Efficiency and Competitiveness of Kosovo …ageconsearch.umn.edu/bitstream/252770/2/Efficiency of...Efficiency and Competitiveness of Kosovo Raspberry Producers Abstract Raspberry

Efficiency and Competitiveness of Kosovo Raspberry Producers

Rina Vuçitërna

Graduate Research Assistant

University of Arkansas

Email: [email protected]

Michael R. Thomsen

Professor



Jennie S. Popp

Professor



Arben Musliu

Professor

University of Prishtina


Selected Paper prepared for presentation at the Southern Agricultural Economics

Association’s 2017 Annual Meeting, Mobile, Alabama, February 4-February 7, 2017

Copyright 2017 by Vuçitërna, Thomsen, Popp and Musliu. All rights reserved. Readers may

make verbatim copies of this document for non-commercial purposes by any means, provided

that this copyright notice appears on all such copies.

mailto:[email protected]




Efficiency and Competitiveness of Kosovo Raspberry Producers

Abstract

Raspberry production is a growing industry in Kosovo. In addition to private investments, this

growth has been supported by grants, subsidies, and direct investment from international donor

organizations and governmental institutions. At present, most of the commercially produced

raspberries in Kosovo are produced on small farms, harvested by farmers and packed manually

by collection centers, and then sold as frozen for the export market. The long-term viability and

continued growth of raspberry production in Kosovo depends on the industry being able to

compete in export markets and hold its own against production regions in Poland, Serbia, and

Russia. Our study measures the efficiency of Kosovo raspberry producers with an aim towards

enhancing industry competitiveness. We collected primary data on raspberry farmers in Kosovo

during the summer of 2016. Using these data, we examine producer efficiency with an input

oriented data envelopment analysis. Our findings suggest that efficiency improves with

production experience and that outreach efforts could emphasize labor management.

Keywords: Raspberries, Kosovo, data envelopment analysis, efficiency measurement

JEL Classifications: Q12, D24

1

Introduction

Agriculture is an important sector for Kosovo, which contributes to the economic growth of the

country. According to a report by the Central Bank of the Republic of Kosovo, in 2014

agriculture’s contribution to the country’s GDP represented 12%. Agriculture is the main source

of exports in Kosovo, however Kosovo remains Europe’s biggest importer of goods per capita.

According to the Green Report published by the Ministry of Agriculture, Forestry and Rural

Development in Kosovo (2014), Kosovo exported €35 million but imported €584 million leading

to a large trade imbalance in agriculture. Kosovo has high expectations from the agriculture

sector, but despite the growth in the agricultural sector since independence, the trade balance in

agriculture is still negative (Simnica, 2016). One of the factors preventing a positive balance of

agricultural trade is an inconsistent climate, which contributes to variability in quality and

quantity (EFSE, 2014). Other factors include high input costs due to diseconomies of scale, low

production levels, difficulties in access to international markets and a small budget dedicated to

agriculture from Kosovo’s government relative to neighboring countries. Nonetheless, raspberry

production has been increasing and performing better than other fruits and vegetables in the

international market for the past few years. At present, 98% of the raspberries are sold in the

frozen international market and only 2% in the domestic market.

The ability to export and profit from frozen raspberries were two of the main motivations for

farmers to start cultivating raspberries. This also motivates international organizations and

governmental institutions to encourage and subsidize raspberry production. Farmers have been

getting help from different organizations in the form of training, advice, and assistance with

different inputs such as plants, irrigation systems, and direct payments. As a result, the number

2

of raspberry farmers in Kosovo has increased from roughly 300 in 2015 to 1000 in 2016. That

said, Kosovo raspberry producers face competition from large exporters of raspberries, such as

Poland and Serbia, which are geographically close to Kosovo and hold a competitive advantage

in the market. Harvesting accounts for 30% of the costs of producing raspberries. In Kosovo,

raspberries are harvested by hand. The most direct way to address this issue and to decrease

production costs and to become competitive in the export market is to introduce raspberry

harvesting equipment. USAID together with the Kosovo Ministry of Agriculture is looking to

support larger raspberry growers through cost sharing mechanism in investing in raspberry

harvesting equipment. Another issue that has an effect on raspberry production cost of is

sourcing high-quality planting material. Due to the incredible demand for planting materials,

there is an increased risk that poorer quality materials will be imported. USAID is working with

the Ministry of Agriculture to ensure that healthy high yielding raspberry planting materials are

imported. Despite these efforts, some poor quality materials have been imported to the detriment

of the industry. The objective of this study is to analyze and measure the efficiency of raspberry

farms in Kosovo. To achieve this goal, an-input oriented data envelopment analysis (DEA), a

non-parametric method, is used.

