MATERIALS AND METHODS

The study is extended for the time period extending from 1970-71 to 2004-05 and it is based exclusively on the secondary data collected from the Statistical Abstracts (1974-75, 1980-81, 1984-85, 1990-91, 1995-96, 2000-01, 2003-04) of West Bengal published by the Bureau of Applied Economics and Statistics, Government of West Bengal (henceforth bureau). District Statistical Handbook is another publication of the Bureau that contains database for each district.

For measuring agricultural growth rate, linear trend equation:

(1)

where, Q_t is the output at period t, t is the time is used. In this equation b gives us an estimate of the absolute increase of agricultural output per unit of time, a the constant appearing in the regression line and u_t the error term in the regression line. However, for measuring the acceleration or deceleration in the growth rate the log quadratic trend equation:

(2)

is fitted to examine the acceleration (or deceleration) in the growth trend. A positive significant value of c indicates acceleration while a negative implies a deceleration.

To measure the growth rates in different sub-periods the following kinked exponential model is used :

(3)

In this equation b₁, b₂ and b₃ are the growth rates for the first, second and third sub-periods respectively. Estimating the trend break equation also tests the significance of the difference between the growth rates in two different sub periods:

(4)

where, b₁^* and b₂^* are the difference between the first and second sub-period growth rates and the difference between the third and second sub-period, respectively.

Coefficient of Variation (CV) has been used to measure the variations in production of crops. There may be variation in the production of crops over the years. Also there are chances that the crop production may vary across the different districts of the state. In this study we have measured the variability in the production of important crops over the three distinct sub-periods across the districts of the state. The difference in CVs of the three distinct periods is tested by Kruskal-Wallis test (Anderson et al., 2008). The hypothesis for the Kruskal-Wallis test can be written as follows:

The Kruskal-Wallis test statistic, which is based on the on the sum of ranks for CVs of each of the districts during the three different sub-periods for any particular crop can be computed as follows:

Kruskal-Wallis test statistic

(5)

Where:

Under the null hypothesis in which the populations are identical, the sampling distribution of W is approximated by a Chi-square distribution with k-1 degrees of freedom.

RESULTS AND DISCUSSION

Sub-Period Growth Rates: Kinked Exponential Method
The pioneer work came from Boyce (1987) who measured the growth rates of crop output in West Bengal agriculture during 1949-80. He observed that the stagnancy of agriculture of the state comes to an end during 1980s and the growth of agricultural rose to its highest level during the decade of eighties. Saha and Swaminathan (1994) estimated growth of aggregate crop output of 6.40% during the eighties was the highest among the Indian states and this spectacular growth of the agricultural sector was widespread in the districts of the state. They argue that the land reforms measures introduced in the state during the early 1980s have significantly contributed to the impressive growth of the agriculture of the state. Chattopadhyay and Das (2000) also claim that the agricultural growth in West Bengal during eighties is higher than the seventies. However, their estimate of the annual growth rate of agricultural production (3.6%) in West Bengal during 1977-78 to 1994-95 is much lower than that (6.4%) estimated by Saha and Swaminathan (1994) during the period 1981-92 to 1990-91. They conclude that agricultural production in West Bengal is still dependent on rainfall and fluctuations in rainfall index significantly positively contribute to the fluctuations in agricultural production in the state. Sanyal et al. (1998) and Mukherji and Mukhopadhyay (1995) also support the view of Saha and Swaminathan (1994) that land reforms programmes is primarily responsible for the breakthrough in West Bengal agriculture. Harris (1992), however, believes that this high growth has taken place in the absence of any reform of the agrarian structure. In his opinion, this was rather technological-'suitable technology' coined with favourable fertilizer-paddy price ratio. To analyze the performance of agricultural sector of the state the main emphasis has been given on the decadal growth rates of the major crops. We measure the sub-period growth rates of area, production and yield of the crops as measured by the kinked exponential growth model. For convenience, the whole period of 35 years (1970-71 to 2004-05) is divided into three sub-periods; first sub-period (1970-71 to 1980-81), second sub-period (1981-82 to 1991-92) and third sub-period (1992-93 to 2004-05). These three sub-periods have special significance to the economy of West Bengal. Clearly, the agricultural development of the state during the period 1970-71 to 1980-81 was not up to the mark, rather it was underdeveloped in nature. However, some institutional and technological changes took place in the state during the early 80s. This institutinal and technological changes had some positive impact on the agricultural development of the state. The productivity and production of all important crops have increased significantly. Lastly, the third sub-period (1992-93 to 2004-05) is the period of LPG (liberalisation, privatisation and globalisation). Clearly, the final sub-period captures the impact of globalisation on the agriculture of the state.

