Economy, Politics or other factors? An exploration of why riots occurred in Cyprus in 1931 using Probit estimation
By Alexander Apostolides, European University Cyprus ()[1]
The crown colony of Cyprus was considered so peaceful by the British that it only had 121 soldiers stationed on an island the size of Lebanon: thus the severe riots that occurred in late 1931 came as a complete surprise. Since then the causes of the riots have been disputed: the disgraced governor blamed the economic conditions at the time for being the catalyst for the explosion of violence, while most Greek-Cypriot historians have argued that the riots were purely the pent up expression of the desire for union with Greece. It is clear that the initial spark was political, yet the prolonged riots could have taken place for many reasons, including the economic effects of the great depression and local grievances.
There have been several attempts to explain why violence occurs including BesleyPersson (2008), but the novelty of this approach lies in the use of microlevel data of a single riot. The paper uses a new dataset that combines geographic information systems (GIS) mapping with a novel dataset for the conditions of Cyprus in 1931, built up through primary and secondary sources to map all the incidences of disturbances during the rioting. Our estimates suggest more needs to be done to be table to clearly ascertain why, riots took place, as the key variables are still not collected in a village level.
Despite that there are still worthwhile conclusions to be made: cities are areas where riots begin, and the distance from the capital where the first riot occurred seemed to play a factor in how the riots spread. Areas with a high proportion of Turkish Cypriots in the population are not so susceptible to riots, suggesting that political factors were very important. Despite that the economic situation as proxied by the type of agricultural production play a factor too. Areas with high Turkish Cypriot consternation clearly suggest that there is an economic undertone to the riots, however we suggest that the negative relationship of Turkish Cypriots population to riot occurrences suggests that nationalism was also an issue. Finally the estimated results are checked for robustness to provide some insights on how the riots should be seen in light of this new research.
Introduction: What causes riots and peasant revolts?
What causes riots? This is a question that has been looked by a wide range of social scientists and economists. Preconditions to revolutions and riots have been proposed as far back as Karl Marx, who considered that history is defined by a series of inevitable revolutions once the rapid change of a structure of society outpaced the existing governmental structure. A historical approach was suggested by Brinton (1938) who suggestedthat there are preconditions that are common throughout most revolutionary incidents, includingdiscontent from all classes of a society, the lack of government support by a sufficiently large group of the population and a state that is close to bankruptcy. The determinants of riots and revolutions have come back in the public interest through the recent Arab spring, with commentators trying to pinpoint commonalities between the causes of the riots in different countries (Rodenbeck, 2013).
Theoretical models on the incidence of violent revolt or civil war have been developed in economics (Grossman1991). The link between an economic crisis and subsequent insurrections has been present in the work of social scientists and historians, and has prominent place in the attempts by economists to explain conflicts (Miguel, Satyanath, Sergenti2004), including economic incentives but also social factors. Yet there are issues in how can one use empirics to test the validity of the theory, when conflict and political violence are eminently exceptional events. As revolutions and riots are rare events, they are difficulties in empirically testing theories that model riots and disturbances.
Quantitative research has tested several theories of political violence by combining violent events across time and/or space, often spanning different continents and several centuries. Yet this can be less than satisfactory if we accept “that attempts at revolution everywhere should not be mere imitations of previously successful upheavals but should, instead, betailored to fit specific circumstances” (Blackey 1974; 191). Blackey’ssuggestion presents a challenge to empirically testing causes of riots, civil wars and revolutions. If all revolutions have specific circumstances, then we might have serious statistical problems in testing economic theories on riots and disturbances: Some explanatory variables might important for some riots but less important to others, causing a variable to be considered as being less important than it really was. It could also be that there are variables that could explain the incidence of a disturbance but they are missing entirely from the databases that are available. There are ways to deal with such issues, but these methods require a substantial data or might involve categorising revolutions. As a result a researcher could be accused of cherry picking data or of expanding the dataset to include events that are very dissimilar in nature.
However a different approach is to build detailed level information for each riot, civil war and revolution. A riot or a revolution is simply an aggregation of many events of disturbances across time: thus information on the series of events can be used to make detailed datasets that might test the predictive power of variables. This approach to riots can combine the empirical and theoretical work of economics with the work of historians who catalogue how riots take place. This approach might give better understanding to the dynamics of riots and revolutions, but it cannot estimate the importance of events that affected the whole population, such as tax increases. Thus a localised approach build up for each revolution and riot can highlight as well as hide causes of disturbances.