The remainder of this paper is organized as follows. The next section provides some background

on DEA and its use in the agriculture sector and other fields. Following this, we describe our

data and the methods used in this paper. We then present the efficiency scores and analyses of

factors that explain these scores. The final section of the paper presents our conclusions and

discusses limitations of the study.

3

Use of DEA to measure efficiency

Efficiency measurement can be done using a parametric approach, nonparametric approach, or

both. The parametric approach assumes a certain production function, parameterizing the

input(s)-output(s) relationship (A. Hadi-Vencheha, 2010). In contrast, in the nonparametric

approach there are no assumptions of any production function. Farrell (1957) did the pioneering

work of introducing the nonparametric approach to the literature. He showed that it is possible to

distinguish efficiency into price efficiency (allocative efficiency), technical efficiency and scale

efficiency. Scale efficiency has been developed by Farrell (1957) and by Charnes, Cooper and

Rhodes (1978) using a linear programming framework. In this study the focus is on technical

efficiency and scale efficiency.

Charnes et al. (1978) built on the nonparametric approach and introduced data envelopment

analysis (DEA), which serves as a method to measure relative technical efficiencies of the same

units operating in similar conditions with the goal of describing an efficiency frontier (Joro et al.,

2015). For example, if a firm or farm is positioned on the efficiency frontier, it means that it is an

efficient unit, and it has an efficiency measure of 1, and if it is below the efficiency frontier, then

it is an inefficient unit and it has an efficiency measure of less than 1. For example, an efficiency

measure of 0.8 means that the firm is 80% efficient.

DEA has been used to measure efficiencies in different fields. It has been used to measure the

efficiency and risk of banks (Nguyen, Nghiem, Roca, & Sharma, 2016; Benites Cava, et al,

2016; Chen, 2015), the efficiency of health care systems (Gouveia, Dias, Antunes, Mota, Duarte,

& Tenreiro, 2016; Kaya Samut & Cafrı, 2016; Shwartz, Jr.Burgess, & Zhu, 2016) and to

4

measure the efficiency of forms of tranportation (Wu, Zhu, Chu, Liu, & Liang, 2016; Guo, Gong,

& Hu, 2015; Omrani & Keshavarz, 2016), etc. Within the agricultural sector it has been used to

measure efficiency in dairy farms (Balcombe, Fraser, & Kim, 2007; Aldeseit, 2013; Mugera,

2013), in wheat farms and production (Chebil, Frija, & Thabet, 2015; A. Hadi-Vencheha, 2010),

and rice production (García Suárez, 2016), to name a few. However, when searching for

literature on the measurement of the efficiency of raspberry farms, we have found no results in

the existing literature.

Even though stochastic frontiers allow the assumption for random noise variables and

inefficiency error components, many authors believe that it leads to biased efficiency scores

(Serra and Goodwin, 2009; Kumbhakar et al., 2007). Thus, Kumbhakar et al. (2007) introduced a

new local modeling method that overcomes the limitations of parametric and nonparametric

approaches, without foregoing their advantages, and this is based on the local maximum

likelihood (LML) method. This method does not require deterministic and stochastic components

of the frontier and it allows for stochastic variables and measurement error when estimating

technical efficiencies (Guesmi et al., 2015).

Given that this is an early study on Kosovo raspberry production, we use standard DEA methods.

However it is worth noting that many researchers have developed DEA further, in order to allow

for error margins and enable DEA to analyze imprecise and/or incomplete data. Furthermore,

relative efficiencies computed with DEA are sensitive to noise in the data, and any outlier or

missing value can cause drastic change in the efficiency measurement of firms (Kao & Liu,

2000). Dealing with imprecise data is important for agriculture because it takes place in an

uncertain environment and therefore the inputs and outputs data can be imprecise. Furthermore,

5

when respondents are asked to answer questions about their farms, there may be some over

reporting or under reporting, since the questions relate to past activities or experiences. To deal

with this problem, one of the methods used is called Imprecise DEA (IDEA), which was first

used by Cooper et al. (1999, 2001). Another method of addressing this issue is called fuzzy

DEA. The first two to introduce or suggest modeling constraints as fuzzy sets to account for

uncertainty and fuzziness of firms were Bellman and Zadeh (1970).