During the first sub-period (1970-71 to 1980-81) it has been observed that the growth rates of production of almost all important crops were either low or even negative. For example, total rice grew only at 0.51% per annum. The output growth rates of all the three varieties of rice were also very low during the 70s. The growth rates of production of Aman, Aus and Boro were 0.52, -4.32 and 2.53% per annum, respectively. The growth rates of production of wheat (-3.18% p. a.), total cereals (-3.23% p.a), jute (-1.74% p.a) and total pulses (-3.16% p.a.) were negative during this period (Table 1). Simple decomposition of output into area and yield indicates that neither area nor yield growth rates were impressive. In most cases, the area and yield growth rates (Table 2, 3) were either very slow or even negative, which in turn forces the output growth rate to be negative. The output of crops such as potato, mustard and total oilseeds were, however, positive during the decade of seventies. For potato, the high growth of area under cultivation was the main thrust behind the output growth rate as around 66% of the production growth came from area expansion and the remaining 34% came from yield growth rate. In case of mustard, the yield growth explains 65% of the output growth rate and the remaining 35% has been contributed by the area expansion. For the total oilseeds, the area expansion explains the majority of the output growth rates. Total foodgrain production which is the sum of total cereals and total pulses grew at a nominal growth rate of only 0.35% per annum with a negative area growth rate of 0.51% per annum.

Table 1:	Sub-period growth rates of production of crops in West Bengal during 1970-71 to 2004-05
Source: Author’s calculation based on BAES data. *p = 0.01, **p = 0.05, ***p = 0.1

Table 2:	Sub-period Growth rates of area under crops in West Bengal during 1970-71 to 2004-05
Source: Author’s calculation based on BAES data. *p = 0.01, **p = 0.05, ***p = 0.1

Table 3:	Sub-period Growth rates of yield of crops in West Bengal during 1970-71 to 2004-05
Source: Author’s calculation based on BAES data. *p = 0.01, **p = 0.05, ***p = 0.1

In fact, a positive yield growth rate (0.86%) resulted the food production to grow at a slow positive growth rate. As already mentioned that the jute production grew at a negative rate during the decade of 70s and this negative production of growth was mainly due to the yield growth of jute which was estimated to be -4.58% per annum.

Thus, we find that during the period 1970-71 to 1980-81, without few exceptions the growth rates of production of most of the crops were slow. Both the area expansion under crops and the yield of the crops were responsible for this slow growth of output. This situation of slow growth, however, changed during the 80s when the growth rates of area, production and yield of all the crops improved significantly. From Table 1, it is found that during the second sub-period (1981-82 to 1991-92), the output growth rates of total rice improved significantly from a negative growth rate observed during the previous decades of 70s to 1.23% per annum during the second sub-period and this increment in growth rate was significant (as tested by the significance of trend breaks). The factor behind the massive increment in the output growth rates of most of the major crops during the eighties in West Bengal was the improvement in the yield growth rates. The yield growth rate of total rice jumped from 0.72% per annum during seventies to 4.07% per annum during the period 1981-82 to 1991-92 (Table 3). During the second sub-period, more than 70% of the output growth came from yield growth. The similar fashion of growth trend has been observed for Aman rice. For Aus rice the growth rate of area was negative (-1.57% p. a) during the second sub-period (Table 2). However, a massive increase in the yield growth rate (6.05%) during the 80s resulted in the impressive growth of output at a rate of 4.47% per annum. In case of Boro paddy quite an opposite scenario has been observed where the area growth rate plays the major role behind its impressive output growth of over 11% per annum during the second sub-period. Boro was introduced on massive scale as HYV crops in the state during the early 80s. As it was a HYV crop, its yield level was already high and its high yield was the main factor for its rapid expansion in the state. As far as wheat is concerned, it is found that the state has been able to achieve a marginal positive growth rate as against the negative growth of the 70s. However, during this period the performance of the state in respect of pulse production was very gloomy. The negative growth rate of the 70s became much more acute during the eighties. However, the impressive growth performance of paddy crops has turned the growth of foodgrain production to an optimum level of 4.52% per annum during the second sub-period. This was one of the best foodgrain production performances among the states of Indian union. The sluggish growth of pulse crops did not affect the overall foodgrain production because the share of pulse crops to the total foodgrain both in terms of area and production was small compared to total cereals. This impressive growth performance has also been observed for the crops such as mustard, Til, total oilseeds, jute and potato. During the period of eighties, the growth of production of all these crops was very much higher compared to the previous period and, in general, the yield growth rate played the most dominant role behind their impressive growth performance during eighties. The production growth rate of mustard jumped from 9.76% during the first sub-period to 12.18% during the second sub-period with more than 80% contribution coming from the expansion of area under mustard crops. The same is true for the total oilseeds.