Our research tried to build from the evidence of detailed data for one incident of riots: the Cypriot riot of 1931. In our case the research is focused on incidences of public disturbance that occurred in a two month period starting from the 21st of October 1931 in Cyprus.
Cyprus was at that time a crown colony of the British Empire. The interwar period between the First and the Second World War show an unusual large number of disturbances in the British Empire, as the global economic instability of the period was combined with the growing nationalistic movements in the colonies. The integration of the colonies within the global economy prior to the establishment of Empire wide duties meant that colonial subjects invested in producing material used in a globalised scale were gravely affected during this period (Brown 1989). Significant riots and protests took place in India, Burma and British West Africa in the 1930s were issues of the collapse of primary product prices, the increase of taxation and a rural credit crunch led to violent confrontation (Rothermund, 1996)
«Τα Οκτωβριανά»: The October 1931 riots and their significance
Cyprus became a British protectorate of Britain in 1878. Under the Cyprus Convention the island remained nominally under Ottoman suzerainty but all power was transferred to the British High Commissioner (Hill 1952). The change did not lead to a rapid modernisation of the outdated administration system: despite some improvement of the judicial system and the introduction of a limited constitution;the emphasis of the British administration was in maintaining the status quo and not in transforming the Ottoman system (Georghallides, 1979). The subsequent occupation of Egypt removed the strategic purpose from the occupation of Cyprus, leaving “a lingering quality of being bereft, verging on pointlessness, which British rule there could never shake off”(Holland and Markides, 2008: 165). The island was annexed in 1914 with the start of hostilities towards the Ottoman Empire; resolution of its status as a crown colony only took place in 1925.
This inaction was partly caused by the fiscal constraints of the colonial administration. Britain was supposed to provide an annual payment of £92,799 sterlingto the Sultan in lieu of the lost tax revenue from Cyprus. The British administration raised most of the ‘tribute’ from taxation in Cyprus[2]. Yet the amount was in excess of the fiscal surplus of the colony, providing a large disincentive for change: any fiscal surplus was whisked off to the British Treasury as payment for the so called “Tribute”(Georghallides, 1985).After insistent protests the ‘tribute’ was partially reduced by a permanent grant-in-aid of £50,000in 1907; yet £42,799 of the ‘tribute’ remained a burden to government expenditure until 1927, and still a payment of £10,000 for imperial defence was expected from 1928 onwards.
The colony had a limited constitution which was revised but essentially remained unchanged when Cyprus became a crown colony in 1925. The Cypriot constitution, given in 1882, was advanced for the time: yet the lack of substantial modernisation with the declaration of the island as a crown colony made it look outdates when compared to the neighbouring crown colony, Malta. It was based on clear “divide and rule principles”: the colonial government had to appease either the majority Greek-Cypriot or the minority Turkish-Cypriot community in order to pass legislation and fiscal appropriation bills (i.e. the budget). As a result the legislative council heightened communal tensions that were already present on the island. For the Cypriots the constitution was ineffective as all power and influence remained with the colonial government: the legislative assembly could not create or even modify laws, as it could only pass or reject laws proposed by the Colonial executive. The executive could also circumvent the council by arranging for a law to be passed as an order-in-council by the British monarch[3].
This Legislative council constituted twelve elected members: six appointees of the High Commissioner, 3 elected Muslim representatives and 9 non Muslim representatives (based on the proportion of the 1881 census). The Greek and Turkish Cypriots elected exclusively by their own communities. Although the numbers were subsequently increased to twenty four with the declaration of Cyprus as a crown colony, the proportions, and thus the divisive nature of the council, remained unchanged. In the case of a tie, the governor, who was also the president of the legislative council, cast the deciding vote. Thus the government could overrule the wishes of the Greek-Cypriot majority as long as it could ensure that the Turkish-Cypriot minority voted in line with the colonial officials.
