2. Key Features of Past Elite Bargains
The success and shortcomings of the 1992 GPA and post-conflict peace-building are reviewed in a longer paper published in early 2018.6 This paper is an update, focusing on resumed targeted armed conflict between RENAMO and the Mozambique government and highlighting what is necessary to make a new elite bargain stick.
The UN oversaw the Rome GPA and, through Security Council Resolution 797, set up United Nations Operation in Mozambique (ONUMOZ) to perform a series of peacekeeping tasks including monitoring and verifying the implementation of the ceasefire.
Map 1: RENAMO presence 1993
— Source: Updated from Vines, A. (1996), Renamo: From Terrorism to Democracy in Mozambique? James Currey, p. 2.
Map 1: RENAMO presence 1993
In order to fulfil its mandate, ONUMOZ was comprised of both civilian and military departments. It was tasked with monitoring the cantonment, disarmament and demobilization of nearly 110,000 combatants from both sides, as well as the creation of a new army and the resettlement of 5–6 million refugees and displaced people. The cost was estimated at $331 million ($1 million per day) until 30 November 1993.7 ONUMOZ was also mandated to support and observe Mozambique’s first multiparty elections in October 1994.
This approach was initially successful. By 1996, 87 per cent of demobilized soldiers had been integrated into society, and most had secured a food supply or small guaranteed income. The total reintegration budget was $94.4 million, of which $35.5 million was allocated to support cash for registered ex-combatants for a period of two years and $33.7 million went directly to demobilized soldiers. In all, some 92,000 soldiers benefited, including about 71,000 from government forces and 21,000 from RENAMO.8
By 1996, 87 per cent of demobilized soldiers had been integrated into society, and most had secured a food supply or small guaranteed income.
However, RENAMO combatants complained that they had been excluded from full reintegration benefits, namely pensions: contributions to their pensions were not deducted from their salaries like government troops, as most were not paid salaries during the conflict up to 1992.9
RENAMO proposed extending pension benefits to its soldiers, but FRELIMO opposed the move and used the pension debate to demonstrate its political strength. This issue resurfaced in the 2003 municipal elections and in the 2004 national elections, with little impact on government policy.10 However, it became one of the drivers of renewed armed conflict by RENAMO in 2013.
Disarmament
When conflict between RENAMO and the government resumed in 2013, many observers were surprised at how easily RENAMO managed to re-arm. According to the Mozambican Force for Crime Investigation and Social Reinsertion (FOMICRES), between three and four million weapons were in circulation at the end of the war in 1992.11 During the 1992–94 peace process, the priority of ONUMOZ was to help RENAMO transform itself into a political party and contest national elections. The UN priority was to dismantle RENAMO’s command and control structures, and also disperse ex-combatants through reintegration initiatives, such as the pay-and-scatter programme. Disarmament was not a priority, and the UN’s Special Representative Aldo Ajello has since admitted that at the time he was concerned that forceful disarmament would undermine the peace process.12
An official mediator in the RENAMO and FRELIMO stand-off, Bishop Dinis Sengulane, concluded that the failure to completely disarm in 1992–94 resulted in many individuals retaining their weapons. However, the fact that arms used in the recent conflict are in much better condition than would be expected after years of disuse, and that some RENAMO fighters are younger than ex-combatants would be, demonstrated that there are new sources of these weapons and that renewed violence is attracting fresh participants.13 EU observers also concluded that the 2014 elections were impacted by failed disarmament, which led to the intimidation of some voters.14
The process of disarmament, demobilization and reintegration (DDR) of RENAMO fighters into government forces has returned as a critical issue with allegations from RENAMO that its members are the victims of discrimination. The International Observer Military Team for the Cessation of Military Hostilities (EMOCHM) was established by the September 2014 agreement between President Guebuza and Afonso Dhlakama.15 It included observers from Botswana, South Africa, Zimbabwe, Kenya, Cape Verde, Italy, Portugal, the UK and the US and was led by a brigadier from Botswana. The team was tasked with monitoring the disarmament and demobilization of RENAMO’s ‘residual forces’ through incorporation into Mozambique’s military (FADM) and the police, or by assisting their return to civilian life. In October 2014, the government said it was prepared to incorporate 300 men from RENAMO’s militia into the armed forces and police (200 and 100, respectively), a figure based on the government’s past contacts with Afonso Dhlakama. For its part, RENAMO failed to prepare a list of fighters it wished to see recruited into the FADM and the police. As such, there was little progress.
