7. Interpretation of the results
7.3. Comparisons with earlier research
Earlier research has given diverse results when studying in different locations and at different times the connection between the unemployment rate and protests or organised violence.
Some studies have indicated that there is a positive correlation between these two factors, but most studies have suggested either no or even a negative relationship. Generally my results from the first research question give scantly support for the hypothesised positive relationship between unemployment and political instability, and taken into account the earlier research, this is maybe not surprising. However, my results provided partial evidence that the higher unemployment rate is connected to the higher level of organised violence in temporal inspection. These results are in line with some studies from the Israeli–Palestinian conflict.
When studying the relationship between unemployment and political violence in Israel–
Palestine, Sayre and Caruso & Gavrilova found evidence that the higher unemployment rate is connected to a higher level of terrorism (Sayre 2009; Caruso & Gavrilova 2012). However, Hassan, Berrebi and Krueger & Malečková have found that Palestinian suicide bombers have not been unemployed or poor (Hassan 2001: 37–38; Krueger & Malečková 2003; Berrebi 2007). This also partially echoes my results from Egypt. The results in hypothesis 1 suggested that during the years, the higher unemployment rate has been connected to the higher level of organised violence, and according to SCAD, the higher unemployment rate is also connected to the higher number of protests. Still, at the individual level the unemployed in Egypt were not more active protesters. White and Thompson did not find a positive relationship between the unemployment rate and political violence, in some cases they found negative relationship (Thompson 1989; White 1989; 1993). But Honaker instead found that if separate
unemployment rates for Catholics and Protestants were used, political violence correlated positively with the unemployment rates (Honaker 2010). This is something that my analyses do not reveal. If some ethnic minorities have been active protesters because of their higher
unemployment rate, this will remain unnoticeable as I did not include a variable for ethnicity in my models.
Some earlier studies have shown at the country level that those Arab countries which experienced higher unemployment rates were not more prone to the revolutions and turmoil of the Arab Spring (Korotayev et al. 2014; Byun & Hollander 2015). My results mainly confirm that the same applies at the subnational level, protests and organised violence have not
concentrated in those governorates with high unemployment.
In relation to the trends of unemployment and protests in the context of the Arab Spring, Ottaway & Hamzawy argue, that over the decade preceding the Arab Spring there was an increase in protests, strikes, and other forms of social protests in the Arab world (Ottaway &
Hamzawy 2011: 2). This rise is actually impossible to observe, when looking at my charts of protests in Algeria, Egypt, Morocco, and Tunisia. This might raise further doubts towards the data of ACLED and SCAD or then this rise was more apparent in some other countries. It is also possible that Ottaway & Hamzawy have not based their argument on quantitative measurements or they have not meant it to be taken this literally. Goldstone noted that governments in Egypt and Tunisia had long created jobs but these subsidies had been cut out in the 2000s (Goldstone 2011: 12). According to Mirkin during the decades preceding the uprisings the numbers of Arabs working as migrant workers in oil exporting Arab countries had decreased and this had blocked an employment opportunity which many had utilised (Mirkin 2013: 8, 17). These developments raised by Goldstone and Mirkin do not seem to appear as increases in the unemployment rates in the ILO data I use. It is likely that these developments mentioned by Goldstone and Mirkin were real but had a minor effect on the unemployment rates at national or regional level.
My results indicating that the unemployed are not active protesters are in line with many earlier studies. Oyefusi and Scacco have found that in Nigeria being unemployed actually often statistically significantly reduces an individual’s willingness to take part in protest (Scacco 2007; Oyefusi 2010). In some cases my results suggest similarly that the unemployed have participated in the protests more seldom than the employed, but mostly I have found no significant differences between the unemployed and the employed. I naturally have a slightly different approach than Oyefusi and Scacco, who study whether the unemployed are willing to protest. If we assume that the unemployed have less political, organisational and other resources than the employed, we would rather expect that the unemployed are more willing to protest compared to how often they finally do protest. Here, instead, the unemployed in Nigeria were mostly statistically significantly less willing to protest, but my results indicated
that there generally were no considerable differences in actual protesting. Taken into account the seemingly notable differences between the countries in the protest attendance of the unemployed, these differences between my results and results from Nigeria are then rather likely to be explained by country specific differences. I did also introduce several studies which did not approach this topic quantitatively but still concluded that the protests of the unemployed have remained rare and small scale incidents.
In relation to the Arab Spring, some studies based on survey data sets have similarly to my results indicated that the unemployed have not been more active, or have even been comparatively passive protesters in the Arab countries (Beissinger et al. 2012: 13; Hoffman &
Jamal 2012; 2013: 28–30; Doherty & Schraeder 2015; Shafiq & Vignoles 2015). These studies have, however, had their focus elsewhere than in the role of unemployment. In addition they have studied just single countries, used data sets which do not actually include information about Arab Spring protest participation or their inspection of the protest
participation of the unemployed has been descriptive. In the case of the 2012 article of Hoffman & Jamal, I also argued that their measurement of unemployment is imprecise.
Beissinger et al. have demonstrated, that between Tunisia and Egypt there were
differences in how different societal groups took part in the uprisings. They suggest that also more generally the characteristics of those who participated the protests varied from country to country (Beissinger et al. 2012). This gets some support from my results in relation to unemployment. It indeed seems that there are differences between countries as to how the unemployed mobilised. Interestingly, also earlier studies done in countries around the world have often indicated similar results from the same countries. Studying partly different periods of time, Sayre and Caruso & Gavrilova found some indication that in the Israeli–Palestinian conflict the unemployment rate is linked to politically motivated violence (Sayre 2009;
Caruso & Gavrilova 2012). In Northern Ireland, instead, several studies exploring different areas and times have found that the unemployment rate is generally not linked to the level of political violence, and if it is, the relationship is negative: higher unemployment predicts less violence (Thompson 1989; White 1989: 1288–1289; 1993: 580–582). In Nigeria, Oyefusi and Scacco have found, using different survey data sets that the unemployed are statistically significantly less likely to participate in a violent protest (Scacco 2007; Oyefusi 2010). This might suggest that more generally, there are differences between countries in how
unemployment is connected to protesting or organised violence. If this is the case, it could at least partly account for diverse empirical results around the world.
Schlozman & Verba found that in the United States the unemployed were considerably
less happy with their achievements in life than the employed, but this dissatisfaction did not lead to political participation (Schlozman & Verba 1979: 346–349). Bagguley interviewed unemployed people in Britain in 1980s. Some of them were involved in the Brighton centre, a centre attempting to organise the unemployed for involvement in collective action, whereas some interviewed did not take part in the activities of this centre. The interviews of the politically passive unemployed highlighted their beliefs that collective action would not improve their matters. Bagguley regarded this fatalism to be the major reason why the unemployed did not protest in the 1980s (Bagguley 1991). My results reflect these findings.
The unemployed are dissatisfied but this does not turn to political activity nor protesting, as protesting actually also is a form of political activity. With my analyses I am not able to identify, whether the unemployed in these Arab countries remained comparatively passive, because they thought that with political activities they would not achieve anything. However, it surely seems one potential explanation.