DEA can be output or input oriented. Mujasi et al. (2016) uses an output orientation with

variable returns to scale (VRS) to measure the efficiency of hospitals in Uganda. Using the VRS

model, the researchers are able to see whether the hospital’s production had increasing returns to

scale, constant returns to scale, or decreasing returns to scale. This is an appropriate model for

their problem, since they have a fixed amount of inputs and managers are responsible in

producing the maximum output. Furthermore, this paper used regression analysis to explain the

observed hospital inefficiencies, regressing the inefficiency scores on explanatory variables

(Mujasi, Asbu, & Puig-Junoy, 2016). They found that some of the hospitals are more efficient

than the others. They recommend that inefficient hospitals should reduce the number of medical

staff and the number of beds to achieve higher outputs.

Lauro, Figueiredo and Wanke (2016) used the input oriented model with constant returns to scale

(CRS) and variable returns to scale (VRS) models. The CRS models assume that variations in

input levels will generate variations in output levels, possibly increasing output and being more

efficient (Lauro, Figueiredo, & Wanke, 2016). According to the authors, CRS efficiency is called

technical efficiency, whereas VRS efficiency is called pure technical efficiency. They used a

bootstrap truncated regression to analyze their efficiency scores. They found that 83% of the

6

schools needed improvement to reach 90% efficiency and that the schools with fewer students

were efficient.

Based on this previous research, our paper will use the non-parametric approach (DEA) to

measure the efficiency of raspberry farms in Kosovo. This paper will use the input oriented

model with variable return to scale (VRS), constant return to scale (CRS) and decreasing returns

to scale (DRS). One reason the input oriented model is more favored for application to Kosovo

raspberry production is that there is limited storage capacity for raspberries in Kosovo. Thus,

Kosovo will need a few more years to increase the storage capacities, in order to increase the

production without decreasing the price.

Furthermore, this paper will follow up by regressing the efficiency scores on farm

characteristics, such as the type of trellis system, irrigation system, whether the farm does water

and soil analysis, the year when the farm first cultivated raspberries, the municipality of the farm

and whether the farm cultivates other crops besides raspberries. Through this analysis, we will be

able to assess how much of the difference in the efficiency scores can be explained by

independent variables available in the dataset.

Data and methods

Kosovo is still in the process of gathering, digitalizing and providing data to the public. As such,

we decided to do a cross sectional study using primary data collected through a survey. The

survey of raspberry farmers took place in Kosovo during the summer of 2016. The survey

represents the major raspberry production regions in Kosovo. These areas include Podujeva,

Prishtina, Prizren, Ferizaj and Lipjan, as shown in figure 1. From these regions, we have

7

successfully surveyed 86 raspberry farmers out of a population of roughly 300 raspberry

producers in Kosovo in 2015. The surveys were conducted with farmers who had started their

raspberry production in 2015 or before, but who had yields during the summer of 2015.

Agriculture in Kosovo is generally compromised of small parcels of land, which in many cases is

a problem for farmers because it increases their costs. In our dataset, the average raspberry farm

was 0.78 hectares. The raspberry varieties cultivated in Kosovo are Polka, Willamette, Meeker,

Tulameen, Mapema and Bliss. Table 1 shows the number of farmers cultivating different

varieties of raspberries. It can be seen that the most cultivated raspberry varieties are Polka and

Meeker. Polka, Willamette and Meeker have shown to give high yields, have high pest resistance

and are mainly used for processed food (Finn, Strik and Moore, 2014). Tulameen is mainly sold

to the fresh market. These varieties have been chosen also to extend the raspberry production

season, so producers can get higher prices for their yield. Descriptive statistics for other variables

in the dataset are summarized in the tables 2 and 3.

As shown in table 2, there is a large variability in yield across the farms included in our survey.

This reflects the fact that many raspberry farms are new and had their first yield in 2015, whereas

others have been operating for a few years. Most of these farms are family businesses. There are

large families in Kosovo, especially in rural areas, with approximately 8 family members per

household. This has not changed since 1948 (Warrander and Knaus, 2010). Thus, it was hard to

calculate the number of family members engaged in the process of producing raspberries or the

value of family members’ labor contributions. The data showed that

66 of the surveyed farmers used both family and non-family labor

8

5 of the surveyed farmers used only non-family labor

48 of the surveyed farmers used only family labor.

In the further calculations, we have 81 observations, out of the sample of 86 raspberry farmers,

because of missing values for non-family labor.

