An Institutional Approach to the Soft Budget Constraint

  • Gun Eriksson Skoog


Before the story of the soft budget constraint can be told, the analytical framework that gives structure to this story must be established. The soft budget constraint is our central metaphor. Institutions form another, which is almost as crucial. We shall define the soft budget constraint as an institution and thereafter use a third metaphor — an invisible-hand explanation — to tell the story of the emergence, persistence and logic of the soft budget constraint.


Budget Constraint Socialist Economy Aspiration Level Behavioural Rule Institutional Approach 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    As noted by Jasinski (1993), footnote 3, p. 3, ‘A full bibliography of what Kornai wrote on the soft budget constraint would have to include almost everything he published since 1979’.Google Scholar
  2. 7.
    The degree of softness may, for instance, be relatively higher in sectors to which the state attaches particular priority. Davis (1989) has shown that the medical system, which was given relatively low priority in the former Soviet Union, faced fairly hard budget constraints, while Ericson (1988) suggests that they were relatively soft in the highly prioritised military sector. Kornai (1992), p. 143, footnote 20, notes that large firms tend to face softer budget constraints than small firms. See also Qian and Roland (1994). For further discussion on the degree of budget softness, see for instance Gomulka (1985) and Scott (1990).Google Scholar
  3. 8.
    This literature is largely inspired by Dewatripont and Maskin (1989), and later versions, eventually published in a revised form (1995); it includes, for instance, Berglöf and Roland (1994), Hardy (1992), Qian (1994) and Qian and Roland (1994), and more recently, for example Huang and Xu (1999), Maskin and Xu (1999). See also Lindbeck and Weibull (1988).Google Scholar
  4. 9.
    As pointed out by and Qian and Roland (1994), p. 1, ‘Time inconsistency is at the heart of the soft budget constraint problem’.Google Scholar
  5. 13.
    Kornai (1990b), pp. 22–23. For a more detailed discussion of the budget-softening characteristics of external assistance, see also Eriksson (1993), pp. 19–21.Google Scholar
  6. 16.
    See Komai (1986c), p. 43, and (1980), p. 308. Scott (1990), pp. 120–122, notes that action in the control sphere is identical to rent-seeking or directly unproductive profit-seeking activity.Google Scholar
  7. 18.
    See Kornai (1986b) and Scott (1990) on the degree of patemalism and its relationship with the soft budget constraint.Google Scholar
  8. 19.
    See Kornai (1990b), p. 35, and (1992), pp. 501–502, which also refers to several empirical studies on the subject.Google Scholar
  9. 22.
    Besides, when losses are compensated for, profits are often taxed away in a discretionary manner. On the ‘levelling’ of profits between Hungarian state-owned enterprises, see Komai and Matits (1984). When the firm is unable to benefit from its profits, incentives to spend all available resources are reinforced. Cf. Hare (1989), pp. 56–57, Kornai (1992), p. 146, and Nagaoka and Atiyas (1990), p. 15.Google Scholar
  10. 27.
    Komai (1979). A debate on the character and interpretation of shortages or excess demand in socialist economies has taken place between Komai and adherents of the so-called disequilibrium school. For a survey of the debate, see van Brabant (1990) and a conference volume edited by Davis and Charemza (1989). A related debate concerns the relevance of the soft budget constraint for the prevalence of shortages in socialist economies. Kornai introduced the concept of the soft budget constraint to explain these persistent shortages. However, other factors have been forwarded as explanations. See Gomulka (1985), Kornai (1985), Hare (1989), Bajt (1991) and Jasinski (1993). Besides, certain authors, notably Gomulka (1985), consider efficiency loss rather than shortages to be the main consequence of the soft budget constraint.Google Scholar
  11. 28.
    Magee and Quandt (1994) explore theoretically how the soft budget constraint may lead to a higher than normal demand for inputs. Cf. Qian (1994).Google Scholar
  12. 29.
    See Xu and Qian (1992) for a theoretical examination of the relationships between the soft budget constraint, investment and innovation.Google Scholar
  13. 32.
    A thorough discussion of shortages and inflation, their causes, including the soft budget constraint, and the relationship between them in the socialist system is provided by Kornai (1992), Chapters 11–12, pp. 228–301.Google Scholar
  14. 31.
    As suggested by empirical studies, discussed in the previous chapter. 3a Hay etal. (1994), pp. 325–326.Google Scholar
  15. 37.
    As suggested by Nagaoka and Atiyas (1990), p. 10; and by Shleifer and Vishny (1994), referred to by Qian and Roland (1994), p. 3, footnote 4.Google Scholar
  16. 43.
