Skip to navigation – Site map

HomeIssues110Forward to the Past ? Agricultura...

Forward to the Past ? Agricultural Restructuring in Bulgaria

Retour vers le passé ? La restructuration agricole bulgare
Mieke Meurs and Angel Bogushev
p. 93-104


Prior to World War II, the majority of Bulgarian farms were very small, fragmented, and produced predominantly for their own consumption. By 1989, however, Bulgarian agriculture was organized in large-scale, mechanized farms, producing for national and international consumption. Many observers expected further radical changes after 1989, with wide-spread distribution of property rights in land, and market liberalization, leading to the emergence of western-style family farms.
In this paper, we outline overall changes in patterns of land use, and then use survey data from the World Bank to examine patterns of agricultural land use by rural households. We find a small group of families engaged in limited commercial farming, and a small group which appears to have returned to subsistence agriculture. But the largest share of landholding households carry on the activities developed under socialism : they produce for personal consumption, to supplement some employment income. In conclusion, we consider some factors which might explain the lack of expansion of market production, in the context of unused land and labor.

Top of page

Index terms

Geographical index :

Top of page

Full text

1 – Introduction

1Prior to World War II, the majority of Bulgarian agricultural producers farmed very small, fragmented farms, predominantly producing for their own consumption. Forty years later, at the end of the socialist period, almost all agricultural land was organized into very large, highly mechanized units, which produced for global markets. Many observers expected further radical changes after 1989, when the vast majority of Bulgarian families received agricultural land that had been collectivized in the 1940s and 1950s and the right to expand production for sale.

  • 1 These were often slightly restructured versions of the old collective farms, although the land is n (...)

2Households were slow to remove their restituted lands from collective production, however. In 1992, 82% of agricultural land was still farmed by cooperative farms1. Over the next two years, this share fell to 67% (BNIS, 1995), but a much larger share of land remained in cooperatives than in other post-socialist countries of Europe. Analysts argued that the private farming might be limited by a weak development of input and product markets (Meurs, 2002 ; Swinnen and Mathijs, 1997). Under such conditions, returns on land might be higher in collective production than in individual farming. This suggested that families might farm more land individually as markets expanded and the feasibility of private farming increased.

3In this paper, we outline changing patterns of agricultural land use in Bulgaria from 1989 to 2003. Using household-level data, we then examine more closely the developments in land use by individuals. Over the period 1989-2003 the share of agricultural land used by individual households increased, leaving a smaller share in large-scale, collective forms of production. Despite the heralding of a major restructuring of Bulgarian agriculture in the early 1990s, however, little has changed in the way families use this land. The vast majority of them have kept their habits as they did under socialism : they farm about half a hectare, producing for self-consumption, thus supplementing income from salaried work. They rent the rest of their land to a cooperative, share-holding company or other farmer. Very few individual farmers earn income from agricultural sales.

4We find only limited evidence of commercially oriented producers among the individual farms, and subsistence farming is an equally important emerging form of production. Commercially oriented households appear to do quite well economically, and it is not clear why more family units do not pursue this strategy. Small landholdings do not seem to explain limited commercial output, as most landholding families use rental markets to adjust their farm size. Availability of labor is a more important constraint, and very few farmers use labor markets to increase their supply of labor. Further research is needed on this issue, but in concluding we discuss a number of possible explanations about this finding.

2 - Radical Restructurings

5In 1946, subsistence farming characterized Bulgarian agriculture. In 1926, 57% of Bulgarian farms were smaller than 5 ha. Farms over 30 ha were rare (Berov, 1977:62-65). Very large units, like those found in Poland or Romania, did not exist in Bulgaria. During the period 1926-1946, fragmentation increased, despite some attempts at land reform (Lampe, 1986). By 1946, 69% of all farms were under 5 ha (Berov, 1977, 62-65).

6This land holding structure was lamented as a cause of slow agricultural modernization (although wars, state policy, and other factors also played a role) and as a factor in poor economic growth. In 1934, there was a metal plow for only every three holdings, and a tractor-pulled plow for every 500 holdings. Most machines were concentrated on the small share of farms with over 10 ha. The majority of households practiced subsistence farming, although many also produced a small surplus for market, and a small group farmed intensively for urban markets (Lampe, 1986:56).

  • 2 In the late 1980s there were experiments with decentralization to smaller units, like production br (...)

7The forty or so years of Communist Party rule and socialist planning significantly transformed the Bulgarian countryside. In the late 1980s the vast majority of agricultural land was organized into large state-controlled units averaging 18 ha (1985)2. Production was highly mechanized, with one tractor for every 32 ha (BNIS, 1988: 258). Bulgarian grain competed successfully on international markets and was an important export.

8Alongside the large-scale cooperatives, small-scale production persisted under central planning. About 5% of agricultural land (mainly in the mountainous regions of central and southwest Bulgaria) was never collectivized. More importantly, most rural households had a “personal plot”, about 0.3 ha of state-controlled land, allocated for personal consumption. This was not exactly subsistence production, however, since the plot production supplemented a salary (or, more often, salaries), and sometimes in-kind payments, from a state employer. These plots were “supplementary farms”, although they did provide goods (fresh dairy products and vegetables) which could often not be purchased in rural stores. Later, the plots were also used to produce a small surplus for sale to the state or, less frequently, on local farmers’ markets. In 1987, the plots accounted for about 14% of agricultural land, while the rest was in large-scale production (BNIS, 1988).

