
Analiza reactiei fermierilor la politica de promovare a fermelor organice utilizand Modelul Probit
Abstract
Conceptul de dezvoltare sustenabila sa ivit īn deceniu saptezeci datorita grijii generale despre ambianta globala a mediului, ca rezultat la poluare si cresterea de surse energetice cum ar fi materiile prime si energia. Sustenabilitatea īnseamna retehnologizarea tehnologica, stiintifica, de mediu, economica si a resurselor sociale īn asa un mod care rezultatul sistemului eterogen sa poata sa fie mentinut īntrun mediu temporal sau echilibru spatial, īn timp ce dezvoltarea sustenabila este o dezvoltare care īntālneste necesitatile prezente fara a compromite abilitatea de generatiile viitoare de a utiliza resursele necesare.
Key words: sustainable, probit model, promoting, organic, farms
Se poate spune ca in Romania suprafata de teren, destinata pentru obtinerea productiei organice este destul de redusa, cu toate ca exista o situatie dezastruoasa ca urmare a nivelului investitional. Ca urmare a cercetarilor publicate de catre OECD in octombrie 1998, situatia terenului restituit este de 91,6 % iar 54,5 % din acesta este suprafata arabila, 10510p1521k sefgmentata in parcele mici, cu o suprafata medie sub 5 ha (2,42 ha), dar suficienta pentru a asigura un venit pentru subzistenta familiei.
Pana ce aceste probleme nu vor fi rezolvate, Romania va avea dificultati mari in realizarea unui potential pentru o agricultura organica sustenabila.
Tinerii ecologisti pentru Romania, sponsorizati de Heinrich Bohl Foundation (Germany), este una din principalele organizatii ce promoveaza agricultura organica in tara noastra. Avand sediul in Bucuresti, ei coordoneaza grupuri de lucru in agricultura sustenabila, in cinci regiuni NUTS ale tari, in scopul gasirii strategiilor de dezvoltare in introducerea si imbunatatirea agriculturii organice in zonele cu productii scazute. Una din cele mai importante realizari a acestui grup este acela de a influenta guvernul sa aprobe legislatia agriculturii organice in compatibilitate cu standardele UE.
Pentru anlaiza reactiei fermierilor la promovarea fermelor organice sa folosit modelul Probit.
Modelul Probit este un un model de programare lineara in care toate activitatile si variabilele sunt codificate cu un X ( se intocmeste mai intai o lista a variabilelor si codificarea acestora).
Valoarea variabilelor de iesire sau outputul va avea valori intre 0 si 1.
Probabilitatea fiecarei variabile de iesire va fi insumata avand o astfel o probabilitate cumulata pe tot modelul.( la fel ca si in cazul interpolarii statistice 0 pentru valoarea cea mai mica sau nesimnificativa 1 pentru valoarea cea mai mare restul prin interpolare). Interpolarea pentru celelalte valori, cuprinse intre 0 si 1 se calculeaza sub forma regulii de 3 simpla dupa cum urmeaza : Din cel mai mare se scade cel mai mic sii dam 100, dupa care din cel mai mare il scadem pe cel carel determinam sii dam X dupa care se determina regula de 3 simpla.
Raportul de sansa pentru organic fata de conventional este egal cu raportul de sansa de insatisfactie īn conventional fata de organic.
In statistics, a probit model is a popular specification of a generalized linear model, using the probit link function. A probit regression is the application of this model to a given dataset. Probit models were introduced by Chester Ittner Bliss in 1935, and a fast method of solving the models was introduced by Ronald Fisher in an appendix to the same article. Because the response is a series of binomial results, the likelihood is often assumed to follow the binomial distribution. Let Y be a binary outcome variable, and let X be a vector of regressors. The probit model assumes that
where Φ is the cumulative distribution function of the standard normal distribution. The parameters β are typically estimated by maximum likelihood.
While easily motivated without it, the probit model can be generated by a simple latent variable model. Suppose that
where , and suppose that Y is an indicator for whether the latent variable Y * is positive:
Then it is easy to show that
[edit]
References
In modelling the choice between statespecific alternatives, is denote farmer decisions by a binary indicator d k for k=1,2 such that d k = 1 if the technology k with current period returns R _{K} is chosen and d_{ K} =0 otherwise.
The value D =1 refers to standard technology and the value D =2 refers to the organic technology. Because the choices are mutually exclusive the condition k 1+k 2 =1 holds.
Current returns are functions of the current state variable in X t (input and output prices, subsidies, fixed inputs and other returns shifters), and a set of parameters in q. The current returns function is continuos and twice differentiable, convex and linearly homogeneous in prices and subsidies, increasing in fixed inputs, output prices and subsidies and decreasing in input prices.
The optional value function, V, for this decision problem then solves:
V (X t, t) = max E t
Where the choicespecific value function have a recursive structure and obey the Bellman equation of the form:
V _{d }(X _{t, }t)= R _{d }(X _{t, }q) + Đ E _{ }
Subject to t< T, V _{d } (X _{t, }T)= R _{d }(X _{t }) for d=1.2
The expectation of the choice specific value function is taken over the distribution of the random returns R in V conditional on x and dk=1. Thus, we simulate directly random but stationary returns (R) of each farm rather than random prices (in X).
The technology choice and the censored choicespecific returns functions can then be written in a standard form) for switching regression (e. g. Maddala, 1983):
I _{t }(X _{t,} q) = R_{2t }(Xt_{.,} q_{ })+ Đ E_{t}  R_{1t }(X t_{,} q_{ } )+ Đ E_{t} + u _{3 t= }X t_{ } + q _{3 }+ Đ E_{t} + u _{3 t }(4_{ } )_{}
Where I _{t }= 0 (standard technology) if I _{t }< 0
I _{t }= 0 (standard technology) if I _{t }< 0
Where the choice specific returns are given by:
R_{1t }(X t_{,} q_{ } ) = X t q t + u _{1 t }(5_{ })_{}
Where R_{1t }if I _{t }= 0 and R_{2t }if I _{t }= 1.
Parameters q 1 and q 2 include zero restrictions in the full parameter set q. The parameters in q 3 are linear functions of the parameters in q. The errors are normally distributed with zero mean. The error in the third (choice) equation is the difference between the errors in the second and first equations (return functions) i.e. u _{3 = }u _{2} u _{1.}
The return function (5) are estimated jointly with the endogenous switching model in (4_{ }).
Data and estimation
This study uses data from Romanian farms over the period of 19982007.These data provide a good opportunity to study farmer response to these policies for several reasons. First, in this period there has been substantial variation in agricultural prices, general subsidies to all farms and subsidies for organic farms. This allows for identification and testing for the affects of prices and subsidies on farmer behaviour. Second, organically grown products have had a market access similar to the standard agricultural products. Farmer choices have, therefore, been guided by economic incentives rather than by institutional constraints. Third, a substantial share of Romanian farmers switched to organic farming in the period under analysis and a sufficiently large number of organic farms is available in the dataset.
The dataset is an unbalanced panel consisting of 948 observations; 169 (17.8 %) observations on organic technology, and 779 (82.2 %) observations on standard technology. The share of organic farms in the sample exceeds their share in the population. This choicebased sampling was corrected for by using the Lerman and Manski (1981) weighting scheme. The sample includes 82 farms that applied organic technology for at least a year.
About a third of them (26 farms) had switched into organic farming before the sampling period. During the sampling period, 56 farms switched to the organic technology for and 11 farms switched back to standard technology. Thus, the relatively short panel data represent farmers having varying experience of organic farming and being at different phases with respect to the maturity of their contracts of farm organically.
Major characteristics of the sample farms, stratified by the production technology, are given in Table 1.
Table 1
Mean values for major farm characteristics averaged over farms and years