The team was given 135 days to complete its task, this period expired in February 2015, and a dispute followed over the renewal of EMOCHM’s mandate. RENAMO sought an extension of 120 days, but the government insisted on 60. EMOCHM’s mandate eventually expired on 15 May 2015, having failed to complete its task. EMOCHM is now generally perceived as merely a political gesture by the government rather than a meaningful concession to RENAMO, which was never intended to be the start of a serious process.16
For RENAMO’s part, its leadership was also reluctant to reintegrate its residual forces and disarm, as targeted violence remains its prime leverage over the government in its continuing effort to extract political concessions. Though there were differing opinions on the RENAMO strategy among its members. RENAMO’s combatants that were assisted by EMOCHM signalled to international observers that they wanted a lasting deal and were fatigued by having been kept on military standby for over 20 years. They also expressed their anxiety over their own lack of skills and were naïve about the resources available to them for support and retraining.17
One of the conclusions that can be drawn from RENAMO’s strategy since 2013 is that maintaining armed men who are prepared to challenge the government has enhanced the party’s political standing in the short term and has resulted in a process to reach a new elite bargain in the long term. In addition, there are two further important lessons from the resumed violence of 2013–14 and subsequent clashes. First, it is clear that DDR efforts were finite, short-term technical processes that did not cover political inclusion. This meant that after a decade of ostensible peace international donors mistakenly concluded that Mozambique had undergone a successful post-conflict transition and that support for NGO efforts in this field was no longer a priority.
Second, disarmament should not have been neglected, and this was a missed opportunity. After ONUMOZ withdrew in December 1994, the opportunity to disarm diminished and subsequently only a small percentage of weapons were given up through official and NGO efforts. As a result, RENAMO has maintained armed men and weapons stockpiles for over 20 years. In 2012, the NGO FOMICRES,18 located large arms caches in five districts in Sofala province, including heavy weapons, but for political reasons it was not able to confiscate or destroy them.19 Furthermore, in January 2014, RENAMO official Rahil Khan warned that the group had arms caches across the country that it could draw upon.20 While RENAMO would never have handed over all weapons to ONUMOZ, more effort during its mandate could have reduced stockpiles, especially if international partners had continued to support disarmament and assist ex-combatants in central Mozambique to find alternative livelihoods.
In January 1998, worried about losing authority over these men, Dhlakama categorically refused to allow his bodyguards to be incorporated into the police.
RENAMO’s armed militia is also the result of a provision in the Rome GPA, which stipulated that the former rebels could maintain bodyguards (who would enjoy police status) as a ‘transitional guarantee’ until elections in 1994. The objective then was for the police to take over these responsibilities, while, Dhlakama’s bodyguards could receive police training if RENAMO named them on a list. Although a list was eventually provided, the government insisted that once trained, these men must obey police orders. In January 1998, worried about losing authority over these men, Dhlakama categorically refused to allow his bodyguards to be incorporated into the police. It seems that a mixture of fear of losing his prime asset (armed men) and concern over his own security drove this decision.
Consequently, from 1994 to 2013, the Mozambican government reluctantly accepted the de facto presence of several hundred-armed RENAMO personnel resident in the Maringué and Cheringoma districts of Sofala province, who occasionally paraded with weapons and intimidated local FRELIMO activists.21 A small group of them also escorted Dhlakama and provided security for his house in Nampula as his ‘presidential guard’. They were dressed in ill-fitting and worn uniforms, with shoes falling off and brandished old weapons.22 The government wanted to disarm this ‘presidential guard’ completely and offered to integrate it in the national police force, but this offer was rejected. A further opening to reintegrate some RENAMO militia fighters emerged from the September 2014 agreement, which temporarily halted hostilities, but RENAMO had to list fighters to take part and participants had to meet fitness standards. That list never appeared. In 2016, integration of RENAMO members and the career progression of those who were integrated after the GPA were key agenda items in talks between the government and RENAMO.
