

CM

cn

oo

CO

Ko

CN CO 00

CD O

oo o

CO q coco LO

coo lo

o oo o

CM

O O)

CO

00

CD CM

CO CO00

oo

o m 00 oo

CN O

o o

LO

o <-

CD CO

<J)

in cn

co o

CO

CM r - CN

cn co

*- o

<o *~9

CO

o

CD oo CO

^ CN CN CD00

CO

CD o

CN eg

w 00

CD

^ CM

CN CD

w 00 CO

00

CO

oo

CM

Contd.

OCD T~00 CO

CO CO

CDO

oo cooo

o CN

CO CD00

t— CD o

r - O

LO00 0000

LO CDO o

CD * -

CO

oo oo

CN

00

CN 00

<3* CD

o rs o

<fr

CN CO CO CMo

CM

<J) OLO LO

CO

LO

o in O CD

o ^s o 00

r— CN 1 00 o 00

CN © 1

d r~

CO00 CMT—

d

O ) 00 00 CD

CN CD r“

CD H 1 «""N 03 00

CM o 1 CN O o

OT— T"™

oi

00 00 ^CM

CN

O

O LO

00

LO CD

°> £

LO ^

co q

CD*

(0 _Q ^-Q(O

.2 «

*o O)

C <Q

— ©

— O

< <

co

4~l

X}o ²

"5

£

o

<1) Oc

®V W

O

2

</) Q . OL. 00

a oo

^ r ic ^ i_ os

© r -

5 : o

— +-l

Q oo i-

to

O I c 10

£ CT>

<0 ■ ■ 3

in

O ’ c LU o

"O ’5 )

C ffl 0) CC

"O m a> i - -t-> (U

</>

a) o

ca h- 4- T3

o a>

</> £

« £

.§ o

■H</>

UJ

CO

a

o

a

<fi

■§ S5

«

I

E I

— aj

</> co UJ Q .

©

§ 1

O) -z

<0 03 CC >

CO

CM

^ CM 05

— v CO * - CM lO o w ^

'tf-

°0 !5

_ o 05

N (O O)

w oi o

00 o

^ CO T -

CN ™ £!

W 00 o

CD

2 05

cn CD 00 O)

LOLO

<D

D5 J= to co co r r o

CQ <

CN CO lO w co oo in o LD r —

o o

or '- ⁰⁵ CD O o j 00 CD COCD

to ^ 00 £

£ co

00 O 00 CO LO

‘ CO

?£ “ >

go CO

h>

00

to -Q O

co 4—>

o

3

•oo

00 f>s

^ rs T- 05

00 O ) to to O) CN CM

O) CN CO LOCO LO

LOCO O)• CO

CM o>LO CO

o

5 I

T-:

r -

H? 05

05 05

00

CD

to X ! O ro .Q u

2

04 00 CD CM CN

O <J) O)

h* Oi

LO U )

00

t - CD CD © CO CO

CM

ay oo

r - ooo o o

in

co

©o o

CM

Tt O

CO oo ID O)

oq

o o

O <J5

oo to S *

05 CM

o 00

CD CD

in (£>m co co co oo oo

CM &

O)

00

in

CDO ^CM

in in co CN in oo o o

cm

co oo oo ^ID

CM O

V“O) in CM

o

00 _CN ^CDO in

inin O)o>

oo o

o in 05

O |

© oo 1

<b •w in

O i

«—* 00

CO

inoo o in

CD ©O)

M (0 W ® o <

Co oD

*DO 2

*ai

>

Contd.

O co

_

LO t—

CM CD

ay CM

o

CO r-

CO

a>

CM

00

ayo

CO CM

o

LO

CM CO

o T- co co

o CO CD r - CM t— CO C ) CT) 00 PN O

00 t -

r*

CM

io O)

r — LO

O)

CM CO

™

CM ^ w LO

CM lo ay

coco co o

o

CO r-

CM

CM CO

CO

; r a>

a>

^ r~>

— 00 ^ g ' - ' ■ ' t o oo '

O to

CD I '' CM 00 O ^ O) I";

cor~ * -

oo

oo or—

CM CMr -

LOO 00

CO

oCM

CM ( O ^ o

CO

O)c CD o ©

^ co D)

<0 0)

© *-

> <°

co

*-*o T>=)

O 2

o

Contd.

