Former Yugoslavia in 1990: Why it had a bad prognosis

By Håkan Wiberg

Originally published here

Former Yugoslavia entered a process of dissolution many years ago, which may be far from completed yet. It took violent forms from 1991; events in 2004 in Macedonia and Kosovo indicate that we did hardly see the end here either. Was the dissolution unavoidable? Was war an inescapable consequence?

I shall attempt to translate these issues into manageable research questions, trying to make various postdictions concerning FY around 1990. There are no natural laws in social science, so the questions will deal with probabilities, asking what were the prognoses with highest likelihood at that time point. No empirical facts are drawn on that were not available at that time; confirmed general propositions are used even if they have only found empirical support later than 1990.



The first question is then: how probable was a dissolution, given the characteristics of FY and the circumstances prevailing some fifteen years ago. There is little quantitative research on when and how states dissolve. One relevant classical finding is Richardson’s (1960: 190f.) that the longer two groups lived under common government, the less likely is a civil war. This does not say anything about peaceful dissolutions; but these are historically quite rare, so this finding actually covers the great majority of cases.

The first problem concerning FY is to define its age: from 1918 or from 1945? In the first case, YU of 1990 was older than two thirds of all states; in the second case, it still belonged to the older half. Its prognosis on the basis of this indicator only was therefore about average, meaning that it was definitely less likely to dissolve than to remain. If we use qualitative analyses instead, the first problem is disagreement: some conclude that it was doomed for a number of reasons, others that it was fully viable. How convincing the pro and con arguments are is a subjective matter, or at least contains large subjective elements.

There had indeed been attempts at dissolving it, temporarily successful in 1941-45. Small armed Croatian groups from abroad failed to get much support in 1968 and were quickly suppressed. The Croatian Spring in 1971 had much more support, initially also in the party leadership, which, however, withdrew when public demands rapidly escalated from cultural autonomy to economic autonomy and from there to secession (eventually claiming large parts of Bosnia-Herzegovina); as a bid for dissolution it failed.

The Albanian uprisings in Kosovo/a in 1968 and 1980/81 demanded recognition as a nation and a republic of their own in FY; after the sanguinary defeat in 1981, largely consensual demands soon escalated to full independence. A series of economic reforms and constitutional compromises had led to less and less of central government; but this is definitely not the same thing as dissolution. In fact, some research (e.g. Galtung 1996: 67f.) indicates the opposite: contested states are more likely to survive as such if they provide for autonomy and confederalisation when demanded rather than stonewalling such demands. In any case, adding qualitative aspects to the quantitative aspect of age gives a different picture: FY in 1990 definitely belonged to the small group of states whose very existence had been repeatedly challenged from within.



Did the dissolution have to be violent if it came? The (postdictive) prognosis in this respect depends on how much we dare extrapolate from mere historical regularities. Peaceful dissolutions of states are historically rare. If we look at the 20th century up to 1990 and disregard decolonisation (where the record is more mixed) there are very few cases where a new state emerged out of an existing one without this being the result of a world war, a regional war or a civil war (Wiberg 1983) : Norway from Sweden (1905), Finland from Russia (1917),

Singapore from Malaysia (1965) and the dissolution of the United Arab Republic in 196..

All these cases are marginal in one of more ways. First, the initial relationship was rather loose. In the two first cases, there was a personal union rather than a common state: Finland and Norway had their own constitutions, parliaments, currencies, etc.. Malaysia was a federation and the UAR a loose association of two republics (in an even looser association with the feudal kingdom of Yemen).

Second, none of the associations was very old: the unions were created by military conquest, in 1814 and 1809, respectively. Malaysia was two years old and UAR three, so one might see them as failed attempts at federation rather than dissolving states. Third, the case of Finland can be seen as a result of WWI resulting in Russia collapsing earlier in 1917 (the Finnish revolution broke out in early 1918, when independence was already established and non-contested). Fourth, no independence declaration was contested, except – to some extent – Norway´s (Sweden eventually agreed after a couple of months of peaceful negotiations). Lenin immediately recognised the Finnish declaration in December 1917, Singapore was actually invited to secede by Malaysia, and the dissolution of the UAR was uncontested. Even after its major constitutional compromise in 1974 (and certainly before it), FY was a “tighter” state than either of the above.

Before 1990, there were some demands for full independence that were shelved after peaceful negotiations and/or referenda, such as Faroe Islands from Denmark in 1946, Quebec from Canada and Scotland from UK a generation ago – in neither of these cases has the issue definitely disappeared. There is a single case of success: Bangladesh from Pakistan in 1971 – but only after India invading to stop the bloodbath of Pakistan’s army (with more than ten million temporary refugees in India). Several attempts were crushed with much violence (Tibet, Biafra, Katanga,….); in other cases, fighting had continued for decades (Myanmar, Sudan, Eritrea, East Timor, …)

So the best prognosis in 1990 would have been that if there were declarations of independence in FY and if they were contested (highly probable), then it was very probable that the result would be a war where they would be defeated. In fact, there was war, but they were not defeated. The most interesting post hoc question (from an analytical point of view) is therefore why not in spite of the heavy a priori odds against.