9Since plot production specialized in items in short supply under central planning (meat, vegetables, and fruit), rural households could sell as much as they produced. Although the state controlled prices, they were high enough (and guaranteed) to motivate many households to produce a small surplus for sale. Expansion of this production was limited, however, by lack of input markets and limits on the size of personal plots (Meurs, 2002:32). Under these conditions, it was easy to imagine a rapid expansion of production for sale if state limits were lifted.

10After the socialist government resigned in 1989, all collective farms were officially disbanded (in 1991) and property rights in land returned to the families which held them prior to collectivization. Although Bulgarian newspapers and international policy makers heralded the imminent emergence of a mass of western-style family farmers (fermeri), shocked rural households (and collective farm managers) immediately reorganized cooperatives based on land holding members (Kaneff, 1996; Creed, 1997; Meurs, 2002). Farms, faced with the post-socialist economic downturn, radically cut back staffing and other inputs, and by 1994, very few members worked on the new cooperatives (Meurs, 2002:47).

11The new landowners did withdraw some land from cooperatives in order to expand household production. By 1994, individual farms had approximately doubled their share of agricultural land, although large-scale production cooperatives continued to farm the majority (67%) of all agricultural land (and 58% of arable land) (BNIS, 1995:259). Individuals farmed an average of about 1 ha, mainly in diverse products for household consumption. Over the next ten years, the share of land used by individual private producers continued to rise, from 33% in 1994 to 47% in 2003 (although the share of arable land rose much less, from 42% to 45%) (BNIS, 1995:259 ; Ministry of Agriculture, 2005).

  • 3 Bulgaria was divided into 28 administrative districts in 1988 (then called okrugs, now oblasts).
  • 4 In a simple linear regression, the share of oblast land in grain production in 1987 explains 37% of (...)

12The shift of land from shareholding farms to private farming proceeded unevenly across the country. In Blagoevgrad and Smolian oblasts3, the share of land in individual farms grew to nearly 100% by 2003, whereas in others, like Veliko Turnovo and Burgas, the proportion farmed individually remained under one-quarter (table 1 ; fig. 1 to 4). The percentage of oblast land held in large, shareholding farms is highly correlated with a history of specialization in large-scale, highly mechanized, grain production4. In mountainous regions with no history of mechanized grain production, land was much more likely to be farmed individually.

Fig. 1 - Oblasts and regions of Bulgaria.

Fig. 1 - Oblasts and regions of Bulgaria.

CAO : R. Darques

Fig. 2 - Land farmed individually, 2003.

Fig. 2 - Land farmed individually, 2003.

Source Ministry of Agriculture and Forestry, 2005:44; BNIS, 1987:547.

CAO : R. Darques

Fig. 3 - Decline in Agricultural Land Use 1987-2003.

Fig. 3 - Decline in Agricultural Land Use 1987-2003.

Source: Ministry of Agriculture and Forestry, 2005:44; BNIS, 1987:547.

CAO : R. Darques

Fig. 4 - Land Farmed in Companies, 2003.

Fig. 4 - Land Farmed in Companies, 2003.

Source: Ministry of Agriculture and Forestry, 2005:44; BNIS, 1987:547.

CAO : R. Darques

Table 1: Land Use Structure (2003), by regions

Land Farmed Individually

Land in Coops

Land in Companies
























































Source: Ministry of Agriculture and Forestry, 2005:44; BNIS, 1987:547.
* 2% of land is held by civil associations and others.

  • 5 The increase in meadows and pasture land is not driven by expanding livestock production. From 1988 (...)

13It is important to note, however, that the increase in the share of land held privately coincided with a significant decline in the total amount of farm land in use in Bulgaria (fig. 3). Agricultural land in use fell by about 15% over the decade 1994-2004/5, while farmed arable land fell by much more, from 4 643 000 ha in 1994 (about the same amount as in 1990) to 2 522 651 ha in 2004/5 (a decline of 46%). Utilized agricultural land is now more likely to be meadows and pastures than in 1987 or 1994, and some previously arable land has been returned to such uses (BNIS, 1995:259 ; Ministry of Agriculture, 2006:26)5. Thus, despite the increase in the share of land held privately, the amount of arable land farmed by private farmers fell about one-third over the period 1994-2004.

  • 6 The share of an oblast’s land under grain in 1987 explains 58% of the variation in companies’ share (...)

14The amount of land farmed by large-scale, collectively owned farms fell by 48%. As a result, cooperatives have declined significantly in size, to an average of 593 ha in 2003 (Ministry of Agriculture, 2005). The land farmed by the cooperatives continues to belong mainly to individual, small-holders who are members of the cooperative (93% of land) (BNIS, 2006). Land in this sector has increasingly been shifted into a new form of shareholding organization, the “company” (table 1, fig. 4), formed by transforming, or splitting cooperatives. Like cooperatives, companies farm mainly land owned by smallholders (Ministry of Agriculture, 2006), but the landholders do not have the same legal rights. In 2004/5, companies accounted for 20% of agricultural land, compared to 34% in cooperatives (Ministry of Agriculture, 2006:26-27). The company farms remain somewhat smaller than cooperatives, averaging 353 ha in 2003 (Ministry of Agriculture, 2005). While cooperatives continue to be more important than companies in controlling land in grain-growing regions, it is in these regions that the growth of companies has been particularly important6.