Ol ^{COLO to07} ooCO o o

CN cn cn

CN05 (N CD O

oco coo oo

coco coco cn

oo ooo oco

CO

d

CN

CN CO COO

CN CNO

CN COT—

cn cd

o

CMCM

O

n

oo oo

r-oo 'I-

CM

o o

cn co cm o co co

o

CO CM

00CO

■<t r->

W 00 Is- oo in oo O Q

o

CM

o o o CO

T— CO CO 0 )

CO in

r -t— LO r * r “i

r - t- oT“ CN

r - ©

d t—

r -

CN to 00CM CO

ooin

a

oo

COo

CM>sf o'S' 00CM

10 £ J3

73 _C

®_W»

o<

co

oD

73O 2

"3

£

A s to the measur e of compar i son of t w o standard errors of the g r o w t h rates for t w o separate periods, it is argued that the s t a nd ar d error e st i ma t ed for the t w o peri ods may be strongly affected by instabi lity t hr oughout the time series and hence cannot be used for i nt er - t e mp or al compar i sons ( B o y c e 1987) .

T h e me t h od suggested by Sen ( 1 9 6 7 ) is to estimate separate t r end for 'peak' and ' t r ough' years and to assess the instability or stabili ty of the var i abl e in terms of the nature of divergence or c o n v e r g e n c e of the trend values. T h e objection against this me t h o d is r egar di ng the definition of 'peak' and 'trough' years. It is a r gued that a small difference in the values of the variable can cause certai n year to be excluded or included and any modification of this criteria seems to be a matter of subjective judgement.

Yet anot her m e t h o d suggested by Boyce ( 1 9 8 7 ) of testing th e c hang es in i nst abi l i t y over time is similar to Glejser's test for het er osc a dastici ty a n d has some ver y distinctive advantages over the earlier me t h od s d i scu sse d above. A c c o r d i n g to this msthod, let us defi ne a statistic

A A

Z, = ( Q t - Q t ) / Q t

w h e r e Q t = ob serv ed val ue at time t and Q t = estimated trend A v a l ue at ti me t.

N o w either the absol ute value of Z t or if one wi shes to put gr eater we i g h t to larger deviati ons the square term of the statistic is regressed agai nst ti me. If the estimated time trend is signifi­

c a nt l y posi t i ve the i nst abi l i t y seems to be increasing and if it is f o u n d to be negati ve the instability seems to be decreasing over t i me . T h e statistical si gni fi cance is tested by the corresponding

t' values.

T h e advant age of this met hod is that one can use the full set of data for the entire t i pi e- per i od w i t h o u t considering any a priori break to measure the changes in instability of the variable. Again,

this method can be applied to any functional specification of the growth process. However, the practical problem of this me t h o d is that the form of the relationship between I Z t I and time is not known and that depends on the nature of the scatter diagrams of I Zt I derived from the fitted relation between Q t and time. N a t u ­ rally, an arbitrary specification of the model in finding out the trend value of I Z t I may ba misleading wh i l e interpre ing the trend values. Since there are no a p riori g r o u n d to take any speci fic form one may test different models to obtain the best fitted model.

Alternatively, on the basis of a scatter diagram of Z ; the form of the regression has to be decided.

Following the later approach w e have, by and large used the measure (1 Z, 1 and Z 2t) suggested by Boyce to study the instability of agriculture in eastern India during 1950- 51 to 1987- 88. The results are presented bel ow.

3.2 Results

In this analysis w e did not attempt to break the entire peri od in two sub-periods as has been done in st udyi ng the g r o w t h rates presented in Section I of this article. Instead, w e have used the entire time-series data wi t hou t any subj ective divi si on of time periods for the reasons described above. A s and w h e n a s t r u c ­ tural breek is suspected w e have introduced d u m m y variables to have a better result. In this analysis, w e have considered on l y two types of functional forms viz. linear and semi - l og. W e have ignored the Gompertz type of relation as it has been f oun d in ou r earlier results that this functional form does not hold g d o d in most of the cases of g r owt h analysis.

Now, regarding the form of the relationship of I Z t I and Z ta wi th time, the nature of the scatter diagrams vi ndi cate for the linear specification to find out the trend value of the test st at i s­

tics. In some cases w e i ntroduced d u m m y variables ( b o t h for intercept and slope) in the linear model and obtained better results.

Presented in Tabl e 4 are the results (significant trends onl y) of our analysis.