Adding the years since 1990 to our data base would not change much. Apart from Czechoslovakia and the USSR, there were two successful independence movements. After the Indonesian invasion of East Timor in 1974, the decades of severe repression may have killed one third of the population; but the success was entirely a result of international diplomatic intervention – and threats of vast economic sanctions – in a situation when only small tatters of the guerilla remained. Eritrea was incorporated by Ethiopia in 1962 in contravention of UNSC Resolution 390 (1950) prescribing considerable autonomy in a federation. After thirty years of bitter war, the Eritrean People’s Liberation Front agreed to assist the Tigray People´s Liberation Front to take over power in Addis Ababa and TPFL agreed to recognise Eritrean independence after doing so; both promises were honoured, and were not contested in the later war between them. Czechoslovakia’s dissolution was uncontested: when Slovakia tried to back up political demands with threats of secession, the Czech government was apparently happy to get rid of Slovakia. The declarations of Estonia, Latvia and Lithuania were contested for about a year, until Russia declared itself independent, thus de facto terminating the USSR. The conclusion would remain the same: a contested unilateral declaration of independence is very unlikely to succeed, unless heavily supported by external intervention.



Let me now rephrase the second question to asking what was the likelihood for war in 1990. This asks for a postdictive prognosis based on all the factors that are known to be correlated with outbreak of domestic war, making secession one (but ominous) factor among others. The analysis will be made in three steps: 1) What is generally known from quantitative research on causes of wars? 2) What can be added to this by adding aspects of the regional context? 3) What can be further added to this by taking into account specific features of FY?

The first question has an immediate complication: several studies have replicated – and none contradicted – the early finding that international wars and domestic wars seem to be different species. A heavy argument for this is that indicators of external conflict and internal conflict have close to zero correlations with each other (Tanter 1966; Finsterbusch 1974)), at least until we introduce third variables (and even then there is no clear pattern – Wilkenfeld 1973). Now one of the hotly debated issues concerning the wars in FY, with strong legal, moral and emotional overtones, is whether they were civil wars all the time, civil war changing to international war by recognitions (that abandoned the traditional criteria for recognitions), international wars from the first day of proclaimed independence, or some other combination? This discussion (where I would not know how to prove whatever stand I might take) can be avoided however: we are considering the prognosis in 1990, which calls for the correlates of domestic wars. To be on the safe side, I shall first also look at correlates of wars in general or international wars.

Since the pioneers (Richardson 1960; Wright 1942; Sorokin 1937), scores of quantitative studies have to relate how frequently a state gets into war to a great number of characteristics of states. Sometimes this was done by “trawling”, running a great number of variables against each other and seeing what correlations turned up. More sophisticated studies used “casting”, testing causal models by looking whether the specific correlations they predicted were in fact there. Results are almost entirely negative. Correlations of single variables with war are almost all so close to zero as to be statistically non-significant. Where they are at least statistically significant, the correlation is almost always so weak that it only accounts for a few per cent of the total variation in occurrence of war, making the variable quite weak as a predictor of war. Among the very few that are stronger than that, even fewer have gained more solidity by being successfully replicated. The variables that satisfy most or all of these desiderata and that therefore permit at least a weak prognosis of war are essentially the following:

A1)On average, great powers go much more to international war than other states (Wright 1942, Wallensteen 1973);

A2) States with many international boundaries are on average more engaged in war than others; if the correlation reveals some causality, it is from boundaries to wars rather than the other way around (Richardson 1960; Weede 1973)

A3) The balance of available evidence speaks for the thesis that states that are “overarmed” in the sense of having higher military preparedness (manpower, expenditures) than is the average for their size go more to war than others – but the relationship is complex and there is sometimes a chicken-and-egg problem of what causes what (Wiberg 1990).

Some studies found additional correlates (Rummel 1979), but nothing to match these three: correlations were weak and/or a third variable was necessary to find them and/or the study was not replicated. Let me just give a couple of examples: 1) A state that is in rank disequilibrium (ranking higher on military and economic strength than on diplomatic recognition) at one time point participates slightly more than average in war 10-15 years later (Wallace 1975); 2) If a democratic state has (some kinds of) external conflict, it gets more (of some kinds of) internal conflict than average afterwards, whereas if an authoritarian state has (some other kinds of) internal conflict, then it gets more than average of (some kinds of) external conflict afterwards. (Wilkenfeld 1973).



So the total picture of causes of belligerence or peacefulness of individual states exhibits pretty little of clear structure (Vasquez 2000). This may be due to wars being so specific and individual that few valid generalisations are possible – and practically all generalisations that can be found in literature are invalid, having no observed significant correlations to get support from. It may be due to looking at the wrong variables, but this is unlikely: very many variables, including all the traditional ideas, were tried in systematic data analysis. And it may be that we have looked at the wrong level, since it takes two to tango. Let us therefore look at how characteristics of pairs of states are related to their getting into war with each other. This was done in several studies, but the picture they give is also relatively hazy. The essential results mainly – but not entirely – mirror those from the single state level (Wiberg 2000, 2002):

B1) An average pair of great powers used to be much more likely to have a war than an average pair of one great power and one minor state – whose likelihood for war in turn is much higher than for an average pair of two small states. In the post-1945 period, however, one part of the picture changes entirely: an average pair of great powers has no war at all, though there is no consensus on why not. In addition, great powers are the only ones to fight others than neighbours; a few exceptions, such as Vietnam and Czechoslovakia, were due to a great power dragging some satellites along (Wallensteen 1973).