However, the state failed to sustain the high growth path as achieved during the eighties. The growth rate of production of almost all important crops declined in the subsequent periods. The area growth rate of Aman rice becomes negative during the decade of nineties and the yield growth rate also reduced significantly. As a result of which the output of Aman rice declined only to 1.05% per annum. During the nineties, the output growth rate of Boro rice declined to 3.35% per annum and this fall in growth rate is significant (Table 1). Considering all the cereals crops together, it is found that due to the significant fall in yield rate, the output of total cereals declined drastically to 2.15% per annum during the third sub-period from 7.18% per annum during the second sub-period. The same is the case for total foodgrain. The state experienced a very high growth rate of production of mustard during the eighties. However, the state has not been able to sustain the tempo of high mustard production growth for a very long period. The mustard production growth rate declined very drastically to only 0.86% per annum during the third sub-period of 1992-93 to 2004-05. For total oilseeds, the yield plays the dominant role behind output growth rate during the eighties and the drastic fall in yield rate (4.60% per annum during 80s to -0.11% per annum during 90s) forced the output to grow at the rate of 0.21% per annum only.

For another important cash crop potato, the performance during nineties was not as good as observed during the previous sub-periods. The output growth rates for potato was 8.84, 7.37 and 3.35%, respectively in the first, second and third sub-periods respectively (Table 1). And the difference in the growth rate between the third (1992-93 to 2004-05) and second sub-period (1981-82 to 1991-92) was significant at 5% level of significance. From the Table 1-3, it is clear that the main reason for this decline in output growth rate is the slowing down of the yield growth rate, which reduced to -0.01% per annum during the period 1992-93 to 2004-05 from 2.35% per annum during the previous sub-period (1981-82 to 1991-92).

Thus, we see that the performance of the agricultural sector in the state of West Bengal during the period of eighties in terms of growth rates of production of important crops is better than that of during seventies and nineties. The expansion of area under cultivation of major crops almost reached a level of saturation. The growth rates of yield of major crops have decreased to a significant level and this the fall in the yield growth rates of the major crops reduces the growth of production of the major crops.

Acceleration and Deceleration in Trend Rates of Growth
To examine the acceleration and deceleration in growth rates of production of some important crops in West Bengal, the log-quadratic trend equation is fitted to the data and the results is shown in the Table 4. The results show that during the period 1970-71 to 1980-81, there were marked decelerations in the growth trend of almost all major crops of the state. However, during the period of eighties the growth trend of crops such as aman and aus rice, total rice, total foodgrain and jute experienced positive acceleration in their growth rates. Total pulse production of the state which experienced deceleration in growth trend during the first two sub-periods registered acceleration in growth in the final sub-period of nineties. During the nineties potato, jute, wheat and boro rice registered significant deceleration in their growth trend. It is to be noted that wheat is a crops which experienced deceleration in growth trend in all the sub-periods. Thus, in general, there seems to be decelerating trend in the growth of major crops in West Bengal since nineties.