    Compare this proposition with the case of perfect information. If the state knew in advance that refinancing of a bad project would not be worthwhile, sunk costs would be genuinely sunk, which according to theory implies that they would have no opportunity value and therefore no bearing on refmancing decisions.Google Scholar
  17. 47.
    Nagaoka and Atiyas (1990), p. 10. Cf. Lin and Tan (1999), who trace the roots of the soft budget constraint to the ‘accountability problem’ of the state.Google Scholar
  18. 49.
    As reflected, for instance, in thematic issues of the American Economic Review (1999), Japan and the World Economy (1996) and the Journal of Comparative Economics (1998).Google Scholar
  19. 50.
    Komai (1998), referred to by Bai and Wang (1999), Maskin (1996) and (1999). 5I By Li (1992) and (1998), and Segal (1998).Google Scholar
  20. 70.
    See, for instance, Verdery (1993) for an analysis of the fall of the socialist system. Kornai (1992), P. 384, footnote 1, provides references on the tensions within the system.Google Scholar
  21. 82.
    On the affinity between ownership forms and co-ordination mechanisms, see Kornai (1990a) and (1992), pp. 447–450 and 497–500.Google Scholar
  22. 85.
    See, for instance, Begg and Portes (1992), p. 8, Raiser (1992), pp. 35–36, and (1993), pp. 266–267, and Zhang and Sjöberg (1992), pp. 33–34. Buch et al. (1994), p. 43, provide additional references.Google Scholar
  23. 86.
    See, for instance, Berglöf and Roland (1994). Nagaoka and Atiyas (1990), pp. 12–16, provide several arguments, but their focus is on the pre-transition period.Google Scholar
  24. 87.
    Besides, given the accumulated debts, it may be difficult to assess a firm’s future profitability potential.Google Scholar
  25. 88.
    Begg and Portes (1992), pp. 7–12 (quotation from p. 11), partly referring to Mitchell (1992).Google Scholar
  26. 89.
    See Kornai (1986c), pp. 46–47, and (1990b), pp. 36–41.Google Scholar
  27. 90.
    As shown in Chapter 1, there are studies suggesting that soft budget constraints prevail in developing countries. Kornai (19906), p. 41.Google Scholar
  28. 97.
    Zhang (1997), pp. 75–109, found that collective - as opposed to non-collective - rural enterprises under local government ownership in China faced basically soft budget constraints despite the hard budget constraints of the local governments, partly because the firms were able to seek soft funding elsewhere.Google Scholar
  29. 98.
    See, for instance, Bardhan (1989), Eggertsson (1990), Gunnarsson (1991), Knudsen (1993), Langlois (1986c) and Raaschou-Nielsen (1988).Google Scholar
  30. 99.
    Raaschou-Nielsen (1988), p. viii; free translation from the Danish. Schotter (1981), p. 11.Google Scholar
  31. 102.
    Later, Schotter (1986), pp. 118–119, makes a distinction between the two views on institutions, by letting them represent two basically different approaches within institutional economics, and he finally defines social institutions as sets of rules.Google Scholar
  32. 104.
    Cf. Hayek (1973), p. 43, according to whom a ‘regularity, of course, means simply that the elements behave according to rules’.Google Scholar
  33. 114.
    Simon (1982b), p. 390; emphasis added. Cf. the position, common within the study of comparative economic systems, that an economic system consists of a set of interdependent, mutually consistent and reinforcing parts (Grosfeld, 1990, p. 7), and Komai’s affinity between ownership forms and co-ordination mechanisms discussed above.Google Scholar
  34. 115.
    Such institutional ‘disequilibrium’ is, according to North (1990), pp. 87–90, passim, characterised by tensions between formal and informal rules. However, in other cases, change in formal rules involves an adaptation to already prevailing formal rules, and thus brings consistency between them.Google Scholar
  35. 126.
    The first type of uncertainty, about ‘which pregiven state will obtain’, is referred to by Langlois (1986b), p. 228, as parametric, while he calls the latter type, uncertainty about ‘which states are possible’, structural. Google Scholar
  36. 127.
    Heiner (1986), p. 60, quotes Hey (1979), p. 232, to illustrate the problem. Often ‘the optimization problems that… agents are supposed to be solving… are so complicated that the economic theorist… probably spent several months finding the solution’.Google Scholar
  37. 141.
    ‘A rule of action is rational if, by following that rule, an agent maximizes his expected utility.’ (Rowe, 1989, pp. 4–5.)Google Scholar
  38. 144.
    ‘Aspirations are expectations - adjusted in the long run to realities - of the result that can reasonably be attained.’ (Simon, 19826, p. 399.) For details, see Simon (1972), p. 168, (1976), pp. 133 and 136, (1978b), p. 10 ff., and (1979), pp. 502–503.Google Scholar
  39. 156.