B2) In general, much trade between two states is associated with less than average war between them, but here is a chicken-and-egg problem (Barbieri 2002). There is one important proviso, however: that the trade is relatively symmetric.If it is strongly asymmetric, i.e. a great power takes a large share of the trade of a minor state, then the probability of war is clearly higher than if is symmetric; and when a great power fights a small power, the latter is very likely to lie in the former’s zone of economic influence and furthermore to be an economic satellite of the bigger (Wallensteen 1973).

B3) If two “overarmed” states (in the above sense) have a militarised dispute, then it is twice as likely to escalate into war as a pair of one “overarmed” and one “underarmed” – whose risk is twice as high as that of two “underarmed” states (Singer 1981). For two “overarmed” states the risk seems to particularly high if they have just had an arms race (Wallace 1979, Smith 1980) but the magnitude of the risk is disputed (Wiberg 1990).

B4) Democracy is particularly interesting. The standard finding from several studies of single nations is that it has no effect: democracies are neither more peaceful, nor more warlike, than other states.Yet when we look at pairs of states, we find a quite different picture: several studies concur that two stable democracies do not get into war with each other (with at most very few and quite marginal exceptions), the so-called double democracy hypothesis (Gleditsch & Hegre 1997).We have already seen that this cannot be because democracies per se are more peaceful – they are not. Two types of explanations have been empirically shown to contribute. One is internal: there is nothing to gain – and something to lose – politically in a democracy by threatening another democracy with war, whereas fighting a war with a non-democracy may increase political support. The other is that democracies are to an especially high degree tied to each other in common organizations with common norms and values. Both seem to have some explanatory value.



So much for international wars. Since we are asking for a prognosis of internal war in FY in 1990, these findings are not of much relevance, unless we make the fiction of seeing its republics as independent states even before any declarations of independence. Once their independence is a clear fact a bit later, some of the findings above may of course be used to gauge the likelihood of war among them; but this must be left for another analysis.

Some reservations must be added. First, we must always be cautious when drawing conclusions from mere correlations to causal relations. Second, there are various categorizations of wars, usually with “internal” or “domestic” or “civil” as one of the categories, but it is often admitted that there is no clear and sharp line between them and international wars: several cases can be classified as both at the same time, or form a special, “mixed” category. Third, “internal conflict” is not a homogeneous category: it is sometimes subdivided into two or even three types, where war between organized forces typifies one, a coup d´etat of the second and riots the third. Yet, collected statistics on wars tends to simply use the total or annual number of casualties as the criterion for inclusion. Fourth, the bulk of quantitative research on correlates of wars used to be on international wars – which for decades have become fewer and fewer in relative terms, nowadays accounting for less than one tenth of all wars. (Gantzel 1997, Eriksson 2003). Finally, given the non-correlation between external and internal war, we should not try to locate causes of the latter by using the correlates of the former. New studies are needed.

An increasing number of major studies on domestic wars has indeed been carried out during the last two decades, and the major results are found below together with assessments as to how they may contribute to the (postdictive) prognosis for FY in 1990. Several of them have to do with economy.

1a) The poorer a state is, the higher is the risk for civil war, with the exception that the risk decreases slightly again for extremely poor states (FitzGerald 2000).

1b) This relationship is strong.

1c) The position of FY used to be in the low risk area, but the economic crisis in the 1970s and 1980s moved it in the direction of the high risk zone.

2a) Long and deep depressions tend to lead to political radicalism of one or more kinds; exactly what kinds depends on the specific circumstances.

2b) The relation is indirect, since radicalism may or may not lead to war.

2c) FY had an extreme position in terms of both the length and depth of the crisis. In terms of GDP per capita (admittedly a rather weak indicator of how people actually live) the average FY worker lost about half his real income during the 1980s and came back to the 1960 level (Schierup 1990).

3a) Recent studies at the World Bank and elsewhere indicate that the higher the proportion of primary goods in the export of a state, the higher risk for civil war (Collier & Hoefler 1998.)

3b) Whereas the finding is relatively strong and replicated, its interpretation (originally in terms of “greed” and “grievance”) is still under dispute and causal relations uncertain (Ross 2004).

3c) FY was not in the high risk zone, but was moving towards it. Different parts of FY had quite different values, with Slovenia at one extreme and Kosovo at the other.

4a) There is some evidence that great regional differences in wealth are associated with a higher risk for domestic war.

4b) The evidence is not systematically quantitative however, so the relative weight in the prognosis may be low.

4c) FY was at the extreme end in Europe in this respect: already the differences between Slovenia and Serbia or Croatia were as big as the greatest differences in Sweden.. Differences clearly grew between the extremes (Slovenia and Kosovo), but the picture is otherwise disputed in this respect. In addition, FY belonged to the few “sandwich” cases (like Spain or USSR) where some of the poorest parts would secede on account (among other things) of being poor and some of the richest on account (among other things) of being rich. The situation was exacerbated by demands from the International Monetary Fund that the development funds going from the richer to the poorer parts of FY be terminated.

5a) The risk of revolution seems to get particularly high if a period of rapid improvement, creating rising expectations, is followed by a movement in the opposite direction. (Davies 1962).

5b) The evidence is anecdotal rather than systematic and has been disputed, so the weight must be rather low.

