Abstract: The purpose of this study was to examine the import of pesticides in Pakistan since February 1980, when import and sale of pesticide was shifted to private sector and subsidy on ground spraying was withdrawn. Various pesticide registration schemes of import have been discussed and number of pesticides registered in each scheme has been reported. The pros and cons of registration schemes have been discussed. An appropriate regression model satisfying basic assumptions has been developed for the import and consumption of pesticides. Using the model estimates, it was concluded that the growth rate of pesticide import was computed to be 35% in 1981 taking 1980 as base year, 30% in 1982 taking 1981 as base year; and for subsequent years, the growth rate of pesticide import has lowered and has reached at 8% in 1999 taking corresponding preceding year as base year. The same growth rate has been calculated till 2002. The agricultural growth rate has been recorded lower than that of pesticide import excepting 1995-96, when highest agricultural growth rate of the decade was recorded to be 11.7% and with pesticide import growth rate of 10%. This indicates that the trend of using pesticide for pest control is higher when compared to agricultural growth trend.
INTRODUCTION
During the post February 1980 period, pesticide consumption has increased from 665 metric tons in 1980-81 to 69,897 in 2002-03 (GoP, 2004). Tariq (2002) reported that in the last two decades, there has been substantial increase in the use of pesticides not only in volume but also in value. Its use has increased by about 70 times (of which about 80% is used on the cotton crop), while cotton yield has increased to 2 times only. Poswel and Williamson (1998) studied the increasing trend of pesticide in Pakistan and stated that ten years ago a quarter of smallholding cotton farmers in Pakistan produced their crop with no use of pesticides. By 1997 the minimum number of applications was four per season with nearly half of all farmers spraying at least seven times.
The problem statement: Over reliance on synthetic pesticides for crop protection has increased since 1980 when new agriculture policy was announced and the import and sale of pesticides were shifted to private sector while other methods of pest control have been overlooked. There is no parity between pesticide import and agriculture growth trends (Tariq, 2002). The present study has attempted to quantify pesticide import trend based upon the assumptions of linear regression models and to relate with agricultural productivity. Besides, schemes of import of pesticides, through which pesticides are imported in the country, have been reviewed.
The specific objectives of the study were as follows:
| • | To discuss the pesticide registration schemes. |
| • | To develop an appropriate regression model for the import of pesticides. |
| • | To study the relationship of the pesticide import with agricultural productivity growth. |
| • | To develop the recommendations for policy makers regarding pesticide registration schemes |
MATERIALS AND METHODS
Secondary data regarding registered pesticides through various schemes and pesticide import during 1980 to 2002 were collected from the Department of Plant Protection, Ministry of Food Agriculture and Livestock, Karachi. Computer packages Excel and Regression Code (R-code) were used to perform analysis and to draw pertinent graphs. Using the time series data, a linear regression model was developed and its assumptions were tested. R-square of the regression model was calculated to know the percentage variance in dependent variable explained by the model. A brief about the linear regression model and its assumptions is given as follows.
Linear regression model: A linear regression model was proposed on the basis of advanced statistical techniques known as aptness of the model through testing the assumption of linear regression. The simple linear regression may be defined as a way to describe the relationship between two variables by calculating a best-fitting straight line on a graph. The line averages or summarizes the relationship. The result is a regression line expressed in a regression equation. The general formula for the regression model is shown as follows:
E(Y/Xi) = β1 + β2Xi |
where Y is effect, response, or dependent variable, Xi is cause or independent variable, β1 and β2 and are unknown but fixed parameters known as intercept and slope, respectively (Gujarati, 2003).
Testing of assumptions of linear regression model: For the appropriateness of the results, assumptions of linear regression model: (1) variances of the errors should be constant and (2) errors should be normally and independently distributed, were tested. To test the assumption of the regression, non-constant variance plot and residual plot were constructed using R-code computer package. In non-constant variance plots, lowest was set at ±1 standard deviation. The variances are assumed to be constant when three lines go parallel. Besides, for constant variances, value of mean score should be non-significant (p>0.05). To test the assumptions that errors should be normally and independently distributed, residual plot’s OLS (ordinary least square) was set at 1 and lowest was also set at 1. When the errors are normally distributed, lowest line and OLS line converges. In case the assumptions are not satisfied, the Box-Cox method is used to find the appropriate transformation (Cook and Weisberg, 1998).