The constitution provided the Cypriots with a council that could only be effective if the communities were united in their opposition to the government, and even then they could only force colonial government to listen, but not to comply, with their opinion. Prior to the First World War the colonial government effectively overruled the council, since the elected Turkish-Cypriot councillors consistently voted with the government. This was in part due to the fear of the fate of the community in the event that the Greek-Cypriots achieved Enosis (as union with Greece was known). The government ensured the loyalty of the Turkish-Cypriots councillors through the use of substantial government patronage that was bestowed on them[4] (Holland and Markides, 2008).
The immediate cause of the riots is related to the action of Legislative Council (henceforth LC) members. The Governor of Cyprus of the time, Sir Ronald Storrs, has ensured that Cyprus was not liable for the Cyprus Tribute back in 1926; yet he intentionally failed to inform the LC a provision to the British Treasury agreeing to stop billing the Cyprus government to the tune of £92,799only if the colony reneged the rights to the amounts accumulated up to that time.
By the middle of 1931, as Table 1 indicates, the Cyprus economy was suffering a severe recession. The recession combined the dramatic decline in agricultural and mining products due to the great depression with the worst drought in Cyprus since records began (Apostolides, 2010). The government deficit as a proportion to GDP was a modest but unprecedented -1.4%. Yet as the Cypriot economy was not able to accumulate substantial reserves due to the fiscal drain of the “Tribute” to the British treasury, the small reserve of £70,000 that the colony maintained was wiped out by the start of 1931.
The government attempted to maintain wages at the same level in order to retain the services of British personnel: yet this was against the wishes of both communities in the council, who wished to see more high ranking Cypriots, which would be a natural consequence of a reduction in the wages of British colonial officials in Cyprus. How this affects Cyprus is mentioned in more detail elsewhere (Apostolides, 2012). Suffice to say here that the LC rejected laws to increase taxation as well as the government appropriation bill through an unprecedented and relatively consistent co-operation of a Turkish Cypriot representative and of the Greek Cypriot community, therefore overruling the government.
The fact that the government was not able to pass the government budget or the additional taxation it requested (11 rises in the span of Ronald Storrs’s governorship, lasting under six years) did not lead to the government acquiescing to the demands of the LC; it chose to ignore the LC by passing laws and government budgets through orders-in-council. The resulting snubbing of the LC, combined with the worsening economic situation (see table 2) made the conditions explosive
Table 1A: Cypriot GDP, Population, Per capita GDP, Government Revenue and Expenditure in constant 1938 Cyprus pounds[5]
Year / GDP / Population / Per capita GDP / Government Revenue / Government Expenditure / Budget Deficit / Surplus / Year / GDP / Population / Per capita GDP / Government Revenue / Government Expenditure / Budget Deficit/ Surplus1921 / £3,732,544 / 310,715 / £12.01 / £418,112 / £415,599 / £2,513 / 1930 / 4,911,844 / 344,042 / £14.28 / £677,372 / £747,559 / £-70,187
1922 / £4,001,854 / 314,253 / £12.73 / £471,461 / £393,196 / £78,265 / 1931 / 4,605,234 / 347,959 / £13.23 / £727,766 / £742,164 / £-14,397
1923 / £4,125,691 / 317,830 / £12.98 / £424,504 / £435,959 / £-11,455 / 1932 / 4,213,399 / 353,982 / £11.90 / £761,123 / £748,050 / £13,073
1924 / £4,029,952 / 321,449 / £12.54 / £435,785 / £393,591 / £42,194 / 1933 / 4,232,438 / 360,109 / £11.75 / £748,102 / £698,357 / £49,745
1925 / £4,342,975 / 325,109 / £13.36 / £489,381 / £453,850 / £35,531 / 1934 / 4,755,859 / 366,342 / £12.98 / £776,723 / £722,396 / £54,327
1926 / £4,169,484 / 328,810 / £12.68 / £444,797 / £463,148 / £-18,351 / 1935 / 5,318,422 / 372,683 / £14.27 / £899,557 / £853,725 / £45,832
1927 / £4,678,897 / 332,554 / £14.07 / £455,657 / £427,200 / £28,456 / 1936 / 5,237,377 / 379,134 / £13.81 / £824,552 / £760,560 / £63,992