5c) Where to locate FY depends on what time periods we look at. With some choices, it lies in the extreme high risk group. It may be argued, however, that the highest risk existed several years before 1990 and that people had by then gotten used to worsening standards of living and developed micro level ways of trying to cope when those at the macro level failed (Schierup 1990).

6a) It a common belief among “federalist” thinkers on the EU that economic integration promotes and even eventually necessitates political integration.

6b) This is not solidly empirically documented. Furthermore, the relation is at least indirect, since lower political integration does not necessarily mean war.

6c) When economic decentralization started in the 1960s, FY went in the opposite direction for a long time: a decreasing proportion of trade went between the republics and an increasing part within them and with foreign centres of economic power (especially in northern Italy and southern Germany). Some countermovement (by necessity rather than preference) seems to have accompanied the deepening crisis in the 1980s however.

There is still a heated debate, both generally and concerning FY in particular, on the extent to which civil wars are related to ethno-national conflicts, the extreme positions being “really not at all” and “first and foremost”, respectively. Given the open or hidden moral and political agendas in this debate, there is no likelihood that it will lead to any consensus in a near future. In any case, some observed demographic regularities are of interest:

7a) In general, the risk of civil war seems to be slightly higher in states that are ethnically heterogeneous, even though there is no consensus on this. In particular, however, the difference in risk is high in former Central/Eastern Europe and the former Soviet Union (Wiberg 1996): the smaller the biggest group in a state, the higher the empirically observed risk for dissolution, civil war, de facto partition or combinations of this.

7b) The relationship is fairly weak in the general case. It is stronger in Europe, where most of the lowest third (on size of biggest ethno-national group) had one or more of the consequences just mentioned, whereas none among the highest third even came close to that.

7c) FY was extreme in Europe on the “biggest group” indicator, with 38 per cent Serbs – and Bosnia-Herzegovina was number 3 with 42 per cent Moslems).

8a) If an ethnic minority group is strongly concentrated to one geographical area and constitutes a considerable majority there, the risk of an outbreak of violence is reduced.

8b) There is clear empirical support for the proposition. (Melander 1999:95f)

8c) For FY, this depends much on what minorities we look at – there is a vast difference between the close correlation between Slovenians and Slovenia at one extreme and the leopard pattern in Bosnia and Herzegovina at the other. Hence, the applicability of this proposition is a complex matter; at best, it can tell us that the risk for extensive war was lowest in Slovenia and highest in Bosnia and Herzegovina.

A final group of factors may be referred to as political, if we need a common label for them:

9a) The younger a state (since independence), the higher risk of domestic war in it (Hegre et al. 2001)

9b) The relationship is moderately strong.

9c) Even if we put the birth date of FY at 1945 (cf. above), it was in the middle risk zone rather than the high risk zone. All the new states into which it might be dissolved, however, would be in the highest risk zone.

10a) Previous domestic war increases the risk for a new war; and the more recent it was, the higher the risk. (Hegre et al . 2001)

10b) The relationship is moderately strong and its character disputed (Walter 2004).

10c) For FY we get two opposite tendencies to balance. On the one hand, FY was at the extreme end in Europe concerning the length and lethality of previous civil war (within the context of the international World War II), with Spain in the 1930s coming closest; this speaks for high risk. On the other hand, in 1990, these wars were more than 45 years ago, which speaks for a more moderate risk. In any case, a possible violent dissolution would mean high risk for domestic war in all the successor states, or at least those where other conditions made it possible (which would really only exclude Slovenia).

11a) War has a complex relationship with democracy and democratization. Stable democracies run the least domestic violence and stable autocracies a slightly higher risk, whereas states located in between these poles run considerably higher risks. It has furthermore been established that these higher risk depend both on position -being in between- and on movement, whether in the direction towards or away from stable democracy (Hegre et al. 2001).

11b) The risk at the middle of a democracy scale is about one and a half times as high as at either of the opposite ends.

11c) FY – as well as its constituent republics and autonomous provinces – was certainly located somewhere in between the opposite ends and therefore ran higher risks than average. In addition, it was at that time clearly in movement, adding even more to the risks. It would take a closer investigation to establish whether it (and, later, its successors) were on their way “uphill” (where more democracy would add to the risk) or “downhill” (where further democratization would reduce the risk of war).

The provisional summary of the prognosis of FY in the light of generalizing quantitative studies of possible predictors and the values of FY on those predictors must be as follows. Several indicators contributed to a negative prognosis: some of them by being relatively strong predictors, some of them by FY having extreme values on the predictor variable. At the same time, there was hardly any indicator contributing in the opposite direction: at best, they would imply that the risk of FY was no higher than average. The prognosis was therefore bad by international comparison and probably the worst in all of Europe. Yet this should be read with precision. It certainly does not say that war was inescapable. In fact, it does not even say that war was more likely than peace – it would take a lot more of model analysis and statistical work to figure that out.



Let me now move to Step 2, looking at how postdictive (and, for that matter, present) prognoses can be based on regional characteristics of South Eastern Europe, including its historical legacy. Both terms require clarification. Exactly how the region is defined and what states to count into it depends on the time period we study as well as on the particular aspects we are interested in.