Proper trellising is critical to increasing fruit and yield. Having a trellis system, showed to be a

frequent practice among the surveyed farmers. Farms in the sample constructed trellis systems

from metal, concrete or wood. In most cases, farmers built their own trellises. It is interesting to

notice that only one surveyed farmer did not have an irrigation system. This is the main

difference between the Kosovo raspberry farmers and those in Serbia. The number of Serbian

growers that are improving their production by using appropriate agro-technical measures and

introducing irrigation systems on their raspberry farms is small. Majority of raspberry producers

in Serbia do not invest in nurturing plantations and are getting low quality and low yields, as low

as 5 tons per hectare, compared to farms that are investing and getting up to 20 tons per hectare

(Keserovic & Magazin, 2014).

Most of the farmers surveyed conducted soil and water analysis at some point during the process

of cultivating raspberries and few of them do these analyses regularly. None of the farmers

surveyed said that they owned a storage unit or freezer. The common practice is that the farmers

send their daily yields to raspberry collection points. These collection points freeze the

raspberries, package them and export them to the international market. As a result, the collection

points set the price of the raspberries in the local market. In 2016, prices paid to farmers were

very low because of the limited capacities of these collection points. Even though, the capacities

9

of collection points have been steadily increasing, raspberry production increased faster, so the

collection points were stuck as a result of an over-abundant supply.

To measure the efficiencies of the raspberry farms in Kosovo, we used the Benchmark package

in R (Bogetoft and Otto, 2015). The two inputs used in this study are days of paid labor per

hectare for pruning and harvesting and the plants per hectare, whereas the output is the yield per

hectare. The planting days were not included as a part of the total labor days because it is an

activity done only once, whereas the other two activities are repeated every year. The total labor

was a result of the working days of non-family members on the farm per hectare. Also, the plants

per hectare were computed using the space between plants and the width between the rows.

Yield is measured in kilograms. Table 4 reports the descriptive statistics for the DEA inputs and

output.

Using the DEA approach helped us identify the top or the most efficient raspberry producers. To

get the efficiency scores, we used an input oriented DEA. The input oriented approach reduces

equiproportionately the use of all inputs (Fare, Grosskopf, Lovell, 1994). We chose to analyze

the efficiencies under three returns-to-scale assumptions: CRS, VRS and DRS.

Scale efficiencies were calculated by dividing the CRS technical efficiency score by the VRS

technical efficiency score. If the farm was efficient or (had an efficiency score of 1) under each

return to scale assumption, it is notated as “Efficient”, if the CRS technical efficiency score and

DRS technical efficiency score are equal to each other, then it means that it is operating under

decreasing returns to scale, which means that the farm is using labor and plants per hectare more

intensively than they should. Otherwise, the farms that are operating under increasing returns to

10

scale could increase efficiency by using labor and plants more intensively.

Results

Table 4 presents the DEA efficiency measurements under the VRS, CRS and DRS assumptions.

There are only three farms that are efficient under the three assumptions, 64 of them are

operating under decreasing returns to scale and 14 of them are operating under increasing returns

to scale.

These results indicate that most of the farms are using more intensive labor per hectare and

plants per hectare than would be scale efficient. Since inputs and outputs are measure per

hectare, those operating under decreasing returns indicate intensity of plants and labor per

hectare are in the range of decreasing returns. One way for these farms to be more efficient

would be to manage better the labor input. This may be as a result of the experience of the farmer

and the year that the farm started to cultivate raspberries. For example, there is a large difference

in yield between the farmers that were established in 2015 and those that were in their second or

later year of production. The presence of a high number of the new farmers is consistent with the

fact that raspberries are becoming the new emerging crop in the country.

Looking more closely at the results, we regressed the efficiency scores on several explanatory.

These are reported in table 6. In the first column of table 6, there are two statistically significant

variables, the year when raspberry cultivation started and the binary variable for the municipality

of Lipjan.

The year coefficient is statistically significant at 1% level and is negative. The negative sign is

expected because all raspberry farms are quite new and have access to comparable technology.

11

The new farms, those established in 2015, are expected to be less efficient because the plants are

not yet fully mature. Moreover, older farms, those will smaller values for year, will be more

experienced and have lower costs as a result of learning economies.

Because many of the farms in the sample are very new and only produced raspberries for one

season, we divide the farms into two subsamples. On that includes the farms established in 2015

and another that includes more established farms, with two or more years of experience in

raspberry production.