Specifications 
Standard farms 
Organic farms 
1. 
Net returns (Euro/year) 
1.6 
1.39 
2. 
Land area (10 ha) 
3.84 
4.21 
3. 
Number of livestock 
3.16 
2.6 
4. 
Machinery and building capital 
2.3 
2.53 
5. 
Labour (1000 hours/ year) 
3.7 
3.2 
6. 
Farmer's age 
43.6 
41.1 
The farm characteristics show that farming is important income source for both organic and standard farmers. An average organic farm in the dataset is even slightly larger than an average farm applying standard technology. Both organic and standard farms are family farms.
Annual values for output and variable input prices indices are obtained from the Yearbook of farm Statistics (Information Centre of the Ministry of Agriculture and Forestry, 1998) and, for given year, they are the same for all farms in the sample. The output price index represents all agricultural output, i.e. output from standard and organic farming. Separate indices for realised prices of organic and standard farming output were not available. However, it may be expected that prices from standard and organic outputs are interrelated, and that relative price movements will take place in both markets simultaneously.
Results
Half (50 %) of the parameters are significant at the 5 % level. The return functions from both standard and organic farming are nondecreasing and convex in prices and subsidies at the observed data points.
The observed and predicted technology choices in the sample are as shown in Table 4. Overall, the model predicts the choice correctly in 878 cases (93 percent of the sample). However, the choice of the organic technology is predicted less accurately than the choice of the standard technology. This is a commonly found result in modelling problems where choice probabilities are also small in the population (e.g. Dorfman1996).
At 1999 prices, the average probability of a switch from standard to organic farming is found to be 12 per cent, whereas the probability of switching back from organic to the standard technology is estimated at 10 per cent. Partial effects of explanatory variables on the variables on the farmers' response cannot be derived analytically from the parameter estimates because these effects also depend on the solution to the dynamic optimisation problem. In other words, the partial effects of the model variables are not policy invariant because they affect the optimisation results. The solution that is adopted in this paper is to compute the partial effects numerically by simulating the model under alternative price, subsidy and farm scenarios.
The results in Table 5 suggest that organic farming have a higher shadow price for land than standard farming, whereas all other fixed inputs receive returns under organic farming that are lower than or equal to those under standard farming. The higher shadow price for land on organic farms can be explained by the premium subsidy that organic farming receive on the land area. The model predicts that a 1 percent output price decrease alone will increase the probability of choosing organic technology by 0.4 percent. The probability of switching to organic farming increases at an increasing rate with increasing premium subsidies of the organic farming. The results also suggest that increasing regular subsidies slightly increase the switch to organic farming.
Table 2
Parameters estimates in the return functions