If we look at the geopolitical aspect, the region has often been ascribed high strategic significance, e.g. by the Commander in Chief Nikola Ljubicic (1977: 249): “Territory of the SFRY is of exceptional strategic significance not only as a Balkan but also as a Mediterranean and Central European area”. Yet this has been varying with time, as have the reasons. Invaders from Asia to Europe always had to pass here, those in the opposite direction often did. With the gradual dissolution of the Ottoman empire, the strategic significance of the Balkans increased. European spheres of interest collided here in “the soft underbelly of Europe” and involved local actors in every major war. The Cold War formula 2+2+2 indicates that Turkey is also definitely a part of the region from this point of view, and sometimes Turkey is seen as a buffer (or intersection) between the European and the Middle East security complex (Buzan 1991: 210). During the last Cold War Peak in the early 1980s President Reagan had issued his National Security Directive on general destabilization of communist regimes in Europe – and FY was not excepted. Yet 1990 was when the Cold War was in its last moments and there are disagreements on how this affected the region: losing significance because of that or keeping it for new reasons?

In economic terms, the region is differently defined. For centuries it had satellite relations to higher developed parts of Western Europe, whether Italian city states , single West European states, the USSR and other parts of COMECON, or, finally, the EU, in relation to which it is weaker than ever before. This was so whether it was politically ruled by the Ottoman or Habsburg empires or consisted of states with at least nominal sovereignty. The main exception is given by much of the Cold War period, when the old relations was broken for a while in the Communist states, but in different ways: dependence on the USSR (Bulgaria, Romania, early Albania), balancing trade partners (Yugoslavia), finding them elsewhere (middle Albania) or attempting autarchy (later Albania). Yet in most of the region, shifting trade patterns and dependence on IMF re-created satellite relations to the West even before the Cold war was over.

Another aspect is that of history and culture, the historical legacy. We should be careful with this term for several reasons. First, as determinants of what happens, the perceived history or historical myths are often more important than “objective history”. If the writing of the latter may change (by new discoveries or interpretations), this is much more true of the former, as exemplified by all the “invention of history” in the last couple of decades in FY (and elsewhere) to legitimize post-communist regimes, nationalist movements, secessionism, etc. Second, “determinant” is not the same as, and does not entail “determinism”. The historical legacy is only one formative force among others, and its relative weight relative to them is also variable. So the notion that “since this has always been so, it is bound to remain so” is wrong on several counts. These reservations having been made, however, there are good grounds not to dismiss history. To understand a conflict it is necessary to know a long stretch of its real history and imagined histories.

One historical legacy of the Balkans is that of the Ottoman empire: how it was created (brutal conquest, but also by dividing and ruling), how it operated and how it disintegrated. Its operation was in one way highly centralist, yet Ottoman rather than Turkish: people from all over the area could rise to high positions, once they were (voluntarily or forcefully) integrated by conversion to Islam and sometimes even if not: the Phanariote Greeks as administrators and priests, local vassal princes (Serbian, Romanian, etc.) who remained Christian, the monotheistic religious institutions enjoying considerable sectoral autonomy in the millet system. In all these cases, however, strict loyalty to the Ottoman empire was required. This was often interpreted as treason by the population – as for that matter was conversion to Islam, which can be clearly seen in the epic Gorski Vijenac by Petar Petrovic Njegos (1948). By the millet system, religious leaders would exert a political influence (“ethnarchos”) far beyond the purely spiritual. Originally, there were only three millet (Christian, Judaic, Zoroastrian), which had as little national characteristics as the Umma of Islam. Yet during the 19th century they proliferated to seventeen, now coming closer to defining nations, at the same time as early Turkish nationalism increasingly undermined the Ottoman character of the empire. The proliferation of millet also made it easier for European great powers to divide and rule by cultivating allies within the Ottoman empire, which contributed to its dissolution and the attempted slicing up of Turkey in the wake of WWI until this was blocked by Kemal Atatürk. (Jung 2001). Among the legacies from this process there is a tendency to define Us/Them distinctions in religious terms, with the ominous implications this has for an area that combines the fault line between Christianity and Islam with that between Catholic and Orthodox Christianity.

This Ottoman legacy in almost all of the region (Slovenia and – largely – Croatia excepted) reinforced an even older Orthodox legacy: the division of Orthodox Christianity in autocephalic churches, which eventually became closely related to the definition of nation – and always were to secular political rule in some way, ranging between caesareo-papism as one extreme and abject subordination to Communist regimes as the other)

Another legacy is a low tolerance for minorities not belonging to the titular nations, whether defined in religious or linguistic terms or both. We see manifestations of this everywhere in recent history: Kurds, were until recently defined as “mountain Turks” in Turkey and their (very different) language forbidden; Macedonians and Albanians in Greece are heavily hellenized linguistically and Turks referred to as Moslem Greeks; Bulgaria tried to bulgarianize the Turks there and Romania to romanianize the Hungarians. What happed in FY after the Cold War (but also long before it) is part of a wider pattern. The term “ethnic cleansing” was invented by the Serb Cubrilovi_ in the 1930s and used by the Croat Ustasha during WWII, in both cases reflecting older ideologists among their own peoples and older processes in the area, as testified by the Carnegie Commission on the wars in 1912-13 (Carnegie 1993). This means that the struggle between civic and ethnic definitions of nation is far from over and can take very violent forms. In fact, it is not over in Western Europe either. The notions of “patrie” and “nation” from French Enlightenment long made France look like a paragon of a civic definition, yet the Dreyfus process came as a nasty chock to French (and other) Jews, and Le Pen is now rebelling against it. Post-war Germany looked like the final triumph of civic over ethnic, but question marks are defined by the ease with which citizenship is acquired by people who were never in Germany before and often do not even speak German (from the Volga region) with the difficulties for those who are born in Germany and speak perfect German (Turks and others). “British” has been the civic term for several generations, but had its limits, as demonstrated in Northern Ireland, Scotland and even Wales. Belgium is in a process of breaking up along ethnic lines and the same may be true for Spain.