RESULTS AND DISCUSSION
Pesticide registration schemes: Prior to 1992, pesticides were registered in Form 1 (one of the forms of Agricultural Pesticide Ordinance) only. Agricultural Pesticides Technical Advisory Committee (APTAC) headed by the Secretary, Ministry of Food Agriculture and Livestock, Islamabad used to decide cases of registration of the pesticides. The procedure for registration was complicated and time consuming especially wait for two years for the results of efficacy of pesticides from research stations. Realizing entrepreneurs’ problems and increasing demands of pesticides by growers, government introduced two new schemes of registration of pesticides. Generic scheme, in Form 16, was introduced in 1992 whereby the pesticide registered in other countries can be registered with its generic name without going thorough pesticide trails. In 1997, a new scheme was introduced for the import of pesticides in Form 17; the pesticide registered with its brand name in other countries can be registered in Pakistan without going through laboratory tests and field trials. Moreover, pesticide registration powers were transferred to the Plant Protection Advisor and Director General, Department of Plant Protection. However, all the policies of pesticides import and sale are framed by the APTAC.
The pros and cons of pesticide registration schemes: The new schemes in Form 16 and 17 have played a very positive role. Supply of pesticides has been ensured round the year. With the introduction of these schemes, numerous pesticide companies have entered in market; and almost all the new companies are known as generic companies. Monopoly of the multinational companies has ceased in the market and the multinationals are facing price competition. First time in the history, the multinationals have reduced prices of their products. Profits made by the local companies are reinvested in the country while the multinationals send them to countries of their origin. However, on the other hand, some stakeholders are of the opinion that registration of pesticides in Form 1 is a standard route of registration of pesticides. But, due to some deficiencies of government agencies especially processing of cases very slowly, the scheme is not too attractive. New registration schemes in Form 16 and 17 were introduced purely on political basis and had nothing to do with fair competition and introduction of new pesticides in the market. For instance, basic aim of registration of pesticides in Form 17 was the introduction of new chemistry in the country. However, practically what happened that pesticides, which were supposed to be rejected in Form 1, were allowed without testing their effects under environmental conditions of the country. With the introduction of these schemes, many pesticide companies have started selling pesticides without having facilities of storage and knowledge about proper handling and usage of pesticides. Their motive is only to make profits, not pest control. There are about 3000 registered pesticides marketed by more than 300 companies in Pakistan. It is very difficult for the Agricultural Extension to monitor these pesticides. Besides, it is also difficult for Chemists to check them efficiently. Hefty profits are offered to the pesticide dealers by the companies to sell ineffective pesticides to farmers. Due to sale of ineffective pesticides the number of pesticide sprays on crops has increased manifold.
Registered pesticides under various schemes: Figure 1 shows that 1244 pesticides, which constitute about 46.3% of all the pesticides (2689) were registered in Form 16, which is known as generic scheme. It is also evident that business of pesticides with generic names in the country is increasing. The second largest scheme was Form 17, in which 36.4% (978 out of 2689) of all the pesticides were registered. The least number (17%) of pesticides was registered in Form 1. Fewer number of pesticides registered under this scheme is evident of its complexities in registration process especially of two years experimental trails at any two research stations.
A linear regression model for import and consumption of pesticides 1981-2002: Table 1 reveals the pesticide import during 1981 to 2002. The table shows that a small quantity of 665 metric tons were imported in 1980, on the other hand a huge bulk of 68,804 metric tons were imported in 2002. During 23 years, a linear trend was observed using regression model (Fig. 2). The regression model is given as under:
Total import = -5762.2 + 2632.1(year) |
The coefficient of year, 2632, indicates that pesticide import is increasing at the constant rate of 2632 per year; every year 2632 tons more pesticides are imported as compared to the previous year. Sufficiently high 0.89, R-square was calculated for the above model, which indicates that about 89% variation in import of pesticide has been explained by the linear regression model.
| Table 1: | Total pesticide import during 1980-2002, predicted import for the same period and percent increase in predicted import |
To test the assumption of linear regression model, non-constant variance plot and residual plot were constructed. Figure 3 shows slight increasing variance since the lines gets wider at right end and p-value (0.039) of the non-variance plot is significant. Besides, Fig. 4 shows that OLS and lowest lines do not converges (when the errors are normally and independently distributed and free from outliers and influential observations, they converge).