In table 6, the second column of results presents regression results from the subsamples of older

farms, those established before 2015. In this regression we can see that the year coefficient is no

longer statistically significant. The hectare coefficient is statistically significant at the 10% level,

meaning that the bigger the farm, the more efficient is their production process per hectare. The

missing coefficients in this column relative to column 1 show that there were no raspberry

farmers surveyed that started cultivating before 2015 in the Municipality of Lipjan, suggesting

that the significant coefficient in column 1 reflects multicollinearity.

The third column of table 6 represents the new farms starting in 2015. These results show that

there are no statistically significant variables. The missing coefficients in this regression show

that all the farms were started in 2015 and so there is no variation in year; there were also no

farms that used river irrigation or were from the Municipality of Gracanica.

Summary results by source of scale efficiency are presented in table 7. It is shown that there is a

variation on the yield per hectare between newly established farms and the old farms because, as

noted above, raspberries do not give high yields in their first year. However, there is only a small

12

change between the increasing and decreasing returns to scale. Thus, there is a small variation in

plants per hectare between the new and old farms and a small variation between the increasing

and decreasing returns to scale. The major variation across the two categories of scale

inefficiency is seen in the labor days per hectare. New farms operating under increasing returns

to scale use almost double the labor days compared to the farms operating under decreasing

returns to scale. We see an even bigger change between the old farms. Old farms operating under

increasing returns to scale use much less labor days than the ones operating under decreasing

returns to scale. This means that raspberry farmers could focus more on managing their labor to

improve efficiency.

Discussion

Overall the study found that there are 64 farms experiencing decreasing returns to scale, 14

experiencing increasing returns to scale and only three efficient farms. In total, 64 farms use

more intensive labor per hectare than they need and their plants are densely planted per hectare

in their farm. Improved utilization of labor could increase the overall efficiency of production.

The 14 farms operating under increasing returns to scale need to use more labor and can plant

more plants per hectare to improve efficiency. Since Kosovo is divided into small parcels, there

may be the possibility of efficiency gains from improved cooperation among small raspberry

farms. This could benefit the farms in terms of specialization and utilization of labor and could

mean a shift from family businesses to partnerships of some kind.

The lack of farmer cooperation is evident and it is needed now more than ever (Zivkov, 2013).

This is one reason that prevents farmers from expanding their operations and activities further.

There is a general lack of trust and many do not believe the state would fairly regulate disputes

13

within a partnership (Zivkov, 2013). There is a need for general awareness of the benefits of

forming cooperatives that could share skills, decrease the risks of the farm and motivate member

farms to succeed (Zivkov, 2013). Forming cooperatives of raspberry farmers will increase

production and Kosovo can be a stronger competitor in the international market.

As a conclusion, there is a need for improving the efficiency of raspberry farmers. The main

finding of this paper is that raspberry farmers should allocate their paid labor more efficiently.

There are some limitations of this study. The main problem we encountered during this study

was in gathering the data. Small farmers may not have accurate records or be in a position to

provide exact answers on the questions. This was especially true of questions regarding the use

of fertilizers and pesticides. Also, this study could be improved if done a few years later by 1)

covering all of the regions in Kosovo, and 2) allowing farmers more time to prepare for the

survey and thus provide more exact answers on the questions, especially regarding the fertilizers

and pesticides. Further, if replicated in a few years, today’s new raspberry producers will have

gained experience in production and a better estimate of overall efficiency may be calculated.

These limitations aside, this paper provides a baseline of raspberry production efficiency, which

could be expanded and elaborated more thoroughly in upcoming studies.

14

References

A.Hadi-Vencheha, R. K. (2010). An application of IDEA to wheat farming efficiency. Agricultural

Economics, 42 , 487–493

Aldeseit, B. (2013). Measurement of scale efficiency in dairy farms: Data Envelopment Analysis (DEA)

approach. Journal of Agricultural Science (Toronto), 37-43.

Balcombe, K., Fraser, I., & Kim, J. H. (2007). Estimating technical efficiency of Australian dairy farms

using alternative frontier methodologies. Applied Economics, 2221-2236.

Bellman, R. E., & Zadeh, L. A. (1970). Decision-Making in a Fuzzy Environment . Management Science.

Benites Cava, P., Pereira Salgado Junior, A., & De Freitas Branco, A. M. (2016). Evaluation of Bank

Efficiency in Brazil: A Dea Approach. Revista de Administração Mackenzie, 17(4), 62-84.

Bogetoft P., Otto L. (2015). Benchmark and Frontier Analysis Using DEA and SFA. [Software].