It is a strength for a country and a blessing for its population if a civic definition is generally accepted or at least (as, e.g., in Switzerland and Finland) takes clear priority before the ethnic. Yet getting there is not easy, ignoring strong ethnic definitions just because of disapproving of them is dangerous, and trying to impose them by political fiat may be suicidal, since precisely this may be seen by the minorities as an attempt by the titular nation to annihilate them, no matter how much constitutions and other documents assert the opposite – they just do not become credible.

When Benjamin Disraeli was referring to “ancient ethnic hatreds” in 1876 to counter demands for British intervention against Turkey or John Major was echoing him in 1993 for similar reasons (Malcolm 1994:xx), there is considerable evidence can be adduced against them. Yet, if we amend “hatreds” to “fears”, we may come closer to widespread Balkan realities. One important point is that such fears (in collective memory, etc.) may have fairly similar behavioural manifestations to those of hatred, once they are provoked; and another point that fears are more easily provoked by the (in his own eyes innocent) behaviour of Alter than if no such legacy exists.

There is another widespread legacy in the region (and a wider one): the demise of Communist regimes (which was in full swing in 1990) and thereby also of Marxist-Leninist legitimizations of the idea of the state, leading to the search for others. As several authors have pointed out, nationalist ideologies were often strong competitors to liberal ones (sometimes even merging with them). Yet here too we find several different cases: Where the state itself was etnically homogeneous and had an old and strong state tradition, this carried no risk for civil wars, at worst for irredentism (which, if too loudly manifested, would also mean losing the chance to join EU one day). More heterogeneous states with weak state traditions constitute the opposite case, with much higher risks. In this respect, FY had bad odds: highly heterogeneous, with short state traditions that had already collapsed once, when Hitler and Mussolini found loyal collaborators in some ethnic groups – and having inside it several groups with state traditions of their own

To put it cautiously, the bad prognosis derived from the first step of the analysis is not counterbalanced by the second step, but rather reinforced by it, even if the qualitative character of the second step makes it even more difficult to put any figures in the prognosis or tell whether a war was more likely than peace. In relative terms, however, the second step reinforces the first provisional conclusion that FY had the worst prognosis in Europe. Let us now see to what extent this may have to be revised when the focus gets even narrower in Step 3 towards a postdictive prognosis.



This step consists in looking at particular features of FY to add to the prognosis based on the two first steps, which tried to answer the question: “What if FY had been an average state and in addition an average Balkan state?”

1. One of the particularities can be summed up from the data presented above: FY had extreme or very high values on several of the variables that are statistically associated with domestic war.

2. Reagan´s general destabilization policy against Communist countries made no exception for FY. In addition, destabilization was also attempted by the West German Bundesnachrichtendienst in collaboration with Croatian exile organizations (Schmidt-Eenboom 1995).

3. Whereas many European states had occasional political disagreements on their constitutions, FY appears unique in this respect with its long series of them. It started already in the planning of its creation (Banac), where Serbs wanted a state of French type and Croats one in Swiss direction. There were repeated constitutional crises in First Yugoslavia and a series of constitutional changes in Second Yugoslavia. And the issues that formally defined the conflict objects at the very end were essentially constitutional: Serbia’s unilateral derogation of the autonomy of Kosovo, the demands of Slovenia and Croatia for de facto and later de jure independence, etc. There remained less and less of central government, the efficiency of what remained was heavily reduced by the need for consensus in important decisions, and in addition it rapidly lost what legitimacy it still had, in particular after the elections in all the republics in 1990.

4. Even after the end of the Cold War, what me may call the “Cold War Filter” for conflicts remained in place in the West, and in particular in USA. This filter can be described as having three axioms:

A. A conflict can have no more than two parties (“becomes too difficult for the readers/viewers”).

B. The parties must be states or something state-like that can be personified by leaders.

C. There must be one Bad Guy – and by virtue of that, the other one is a Good Guy.

Whereas some search may identify a few conflicts that are not too badly represented even after passing through this filter, this is most decidedly not true for FY. First of all, the complex conflict pattern consisted of sub-conflicts – and these in turn usually had three or more parties: Serb/Croat/Slovene in the north, Serb/Croat/Moslem in the center, Serb/Albanian/Macedonian in the south. The important problem with axiom A for FY is that multi-party conflicts have a strategic logic that differs entirely from that of two-party conflicts, with shifting coalitions as a frequent pattern. Shifting coalitions was a pattern long before conflicts got militarized (Ramet 1992). The main problem with axiom B is that it focuses on (stereotyped presentations of) the personalities etc. of single leaders to the detriment of understanding the more fundamental conflict dynamics, including the issue of to what extent the leaders were actually driving or largely drifting along with these dynamics. And the problem with axiom C, if I permit myself to make value judgments, may not so much be the appointment of the Bad Guy as the amazing procession of Good Guys that were imagined by this filter by the sole virtue of somebody having to fill that role. Whether or not Western politicians were sufficiently uninformed to believe in the mass media versions, by the logic of dominant discourses they had to speak and act as if they believed in them.