Specification 
Return to standard farming (R1) 
Return to organic farming (R2) 

1. 
Linear terms 





Intercept 
1.6 
2.3 
0.23 
0.23 

Lagged choice 
0.21 
0.000 
3.31 
0.000 

Output price 
0.069 
0.710 
0.349 
0.000 

Regular subsidy rate for land 
0.56 
0.29 
0.42 
0.60 

Land area 
0.344 
0.039 
0.39 
0.183 

Number of animal units 
0.12 
0.028 
0.0223 
0.144 

Capital 
0.0006 
0.041 
0.009 
0.108 

Labour 
0.35 
0.075 
0.25 
0.305 

Regional yields average 
1.31 
1.36 
2.41 
0.000 

Premium subsidy for transform in organic 
 
 
0.209 
0.000 

Dummy for arable farming 
0.323 
0.142 
0.367 
0.40 

Dummy for cattle farm 
0.244 
0.112 
0.212 
0.51 

Quadratic terms 





Output price 
0.260 
1.16 
0.406 
0.000 

Regular subsidy rate for land 
0.304 
0.104 
0.266 
0.389 

Land area 
0.0142 
0.0051 
0.0139 
0.237 

Number of animal units 
0.0031 
0.0025 
0.0016 
0.0221 

Capital 
0.0002 
0.0056 
0.0029 
0.0187 

Labour 
0.0444 
0.0153 
0.0323 
0.0663 

Regional yield average 
0.393 
0.406 
0.696 
0.000 

Premium subsidy rate for organic farming 
 
 
0.150 
 
Table 3
Observed and predicted frequencies

Observed 
Predicted 

Standard 
Organic 


Standard 
770 
9 

Organic 
61 
108 
Table 4
Shadow prices for fixed inputs in the return of farms (R)