15Although the shareholding sector has undergone significant restructuring, the individual farm sector appears to have restructured little since 1994, or even 1989. Instead, farms look a lot like the personal plots which existed during the socialist period.

16In the remainder of the paper, we examine the individual farming sector in more detail. In particular, we look for evidence that a nascent commercial farming sector (fermeri) may be disguised in the aggregate data. We also ask what role these small farms, which control almost half of agricultural land, play for the households that farm them. Are these small farms, as they were under socialism, supplemental to the income of salaried rural workers? To what extent might they instead reflect a return to the predominantly subsistence agriculture of the pre-war period ? And, are there clear patterns in which types of households use their land in these distinct ways ?

3 - Individual Farming in Bulgaria (2003)

  • 7 The authors wish to thank the Bulgarian National Statistical Institute for permission to use this d (...)
  • 8 Land was restituted on the basis of pre-war ownership, and sales of agricultural land remain limite (...)

17To examine more closely the individual farm sector, we use the 2003 Bulgarian Integrated Household Survey conducted by BBSS Gallup International under the supervision of the World Bank. Using a stratified two-stage cluster design, a nationally-representative sample was drawn by the National Statistical Institute (NSI)7. They interviewed 3 023 households. The BIHS03 questionnaire contains comprehensive information on land use, agricultural production and consumption, and employment. Because we are interested in changing patterns of land use, we include in our analysis only households which report owning land. There are 1 789 such households8.

18The BIHS data presents an overall picture of highly fragmented individual farms. Sample households own an average of 1.2 ha. Over fifty percent of households have less than 0.5 ha, while very few own over 5.0 ha.

  • 9 Some of the 149 families which did not work any of their land did raise some livestock. However, si (...)

19Not all households farm their land. In our sample, about a quarter farmed none. One possible reason is that they do not live where the land is9. Fourteen percent of land-owning households in our sample did not own any land where they live, about half of the non-farming landholders. Another quarter of non-farming owners had only a yard in their area of residence. The most common use of non-local land was to leave it idle (39% of non-farming households), while others (34%) put land in a cooperative and the rest rented land to others.

20Below, we focus on the 1 367 households which own some land in their area of residence. We consider separately those that choose not to farm any land compared to those that farm. We then consider selected subgroups of farming households, in order to evaluate whether the average characteristics of farming households disguise important new trends in land use. In table 2, we summarize the characteristics of the selected subgroups of households.

Table 2: Characteristics of Landholding Households, All Households and Sub-Groups.


All Farming Households

of which :

Commercial Farm

Subsistence Farm

Supplemental Farm


n=1218, 100%

n=139, 11.4%


n=310, 25.4%

«Pre-war» *

n=120, 9.8%

n=649, 53.3%

Owned land (ha.)







Farmed land (ha.)







Working age members







% owning tractor







% w/ag expenses







Ag sales income







Job income







Pension income







Total cash income







Source: BLSMS, 2003.
* The «subsistence households» mentioned in the text refer to this restricted category only.

3.1 - Non-Farming Households

  • 10 Only a very small number of non-landowning households reported cultivating any land (19 households) (...)

21There were 149 households (11% of the sample) with land in their area of residence that chose not to farm any land10, and they were evenly distributed across the regions of Bulgaria. These families owned slightly less land than the average household (0. 10 ha).

22Families might not farm their land if they could earn more in other endeavors, or if others would pay them more for their land than they could earn farming it themselves. In fact, on average about half of these families’ land was placed in a cooperative, while about a fifth was rented to others. But most of the rest was left idle.

  • 11 1 lev = about 1.95 Euro in 2003

23Perhaps members of these households had more lucrative uses for their time, making leaving the land idle more profitable than farming it. Non-farming households were over-represented in the South-West region, which includes the capital city, Sofia, where work might also be most available (fig. 1). Their average income of 3 316 leva  was approximately equal to average income for the sample as a whole (table 211).

24Thirty-six percent of families using no land reported having no working age member (15-65 years). In this case, lack of labor may have made farming impossible. An unexpected finding was that a significant minority (14%) of the non-farming households report having working age members but no employed members. These 34 families thus have both unused land and labor, and very low incomes (an average of 1 449 leva), almost half of which comes from pensions. Farming appears a possible means of improving economic conditions for these families, but was not chosen. Perhaps impediments to farming (lack of machinery, skills, credit) outweighed the income advantages.

3.2 - Farming Households

25The remaining 1 218 households which own land where they live farmed at least part of their land. These families owned an average of 1.3 ha where they lived (and a very small amount elsewhere, table 2), and landholding size remains uniformly small. As in the prewar period, half of the households owned less than 0.5 ha, and 80% owned less than 2 ha. Very few possessed more than 10 ha.

26Landowners farmed just under half their land themselves, just slightly more than in 1986 (Meurs, Kouzhouharova, and Stoyanova, 1992). Farming landowners put an average of 0.4 ha in a cooperative and rented about 0.2 ha to other farmers. They left a small amount idle. Households have small plots, but prefer even smaller ones.

27Rural households are also small. On average, landowning families were composed of almost three members, of which two were of working age and one was a pensioner, who might also be available for farm work.