5. FY was not the only ethnically heterogeneous state in the region, but it was by far the most heterogeneous one by any criteria. This in itself contributes to a bad prognosis on statistical grounds, as shown above. In addition it is a matter of the demographic distribution of groups. The “leopard patterns” of ethno-national groups in Bosnia-Herzegovina and some other areas has been shown to be associated with higher risks of escalation to violence than when each group is fairly concentrated to one area where it is the big majority (Melander 1999:95f)

6. A final and fatal particularity was that for various reasons the conflicts in FY attracted a high degree of great power interest. In 1990, there were many things going on at the same time while the Cold War was being written off. EU was in the process of adapting to the new position of the united Germany at the same time as mass media pressure and an efficient Croatian propaganda machine made the German government drift helplessly drifting into stands on the FY conflicts that made these the first demonstration of the new German power position, but were deeply controversial in many other states in EU (and outside it). This made FY the arena of an internal power struggle in the EU in a critical period. Even if Germany eventually largely won by bribing the others in different ways, the victory meant making demands that were unlikely to be satisfied without a war – which Germany for historical reasons could not fight and nobody else was willing to fight for it.

At the same time the end of the Cold War also created great uncertainty concerning transatlantic relations: would the USA be in a stronger position by its claim to have “won” the Cold War – or in a weaker position by virtue of the weight of military power relative to economic power going down, when the former was in far less demand? The Soviet Union was in its death throes, the first declarations of independence already proclaimed. It -and later Russia- was in great confusion concerning its future doctrine and for a while believed that close cooperation with the West would bring desired rewards. So the FY also became an arena of transatlantic contradictions, where the eventual US victory meant (concerning Bosnia-Herzegovina) demands which were unlikely to be satisfied without a war – which UNPROFOR refused (and did not have the resources) to fight, USA would not put land troops into and no alliance partner was willing to fight in their absence.

Because of all this, the actors in FY were bent to believe that they had some bargaining cards in terms of potential external support, but they were no better than others in guessing who would intervene when and how in favour of whom. Wishful thinking actually made their guesses worse, in the direction of greatly overestimating the values of these cards; and this in turn made them more likely to escalate their demands, less likely to be able to find necessary compromises and more likely to get into a war by accident or even by intention.



Adding Yugoslav particularities did nothing to dispel the somber (postdictive) prognosis based on the other sets of premises, but, once more, rather the opposite. It is fair to say that FY had a far worse prognosis than any European country at that time. Let me return once more to what this means and does not mean. It does not mean that war was inescapable, only that the risk was high and continuously increased due to internal dynamics and international postures. There has been much “iffy” history written, claiming that the war could have been avoided, if only… And there is hardly consensus on a single “if only” clause, some of which are proposed in other chapters of this book. When I do not enter this debate here, it is for lack of space – it would become a chapter in itself. It is a very important debate, we need more of it and scholars from FY may in many respects have more valuable contributions to make than outsiders.

What the conclusion does mean, however, is that any attempt to find THE cause of the wars is likely to be futile. The situation was over-determined, with the implication that eliminating any single causal factor would reduce very little of the total risk of war. Personally, I am not yet convinced by any of the “if only” suggestions, either because they seem insufficiently argued or because the proposal itself calls for new “if only” clauses. A convincing “if only” construction is likely to have to combine several things at various levels at the same time.



This article has treated the likelihood for an outbreak of war. This is far from the whole story. To analyse the continuation of war is a different thing. In the first book in English on the Yugoslav wars , one of the authors (Vrcan 1993) quotes the Lebanese sociologist Ahmed Beidoun looking back at the experiences of his country and stating that whereas in the beginning the war there was about something, it eventually became war for war´s sake, since so many actors had gotten something to gain (economically, politically, etc.) from the war continuing. Several scholars have followed this line of analysis – which would also call for a chapter of its own. Let me limit myself to the remark that in order to end a war, it may not be enough to remove its original causes when new causes have been added during its course.



Barbieri, Catherine 2002. The Liberal Illusion: Does Trade Promote Peace? Ann Arbor: University of Michigan Press.

Buzan, Barry 1991. People, States and Fear. Second edition. London: Harvester Wheatsheaf.

Carnegie Endowment 1993. The Other Balkan Wars. A 1913 Carnegie Endowment Inquiry in Retrospect with a New Introduction and Reflections on the Present Conflict by George F. Kennan. Washington: Carnegie Endowment for International Peace.

Collier, Paul & Anke Hoefler 1998. “On Economic Causes of Civil Wars”, Oxford Economic papers, vol.50, no. 4.

Davies, James C. 1962. “Toward a Theory of Revolution”, American Sociological Review, vol. 27, no.1.

Eriksson, Mikael; Margareta Sollenberg & Peter Wallensteen 2003. “Patterns of Major Armed Conflicts, 1990-2002”, SIPRI Yearbook 2003. Oxford: Oxford University Press.

Finsterbusch, Kurt 1974. „Theories of Domestic and International Conflict and Their Relationship”, in Joseph D. Ben-Dak, ed., The Future of Collective Violence. Lund:Studentlitteratur.