The principal achievement of the GPA was the conversion of RENAMO from an almost entirely military force into a predominantly political organization23 supported by a UN Trust Fund that, in the run up to the 1994 election, provided some $17 million. The Trust Fund played an important role in incentivizing RENAMO to forego violence and was referred to as an ‘effective insurance policy against failure’.24
RENAMO’s electoral fortunes
RENAMO has contested all five presidential and parliamentary elections since the civil war ended in 1992. The October 1994 elections enjoyed high voter turnout of above 85 per cent and there was little violence in the low-key campaign, although there was some intimidation by both sides in their stronghold areas. The south and north voted mostly for presidential candidate Joaquim Chissano and FRELIMO, while the central provinces of Manica and Sofala voted for RENAMO, indicating the role of regional and ethnic politics in the elections. The results also reflected inequality and poverty, with a strong sense that FRELIMO had channelled national resources to its hinterland, the far south of Mozambique, where it dominated among the electorate. In the strategically important provinces of Nampula and Zambézia, where 41 per cent of the electorate were registered, RENAMO had the advantage, but it was close.
The results of all five elections up to 2014 show a general clear and consistent pattern. FRELIMO has commanded a majority overall and dominated in the capital, the south and the extreme north, while RENAMO remained strong among voters in the centre and north. However, RENAMO was visibly weakened after the 2009 elections that were dominated by FRELIMO, which won 75 per cent of the votes and had majorities in all former RENAMO strongholds. Positions in the National Assembly were allocated to parties in proportion to their number of parliamentary seats.
Figure 1: Presidential elections vote share (%)
— Source: República de Moçambique Comissão Nacional de Eleições (CNE) [Republic of Mozambique National Elections Commission]; Secretariado Tecnico da Administracao Eleitoral (STAE) [Election Technical Secretariat], http://www.stae.org.mz/.
Figure 1: Presidential elections vote share (%)
After 1994, and especially after the 1999 election results – where RENAMO’s support held up and Dhlakama came close to challenging Chissano’s share of the vote – FRELIMO concluded that RENAMO posed an electoral risk.25 In response, FRELIMO aggressively countered RENAMO, at times this involved intimidation and harassment of its supporters, especially during electoral cycles. As a result, there have been electoral irregularities in the 1999, 2004, 2009 and 2014 elections.
RENAMO also contested the 2003 municipal elections for control of 33 municipalities, and successfully returned mayors in Beira, Ilha de Moçambique, Marromeu, Nacala and Angoche and achieved some 40 per cent of the vote in Nampula, Quelimane and Chimoio. However, in the 2008 municipal elections, RENAMO did not win control of a single municipality. Four of its former municipalities were won by FRELIMO and Beira was won by Daviz Simango (standing as an independent candidate). Dhlakama boasted at the time that he would swear the defeated RENAMO candidates into office to run parallel municipal administrations, but did not carry out his threat. After RENAMO’s defeat in the 2009 presidential and parliamentary elections, Dhlakama regularly threatened to hold nationwide demonstrations against what he claimed were fraudulent election results, but no RENAMO demonstrations were staged. Dhlakama also announced that the RENAMO deputies elected in 2009 would boycott the new parliament, but all the RENAMO deputies, including their secretary-general, defied him and took up their seats, anxious to claim their allowances.
Dhlakama’s strategy between 1994 and 2013 was to regularly obstruct parliament or circumvent it by seeking high-level bilateral negotiations between political party leaderships. There was little vision beyond oppositionist politics, and RENAMO’s falling electoral share until 2014 reduced Dhlakama’s bargaining power.26
From May 2009, Dhlakama relocated permanently from Maputo to the northern city of Nampula, and in October 2012, Dhlakama left Nampula for Satunjira, Gorongosa, in central Mozambique, near Casa Banana, a guerrilla base that served as RENAMO’s headquarters during the early 1980s. The timing of this latter move was significant: Satunjira was a former RENAMO military base and Dhlakama scheduled his arrival to commemorate the anniversary of the death of RENAMO’s founder, André Matsangaissa, killed during military action near Satunjira by FRELIMO on 17 October 1979.
Return to armed conflict
As the above sequence of events show, Dhlakama was isolated in Nampula and lacked resources for patronage, which meant that his core supporters in central Mozambique were increasingly impatient. His return to central Mozambique reflected a calculation that his only viable option to strengthen his grip on power and his negotiating hand was to return to targeted armed violence. His economic interests, including artisanal gem mining (Tourmaline) in central Mozambique, also came increasingly under pressure from government officials in 2012, heightening the squeeze on his resources.