28The landowning households have very few machines. Mechanization declined dramatically in the agricultural sector as a whole over the period 1988-2003, with the number of combines in Bulgaria falling by half, and land per tractor rising from 32 ha to 78 ha, even as land in use fell. There was one tractor for every 15 of the individual farms in 2003, and 42% of these belonged to the tiny share of farms over 10 ha (Ministry of Agriculture, 2003).

29In our sample, less than 3% of households reported owning a tractor. No one reported owning a combine. Of course, machine services can be purchased, but only about one third of households did use some machine services.

30Producers spent limited amounts on other purchased inputs. Twenty-seven percent reported spending no money at all on agricultural inputs. The most frequently mentioned expenses were feed for livestock, followed by seed, veterinary services, fertilizer, and herbicides. Few households announced spending on hired labor, machines or land. An average household spent 259 leva in 2003 on purchased agricultural inputs.

31Agricultural production was highly diversified. Most holders grew vegetables (88% tomatoes, 61% potatoes). About half grew fruits (especially apples and grapes), fewer grain crops (24% corn, 9% wheat) although these could be used to feed household livestock, which were owned by the majority of farming families.

32Most households (87%) reported no sales of any crop. The most commonly sold product was potatoes (4% of households), followed by tomatoes and grapes. Only 8% declare selling livestock, while almost a quarter sold a livestock product, most often milk. In most cases, sales from agricultural products were an insignificant share of household incomes. Pensions accounted for a third of the household income.

33Overall, the upper characteristics suggest a continuation of the socialist-era practices. Households farm very small plots and, appropriately, own few machines. Individual producers grow products for their own use, although a small group also produces a small surplus for sale, especially milk, as they did during the socialist period.

3.3 - One kind of commercial farming activity

34Perhaps, however, these average figures disguise emerging new types of farms. We consider two possibilities : households that produce mainly for the market (“commercially oriented farms”) and households which rely entirely on their small plots for survival, but do not generate cash income (“subsistence farms”).

35Distinguishing a commercially oriented farmer is not simple. At a minimum, such units must sell some agricultural output. Only 13% of our sample does so. But beyond this requirement, commercial farms could take very different forms. Very small farms could grow vegetables intensively for the market, with little land or machinery.

  • 12 We also considered defining commercial farming households as those earning substantial income from (...)

36In order not to exclude any commercially oriented producers, we use a very broad definition. We include farms that produce at least one item mainly for sale (as measured by sale of half or more of the production of any one product)12. By defining a commercial orientation in this way, we hope to capture the intent to produce for sale. Sales of a smaller share of production seem more likely to be sales of a surplus of goods produced mainly for own consumption. Production of one good mainly for sale does not indisputably define a commercially oriented farm. The data suggest, however, that farmers meeting this definition are indeed distinct from the average producer, as described above.

37There were 139 farms producing at least one crop mainly for sale (11% of the sample). In most cases, only one product was concerned : tomatoes, followed by potatoes, wheat, and cabbage are the most commonly marketed crops. Those commercial producers sold an average of about 2 000 kg of wheat and tomatoes, about 1 000 kg of potatoes and 400 kg of cabbage. The vast majority of them sold less than this amount, however.

38Commercially oriented farms earned this way an average of 1 345 leva in 2002, about a quarter of their household income. But the distribution of agricultural income was heavily skewed to the low end, with three-quarters of such households earning less than this amount. The total cash income of these commercial farmers (5 696 leva) was much higher than that of the standard landholding family. For a quarter of them, the selling activities provided more than one-third of their income.

39These farms were also distinct from the other units in their use of agricultural inputs. They had more agricultural machinery : 6% owned a tractor, 3% a planting machine, 7% owned a truck, 8% owned a mill, and 5% had a mechanical plow (tiller). They were more likely to purchase inputs (90%), including the mix of inputs purchased by other. Distinctly, a significant number also hired in workers (13%) and land (15%), and paid for irrigation and transport.

40Like other landholders, these farmers owned an average of 1.5 ha, but more holdings (almost 25%) were above 2 ha. These households cultivated more of their own land than the average landholding household, and they also rented in another 0.3 ha, but they still farmed less land than they owned. Farms remained very small, with 90% of farms being under 2 ha. Like other households, they did not cultivate all their land. They worked about 1.3 ha of their own land. The rest was about evenly divided between a cooperative and other renters. Very little was left idle. These farmers use rental markets to restructure their landholdings.

41Commercially oriented households were not evenly distributed across the country. These farms, which are not oriented toward grain production, are under-represented in the grain growing regions of the northern part of the country. Instead they are clustered in the South-West region around the capital city, Sofia and in the adjoining South-Central region, where garden vegetable and tobacco production have historically been more common.

42Finally, to support such activities, the households of this category had more members (an average of 3.3), and more working age members (2.2). All these data confirm the work of Adrian Smith (2000) who found that “work rich” households tended to engage in higher levels of agricultural production than households with lower employment incomes.

43Further expansion of their agricultural production seems possible, as these producers do not farm all their land. Possible reasons for this failure to expand are explored in the conclusions.

3.4 - The subsistence farming units

  • 13 Unemployment remains very high in some parts of Bulgaria. In 2003, municipal unemployment rates ran (...)
  • 14 The definition does not rule out occasional work for cash, or exchange of agricultural production o (...)