Table 4:	Growth and acceleration in crop output in West Bengal using Log Quadratic form, 1970-71 to 2003-04
Source: Author’s calculation based on BAES data. *p = 0.01, **p = 0.05, ***p = 0.1

Table 5:	Coefficient of variation of production of major crops in West Bengal

Variability of Production of Major Crops
Using the coefficient of Variation (CV) the variability of production of major crops in West Bengal during 1970-71 to 2004-05 has been examined. An attempt has also been made to examine the contribution of area, yield variability in the variability of the total production of the selected crops separately for three periods viz. 1970-71 to 1980-81, 1981-82 to 1991-92 and 1992-93 to 2004-05. Further, the difference in coefficient of variation values of three sub-periods based on the districts of the state for each crop was tested by Kruksal-Wallis rank test.

The variability of production of major crops is reported in Table 5. By looking at the production variability of crops as measured by the coefficient of variation at the state level the following important outcomes emerge:

Decomposition of Production Variability
To examine the contributions of area, yield and their interactions effects on the variability of total production, we use the following decomposition scheme.

We know:

Table 6:	The decomposition of output variability into its various components during 1970-71 to 1980-81
Source: Author’s calculation based on BAES data

Table 7:	The decomposition of output variability into its various components during 1981-82 to 1991-92
Source: Author’s calculation based on BAES data

Table 8:	The decomposition of output variability into its various components during 1992-93 to 2004-05
Source: Author’s calculation based on BAES data

Where:

The decomposition of production variability in to its different components is done for the three sub-periods for each of the crops and the results are shown in Table 6-8. The major outcomes of this production variability decomposition are outlined as follows:

Test of Difference in the Variability of Crop Output: Kruksal-Wallis Test
The significance of the difference in C.V values of three sub-periods based on the fifteen districts for each crop was tested by Kruksal-Wallis test. The results are displayed in Table 9. The result indicates that there was no significant difference in the variability of production of total pulse among the sub-periods. However, it was observed that the sum of ranks for period III and period II were almost equal and much higher than that for the period I, implying that variability in total pulse production was less in the 70s as compared to the 80s and 90s. For all other crops, except aus rice, the K-W test statistic was found to be significant, implying that the crop production variability differ significantly in all the three sub-periods.

Table 9:	Estimated Kruskal-Wallis test for variability in West Bengal agriculture
Source: Author’s calculation based on BAES data. R1 = Sum of ranks in period I (1970-71 to 1980-81); R2 = Sum of ranks in period II (1981-82 to 1991-92); R3 = Sum of ranks in period III (1992-93 to 2004-05)

Furthermore, because the CV ratings are higher in the second sub-period for the crop jute, total cereals, total food grain and total rice, it can be said that the high growth rate of 80s was associated with high degree of variability. For wheat and potato the period of 70s was found to have highest variability in production. In the decade of 90s the variability of production was found to be stabilized for all crops. During this decade the growth rates of production of all major crops were also found to stable and reached a saturated level.

Thus, in short it can be said that except Aus rice and total pulses, there is significant variation in the production of all other crops across the districts in all the three sub-periods. The high growth path of the decade of eighties, in general, is associated with the high degree of variability. However, the variation was found to be stabilized during the decade of nineties.

CONCLUSIONS

The study makes it clear that the agriculture of the state had been able to boost its performance during the decade of eighties, at least in terms of growth rates production of major crops, mainly for tremendous increment in the yield growth rates of the crops along with expansion of area under cultivation. The effective introduction of HYV technology coupled with successful implementation of land reforms programme at the very grass root level set the path of agricultural development in the state of West Bengal. However, this scenario of impressive growth performance did not sustain for a very long period of time. The fall in the yield growth rates of crops reduces the production growth rates during the era of globalisation which have been started in the early nineties. During this period, crops such as Boro rice, total foodgrain, mustard, jute, potato experienced fall in their growth rates to a significant extent with marked deceleration in growth trend. Again so far as variability of crop production is concerned it can be concluded that the higher growth path is associated with the high degree of production variation across the districts of the state during eighties. The degree of variation was found to be stabilized during the nineties. The decomposition of variability clearly reveals that yield variability of crops, in general, is the prime factor behind crop production variability.

The cultivation of the same crop on the same piece of land over a long period and non-optimum doses of chemical fertilizer might cause the soil fertility to decline in the state. This is also very much prominent from the stagnation of yield level of most the important crops in West Bengal in the recent years. And this stagnation or slowing down of yield growth rates of major crops cause the total agricultural output to grow at a slower rate in the recent times. This is the main cause of concern of today, especially with respect to the food security issue of the state.