    Case-by-case maximisation entails the risk of making the wrong choice, as noted by Vanberg (1993), pp. 177178, referring to Heiner (1983), (1987) and (1990), whereas rule following entails the risk of missing preferred exceptions.Google Scholar
  40. 169.
    Cf. Vanberg (1993), p. 172, who distinguishes between rule following as ‘an attribute of individual human behaviour’ and ‘the role of rules in coordinating human social interaction’.Google Scholar
  41. 170.
    Langlois (1986b), p. 237, recognises that ‘institutions have an informational-support function’. Cf. Nelson and Winter (1982), referred to by Knudsen (1993), p. 293. ‘Routines are repositories of knowledge concerning how a firm can “do” different things.’ See also Schotter (1981), p. 109.Google Scholar
  42. 173.
    The uncertainty facing decision makers due to the C-D gap can also be seen as a reflection of high costs of information (and computation), and rule following is then a solution to this problem. Hence, mles reduce information costs. Cf. transaction cost economics: it is costly to transact essentially because information is costly. See Eggertsson (1990), p. 15, according to whom the resulting transaction costs include search and bargaining costs, costs associated with the making, monitoring and enforcement of contracts and with the protection of property rights. Român (1995), pp. 33–36, distinguishes between co-ordination costs (for acquiring, processing and exchanging information) and motivation costs (for specification, observation, verification and enforcement of an agreement), and briefly surveys transaction cost economics (pp. 25–26.) See also North (1990), pp. 27–35.Google Scholar
  43. 175.
    For instance, by Schelling (1960), referred to by Knudsen (1993), p. 287.Google Scholar
  44. 186.
    Within the production sphere, for instance, institutions affect, according to North (1990), pp. 64–66, both transaction costs and traditional production costs - the latter by influencing the technology employed.Google Scholar
  45. 187.
    For examples on how the ownership structure of firms may influence economic outcomes, see Eggertsson (1990), Chapter 5.Google Scholar
  46. 189.
    Cf. North (1990), p. 9, who talks about efficient and inefficient institutions. In North (1993), p. 252, he defines inefficient institutions (property rights) ‘simply as rules which do not produce increases in output’. This example reflects the problem associated with the meaning and use of these concepts once institutions are introduced into economic analysis. For a brief presentation of the arguments and a discussion, see Eggertsson (1990), pp. 2025. Demsetz (1969), referred to by Eggertsson (1990), p. 21, for instance, argues that talking about inefficiency and similar concepts is ‘misleading and ambiguous unless the outcome that they describe can be improved upon’. Given institutional interdependence, ceteris paribus assumptions thus make little sense. Cf. Vromen’s (1995, p. 174) interpretation of Menger (1985). ‘The welfare effects of some institution thus cannot be judged in isolation. The effects depend crucially on the other institutions that exists in society… the whole of which they are part.’Google Scholar
  47. 190.
    As argued in Chapter 1, this would amount to a purely functionalist explanation. The issue will be further discussed in the following section.Google Scholar
  48. 191.
    Knudsen (1993), p. 288, referring to Elster, without specifying the source further.Google Scholar
  49. 195.
    The distinction between formal and informal institutions is partly arbitrary. A formal institution may, in fact, be a reflection of how an informal institution, spontaneously evolved in society, at a certain point in time gains official recognition by being formalised, for instance, into a law. Cf. Sugden (1986), p. 5.Google Scholar
  50. 199.
    Cf. Rowe (1989), p. 5, according to whom ‘social institutions are in fact nothing more than agents rationally following rules of action, and being believed by other agents to do so’.Google Scholar
  51. 202.
    Classical references on path dependence (of technological change) are, for instance, Arthur (1988) and David (1985). For brief presentations, see Knudsen (1993), pp. 290–291, and North (1990), Chapter 11, passim. Google Scholar
  52. 204.
    Alchian (1950) is a classical reference. Others are Hayek (1978) and (1988), and Nelson and Winter (1982).Google Scholar
  53. 205.
    Knudsen (1993), p. 278. See Vromen (1995), pp. 119–121, for a discussion of the dissimilarities between natural selection and adaptive learning.Google Scholar
  54. 207.
    Cf. Vromen (1995), p. 107, who distinguishes between natural selection and adaptive learning as two ‘evolutionary “feedback” mechanisms’.Google Scholar
  55. 216.
    Bush (1988), p. 153, who notes that the direction of causality may also be the opposite: institutional change accelerating the process of echnological development.Google Scholar
  56. 226.
    Hence, he satisfies himself with less. Similarly, if he achieves or surpasses his aspiration them upwards. He learns. And he wants more.Google Scholar