It is evident from the figures in T a bl e 4 that instability in area, pr oduct i on and yi el d of rice and total foodgrains in most of i hc oases are i ncr easi ng over time. In W e s t Bengal trend of insta­

bility in acreage of rice is negative but the intercept dummy shows a si gni fi cant u p w a r d shift of the magni t ude of fluctuations during the post break peri od. Acr eage of total f ood cultivation in West Bengal b e come mor e unstable (in most of the cases) during the later phase of the time series whi l e in the first part of the break the trend is negati ve but not very significant. Yield instability of b ot h rice and toial f oodgr ai n is found to be true for the entire time per i od r esulti ng ther eby in the instability in rice and foodgrains pro­

d uc t i on in the states of eastern India. Instability in area, produc­

ti on and yi el d of rice and total food cr ops are much pronounced in Or i ssa than the other states as in most cases the trend value of the variables are positi ve and signi fi cantly hi gh. When the dummy var i abl es are i nt roduced, the acreage of rice in Orissa shows a d e cl in i n g trend in the p ost - b re a k period wi t h a significant jump in the ma g n i t u d e of f luctuations. ( T h i s is also true in the case of f ood gr a i n s ou t pu t and yi el d )

As to the results of analysis for wh ea t crop, the post-green r evol uti on enterprise- most of the regions under study show a d e c l i n i n g ttend of i nstabi lity of pr oducti on and yield which con­

f i r ms o n e ’s expectati on. Al t h o u g h instability in acreage of wheat c r o p in Orissa s h o w s a decl ining trend, the yield instability seems to be rising in the p r e- b r ea k period wh i l e it is significantly nega­

tive in the p ost - br eak period.

T o sum up, the results of our analysis clearly show the i ncr easing i nstability in the g r owt h of area, production andyi el dof f o o d cr ops and rice in particular in the states of eastern India dur­

i ng the last 38 years. T h i s exercise also confirms the proposition that there has been n o break in the pr oducti on o f fo o d g ra in s during 1 9 5 0 - 5 1 to 1 9 8 7 - 8 8 . ( T h e exception is noticed in the case of O r i s s a ) . In Or i ssa maj or agricultural break-through in rice and f ood gr a i n s p r odu ct i on is noti ced duri ng the late fifties, much before the i nt roducti on of n e w seed- fertilizer- irrigation technology in

different parts of India. Thi s break, however, coul d not help to reduce the fluctuations of agricultural production during the p o s t ­ break period. T h e picture of wheat cultivation is almost some in regard to instability in all the states of eastern India barring West Bengal and India in general.

I I I

4. S u m m a ry and C o n clu s io n

The basic thing w h i c h w e intend to establish here is that it is not enough to study the agricultural g r o w t h wi t hou t consi der i ng the variability of the variables associated wi t h the g r o wt h rates. It is sometimes observed that the gr owth of output of a region or of a country increased at an increasing rate, but this increase has br ought in its wake considerable instability in output. T h i s obviously affects the i ncome and empl oyme n t stability of the country. T h u s from the v i e w point of policy implications the gr owth of agriculture shoul d be vi ew e d in terms of an integrated framework of gr owt h wi t h stability.

Wi th such considerations in mi n d w e have attempted here to study the g r owt h of agriculture in eastern India along wi t h an analysis of fluctuations in foodgrains output and in the output of other major crops. T o explain the relationship bet ween g r o w t h and instability of agricultural producti on, an attempt has also been made to study the g r owt h and instability of acreage and yield of the major crops g r own in this region.

T he results of our analysis provide some interesting insi ghts i nto the problem. O u r foregoi ng observati ons suggest that a sharp increase in the g r o w t h of agricultural producti on had not been possible during the period 1950*51 to 1 9 8 7 - 8 8 in eastern India mai nly because of the dimi ni shi ng/constant rate of g r o w t h of c r opped area of the maj or crops cultivated in this regi on.