44At the other end of the spectrum is a group which we will call “subsistence farms”. Like commercially oriented farms, subsistence farms could be defined in many ways. We keep a narrow definition. These are households which report no member regularly employed13 and no sales of agricultural products. They depend almost entirely on their farm for survival14.

45There are 430 households in the sample which meet this definition. Three hundred and ten of them contain no working age members, however. These retired persons own more land than other households in the group, but they farm very small amounts (0.3 ha) and leave about the same amount idle. The majority of their land is put into cooperatives (0.6 ha). Retirees own almost no agricultural machinery. They survive mainly on pension payments, which account for 96% of their low incomes (1 739 leva). We exclude them from our analysis in order to focus attention on the possible emergence of households with working age members who depend entirely on their own (non-commercial) production for survival, seeming to return to the practices of pre-war agriculture.

46The remaining 120 subsistence farming households have an average of 3.3 working age members, in addition to 0.6 pensioners. They owned slightly less land than average (1 ha), and farmed the same amount as the average (0.6 ha, table 2). Most land not used by the households was rented to others (0.3 ha), or placed in a cooperative (0.2 ha). Subsistence farming households left almost no land idle.

47Despite farming more land than average, without a salary income subsistence farming households were less likely to spend money on purchased inputs (35% of households did not) ; on average they spent under 100 leva on inputs.

48As noted above, few landholding households sold any production. In this sense, subsistence farmers (as we have defined them) are not different from the overall group of landholders. What makes the subsistence farming group distinct is their dependence on farming -the absence of an employment income from outside agriculture. With no regular job and no marketed production, subsistence farming households had the lowest per capita cash incomes of any group, of which a large part came from pensions (and the rest from occasional work).

49While these families appear to have surplus labor, they rent about 0.6 ha to cooperatives and other farmers. Perhaps those renters offer higher returns than the households could hope to earn themselves, either because the households cannot access needed capital, or because the other farms enjoy economies of scale not achievable by these farms.

50Limited access to markets does not seem to be the main reason for these farms’ lack of production for sale. Subsistence farmers were under-represented in the grain-growing region of the North-West, and slightly over-represented in the South-East, which is far from most urban markets, but otherwise they were not clustered in any particular region.

3.5 - Supplementary farms: the inherited socialist-period model

51Although the commercially oriented and subsistence farms do appear distinct from the socialist-period model of supplementary farms, a much larger share of the sample (53%), seemed to have changed little from that model. These households had at least one income from regular employment outside agriculture, which they combined with agricultural production mainly for personal use.

52As these households look most like the average landholding household described at the beginning of this section, we only briefly review their characteristics. They own an average of 1.2 ha, of which they farm just under one half. The largest share of the remaining land (0.4 ha) is placed in a cooperative, while about equal amounts are rented to others and left idle. Supplementary farm households had about 3.3 working age members, in addition to 0.3 pensioners.

53Like other units, supplementary farming households own few machines. Using their off-farm income, these households are more likely than subsistence farmers to spend money on agricultural inputs, and they spent an average of just under 300 leva. They buy the same mix of inputs.

54Like subsistence households, the supplementary households grew corn, to feed livestock, as well as garden vegetables and grapes. They were much more likely to produce milk and milk products than were subsistence households, however, allowing them to diversify their consumption. The majority (70% of them) did not sell any production. When supplementary farming households did market production, milk was the most frequently sold product, followed by potatoes and sheep.

55Non-farm income contributes importantly to the well-being of these families. The supplementary farming group report an average of 3 032 leva in job income, plus an average of 883 leva in pensions, and a tiny amount (21 leva) of income from agricultural sales. Total reported cash income was 3 936 leva, placing this category below commercial farming households in cash income, but well above the other group. While these households did have land which they did not use, they mainly rented it to others. Their household members were fully employed.

4 - Examining Underlying Factors

56In this section, we use simple regression analysis to examine patterns of household characteristics, location, and land use. Previous work on this topic (Meurs, 2002 ; Swinnen and Mathijs, 1997) suggested that resource constraints, market conditions and access, and the history of local agricultural specialization influence household choice of how to use privatized land.

57Some households may not have the resources needed to farm all their land efficiently. Poorly developed land markets may make it difficult to adjust landholdings, limited labor markets may limit the labor recruitment during peak periods, and weak credit markets may restrain the capital substitution for labor, or capital acquisition needed to complement existing labor or landholdings.

58Location can also play an important role. Proximity to urban markets may reduce costs and increase awareness of market opportunities. Agricultural traditions and infrastructure differ significantly across the regions of Bulgaria, in part due to specialization under socialist planning. Grain production has traditionally dominated certain regions (in the plains, but these regions are often also farther from urban centers), while other areas (more mountainous and/or closer to urban markets) have strong traditions of vegetable growing or livestock grazing. These differences were important in explaining patterns of de-collectivization of land during the early 1990s (Meurs, 2002).

59Finally, analysts have recently argued that state-supplied incomes, such as pensions or social welfare payments, may reduce household incentives to work (Haan and Myck, 2007 ; Jones and Kotzeva, 1998). Given the significant levels of pensions received by the households examined above, we investigate whether higher-pension households are less likely to farm for cash (social welfare payments are insignificant for the landholding households).