  57. 227.
    Vanberg (1993), p. 174, refers to Commons’ view, noted by Biddle (1990), p. 37, that changes in the environment, our habits tend to be adjusted only after some delay.Google Scholar
  58. 228.
    Innovative behaviour leading to institutional change may also involve deliberate action, instance, to alter formal rules or renegotiation of contracts. Cf. North (1990), pp. 86–87.Google Scholar
  59. 232.
    On other occasions, he refers to the budget constraint as a behavioural regularity or a behavioural characteristic of the decision maker (the firm), and to the degree of softness of the budget constraint as a behavioural pattern. See Kornai (1986c), p. 36, Kornai (1990b), p. 21, and Kornai and Matits (1984), p. 225.Google Scholar
  60. 233.
    Cf. Kornai (1985), Figure 1, p. 50, where he defines the soft budget constraint as ‘expectation of external assistance in case of financial trouble’.Google Scholar
  61. 236.
    The strength of these expectations may vary with, for instance, the ease and speed by which financial accommodation occurs. Cf. Scott (1990) and Gomulka (1985).Google Scholar
  62. 237.
    Cf. Komai (1979), p. 807, according to whom the soft budget constraint does not result from a single event of bail-out, but when this behaviour becomes frequent and widespread, the firm’s expectations are strengthened and its budget constraint becomes soft.Google Scholar
  63. 238.
    The two parties involved in the repeated interaction are allowed to vary to some extent, for instance within the categories ‘the state’ and ‘the state-owned firms’. For instance, various state agencies may grant assistance to one state-owned firm, which repeatedly receives extemal finance, or one state agency may repeatedly grantGoogle Scholar
  64. 242.
    Just as he notes that the degree of budget softness can only be measured ordinally, not cardinally. (Kornai, I986c, p. 45.) A theoretical discussion on how to measure the degree of budget softness is provided by Gomulka (1985) and Scott (1990). For an example of how the soft budget constraint can be ‘observed’, see Kornai and Matits (1984), who examined financial data for Hungarian state enterprises for the period 19751980. A number of indicators were defined, related to different measures of profit, subsidies and investment.Google Scholar
  65. 246.
    Our reasons are pragmatic, motivated by this particular inquiry. We do not take a stand on principle in the controversy between advocates of satisficing versus optimising assumptions. For a presentation of the arguments and a discussion, see for instance Langlois (1986b), pp. 225–230.Google Scholar
  66. 250.
    Langlois and Csontos (1993) argue that this method eliminates part of the controversy between ‘neoclassical optimizers’ and ‘boundedly rational satisficers’ (p. 114).Google Scholar
  67. 254.
    As reflected in various Annual Reports by the Tanzania Audit Corporation (TAC), whose definition of commercial parastatals is, however, more narrow than ours. The TAC definition excludes, among other firms, many agricultural parastatals. See, for instance, Tanzania Audit Corporation (1992), pp. 26–55, passim. Google Scholar
  68. 256.
    The organisational structure, and the relationships within the bureaucracy of the state, the party and the parastatals, will be discussed in greater detail in Chapter III.Google Scholar
  69. 260.
    According to Alchian (1950), a certain kind of behavioural pattern may prevail regardless of the actors’ intentions, because the economic system acts as a mechanism that selects this behaviour as successful. For a discussion of attempts in the literature ‘to show that economic phenomena can be explained without any reference to preferences’, see Langlois (1986b), pp. 230–241 (quotation from p. 232).Google Scholar
  70. 262.
    Cf. North (1990), p. 25, who notes that in a strict socio-biological model, what motivates human action is maximisation of the survival potential.Google Scholar
  71. 263.
    As recognised not least by Etzioni (1988). See also, for instance, North (1990), pp. 20–22, and Williamson (1986), pp. 177–178.Google Scholar
  72. 264.
    Niskanen (1971) suggests that bureaucrats strive for budget maximisation because larger budgets increase these ‘three Ps’. (Godana, 1991, p. 30.)Google Scholar
  73. 273.
    Bates (1995), pp. 39–47, passim. See also Bates (1981), p. 2, referred to by Vromen (1995), p. 120.Google Scholar
  74. 269.
    Langlois (1986b), p. 236, adheres to the view held by both Popper (1965), 147, that it is precisely the unintended or undesigned social repercussions the social sciences should aim at explaining.Google Scholar

Copyright information

© Springer Science+Business Media New York 2000

Authors and Affiliations

  • Gun Eriksson Skoog
    • 1
  1. 1.Stockholm School of Economics — EFI The Economic Research InstituteSweden

Personalised recommendations