Cooper, W., Park, K. S., & Yu, G. (1999). Models for Dealing with Imprecise Data in DEA. Management

Science.

Central Bank of Republic of Kosovo. (2014). Report on Macroeconomic Development. Prishtina: Central

Bank of Republic of Kosovo.

Charnes, A., Cooper, W., and Rhodes, E. (1978). Measuring the Efficiency of Decision Making Units.

European Journal of Operational Research.

Chebil, A., Frija, A., & Thabet, C. (2015). Economic Efficiency Measures and its Determinants for

Irrigated Wheat Farms in Tunisia: A DEA Approach. New Medit: Mediterranean Journal of

Economics, Agriculture and Environment, 14(2), 32-38.

EFSE. (2014). Agricultural Funding in Kosovo. Prishtina: European Fund for SouthEast Europe.

Fare, R., Grosskopf, S., & Lovell, C. K. (1994). Production Forntiers. NY: Cambridge University Press.

Farrell, M. (1957). The Measurement of Productive Efficiency. Journal of the Royal Statistical Society.

Finn, C. E., Strik, B. C., & Moore, P. P. (2014). Raspberry Cultivars for the Pacific Northwest. A Pacific

Northwest Extension Publication.

García Suárez, F. (2016). Data Envelopment Analysis (DEA) methodology: an application to rice

production in Uruguay. Agrociencia, 99-112.

Guesmi, B., Serra, T., & Featherstone, A. (2015). Technical efficiency of Kansas arable crop farms: a

local maximum likelihood approach. Agricultural Economics, 46, 703-713.

Guo, W., Gong, D. Q., & Hu, J. Z. (2015). Applying Data Envelopment Analysis (DEA) approach to

analyse investment efficiency of transportation projects. Advances in Transportation Studies.(2),

139-150.

Kao, C., & Liu, S.-T. (2000). Data envelopment analysis with missing data: an application to University

15

libraries in Taiwan. Journal of the Operational Research Society.

Kaya Samut, P., & Cafrı, R. (2016). Analysis of the Efficiency Determinants of Health Systems in OECD

Countries by DEA and Panel Tobit. Social Indicators Research, 129(1), 113-132.

Keserovic, Z., & Magazin, N. (2014). Fruit Growing in Serbia- State and Prospects. Utilisation of the

Census of Agriculture 2012 data in analysing status of agriculture and agricultural policy

making in the Republic of Serbia. Novisad. Serbia.

Kumbhakar, S. C., & Tsionas, E. G. (2007). Scale and efficiency measurement using a semiparametris

stochastic frontier model: evidence from the US commercial banks. Empirical Economics, 585-

602

Lauro, A., Figueiredo, O. H., & Wanke, P. F. (2016). Efficiency of Municipal Schools in Rio De Janeiro:

Evidence From Two-Stage Dea. Journal of Economics and Economic Education Research

Ministry of Agriculture, Forestry and Rural Development. (2014). Green Report. Prishtina: Ministry of

Agriculture, Forestry and Rural Development .

Mugera, A. W. (2013). Measuring technical efficiency of dairy farms with imprecise data: a fuzzy data

envelopment analysis approach. Australian Journal of Agricultural and Resource Economics, 57,

501-519

Mujasi, P. N., Asbu, E. Z., & Puig-Junoy, J. (2016). How efficient are referral hospitals in Uganda? A

data envelopment analysis and tobit regression approach. BMC Health Services Research.

Nguyen, T. P., Nghiem, S. H., Roca, E., & Sharma, P. (2016). Bank Reforms and Efficiency in

Vietnamese Banks: Evidence Based on SFA and DEA. Applied Economics , 48(28-30), 2822-35.

Omrani, H., & Keshavarz, M. (2016). A Performance Evaluation Model for Supply Chain of Shipping

Company in Iran: An Application of the Relational Network DEA. Maritime Policy and

Management, 43(1-2), 121-35.

Serra, T., & Goodwin, B. K. (2008). The efficiency of Spanish arable crop organic farms, a local

maximum likelihood approach. Production Analysis, 113-124.

Simnica, M. (2016). Benefits of the Stabilization and Association Agreement in the Agriculture sector.

Prishtina: Ministry of Trade and Industry.

Warrander, G., & Knaus, V. (2010). Kosovo. The Globe Pequot Press

Wu, J., Zhu, Q., Chu, J., Liu, H., & Liang, L. (2016). Measuring energy and environmental efficiency of

transportation systems in China based on a parallel DEA approach. Transportation Research, 48,

450-472.