Specification 
Returns to standard production 
Returns to organic production 
1 
Land area 
0.27 
0.32 
2 
Number of animal units 
0.12 
0.03 
3 
Capital 
0.00 
0.00 
4 
Labour 
0.19 
0.15 
Conclusions
The organic share in the total agricultural
surface in the
In the nineties, however, the growth increased considerably. Between 1993 and 1997 an average of 60 farms per year was converted. In 1998 and 1999 more than 200 farms was converted per year, which is the equivalent in growth of more than 25 percent per year. In the last two years the growth rate dropped to 14% in 2000 and 8% in 2007.
The econometric estimation of farmers' choices between standard and organic technologies involves the solution to a stochastic dynamic optimisation problem and the econometric estimation of structural parameters of the underlying behavioural equation. This paper combines the numerical Dynamic Programming Routine and the estimation of a Probittype switching model to explain the decisions of Romanian farmers to switch between organic and standard farming technologies.
The results suggest that economic incentives play an important role in the farmer's decision to choose between organic and standard farming technologies. The likelihood of switching to organic farming responds to changes in output and input prices, and to subsidy rates. Therefore, agricultural price and subsidy policies have good prospects for guiding farmers technology choices in the direction preferred by society. It was found that incomeneutral policy reforms that decrease output prices and compensate for the resulting farm income losses by direct subsidies would substantially increase farmer incentives to switch from standard to organic farming.
The results also indicate that farms with low
returns to standard farming i.e. farms with low yields per hectare have a
stronger incentive to switch than farms with high returns to standard farming.
This implies that the subsidy programme aimed at encouraging farmers to switch
to organic farming in
Both response probabilities and the shadow prices for fixed inputs in standard and organic farming show that the Romanian programme for stimulating organic farming provides an incentive towards extensification. Farmers who have large land areas and, thus, good possibilities to practise extensive farming technologies, are more likely to switch to organic farming. Specialisation in either livestock production or arable farming increases returns to standard farming and, thus, decreases probability of switching to organic farming.
The framework that was used in this paper provides a first step in analysing the factors underlying farmers' (revealed) decisions to switch from standard to organic farming. To simplify the analysis, it was assumed that quantities of inputs such as land and labour are taken as given by the farm operator over the sample period. However, in the long term, it may be expected that the quantities of these inputs will adjust and that. Future research should also asses the role of noneconomic factors in the decision between standard and organic farming. It may be expected that the adoption of organic farming technology is affected by social (the social acceptance of organic technologies) and psychological factors (such as farmers' attitudes towards and knowledge of environmental problems).
Bibliography
1. Chambers, R. G., 1988 
Applied Production Analysis. A Dual Approach.
2. Dorfman, J. H., 1996  Modelling multiple adoption decision in a joint framework. American Journal of Agricultural Economics, 78.
3. EIM/Haskoning, 1999  Study on a European Union wide regulatory framework for levies on pesticides, EIM/Haskoning, Zoetermeer.
4. Heast, C., 2000  Data obtained from sheets presented by Heast during Aurelia conference in The Romania.
5. Hiuang, C. L., 1996  Consumer preferences and attitudes towards organically grown produce. European Review of Agricultural Economics, 23.
6. Romero, C., Rehman, T.,
1989  Multiple criteria analysis for agricultural decision . In: Developments
in Agricultural Economics 5. Elservier,
7. Van Keulen, H.,Wolf, J., 1986  Modelling of Agricultural production: Weather, Soils, and Crops. Simulation Monographs.Pudoc, Wageningen.
Politica de confidentialitate 