| Fig. 1: | Pie chart showing number of pesticides registered with Department
of Plant Protection, Karachi under various schemes of registration |
| Fig. 2: | Graph of the linear regression model for pesticide import during 1980-2002 |
| Fig. 3: | Non-constant variance plot for regression model of pesticide import during 1980-2002 using original values |
| Fig. 4: | Residual plot for regression model of pesticide import during 1980-2002 using original values |
| Fig. 5: | Suggested transformation for regression model of pesticide import during 1980-2002 |
p-value (0.000) of the test of curvature is highly significant; therefore, Box-Cox method was applied to find the appropriate transformation. Figure 5 suggests square root transformation since Lambda Hat is equal to 0.50. The square root transformation indicates that pesticide import is increasing at faster rate than simple linear regression model, when a trend is estimated by adding a fixed quantity to the preceding terms to obtain next term. The square root transformation was applied to total import of pesticide during the period. Based upon transformed values, the proposed regression model was as under:
The trend of the regression model is shown in Fig. 6. The R-square was reported to be about 0.94, which shows that about 94% variation in pesticide import during 1980-2002 has been explained by the model. Figure 7 and 8 are evident that the assumptions of the linear regression are satisfied with the proposed transformation because p-values of non-constant variance and test of curvature are non-significant.
| Fig. 6: | Graph of regression model for pesticide import during 1980-2002 using transformed values |
| Fig. 7: | Non-constant variance plot for regression model of pesticide import during 1980-2002 using transformed values |
| Fig. 8: | Residual plot for regression model of pesticide import during 1980-2002 using transformed values |
Using the model estimates, it was concluded that in the initial years, the growth rate of pesticide import was computed to be 35% in 1981 taking 1980 as base year (Fig. 9), 30% in 1982 taking 1981 as base year; and for subsequent years, growth rate of pesticide import has lowered and has reached at 8% in 1999 taking corresponding preceding years as base years. The same growth rate has been calculated till 2002. The agricultural growth rate has been recorded lower than that of pesticide import excepting 1995-96, when highest agricultural growth rate of the decade was recorded to be 11.7% and pesticide import growth rate was 10%.
| Fig. 9: | Growth in pesticide import (%) using regression model of transformed values and agriculture |
The above model shows that consumption of pesticide does not follow a simple linear regression model; however, it follows square root transformation (quadratic type relationship having both the estimates with positive sign). This finding does not coincide with that of NFDC (2002) who have proposed linear trend in the consumption of pesticide in Pakistan. The best reason behind the linear trend proposed by NDFC could be that the assumptions of the regression model were not tested while in the present study assumptions of the regression model were tested. First simple linear regression model was applied, but assumptions did not satisfy while the proposed model (based upon transformed value) satisfied the assumptions and gave maximum R-square (0.94).
No relationship was inferred between the estimated growth rates in pesticide import and agriculture. Higher growth trend in pesticide import was obvious than that of agriculture. The same problem is discussed by Tariq (2002) who reported that increase in the use of pesticide has not increased the cotton (on which 80% of pesticide is applied) production, but has aggravated the insect situation through indiscriminate and reckless sprays.
CONCLUSIONS AND RECOMMENDATIONS
Square root transformed regression model tested upon assumptions was proposed for the import of pesticide in the country which reveals that the import of the pesticide in the country is much higher than ordinary linear regression model estimates. Growth trend of pesticide import is higher when compared to agricultural growth trend. Increased pesticide import would ultimately increase the number of sprays on the crops, which will cause to further aggravate health hazards and environmental degradation. Besides, the cons of the new registration schemes appeal the policy maker to reconsider the registration schemes of pesticides in Form 16 and 17. Likewise, NFDC (2002) has suggested, Relook at issue of Form 17 whereby pesticides approved in one country does not require any testing in Pakistan. Such a free leverage has created problems. It is, therefore, suggested that the import of pesticide in Form 16 and 17 may be restudied. Heavily reliance on pesticides for pest control may be discouraged because it is associated health and environmental risks especially when the thousand of tons of pesticides are allowed to import without checking their effects in local conditions. Alternate methods (envisaged in Integrated Pest Management, IPM) of pest control may be encouraged. In this regard, agricultural extension activities like farmers field schools may be continued and fully supported so that the pesticide import bills as well as health and environmental risks associated with the overuse and misuse of pesticides could be reduced.