Zivkov, G. (2013). Association of farmers in the Western Balkan countries. FAO Regional Office for

Europe and Central Asia.

16

Table 1. Raspberry varietiesa

Raspberry varieties Number of farmers cultivating

Polka 68

Meeker 21

Willamette 7

Tulameen 4

Mapema 2

Bliss 1

a Producers often produce multiple varieties. For this reason the numbers in the table do not sum to the

number of the farmers in the sample.

17

Table 2. Summary of continuous variables

Variables Mean St. deviation

Hectares 0.78 1.20

Total yield (kg) 5576.51 8274.42

Price received in euro/kg 1.93 0.24

Days of non-family workers

for pruning

0.84 2.64

Days of family workers for

pruning

3.45 3.60


for harvesting

36.47 46.56


harvesting

67.67 35.14


for planting

1.48 4.24


planting

3.38 3.55

18

Table 3. Summary of farm technology variables

Variables Yes No

Trellis system 59 27

Irrigation system 85 1

Soil Analysis 66 20

Water Analysis 47 39

19

Table 4. Descriptive statistics for DEA inputs and output

Inputs Minimum Mean Maximum St. deviation

Plants per hectare 666 8577 70833 11321.05

Total labor days 0 60.45 900 118.04

Output

Yield per hectare (kg) 285.71 7510.10 23000 5269.43

20

Table 5. Results of efficiency measurements of 81 farmsa

Nr CRS VRS DRS SE Nr CRS VRS DRS SE Nr CRS VRS DRS SE

1 0.5211 0.538 0.5211 DRS 28 0.3864 0.567 0.3864 IRS 55 0.3557 0.654 0.3557 DRS

2 0.4624 0.587 0.4624 DRS 29 0.1660 0.703 0.1660 DRS 56 0.2835 0.625 0.2835 DRS

3 0.0272 0.394 0.0272 DRS 30 0.4940 0.713 0.4940 DRS 57 0.3689 0.690 0.3689 DRS

4 0.0818 0.394 0.0818 DRS 31 0.8783 0.916 0.8783 DRS 58 0.3952 0.545 0.3952 DRS

5 0.0516 0.394 0.0516 DRS 32 0.3960 0.671 0.3960 IRS 59 0.2766 0.732 0.2766 DRS

6 0.0454 0.328 0.0454 DRS 33 0.3293 0.649 0.3293 DRS 60 0.9288 1 0.9288 DRS

7 0.5355 0.567 0.5355 IRS 34 0.5911 0.799 0.5911 IRS 61 0.2928 0.660 0.292 DRS

8 0.0108 0.328 0.0108 IRS 35 0.8405 0.979 0.8405 IRS 62 0.0889 0.385 0.0889 DRS

9 0.5434 0.642 0.5434 IRS 36 0.2049 0.236 0.2049 DRS 63 0.4565 0.693 0.4565 DRS

10 0.6705 0.671 0.6714 DRS 37 0.9486 1 0.9486 DRS 64 0.2470 0.720 0.2470 DRS

11 0.5434 0.642 0.5434 IRS 38 0.6645 0.733 0.6645 DRS 65 0.3705 0.769 0.3705 DRS

12 0.4205 0.645 0.4205 DRS 39 0.5668 0.877 0.5668 DRS 66 0.6127 0.888 0.6127 DRS

13 0.0197 0.428 0.0197 DRS 40 1 1 1 -- 67 0.1965 0.606 0.1965 DRS

14 0.0292 0.394 0.0292 DRS 41 0.1633 0.53 0.1633 DRS 68 0.1544 0.385 0.1544 DRS

15 0.0227 0.394 0.0227 DRS 42 0.5797 0.656 0.5797 IRS 69 0.1776 0.545 0.1776 DRS

16 0.1897 0.448 0.1897 DRS 43 0.1500 0.694 0.1500 DRS 70 0.6858 0.711 0.6858 IRS

17 0.2075 0.627 0.2075 DRS 44 0.3890 0.743 0.3890 DRS 71 0.1753 0.585 0.1753 DRS

18 0.5937 0.774 0.5937 DRS 45 0.7378 0.871 0.7378 DRS 72 0.2941 0.748 0.2941 DRS

19 0.5070 0.524 0.5070 DRS 46 0.2917 0.532 0.2917 DRS 73 0.4510 0.674 0.4510 DRS

20 0.5882 0.853 0.5882 DRS 47 0.1805 0.598 0.1805 DRS 74 0.2305 1 0.2305 DRS

21 0.3799 0.390 0.3799 IRS 48 0.1976 0.545 0.1973 DRS 75 0.3321 0.706 0.3321 IRS

22 1 1 1 -- 49 0.2371 0.716 0.2371 DRS 76 0.4841 0.699 0.4841 IRS

23 0.5882 0.853 0.5882 DRS 50 0.1863 0.587 0.1863 DRS 77 0.3952 0.778 0.3952 DRS