60We use two probit models to investigate the factors correlated with a household’s practice of producing mainly for sale or farming only for subsistence. Because of the necessarily subjective nature of our definitions of commercial and subsistence farming, we also examine factors related to the sale of any production at all in a third probit model. We include in the regression analysis all landholding households engaged in farming, the 1 218 units discussed above.

  • 15 As noted above, analysts have also argued that credit constraints may limit production. Hertz (2007 (...)

61In the regressions, we examine the impact of landholding size and supply of working aged household members (resources available for farming) on farm orientation15. The previous descriptive analysis suggests that the families use rental markets to adjust their landholdings, and that the households generally prefer to farm less land than they own. We do not, therefore, expect landholding size to be a significant factor in farm orientation. Households have few working age members on average, however, and few of them use labor markets to adjust their labor inputs into agriculture, so the availability of labor may play a significant role in household choices of how to drive their activities.

62We also include in the regression the share of land in the oblast which was under grain in 1988, to examine the impact of local agricultural traditions and infrastructure on how households use land. The large, collectively owned farms persisting in grain-growing regions may enjoy economies of scale, offering higher returns and making other commercial uses of land less viable.

63To further examine the impact of local conditions, we include dummy variables for six regions (Ministry of Agriculture, 2005, fig. 1). These region variables may capture access to both urban markets and off-farm employment, as well as characteristics of physical geography. As is the case for dummy variables, one region is omitted from the regression. Coefficients on other regional variables measure the impact of being in that region, as compared to the omitted region. Finally, to examine whether pension payments create disincentives for market activity, we include the number of pension-aged people in the household in the regression.

  • 16 Because the data were collected using a cluster method, we run the regression using Robust Cluster, (...)

64We run the regressions as d-probit in STATA. In this analysis, the outcome variable takes a value of one if the household has specified orientation (commercial farming, for example) and zero otherwise. The regression results report the marginal effects on household and local characteristics on farm orientation. Specifically, they estimate the impact of a one-unit change in household or local characteristics on the likelihood that an average household will have the specified orientation16. For all regressions, we find variables which have a statistically significant relationship to the orientation of the household. But much of the variation in household orientation remains unexplained, as seen in the low pseudo-R2. Other factors also play a significant role in the household orientation. The results are presented in table 3.

Table 3: Factors Associated with Farm Type, Landholding Households, Bulgaria, 2003.

Dependent Variable

Commercial Farms


Pre-war Subsistence Farms***


Farms with Any Sales n=1218

Pseudo R2




Independent Variables

Marginal Effect


Marginal Effect


Marginal Effect


Land Owned (ha)







Working Age Members














% oblast land in grain







Region 1







Region 2







Region 3







Region 5







Region 6







*Variable has a significant relationship to outcome. Probability that relationship is insignificant <.10.
**Variable has a significant relationship to outcome. Probability that relationship is insignificant <.05.
***Excludes retirement farms.

65Looking first at the factors which distinguish commercially oriented producers from landholding households as a whole, the presence of an additional household member significantly increased the likelihood of commercial production. For the regional dummies, STATA identified Region 4, the South-East region, as the appropriate reference category. Compared to producers in Region 4, location in Regions 5 and 6 (the rest of southern Bulgaria, closer to the urban centers of Sofia and Plovdiv and areas traditionally less specialized in grain production) had a large, positive impact on a producers’ likelihood of engaging in production mainly for sale, whereas location in the traditional grain growing Region 3 decreased the likelihood of individual commercial production. Controlling for region, the share of land traditionally used for grain had an unexpected positive impact of the likelihood of producing mainly for sale. Perhaps this oblast-level variable captures within-region differences in the suitability of land for agriculture (slope). As suggested by the analysis above, the size of the households’ landholding was not significantly related to commercial production.

66The regression finds few factors correlated with the likelihood of engaging in subsistence farming. Only location had an impact. Compared to Region 4, location in most other regions reduced the probability of engaging in subsistence farming by about 5%. As noted above, Region 4 is relatively distant from urban markets (other than the city of Burgas). It is an area with a relatively mixed ethnic population, including significant representations of Roma and muslims, and with some areas devoted to tobacco farming. Neither land nor labor available seemed to affect the decision to farm only for subsistence. Further, pensions do not appear to encourage withdrawal from labor markets and limitation of production for sale.

67Finally, considering households which report any sales (compared to households which sell nothing), we again note the centrality of household labor supply. An additional working aged member significantly increased the likelihood of selling some production. As was the case with commercial farmers, location in Region 3, the North-East, had a big negative impact on the likelihood of producing something for sale, while location in Region 5 increased it. Again, landholding size and the availability of pension payments pay no role in the likelihood that a household will market some production.

68Overall, the regressions provide a few important insights into the limitations on the expansion and commercialization of private farming in Bulgaria. Not surprisingly, given the aggregate data on declining land use in agriculture, and the evidence that most households use rental markets to adjust their landholdings, there is no evidence that small land-holding sizes affect household choices to engage in market production. The most consistent and important factor in likelihood of market production is the presence of additional available labor. Pensions, as shown by the number of pension-aged members in a household, do not appear to provide a disincentive to produce a surplus for sale.

5 - Discussion

69The Bulgarian agricultural sector has undergone important changes since the collapse of the Jivkov government in 1989. Property rights in land have been restituted, and the share of agricultural land in collectively-owned farms has declined from 86% to 47%. At the same time, the amount of agricultural land in use in Bulgaria has fallen and machine power available in agricultural has also declined dramatically (Ministry of Agriculture, 2005).