Armed attacks started in April 2013 and, in June 2013, the government introduced military convoys to protect a 100-km stretch of road between the River Save and Muxungué (until 28 August 2014).27 Political tensions deepened on 21 October 2013 when FADM occupied the Satunjira base after RENAMO had congregated there again to commemorate the anniversary of the death of Matsangaissa. Dhlakama fled to another base deep in the Gorongosa region and the FADM moved on to occupy another RENAMO base, Maringué, which had remained a location of armed RENAMO men with tacit acceptance by the government since 1994.
RENAMO then launched low-intensity targeted attacks, aimed at strengthening its negotiating position. This approach differed from their objectives in the early 1990s of holding towns and territory, degrading government infrastructure and causing massive displacement. In early May 2014, as peace talks progressed, RENAMO announced a ceasefire along the EN1 highway to assist election registration efforts in central Mozambique for the October national elections. The most recent RENAMO armed incident in 2014 was on 1 July in Condue, Sofala province along the railway.
This was followed by RENAMO and the government agreeing upon a ceasefire at the 74th round of negotiations on 24 August 2014 in Maputo. Finally, Dhlakama agreed to leave his hiding place in central Mozambique and be escorted by an international delegation of foreign diplomats accredited in Mozambique back to Maputo on 4 September, and on 5 September President Guebuza and Dhlakama met in Maputo and formally signed the agreement to end hostilities. During 2013, at least 60 people were killed and more than 300 injured. There are no accurate comparable figures for 2014, but observers believe the total number of casualties was less than 100, most of which were injured, while in 2015 the figure was some 20 killed and 50 injured.28
Map 2: Comparison of RENAMO Attacks 2013–14 (left) and 2015–16 (right)
— Source: Adapted from @Verdarde, AIM, ACLED and author interviews.
Map 2: Comparison of RENAMO Attacks 2013–14 (left) and 2015–16 (right)
In early 2015, RENAMO again threatened conflict if it did not obtain concessions from the government, following two face-to-face meetings between Dhlakama and President Nyusi. The president signalled he would support a parliamentary bill for autonomous provincial governments if RENAMO ended its boycott of parliament and submitted it to the National Assembly for debate. In the short term, this provided hope of a new deal but Nyusi failed to deliver on his promise and FRELIMO parliamentarians rejected the draft bill. On 2 April 2015, the FADM and RENAMO forces exchanged fire in Guija district, Gaza province. Soon after, Dhlakama confirmed the incident saying that 150 RENAMO troops had moved south (many of them aged 40 and over).
Accurate numbers of how many armed men Dhlakama remobilized are not available but the figure is probably in the hundreds. Many of these are ex-combatants of the 1977–92 conflict from central Mozambique, although some younger participants seem to have been drawn to RENAMO by the fighting. RENAMO repeated its low-intensity guerrilla tactics of the past such as ambush, hit and run, and disruption of infrastructure. RENAMO’s ability to cause disruption was aided by weak government forces, unable to respond efficiently with counter-insurgency operations in central Mozambique. RENAMO tried to spread its military operations outside central Mozambique and Tete, but this was less successful and better contained by government forces.
Impact of municipal politics
RENAMO’s boycott of the 20 November 2013 municipal elections, due to an electoral law dispute, back-fired spectacularly. At the local level, in Quelimane and Nampula, RENAMO supporters tactically voted for the Democratic Movement of Mozambique (MDM) – a splinter party of RENAMO. But FRELIMO won 49 mayoral seats and the MDM gained four, including the large cities of Beira, Quelimane and Nampula. MDM managed to secure 365 (30 per cent) of 1,216 municipal assembly seats overall and its candidates took more than 40 per cent of the vote in 13 municipalities, including in the FRELIMO heartlands of Maputo and Matola, a feat never achieved by RENAMO. This was the first time that the MDM contested municipal polls nationwide, and the results show the party can campaign at the national level and attract support from urban areas outside Beira and Quelimane.29
Figure 2: FRELIMO share (%) of total valid votes in municipal elections, 2003–18
— Source: República de Moçambique Comissão Nacional de Eleições (CNE) [Republic of Mozambique National Elections Commission]; Secretariado Tecnico da Administracao Eleitoral (STAE) [Election Technical Secretariat], http://www.stae.org.mz/.