24 0.3087 0.446 0.3087 DRS 51 0.3271 0.592 0.3271 DRS 78 0.2134 0.645 0.2134 DRS

25 0.1401 0.392 0.1401 DRS 52 0.2205 0.561 0.2205 DRS 79 0.3098 0.703 0.3098 DRS

26 0.1646 0.714 0.1646 DRS 53 0.3794 0.731 0.3794 IRS 80 0.7035 0.728 0.7035 DRS

27 1 1 1 -- 54 0.1888 0.540 0.1888 DRS 81 0.1976 0.597 0.1976 DRS

a The two dashes ‘—‘ in table 5 represent the farms that are efficient under the three assumptions. The

three assumptions are returns to scale (CRS), variable returns to scale (VRS) and decreasing returns to

scale (DRS).

21

Table 6. Regressions explaining efficiency scores under the constant returns to scale assumptiona

All Farms Old Farms New Farms

(Intercept) 134.6*** -92.36 0.3

(-45.7) (-93.27) (-0.29)

Year -0.07*** 0.05

(-0.02) (-0.05)

Hectares 0 0.15* 0

(-0.03) (-0.08) (-0.02)

Irrigation with Well 0.14 0.18 -0.18

(-0.27) (-0.32) (-0.14)

Irrigation with River 0.31 0.08

(-0.27) (-0.33)

Trellis -0.08 -0.17 -0.01

(-0.07) (-0.13) (-0.08)

Soil analysis -0.02 0.13 0.04

(-0.1) (-0.19) (-0.09)

Water analysis -0.03 -0.06 -0.09

(-0.07) (-0.14) (-0.07)

Polka 0.03 0.03 0.17

(-0.13) (-0.13) (-0.2)

Municipality of Lipjan 0.40* 0.37

(-0.24) (-0.21)

Municipality of Podujeva 0.09 -0.12 0.05

(-0.2) (-0.34) (-0.19)

Municipality of Prishtina 0.27 0.4 -0.08

(-0.2) (-0.34) (-0.21)

Municipality of Prizren -0.18 0.48 0.04

(-0.28) (-0.49) (-0.09)

Municipality of Gracanica 0.24 0.25

(-0.27) (-0.41)

Municipality of Shterpce -0.02

(-0.25)

0.47

(-0.41)

Other crops cultivated -0.05

(-0.07)

-0.01

(-0.11)

-0.11

(-0.09)

R square

Adjusted R squares

0.33

0.16

0.43

0.07

0.57

0.39

Number of observations 76 37 39

RMSE 0.24 0.26 0.17

22

a Table 6 represents the results of three different regressions. The dependent variable on the three first

regressions is the technical efficiency under the constant returns to scale assumption of the total number

of observations, the total of 81 farms. The second column of results includes only old farms that have

started before 2015 and the third column of regression results include only farms that have started in

2015.

The values in the brackets represent the standard deviations.

The stars represent level of statistical significance as follows

*** represent 1% level of statistical significance;

** represent 5% level of statistical significance and

* represent 10% level of statistical significance.

23

Table 7. Aggregate results by source of scale efficiencya

Source of Scale

Efficiency

Year

Established

Yield per

Hectare

Plants per

Hectare

Labor Days per

Hectare

DRS 2015 4134.77 10869.07 132.16

IRS 2015 2786.02 11033.82 220.18

DRS < 2015 9933.47 12631.05 519.48

IRS < 2015 12443.18 12625 117.05

a New farms are farms established in 2015 and old farms are the ones established before 2015. Source of

scale efficiencies are decreasing returns to scale (DRS) and increasing returns to scale (IRS)

24

Figure 1. Map of Kosovo and the number of surveys done in each municipality

Efficiency and Competitiveness of Kosovo …ageconsearch.umn.edu/bitstream/252770/2/Efficiency of...Efficiency and Competitiveness of Kosovo Raspberry Producers Abstract Raspberry

Documents