70Despite these important changes, Bulgaria has not seen the restructuring of landholding into commercially oriented family farms that many expected would follow the extension of property rights and markets. Instead, most Bulgarian households continue to use their land the way they did under socialism : to grow vegetables and fruit for home use, supplementing off-farm employment. But whereas before 1989 this pattern of supplementary farming accounted for 14% of agricultural land, it now accounts for 53%. A much smaller share of land is now farmed in large-scale, mechanized farms.

71Careful examination of the household data highlights two groups distinct from the broad pattern of continued supplemental farming : a few (11% of our sample) commercially oriented households, and a small group (10%) of subsistence farms. The former devote significant resources to agricultural production for sale and earn an important share of income from these sales. These households, focused mainly on vegetable farming, are most likely to be located in the southern regions of the country, rather than in the historically grain-growing regions of the north. Subsistence farms practice very low-input agriculture and produce nothing for sale, despite having no member regularly employed in off-farm work and very low incomes. Subsistence households are distributed more evenly across plans and non-plains regions.

72With commercially oriented farm households earning higher than average incomes, two questions arise. First, why do these families not expand their activities onto a larger share of their owned land ? And second, why do subsistence households (and non-farming landowning households without alternative income sources) not use available land and labor to produce for sale ? Small landholding size does not appear to play an important role in the decisions of how to use the land, as households use rental markets to adjust the size and location of their farm. Farms do not yet use hired labor regularly, and the number of adults in a household does appear to play a role in land use decisions, but subsistence farming households seem to have labor available to intensify production.

73A full analysis of why households do not use available land and labor to expand or intensify agricultural production would require a much more detailed examination of returns to these resources in agriculture and alternative uses than is possible here. In conclusion, however, we offer three possible reasons for this.

74First, families may face credit constraints. Hertz (2007) found that credit constraints did not appear a limiting factor in agricultural production, given that agricultural households borrow for consumption. But perhaps households find it easier to borrow find that the commercially-oriented households also have the highest employment incomes, which may permit greater investment in agricultural production. Without production credit, lower income families may be unable to finance the capital or inputs needed to make expansion and intensification profitable.

75Second, perhaps some Bulgarian agricultural households are not “maximizers,” seeking to the most household income possible given their resources. Perhaps they satisfy themselves (Simon 1988), or seek to achieve a benchmark of “normal” consumption. Having lived many years on a small agricultural worker’s salary and the output of small plot, families may accept low levels of income as normal and seek only to maintain those.

76Third, returns to agricultural are low. This is evidenced by the large-scale withdrawal of land from agricultural use, the recent expansion of meadows and pastures in the place of arable land, even as livestock herds decline, and the radical fall in the use of purchased inputs. Returns may not outweigh costs of expanded or intensified production.

77Further research may begin to clarify why agricultural land use has declined in Bulgaria, why apparently profitable commercially oriented individual farms do not expand onto more land, and why more individually farmed land is not used for commercial production. What is clear now is that the expected broad emergence of “farmers” still has not occurred in Bulgaria. Significant restructuring may still lie in the future for the Bulgarian corner of southeastern Europe.

Top of page


Berov Liuben, (1977), « Социалната структура на селото в периода между двете светови войни [Village Social Structure in the Balkan Countries between the World Wars] », in Трудове на Висшия Икономически Институт Карл Маркс 4 [Works from the Higher Economics Institute Karl Marx 4], p. 62-65.

BNIS (Bulgaria National Institute of Statistics), Statistical Yearbook of the Republic of Bulgaria, Sofia, BNIS : various years.

BNIS (Bulgarian National Institute of Statistics), Regions, Districts and Municipalities in Bulgaria (CD rom), Sofia : 2001, 2002, 2003.

Creed Gerald, (1997), « An Old Song in a New Voice : Decollectivization in Bulgaria » in David Kideckel, ed., East-Central European Communities : The Struggle for Balance in Turbulent Times, Boulder, Westview.

Eurostat on-line statistics.

Haan P., Myck M, (nov. 2007), Safety Net Still in Transition : Labour Market Incentive Effects of Extending Social Support in Poland, Institute for the Study of Labor, IZA DP No. 3157.

Hertz Thomas, (2007), The Effect of Non-Farm Income on Investment in Bulgarian Family Farming, FAO Working Paper, American University Working Paper.

Jones D., Kotzeva M., (1998), « Work incentives and other effects of the transition to social assistance in the transition economies : Evidence from Bulgaria », Empirical Economics, 23:237-261.

Kaneff Deema, (1996), « Responses to “Democratic” Land Reform in a Bulgarian Village » in Ray Abrahams, ed., After Socialism : Land Reform and Social Change in Eastern Europe, Providence, Berghahn Books.

Lampe John, (1986), The Bulgarian Economy in the 20th Century, New York, St. Martin’s Press.

Lampe John, Jackson Marvin, (1981), Balkan Economic History, 1550-1950, Bloomington, Indiana University Press.

Meurs Mieke, (2002), The Evolution of Agrarian Institutions, Ann Arbor, University of Michigan Press.