Figure 2: FRELIMO share (%) of total valid votes in municipal elections, 2003–18
The municipal election results convinced RENAMO and FRELIMO that they needed to return to negotiations to ensure they neutralized MDM’s growing support base, and the 2014 presidential and legislative elections re-emphasized that RENAMO and FRELIMO were the primary political players in Mozambique. Both also showed they could control their supporters as there were only isolated violent incidents during the 2014 elections. Given that the elections followed 18 months of targeted armed violence, this was an achievement and shows that peace can prevail when there is leadership and political will. However, electoral participation was low, at around 48.6 per cent, compared with 87.87 per cent in 1994, continuing a long-term trend of political indifference. More than half of registered voters (55.27 per cent) did not participate in the 2009 elections.
Electoral participation in 2014 was low, at around 48.6 per cent, compared with 87.87 per cent in 1994, continuing a long-term trend of political indifference. More than half of registered voters (55.27 per cent) did not participate in the 2009 elections.
Nevertheless, the election results in 2014 showed that support for FRELIMO was declining and that RENAMO was winning over voters. It also confirmed to Afonso Dhlakama that his strategy of returning to targeted armed conflict had been rewarded, not only had the government re-opened negotiations directly with RENAMO but the party had made gains, increasing its number of deputies. A few years earlier memories of the previous civil war would have been stronger among the electorate and a return to conflict might have been punished but, as is often the case in politics, timing matters.
As occurred after previous elections, the judiciary rejected opposition claims of rigging at the 2014 elections on technicalities.30 Although there is insufficient evidence to conclude that rigging took place on a scale that would affect the overall result of a FRELIMO presidential victory and parliamentary majority, it is clear the electoral commission did commit some level of fraud and did not perform its duties satisfactorily. The opposition parties are also at fault for being unable to provide credible evidence of widespread fraud despite having deployed electoral observers across the country. The resulting suspicion, conspiracy and allegations have contributed to deepening political tensions between FRELIMO and opposition parties. Lessons from the 2014 elections are that the electoral legislation should be amended to provide a clear system of complaints and appeals, and that judges, electoral managing bodies and political parties need training in how to use these procedures. There also needs to be efficient training on counting and tabulation procedures.
While RENAMO has rejected all election results since 1994, due to the 2013–14 outbreak of armed violence, it was important for reconciliation that the electoral process was at least perceived to be better than it had been in the past. RENAMO’s strong performance, with Afonso Dhlakama winning the majority of the vote in five provinces (Nampula, Zambézia, Tete, Manica and Sofala), was surprising not least because of his late start to campaigning. It strengthened Dhlakama’s position in the party and there were no longer calls for him to step down. RENAMO also concluded that calculated armed violence has restored greater parity with FRELIMO, brought about concessions and marginalized the threat posed by MDM.
Figure 3: FRELIMO share of votes municipal and National Assembly elections (%)
— Source: República de Moçambique Comissão Nacional de Eleições (CNE) [Republic of Mozambique National Elections Commission]; Secretariado Tecnico da Administracao Eleitoral (STAE) [Election Technical Secretariat], http://www.stae.org.mz/.
Figure 3: FRELIMO share of votes municipal and National Assembly elections (%)
The October 2018 municipal elections were generally peaceful, and the results further confirmed that even if RENAMO had been disadvantaged by the death of Afonso Dhlakama, the electorate were increasingly voting for the party in protest against FRELIMO. Although FRELIMO won 44 out of 53 municipalities, there were at least five other municipalities in which electoral irregularities denied RENAMO victory.
7 Synge, R. (1997), Mozambique: UN Peace-keeping in Action, 1992-1994, United States Institute of Peace.
8 Vines, A. (2013), ‘Renamo’s Rise and Decline: The Politics of Reintegration in Mozambique’, International Peacekeeping 20 (3), pp. 375–393.
9 Schafer, J. (1998), ‘“A Baby Who Does Not Cry Will Not Be Suckled”: AMODEG and the Reintegration of Demobilised Soldiers’, Journal of Southern African Studies 4 (1), pp. 207–222.
10 McMullin, J. (2004), ‘Reintegration of Combatants: Were the Right Lessons Learned in Mozambique?’, International Peacekeeping 11 (4), pp. 625–643.