1928 / £4,574,781 / 336,340 / £13.60 / £507,504 / £483,490 / £24,014 / 1937 / 6,226,713 / 385,697 / £16.14 / £956,656 / £835,170 / £121,487
1929 / £5,189,706 / 340,169 / £15.26 / £637,223 / £602,063 / £35,160 / 1938 / 6,544,460 / 392,373 / £16.68 / £1,023,230 / £908,024 / £115,206
Table 1B: Growth rate of Cypriot GDP, Per capita GDP, Government Revenue and Expenditure
Year / GDP / Per capita GDP / Government Revenue / Government Expenditure / Year / GDP / Per capita GDP / Government Revenue / Government Expenditure1921-1922 / 7.2% / 6.0% / 12.8% / -5.4% / 1930-1931 / -6.2% / -7.3% / 7.4% / -0.7%
1922-1923 / 3.1% / 1.9% / -10.0% / 10.9% / 1931-1932 / -8.5% / -10.1% / 4.6% / 0.8%
1923-1924 / -2.3% / -3.4% / 2.7% / -9.7% / 1932-1933 / 0.5% / -1.3% / -1.7% / -6.6%
1924-1925 / 7.8% / 6.6% / 12.3% / 15.3% / 1933-1934 / 12.4% / 10.5% / 3.8% / 3.4%
1925-1926 / -4.0% / -5.1% / -9.1% / 2.0% / 1934-1935 / 11.8% / 9.9% / 15.8% / 18.2%
1926-1927 / 12.2% / 11.0% / 2.4% / -7.8% / 1935-1936 / -1.5% / -3.2% / -8.3% / -10.9%
1927-1928 / -2.2% / -3.3% / 11.4% / 13.2% / 1936-1937 / 18.9% / 16.9% / 16.0% / 9.8%
1928-1929 / 13.4% / 12.2% / 25.6% / 24.5% / 1937-1938 / 5.1% / 3.3% / 7.0% / 8.7%
1929-1930 / -5.4% / -6.4% / 6.3% / 24.2% / 1926-1931 / 2.0% / 0.9% / 10.3% / 9.9%
Note: Deflated using the CPI provided in appendix H of Apostolides, A. (2010) Source: Cyprus, Statistical (Blue) Books, (1921-1938) “Government Revenue and Expenditure”
Table 2: Real GDP decline in Cyprus compared to other countries in constant prices (Index; 1929=100)
Source: Apostolides 2010; Maddison 2010
In fact the economic situation of Cyprus was affecting the majority rural population (78% of the population according to the 1931 census). Cypriot communities (majority Greek/Christian and minority Turkish/Moslem) were overwhelmingly resident in rural areas and working in the primary sector[6]. They were also scattered geographically in a mosaic of communities; thus the effect of the combined depression and drought affected both communities relatively similarly.
The depression of Cyprus might have affected the different minorities in a similar way, but it does not affect all geographical areas equally. Agricultural production in Cyprus is very structured due to environmental conditions. The presence or not of height and water delineated what the villagers could or could not produce. As table 3 shows different types of agricultural production: crops dependent on rainfall such as grain, carobs and olives were much more seriously affected that those areas where aquifers were based on irrigation.
Table 3: Real value added of agricultural production in Cyprus (Index; 1929=100)
Source: Apostolides 2010
The downturn in agricultural production led a credit crunch in the informal credit system. Cyprus farmers were burdened with debt they could not repay, which was similar situation as in other East Mediterranean states; the combined drought and depression rapidly increased the amount of non-performing loans, as indicated by the increase in forced sales of mortgaged rural property. Despite the rapid growth of the cooperative saving banks prior to the great depression, the faulty structure of the sector ensured that the cooperatives could not provide relief, leading to a serious retrenchment of the available rural credit markets (Apostolides, Vrachimis, Vostis, 2013)
By late 1931 the situation was worsening both politically and economically volatile. The government reserve was exhausted and the government refused to even discuss the demand by the LC to dramatically cut the wages of British officials to level desired by the Cypriots. Instead it passed substantial increases in duties and excise, increasing taxation both in luxury and in staple consumption goods (the government monopoly on salt was particularly hated). The British Colonial office considered that the increases were not enough to ensure the island would not need to be bailed out by a “grant-in-aid” by the British government and it was delaying the confirmation of the Order-in-Council that would pass the government budget and tax increases. This severely weakened the bargaining hand of the Governor, who could not pass any legislation by the LC unless the grievances were acted upon, but he could not count on the Colonial office accepting that there would not be additional taxation demands in 1932.