Meurs Mieke, Kouzhouharova Veska and Stoyanova Rositsa, (1992), « Bulgaria : From Cooperative Village to Agro-Industrial Complex, the Rise and Fall of Collective Agriculture », in Mieke Meurs, ed, Many Shades of Red : State Policy and Collective Agriculture, Boulder, Rowman and Littlefield.

Ministry of Agriculture, Agricultural Census in Bulgaria. 2003 Results, (2005), Sofia, Ministry of Agriculture.

Ministry of Agriculture, Structure of Agricultural Holdings in Bulgaria, (2006), Crop Year 2004/2005, Sofia, Ministry of Agriculture.

Smith Adrian, (2000), « Employment restructuring and household survival in post-communist transition : rethinking economic practices in Eastern Europe », Environment and Planning A, vol. 32, p. 1759-1780.

Simon Herbert, (1988), « Rationality as a Process and Product of Thought », in Decision Making : Descriptive, Normative, and Prescriptive Interactions, David E. Bell, Howard Raiffa, and Amos Tversky, eds., Cambridge, Cambridge University Press.

Swinnen Johan, Mathijs Erik, (1997), « Agricultural Privatization, land reform, and farm restructuring in Central and Eastern Europe : A Comparative Analysis », in Johan Swinnen, Allan Buckwell and Erik Mathijs eds., Agricultural Privatization, land reform, and farm restructuring in Central and Eastern Europe, Aldershot, Ashgate.

WDI (World Development Indicators), (2000), CD Rom, Washington, DC, The World Bank.

Top of page


1 These were often slightly restructured versions of the old collective farms, although the land is now owned by the cooperative members (Meurs, 2002; Kaneff, 1996).

2 In the late 1980s there were experiments with decentralization to smaller units, like production brigades. Farms, however, remained large-scale.

3 Bulgaria was divided into 28 administrative districts in 1988 (then called okrugs, now oblasts).

4 In a simple linear regression, the share of oblast land in grain production in 1987 explains 37% of the variation in share of land held individually.

5 The increase in meadows and pasture land is not driven by expanding livestock production. From 1988-2003, numbers of all types of livestock fell dramatically, with the exception of goats (BNIS, various years).

6 The share of an oblast’s land under grain in 1987 explains 58% of the variation in companies’ share of land in a simple linear regression.

7 The authors wish to thank the Bulgarian National Statistical Institute for permission to use this data.

8 Land was restituted on the basis of pre-war ownership, and sales of agricultural land remain limited. Our data do not allow us to distinguish well which households live in urban areas, or the distance between residence and owned land, so included in this group are both urban households, and rural households which own land where they do not reside.

9 Some of the 149 families which did not work any of their land did raise some livestock. However, since we are mainly interested in land use, families which raise livestock but do not farm any land will not be included in this analysis.

10 Only a very small number of non-landowning households reported cultivating any land (19 households), and these were excluded from the analysis.

11 1 lev = about 1.95 Euro in 2003

12 We also considered defining commercial farming households as those earning substantial income from agriculture. Using the benchmark of 30% of household income, nearly all households which met the income definition also met the “production mainly for sale” definition, while the reverse was not true.

13 Unemployment remains very high in some parts of Bulgaria. In 2003, municipal unemployment rates ranged from 5% to 60% (BNIS, 2003).

14 The definition does not rule out occasional work for cash, or exchange of agricultural production or labor with neighbors.

15 As noted above, analysts have also argued that credit constraints may limit production. Hertz (2007) found little evidence that credit constraints limited agricultural production by households in Bulgaria, however. We do not attempt to measure the impact of credit constraints because credit, more than household size or landholdings, is as likely to be driven by production choices as it is to cause them.

16 Because the data were collected using a cluster method, we run the regression using Robust Cluster, in order to correct for possible correlation in the error terms.

Top of page

List of illustrations

Title Fig. 1 - Oblasts and regions of Bulgaria.
Credits CAO : R. Darques
File image/jpeg, 68k
Title Fig. 2 - Land farmed individually, 2003.
Caption Source Ministry of Agriculture and Forestry, 2005:44; BNIS, 1987:547.
Credits CAO : R. Darques
File image/jpeg, 44k
Title Fig. 3 - Decline in Agricultural Land Use 1987-2003.
Caption Source: Ministry of Agriculture and Forestry, 2005:44; BNIS, 1987:547.
Credits CAO : R. Darques
File image/jpeg, 44k
Title Fig. 4 - Land Farmed in Companies, 2003.
Caption Source: Ministry of Agriculture and Forestry, 2005:44; BNIS, 1987:547.
Credits CAO : R. Darques
File image/jpeg, 44k
Top of page


Bibliographical reference

Mieke Meurs and Angel Bogushev, « Forward to the Past ? Agricultural Restructuring in Bulgaria », Méditerranée, 110 | 2008, 93-104.

Electronic reference

Mieke Meurs and Angel Bogushev, « Forward to the Past ? Agricultural Restructuring in Bulgaria », Méditerranée [Online], 110 | 2008, Online since 01 January 2010, connection on 05 July 2022. URL : ; DOI :

Top of page

About the authors

Mieke Meurs

PhD Program Director and Professor of Economics, American University,

Angel Bogushev

American University

Top of page


Tous droits réservés

Top of page
  • Logo Presses Universitaires de Provence
  • OpenEdition Journals
Search OpenEdition Search

You will be redirected to OpenEdition Search