11 Reisman, L. and Lalá, A. (2012), ‘Assessment of Crime and Violence in Mozambique & Recommendations for Violence Prevention and Reduction’, Open Society Foundations Crime and Violence Prevention Initiative & Open Society Initiative for Southern Africa, https://issuu.com/osisa/docs/cvpi_mozambique_report_-_final_engl (accessed 6 Jul. 2019).
12 Author interview with Aldo Ajello, Rome, 17 October 2012. Others consider this as one of the flaws and that this was a lesson learned by the UN Department of Peacekeeping Operations (DPKO) for future actions; author interview with Owen Greene, University of Bradford, 2 March 2017.
13 Sengulane, D. (2014), ‘Restoring Peace in Mozambique’, speech at Chatham House, http://www.chathamhouse.org/event/restoring-peace-mozambique (accessed 6 Jul. 2019). There were not many fresh, young combatants involved in this armed conflict, but small numbers joined RENAMO for training in 2013–16 and more joined in 2019.
14 The EU Election Observation Mission’s final report noted that RENAMO had not disarmed before the elections, which ‘bore an influence on the electoral campaign’. See EU Election Observation Mission (2014), ‘Mozambique: Final Report – General Elections, 15 October 2014’, p. 4.
15 EMOCHM (2014), ‘Termos de Referência da Equipa Militar de Observação Da Cessação Das Hostilidades Militares’, http://www.open.ac.uk/technology/mozambique/sites/www.open.ac.uk.technology.mozambique/files/files/Renamo%20dialogo%20acordos.pdf (accessed 6 Jul. 2019).
16 Author interview with Jamie Martin, ex-EMOCHM – Beira observer, Salisbury, 24 February 2017.
17 Ibid.
18 Littlejohn, G. (2015), ‘Secret Stockpiles, Arms Caches and Disarmament Efforts in Mozambique’, A Working Paper of the Small Arms Survey, 21 September 2015, http://www.smallarmssurvey.org/fileadmin/docs/F-Working-papers/SAS-WP21-Secret-Stockpiles.pdf (accessed 6 Jul. 2019)
19 FOMICRES is led by ex-RENAMO members and was able to access areas of central Mozambique other NGOs could not. Over the last decade FOMICRES suffered from funding difficulties due to international attention shifting from post-conflict assistance. Author interview with Gary Littlejohn, Harrogate, 2 March 2017.
20 Agência Lusa (2014), ‘Assesor da Renamo Garante Que Homens do Partido Têm Armas em Tudo o País’, 10 January 2014, http://noticias.sapo.pt/internacional/artigo/assessor-da-renamo-garante-que-homens-do-partido-tem-armas-em-todo-o-pais_17143831.html (accessed 6 Jul. 2019).
21 AllAfrica (2005), ‘Mozambique: Dhlakama Refuses to Disband His Guard’, 6 October 2005, https://allafrica.com/stories/200510060517.html (accessed 16 Jul. 2019).
22 Author witnessed them line up as guard of honour after his meeting with Afonso Dhlakama, Nampula, 23 September 2010.
23 Synge, R. (1997), Mozambique: UN Peace-keeping in Action, 1992-1994, United States Institute of Peace, p. 148.
24 Ibid.
25 Author interviews with FRELIMO officials, Maputo, 22 September 2010.
26 Machietto, R. (2016), Beyond Peacebuilding: the Challenges of Empowerment Promotion in Mozambique, Palgrave MacMillan, p. 268.
27 Informal mediation started in July 2013, when Anglican bishop Dinis Sengulane and academic Lourenço de Rosário met Dhlakama in Satunjira and agreed to his key conditions: reform of electoral law; the depoliticization of the civil service; reform of the military; and more equitable wealth sharing.
28 Author interviews with officials, Maputo, January 2018.
29 For an analysis of the impact of MDM on Afonso Dhlakama, see Vines, A. (2013), ‘Renamo’s Rise and Decline: The Politics of Reintegration in Mozambique’.
30 Vines, A., Thompson, T., Jenson, S. K. and Azevedo-Harman, E. (2015), Mozambique to 2018: Managers, Mediators and Magnates, Chatham House Report, p. 18, https://www.chathamhouse.org/sites/default/files/field/field_document/20150622Mozambique2018VinesThompson
KirkJensenAzevedoHarman.pdf (accessed 6 Jul. 2019).