A l t hou g h the rate of g r o w t h of producti vity of land in rice seemed

TABLE4 Measuresofinstabilityin termsofI Zt I andI Z2

© >*

Q_

O E

c: o

c

C O rj ¹— '

"a Q_Q)

U E

0) E S T 4—' r j

c _“O

"O c

CD oo

h - a0) o

© s—"

■*—*

c

if )

i O ^O

* o » + - c 4-*

C/9 c

© o a )

o 4—*

if )

*—•

CO C O k_ Q . CO a ) 4->

»— if ) o ** w

if ) C J

CO LO

• —

</>

2

©

\ - ^C/5

*+—

o — ©

CO

T3 C

»*—

o c

o c

D> ^CO

0) © +-* ^{k .}

O . CO CO

h - w o O >

l o ©

c _ Q CO

O)_•— ^CO_w ^w

V . CO

O >

a < N

o ' w '

k_

o

4->©

5

CO 'w '

C O

COin o

N

co<D C

©o

a>

5

05 — ~

O cnJ 6 C i

C O

o 9 d ?

CO00 o

>*

£ E

CD<D £

c

J 5 oo o ^CN

N

cd _ r-s

* - p U a>

o oo o O 9 ■ q cn O W o w

CO

CDc

CO

CO

LOO

COO

cca>

c

N

coa>

c

©o a:

CO05 COO o

co©

c

N

CDO ECD

00

0)o

FoodA Semi log I Zt I Linear -0.03640 0011 (

2

21

<y>

oo

co

LO

co

CM (D© C

N

CD©

C

O ^

r - LO

o ^

o CM

O ^

O CO

o o w

o CO

o T“ o ID CO

*“• CN o o o 00 o 1^ . o

o O o o o 00 o o

o c i o t © o■ o r -‘ o I o CNi

w

oo CN CD CN in

00 O 00 O

CN CO o CN CO CN

o O o O «*•

o O o o o d

E E - 2

m Q

(0© SZ C ~

3 5

N

(0©

c N

*DO O

CD0) c

N

O)o E©

</>

•DO u.o

©(D c

N

.CCO

£

CD ^ r - CO O CD

O *

lO

>

~ E E

* _{m Q}

Z

CO© £ c *“*

J i

(0© JZ

£

N N

<o

sz

o coco O CN

O ' LO

CD co o

N

O)o

>»

E _ 3E E Q W tj 5

CO® -C

OO o

•sfo

CN

O CD O CO

o • c

o0

cn

<N cn

o OT- o C-

co r-' ^

° r-' _

° ccr o

o ° .

o C o

oo

r-' -r-

° 5 oo _

co cn o w

9 co

iD

CO CMO

LO

oo o

oo o o

OCM

00 5 o

CD (C

©

c

>~ £ ■*—»

E 03 CN

E O) CO CD Cv 3

Q c +-» a>

> . *o

CO (1) <0 (0c (0

a> .c a>

■a

“ CD

c _ j

.*—»

5

D O

T~

CD c

- J

N

> - E

03 ^

Q Q

C - j *£

N N N

05O

> - E

£

"Z ^ E Q

<15 .C -

</>

5 N

CO0) C

N

CM

©o ©

o

ou CC

<a>

o

-0. 05 25 0. 00 10 (3 .4 8)

00

(0

co

COo oq O CO

<0 O)

CO

r - 00^{^} O^{h *}

r“ r> CM CM O

O CO O CO O

o 6

CO<w» 9 CM

O o

i

00 in

in

in

o o o

o o o

(0<0

c

N N

E E3 Q

O ) <«—s

0) CN in

r^» o _{r -}

CM o _CM

Sw

or —

o

N s- E

*L E

CD Q

■S x:

N

oo —.

O 00 O 00

9 CN

o

•F⁺ o05

co

o

N

>~

05 ^£ O £3 E Q to t :

N

CN

co

N

E S ETO 13

S q co a

COa>

.c

cn

coo

10

O 05

o co

oo TT

§ «?

CDr^*

oo oo o O I

*3-CM

CD

o O)CM

00

CM CO

o °?

o a . 00 rj- o

CO

00o

<*> ^

o "

o « 6 £S

CO

CM

CD <o

0)c

&Q £

CO /-V

Q UC r-

m n

hi M e«

Nf

N

N

OO

CM

>

Is i

s =>

E Q

CDO

C10O

CO

-0.22390.0047-0.0018 (2.79) (-3.93)

FoodP Linearb ) Z, ) Linear -0.0893 0.0026 withDummy (2.65) CD

^ h- ^ co ^ r*^ in

o

o

o CO o o

o *“

CNJ

S « O W o —

9

00 o>

in

CD₀₀ CO_CO

o

o

o

o d

d d d

N N N

N

05 _p

E

© CO

>.

E E3 Q

N _N

n>

_o

E<o V)