FitzGerald, Valpy 2000. “Global Linkages, Vulnerable Economies and the Outbreak of War”, in E.W. Nafziger & Raimo Väyrynen, eds., War and Destitution: The Prevention of Humanitarian Emergencies. London: Macmillan.

Galtung, Johan 1966. Peace by Peaceful Means. Peace and Conflict, Development and Civilization. London: Sage

Gantzel, Klaus J. 1997. “War in the Post-World War II Period: Some Empirical Trends and a Theoretical Approach”, in David Turon, ed., War and Ethnicity. Global Connections and Local Violence. San Marino: University of Rochester Press.

Gleditsch, Nils Petter & Håvard Hegre 1997. „Peace and Democracy: Three Levels of Analysis”, Journal of Conflict Resolution, vol. 41, no.2.

Hegre, Håvard; Tanja Ellingsen, Scott Gates & Nils Petter Gleditsch 2001. ” Towards a democratic civil peace?” American Political Science Review, vol. 95, no.1.

Jung, Dietrich with Wolfango Piccoli 2001. Turkey at the Crossroads. Ottoman Legacies in a Greater Middle East. London: Zed.

Malcolm, Noel 1994. Bosnia. A Short History. London: Macmillan.

Melander, Erik 1999. Anarchy Within. The Security Dilemma Between Ethnic Groups in Emerging Anarchy. Uppsala: Department of Peace Research.

Ljubicic, Nikola 1977. Total National Defence – Strategy of Peace. Belgrade: Socialist Thought and Practice.

Njegos, Petar Petrovic 1948.Gorski Vijenac. Beograd: Jugoslovenska Knjiga (first ed. 1847).

Ramet, Sabrina P. 1992. Nationalism and Federalism in Yugoslavia, 1962-1991. Bloomington: Indiana University Press.

Richardson, Lewis F. 1960. Statistics of Deadly Quarrels. Chicago: Quadrangle Books & Pittsburgh: The Boxwood Press.

Ross, Michael L. 2004. “What Do We Know About Natural Resources and Civil War?”, Journal of Peace Research, vol. 41, no. 3.

Rummel, Rudolph J. 1979. Understanding Conflict and War, vol. 4: Ear, Power, Peace. Beverly Hills, CA: Sage.

Schierup, Carl-Ulrik 1990. Migration, Socialism and the Inernational Division of Labour. Aldershot: Avebury.

Schmidt-Eenboom, Erich 1995. Der Schattenkrieger. Klaus Kinkel und der BND. Düsseldorf: ECON.

Singer, J. David 1981. “Accounting for International War. The State of the Discipline”, Journal of Peace Research, vol. 18, no. 1.

Smith, Theresa Clair 1980. “Arms Race Instability and War”, Journal of Conflict Resolution, vol. 24, no. 2.

Sorokin, Pitirim A. 1937. Social and Cultural Dynamics, vol. III. New York: American Book Co.

Tanter, Raymond 1966. “Dimension of Conflict Behavior Between and Within Nations”, Journal of Conflict Resolution, no. 1.

Vasquez, John A., ed. 2000. What Do We Know About War? Lanham,MD: Rowman & Littlefield.

Vrcan, Srdjan 1993. “A European Lebanon in Making or a Replica of Pakistan/India?”, in Tonci Kuzmanic & Arno Truger, eds., Yugoslavia War. Schlaining: Austrian Study Centre for Peace and Conflict Resolution & Ljubljana: Peace Institute.

Wallace, Michael D. 1975. “Power, Status and International War”, in Manus Midlarsky, ed., On War. New York: Free Press.

Wallace, Michael D. 1979. “Arms Races and Escalation”, Journal of Conflict Resolution, vol. 23, no. 1.

Wallensteen, Peter 1973. Structure and War. International Relations 1920-68. Stockholm: Raben & Sjögren.

Walter, Barbara F. 2004. “Does Conflict Beget Conflict? Explaining Recurring Civil War, Journal of Peace Research, vol. 41, no.3.

Weede, Erich 1973. “Nation-Environment Relations as Determinants of Hostilities among Nations”, Peace Science Society (International) Papers, vol. 20.

Wiberg, Håkan 1983. “Self determination as an international issue”, pp.43-65 in Ioan M. Lewis, ed., Nationalism and Self Determination in the Horn of Africa. London: Ithaca Press.

Wiberg, Håkan 1990. “Arms Races – Why Worry”, in Nils Peter Gleditsch & Olav Njølstad, eds., ARMS RACES – Technological and Political Dynamics. London: Sage.

Wiberg, Håkan 1996. “Identity, ethnicity, conflict”, in Simon Bekker & David Carlton, eds. Racism, Xenophobia and Ethnic Conflict. Durban: Indicator Press.

Wiberg, Håkan 2000. “Patterns of War and Peace: How Much Can We Trust Trends?”, in Young Seek Choue, ed., Will World Peace Be Achievable in the 21st Century?. Seoul: Institute of International Peace Studies, Kyung Hee University.

Wiberg, Håkan 2003. “Peace, Security and South Eastern Europe”, in Anton Grizold, ed., Security and Cooperation in Southeastern Europe.Ljubljana: Faculty of Social Sciences.

Wilkenfeld, Jonathan 1973. “Domestic and Foreign Conflict” in Jonathan Wilkenfeld, ed., Conflict Behavior & Linkage Politics. New York: McKay.

Wright, Quincy 1942. A Study of War. Chicago: Chicago University Press.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: