A Geography and Ethnic Diversity

Mozambique, with a population of around 30 million (INE 2019), covers a total land area of more than 799,380 km² (FAO 2016). It is located on the East Coast of Southern Africa and borders six other countries: Tanzania in the north; Malawi, Zimbabwe, Zambia, and Swaziland to the west; and South Africa to the south (see Figure 2.1).

Figure 2.1 Mozambique in Southern Africa

Source: Authors’ construction based on Natural Earth Data (available at: www.naturalearthdata.com/about/terms-of-use/).

Mozambique’s total border length is around 4,445 km and its approximately 2,515 km of generally sandy coastline to the east faces the Indian Ocean and includes a number of islands (FAO 2016). Some twenty-five main rivers flowing east towards the Indian Ocean cross Mozambique. The largest and historically most important river is the Zambezi, whose 820 km Mozambican section is navigable for 460 km. The road infrastructure illustrates the lack of integration in the north–south dimension (see Figure 2.2), a characteristic to which we return in what follows.

Figure 2.2 Road infrastructure and rivers

Source: Authors’ construction using © OpenStreetMap contributors (available at: www.openstreetmap.org/, www.opendatacommons.org/) for roads and the Regional Centre for Mapping of Resources for Development (RCMRD) (available at: https://opendata.rcmrd.org/) for waterways.

The country is made up of eleven provinces, and the major urban centres are generally situated along the coast, the most important being Maputo, the capital, in the far south. Other cities include Beira, in the middle of the country, Quelimane, north of Beira, Nampula, yet further north, and Cabo Delgado in the north-east corner (see Figure 2.3).

Figure 2.3 Administrative map of Mozambique

Source: Authors’ construction based on Natural Earth Data (available at www.naturalearthdata.com/about/terms-of-use/).

Before the arrival of Vasco da Gama in 1498, Bantu-speaking tribes had occupied the territory of modern Mozambique since the first centuries CE. The Arabs started influencing from approximately 1100 CE and, while the effects of the encounter between the local population and Arab traders differed between the existing tribes, the resulting ethnic diversity and Arab influence remain present to this day. Many of the conflicts and contradictions between the different parts of Mozambique that exist today also have deep historical roots – a point to which we will return in what follows.

The major ethnic groups (and subgroups)Footnote ² in Mozambique, some of which are linked to ethnic groups in neighbouring countries, are: Yao, Nyanja, Makonde, Mwani, and Makua-Lomwe in the north and centre; Nsenga, Pimbwe, Chuwabo, Nyungwe, Maravi, Sena, and Shona-Ndau in the centre; and Chopi, Bitonga, Tswa, Shangana, and Ronga in the south. To these must be added Portuguese descendants and a ‘mestiço’ minority of African/Portuguese descent. Finally, there is a community of Indian Asian origin from Pakistan, India, and various Arab countries.

Importantly, these geographic characteristics imply that economic transaction costs are exceedingly large, and regional and ethnic divides have resulted in both economic fragmentation and differences in access to political power (for details, see Grobbelaar and Lala Reference Grobbelaar and Lala2003). In broad terms, economic as well as political power concentrated in the south even before independence, and did so because of the South African economy. The capital was moved from Ilha de Moçambique to Lourenço Marques (renamed Maputo after independence) in the early 1900s, following the discovery of gold in South Africa. The significance of the move of the capital to the extreme south of a very large country because of the important economic links with a much larger neighbouring economy cannot be overemphasised.

B Colonial Legacy

Historically, Mozambique was particularly attractive for the trade of gold, ivory, and slaves, as well as other metals and hides, and, even though their occupation was limited at first, the Portuguese gradually established commercial influence from 1505, adopting protectionist measures as well as clientelist and corrupt practices. The Portuguese importance grew in the following centuries as the colonial system was progressively developed,Footnote ³ and it intensified in the mid-1880s, backed by investments by foreign capital and companies, especially British, Rhodesian, and South African.

The Portuguese colonial presence and the control of private – often foreign, Anglophone – companies over Mozambique’s basic infrastructure had tremendous consequences. Portuguese settlers effectively colonised Mozambique,Footnote ⁴ and its administrative and bureaucratic institutions, as well as the army, depended on instructions from Lisbon. Existing infrastructure ran east–west and the country remained not only ethnically but also economically disintegrated in the north–south dimension. Moreover, there was a weak and biased colonial education system, resulting in virtually no training of indigenous people or recruitment of African-Portuguese, though with some prominent exceptions.

Until around the mid-1950s, Mozambique was largely limited to supplying cheap raw materials to Portugal and cheap labour and transport services to neighbouring countries. The traditional rural sector received no government support and remained underdeveloped. Additionally, a series of legal measures aggravated the economic and social differences that divided the population. In the 1960s, there was increasing migration of settlers to Mozambique, expanding the economy and the internal market. By 1964, industrial production for the internal market exceeded production for export (Leite Reference Leite1989). But despite the fact that the economy had started to become less orientated towards continental Portugal, the struggle for independence began in that same year.

Serious challenges mounted (see Section C) and outright chaos broke out at the time of independence in 1975, when a majority of the Portuguese population left the country. Within weeks, administrative structures had collapsed and become empty shells,Footnote ⁵ and the Mozambique Liberation Front (FRELIMO), which, although a well organised liberation movement, had no experience of running a country, took over.Footnote ⁶ Its focus was on dismantling colonial institutions.

C Independence, Geopolitics, and War

The decade before independence was marked by military struggle. Although FRELIMO had been formed in 1962 as a constitutionalist and non-violent movement, it launched a military offensive in 1964 and continued to do so into the 1970s, initially taking control of parts of the north and progressively advancing towards the centre of the country. Largely due to the overthrow of the Portuguese government in April 1974, official talks about Mozambican independence started, and they led in September of the same year to the signing of the Lusaka Agreement, which established a transition government with a majority of FRELIMO members. Formal independence and the foundation of The People’s Republic of Mozambique as a one-party socialist state happened on 25 June 1975 under the leadership of FrelimoFootnote ⁷ and President Samora Machel (see Figure A2.1 for a timeline of the main events since independence).

As was the case in Vietnam in the mid-1970s after its defeat of US forces, the atmosphere in Mozambique was ‘upbeať among the Frelimo ‘comrades’. Having ‘defeated’ the Portuguese colonial power, Frelimo ambitiously declared, in jubilant tones, that the coming decade would see victory over under-development. While it was over-optimistic and, in retrospect, naive to compare victories achieved on the battlefield with challenges faced in the development field, it is pertinent to recall that the course of action taken appeared at the time self-evident to many, including the leadership in both Vietnam and Mozambique, with both economies being poor, exploited, and destroyed.Footnote ⁸

During the immediate post-independence period, national reconstruction and consolidation were in focus. However, a series of factors undermined these efforts. The first was the serious economic challenges that marked this period. As alluded to above, approximately 80 per cent of the Portuguese population had left the country. This created extreme labour shortages in public administration and the secondary and tertiary sectors, as well as having a devastating effect on agriculture, which depended on settlers for input supplies and the marketing of output, including extensive credit provision, often in kind. Consequently, living and educational/managerial standards were dismal to begin with in the newly independent country.

Second, after its transformation into a Marxist-Leninist party in 1977, Frelimo initiated a ten-year plan (1981–90) with the goal of overcoming under-development. Socialist policies were indeed the norm at the time in newly independent African countries and a logical reaction to the past. However, the strategy, conceived in orthodox central planning terms, was out of touch with the reality on the ground. Import substitution, forced mechanisation of agriculture, and resettlement of large numbers of people were at the core of the strategy. Similarly, only scant attention was paid to small farmer needs and their requirements in terms of input supplies and output outlets. Accordingly, misconceived policies were introduced, including a hostile view of the private sector. This bias had long-lasting effects, and rather than winning support among the peasantry, Frelimo was losing it.Footnote ⁹ There was room for voice and participation but only inside the party structures. At the same time, Frelimo developed authoritarian and coercive methods, in large measure in response to the context of a brutal war instigated by outsiders.

Third, independence came suddenly as well as late to Mozambique as compared with other African countries, and regional (Rhodesia and South Africa) and global (Cold War) conflicts left the new country with very little room for manoeuvre, either politically or economically. This turned out to be very costly indeed, illustrated by the heavy economic and military impact of Mozambique’s decision to enforce United Nations sanctions against Rhodesia in 1976 (UN 1976).

Mozambique’s independence posed threats to Rhodesia and South Africa, which feared a communist onslaught. They therefore supported and financed Renamo (the Mozambican National Resistance, originally designated MNR), which opposed Frelimo’s socialist orientation.Footnote ¹⁰ Renamo gathered Frelimo dissidents under the leadership of André Matsangaíssa (1977–79) and Afonso Dhlakama (1979–2018). The Renamo leadership came from the centre of Mozambique, but the movement was also constituted by members from elsewhere in the country. Frelimo has over the years pursued a policy of having representatives from all relevant ethnic groups in high positions in the party and in the government, yet prominent representatives from the centre and the north have felt ostracised.

After the independence of Zimbabwe (former Rhodesia) in 1980, the apartheid regime continued skilfully to exploit both historical and current ethnic contradictions and grievances and turned Renamo into a significant military force capable of disrupting and sabotaging large parts of the country. The Mozambican government responded by increasing repression and by targeting the supply lines of Renamo with a stronger and more sophisticated military presence and firepower. This strategy failed and the war escalated, in the midst of the regional and global Cold War. With the signing of the Nkomati Accord of non-aggression between Mozambique and South Africa in 1984, President Samora Machel hoped for some reprieve, but large-scale sabotage and destruction of infrastructure and killings continued, with support from South Africa.Footnote ¹¹ Backed by the military might of South Africa, Renamo’s ruthless warfare slowly but surely undermined Frelimo’s nation-building efforts. A stubborn but ineffective strategy by Frelimo to criminalise insurgents and Renamo supporters, not recognising the need for political settlements, added to the complexity of the situation.

Fourth, another consequence of the late timing of Mozambique’s independence was that the country had no more than a five-year ‘settling-in’ period before laissez-faire ideological winds of change swept across the world around 1980, including the election of President Ronald Reagan in the US and Margaret Thatcher as Prime Minister in the UK. These changes influenced the policy stance of international financial institutions, including in particular the International Monetary Fund (IMF) and the World Bank, and this turned out to be critical in the mid-1980s in Mozambique (see Tarp Reference Tarp1993).

D The Bretton Woods Institutions and a U-turn in Economic Strategy

The fragility of former alliances in Eastern Europe and elsewhere in the global arena (including those between the USSR, East Germany, Cuba, and China), the ongoing war, and an economic state of affairs that reached a historical low point in 1986 forced Frelimo to change course. Mozambique had no viable option but to turn increasingly to Western donors (with the Nordics having been present from very early on) for external finance, and the donors came to play a decisive role, imposing free-market policy conditionality on a war-torn economy. The Nkomati agreement, signed by President Samora Machel and the Prime Minister of South Africa, P. W. Botha, signalled the beginning of negotiations for support from the Bretton Woods institutions and the Paris Club creditors,Footnote ¹² though concrete policy initiatives only materialised after Samora Machel’s death in a plane crash in South Africa in 1986. The government launched the five-year Economic Rehabilitation Programme (PRE) in 1987 and in effect turned the economic strategy pursued so far on its head.Footnote ¹³ The implications were wide ranging.

First, the structural adjustment programme (reflecting a standard package of orthodox policy reforms that took no account of local circumstances) was ill conceived for a country at war, as had been the ten-year development plan, and the short-, medium-, and longer-term economic effects were dire. No supply response from agriculture was possible due to the overwhelming effects of the war; and subsidised industries in the cities had to lay off workers, adjust wages, and eventually expect privatisation or closure.

Second, a key element of the PRE was the privatisation of state-owned enterprises. Diogo (Reference Diogo2013: 18) notes: ‘To make the transition from a centrally planned economy to a market economy, Mozambique privatized more than 4,000 companies, about 80 pct. of them benefitting Mozambicans and on highly advantageous conditions.’ This was done presumably over a few years and without a transparent and well-functioning financial market. Moreover, in the longer term, the PRE introduced a deep-seated structural and highly problematic economic imbalance between Frelimo and Renamo as privatisation went ahead while Renamo guerrillas were still in the bush. The drive to avoid foreign takeovers ran as a red thread through the process, instituted through a national ownership requirement. There was, however, no indigenous entrepreneurial class, given the weakness in the domestic politico-judicial system and the private sector. So assets and power transferred instead to a group of individuals in the public service and to Frelimo party officials, who managed to secure favourable bank loans (see Diogo Reference Diogo2013).Footnote ¹⁴

Sumich and Honwana (Reference Sumich and Honwana2007: 19) add:

Thus, as a consequence of policy conditionalities, a ‘new’ political and business elite (which overlapped in very large measure with the ‘old’ elite) started emerging that commanded not only political power, as in the early days after independence, but also economic power – and this often without the necessary management training and capacity. Importantly, Renamo associates were in no position to benefit, as they were still in the bush when assets and opportunities were privatised.

E Peace and Multiparty Elections

Peace returned with the Rome General Peace Accords in 1992, made possible by the political transition from apartheid in South Africa. In 1994, the first multiparty elections led to the election of President Joaquim Chissano, leader of Frelimo, as President of the Republic of Mozambique. There were signs of economic recovery after peace re-established itself, with impressive growth rates, a significant decline in poverty, and the creation of new infrastructure, particularly in the social sectors, supported by the donor community.

Aid flows surged and Mozambique became widely perceived as a donor darling. However, the external assistance often came with explicit and implicit conditionalities, such as a heavy reliance on external consultants rather than promoting capacity building of nationals. Moreover, the impressive growth, in large measure due to returning refugees, did not lead to a fundamental transformation of economic structures (see Section III). Finally, while some Renamo commanders and soldiers were integrated into the new Mozambican army, no full disarmament took place. Integration into the police force was especially difficult, and the Renamo leader, Afonso Dhlakama, maintained a strong personal protection force.

On the political front, Sumich and Honwana (Reference Sumich and Honwana2007: 18) stress that

The polarisation between Frelimo and Renamo continued, and while the ideological differences weakened, regional and ethnic contradictions came more into focus.

F Dominance of Frelimo

Frelimo maintained political power in the country after the closely contested general election in 1999. Joaquim Chissano continued as president, and in the Assembly of the Republic, Frelimo held 133 and Renamo 117 seats. Shortly afterwards, Frelimo made an important strategic move ‘when the government after six years of debate, in an effort to broaden its social base and weaken Renamo, issued a decree creating “community leaders”, which meant that former “régulos” [traditional chiefs] could hold office at the local level’ (Sumich and Honwana Reference Sumich and Honwana2007: 17). Accordingly, the Frelimo government actively sought ways to engage with and remunerate the traditional chiefs by attributing formal duties to them. The system was quite transparent and deliberately visible as chiefs also got uniforms for official occasions and visits. Such relations with traditional authority count a lot in an election, when many voters seek guidance from their local chiefs. While this was a remarkable move considering Frelimo’s history as a socialism-inspired liberation movement, it helped Frelimo remain the dominant political power in the subsequent elections.

In 2002, Frelimo elected Armando Guebuza as its Secretary General and presidential candidate. Guebuza moved quickly to strengthen the party structures in all spheres of the public administration and created cells at all levels. He also deliberately chose a government of politicians over technocrats and soon commanded a structure not much different from the one in place under the one-party state. He won a comfortable victory in 2004 and became President in 2005, backed by an overwhelmingly superior financial and organisational machinery amid widespread abstention.

In 2009, a new political party started competing in the parliamentary and presidential elections. The Democratic Movement of Mozambique (MDM) started up as a splinter group of Renamo, led by the mayor in Beira, Daviz Simango, son of one of FRELIMO’s first vice-presidents during the liberation war. The Simango family, from the centre of Mozambique, had joined Renamo expecting to create a strong political opposition to Frelimo. But Renamo, dominated by Afonso Dhlakama, was not a democratic party. Young and capable members therefore broke away, aiming to pursue democratic political process. Unlike earlier splinter groups, MDM enjoyed widespread popular support, and expectations were that it would win a significant part of the votes, ending the bipolar impasse in Mozambican politics.

However, in what many observers see as a concerted effort to sabotage MDM by Frelimo and Renamo party members of the electoral commission, MDM’s lists of candidates were rejected in 7 of the 11 provinces (including Maputo city), leaving the party with no possibility of winning any significant part of the votes. Among the reasons given were missing details in the candidates’ dossiers; the whole provincial list was dropped if just one of the candidates, even if it was a substitute candidate, failed to submit all the necessary documents, such as an original birth certificate. Some candidates’ dossiers were even ‘losť in the election commission’s offices. Donors protested to President Guebuza before the election but to no avail. Nevertheless, Renamo suffered significantly from MDM’s appearance on the political scene and was reduced to only 49 seats in Parliament, while Frelimo was taken by surprise when it lost important areas in Maputo to the new party.

While traditional party politics seemed locked in a Frelimo hegemony, 2008 and 2010 witnessed a surge of demonstrations and popular uprisings in the largest cities. Discontent was ignited by the worsening living conditions among the lower middle class (due, for example, to rising food prices – see Pinstrup-Andersen Reference Pinstrup-Andersen2014). Moreover, on both occasions Frelimo and the government underestimated the strength of the reaction. The events, especially in Maputo, had tragic consequences, with police shooting resulting in civilian deaths, including children, and gave rise to a debate within Frelimo about leadership and inclusion. Eventually, President Guebuza controlled the situation partly by force and partly by giving concessions, delaying increases in the price of water, electricity, and public transport (fuel).

The Guebuza regime ended in violent confrontation with Renamo,Footnote ¹⁵ which no longer believed that the government was willing to give in to its demands, including increased decentralisation. A last-minute peace deal paved the way for an election in 2014, but even within Frelimo, the issue of succession created problems. Guebuza held on to power as long as he could, and although the Frelimo candidate Filipe Nyusi won the election, he faced massive challenges from his first day in office, as discussed below.

G Recent Events

The first of Nyusi’s challenges was that the settlement with Renamo was short-lived. Renamo argued that all the problems it had pointed to remained, and the hardliners in Frelimo, who had gained strength under Guebuza’s ten-year rule, argued that it was better to solve the Renamo challenge militarily and decisively. So violent conflict escalated and soon central parts of the country were at war. Broken power lines, sabotaged railway lines, and army convoys protecting vehicles on the main roads were once again the order of the day in Mozambique in 2016.

The Nyusi government then chose to negotiate with Renamo and give in to some of its main demands. These included decentralisation and the election of provincial governors. This move was sustained even after the death of Renamo’s leader, Afonso Dhlakama, in May 2018, which led to a considerable weakening of the party under its new leader, Ossufo Momade. A 14-page constitutional amendment was approved by which elected bodies were created at provincial, district, and municipality levels.

A second challenge was the so-called ‘hidden debť scandal, which had become symptomatic of the increase in corruption Mozambique had experienced over the last couple of decades. While they recognise that the figure is open to contestation, Williams and Isaksen (Reference Williams and Isaksen2016) note in their introduction: ‘The average annual cost of corruption to Mozambique was recently estimated to be up to US$4.9 billion for the period 2004–2014.’ The authors, who provide a detailed account (on which the following is based),Footnote ¹⁶ point out that in the final years of Guebuza’s second term, three semi-public entities took out over US$2 billion in loans from private foreign banks without submitting them to the Assembly for approval, as required by the Mozambican constitution, even though they greatly exceeded the limit placed on government borrowing by the relevant annual budget appropriation bill. Williams and Isaksen (Reference Williams and Isaksen2016) pointedly note:

The sources of the loans included Credit Suisse and the VTB Bank, and it is clear that Ematum, Proindicus, and MAM (Mozambique Asset Management) were owned and controlled by a very small group of individuals and that they were very closely linked to the security sector. When the existence of the loans became public, the IMF suspended its support to Mozambique, and much foreign aid, including all direct support to the state budget (already on a downward trajectory), was frozen or significantly reduced. This drastically restricted the fiscal space of the government and led to an economic crisis that lasted throughout the first term of President Nyusi, and which continues to impact.Footnote ¹⁷ The subsequent legal process led in 2019 to the US indictment of former Minister of Finance Manuel Chang and others (see MNRC 2019). While the case continues in the courts, Mozambique’s relations with the donor community have started to normalise. A central element in this has been the re-establishment of relations with the IMF,Footnote ¹⁸ illustrated in the approval on 19 April 2019 of IMF emergency assistance following Cyclone Idai, which hit Mozambique earlier in the month, causing massive damage in the centre of the country, including the city of Beira.

A third challenge has been the deteriorating security situation and armed insurgency in the north-eastern province of Cabo Delgado, where Islamic militants have attacked key towns, killing or uprooting well over 100,000 people. The roots of the conflict are still debated, but seem to include a series of local grievances related to social exclusion, combined with interventions from international groups involved in criminal activities such as ivory and heroin trafficking. The conflict is dangerous in itself but also has the potential to spill over to other parts of the country. Importantly, Cabo Delgado is close to the major gas fields.

The fourth challenge, the discovery of huge gas resources after the turn of the millennium, is also an opportunity to turn the country around, provided appropriate institutions and policies are put in place. The aforementioned hidden debt deals undertaken by the Guebuza regime were likely associated with the decision to start developing Mozambique’s massive oil and gas reserves at a time when institutions that provide for various checks and balances were still weak. The creditworthiness of Mozambique stemming from the hype of the gas resources was enough to nourish fraudulent deals, and long before gas exploitation was to begin, speculation and rent seeking was rife. As history and experiences elsewhere in Africa and beyond document, benefiting from natural resources is a most challenging political exercise (see Roe Reference Roe2018). Meanwhile, on the economic front, Mozambique has started engaging with some of the world’s largest companies to find the best way of developing the oil and gas sector to the benefit of the country’s mostly poor people.

Finally, Mozambique has, like the rest of the world, struggled to address the impact of the COVID-19 pandemic. The first COVID-19 case was reported on 22 March 2020 and a state of emergency was declared on 31 March. The COVID-19 crisis and its associated economic effects came on top of the hidden debt scandal and the impacts of Cyclones Idai and Kenneth, highlighting that Mozambique remains highly vulnerable to climatic and policy-induced shocks.

A Structural Shifts

On the return of peace in the early 1990s, Mozambique was predominantly an agricultural economy. As shown in Figure 2.5, in 1996/97 more than 80 per cent of the workforce identified agriculture as their primary occupation, virtually all of which was smallholder or peasant agriculture, not larger-scale commercial activities.

Figure 2.5 Sectoral trends in employment (1997–2015)

Source: Authors’ estimates based on household survey data series 1996/97, 2002/03, 2008/09, and 2014/15 (see MEF/DEEF 2016); private services include finance; public services include health, education, and public administration.

At the same time, whilst important in terms of total production, agriculture has never represented more than half of aggregate real value-added. Figure 2.6 shows that even in the early 1990s, agriculture contributed only about 40 per cent to GDP, while service activities (public and private) contributed approximately another 40 per cent. As we discuss below, this clearly indicates that agriculture has been and remains a very low productivity activity on average.

Figure 2.6 Sectoral contributions to levels of real value added (1991–2019)

Note: Figure refers to the percentage contribution to production value-added in constant 2014 prices; public services include health, education, and public administration.

Source: Authors’ estimates based on unpublished data from National Institute for Statistics.

Over time, there has been a limited shift in the structure of the economy away from agriculture. According to the most recent household survey (2014/15), which constitutes the most reliable source of data on the labour force, around two-thirds of workers were primarily involved in agriculture and the sector contributed around one-quarter of total value-added. Labour movements out of agriculture have been largely towards the private services sector, including commerce and transport, which now represents 22 per cent of all employment. However, the value-added of private services has remained flat over time, indicating that new workers in this sector tend to be less productive than existing workers. Figure 2.7 clarifies this insight, showing labour productivity differentials across sectors relative to the economy-wide average. Labour productivity in agriculture has been about half (-.8 log points) that of the average worker, while the productivity of workers in private services declined over the two decades from 1991 to 2019 from nearly 5 times to just 1.5 times that of the average worker.

Figure 2.7 Labour productivity differences by sector (1997–2015)

Note: X-axis is the natural logarithm of the ratio of labour productivity of the indicated sector to aggregate labour productivity; thus, values below 0 show lower than average productivity; private services include finance.

Source: Authors’ estimates based on household survey data series 1996/97, 2002/03, 2008/09, and 2014/15 (see MEF/DEEF 2016).

As discussed in detail by Jones and Tarp (Reference Jones and Tarp2015), who provide a full decomposition of within- and between-sector productivity changes over the same period, the shift of workers has been largely out of very low-productivity agriculture to informal services activities. Thus, the contribution of the services sector to growth since 2003 has been dominated by a sectoral reallocation effect, as even informal services are more productive than agriculture, but not a within-sector productivity effect. In contrast, agriculture has shown fairly robust moderate within-sector productivity increases (over 1 percentage point per year), despite losing workers in relative terms.

The same figures also point to a ‘missing middle’. Looking over the full time period, secondary sectors (manufacturing, construction) have remained small, both in terms of their contribution to GDP and in terms of employment. For example, in both 1997 and 2015, just 3 per cent of workers were in the manufacturing sector, and the value-added of the sector was equal to 8 per cent of GDP at the start and 9 per cent at the end of the period. In contrast, mining (extractive industries) has achieved very rapid growth of output, expanding to 8 per cent of real GDP, but with almost no change in its employment share. We return to this point below.

B Changing Growth Drivers

The changing structure of the Mozambican economy, in terms of both output and employment, hints at how a series of different drivers of growth have been operative over the post-war period. Three main growth phases are apparent. First, as further supported by Figure 2.8, from the mid-1990s onwards, Mozambique successfully took advantage of the peace dividend, achieving post-conflict stabilisation and recovery (see Tarp et al. Reference Tarp2002). In this period, strong rates of growth across multiple sectors, including agriculture, reflected a combination of a return of displaced people to their homes, the rebuilding of private and public infrastructure, supported by foreign aid and private investment (domestic and foreign).

Figure 2.8 Sectoral contributions to real value-added growth, smoothed (1991–2019)

Note: Manufacturing includes construction and utilities; growth refers to production value-added in constant 2014 prices; non-linear five period filter used.

Source: Authors’ estimates based on unpublished data from National Institute for Statistics.

A second phase, starting around the turn of the millennium, is associated with strong growth in the manufacturing sector, dominated by large-scale capital-intensive investments, particularly in the Mozal aluminium smelter.Footnote ¹⁹ This period, which lasted until the late 2000s, also saw robust service sector growth, alongside a sustained and substantial contribution of agriculture (at 2 percentage points), or more than a quarter of aggregate growth per annum.

The late 2000s, however, saw a shift to a third phase, characterised by an emphasis on the natural resources sector (mining), driven by large increases in foreign direct investment (FDI) into coal and then natural gas field developments. This coincided with an expansion of the finance (banking sector), which increased in size from 2 per cent of GDP in the mid-2000s to over 6 per cent in the latest period (see Figure 2.6). The same period also saw a trend decline in the contribution of agriculture to growth, but strong private services growth.

To get a sense of the magnitude of the recent shift in the sectoral emphasis of growth towards extractives, which has also arguably become a focus of public policy, Table 2.1 provides a summary of changes in consumption and investment components (reported as shares of GDP), as well as sources of investment funds.

Overall, we note that Mozambique has always been a consumption-heavy economy. Total consumption has consistently equalled around 90 per cent of GDP and, as discussed below, the public sector has accounted for an increasing share of total consumption. This is not to say that investment has been low; on average it has been sustained at about 30 per cent of GDP, with public investment comprising roughly one-third of total investment.

Critically, investment has been financed largely from abroad, indicated by the very large volume of foreign savings (which is the mirror of the current account deficit). During the first period shown in the table (1999–2004), which roughly reflects the beginning of the second growth phase, private foreign investment and foreign aid to the public sector (grants and loans) amounted to nearly 20 per cent of GDP, both in equal measure. In the second half of the 2000s, foreign investments in the manufacturing sector tailed off but aid remained strong. In the last decade, however, FDI jumped to over 20 per cent of GDP as Mozambique opened up to investment in natural resources extraction, and these flows quickly came to overshadow the contribution of foreign aid, which has declined substantially in relative terms (see Figure 2.9).

Figure 2.9 ODA as percentage of GNI, 1990–2018

Source: Net ODA received (% of GNI) from World Bank (2020) data.

Three conclusions can be drawn. First, much of Mozambique’s growth has been fuelled by significant inflows of foreign exchange, both public and private, which have generated spillovers either directly into consumption or indirectly into income through investment. Second, however, the pattern of these inflows has shifted increasingly towards capital-intensive natural resource investments, which provides a proximate explanation for the widening labour productivity deficits across sectors (Figure 2.7) and which by definition also suggests that the growth process has lacked ‘inclusivity’. Third, Mozambique has not discovered a consistent (sustained) domestic engine of growth. Rather, not only have growth drivers shifted depending on the dominant source of foreign savings, but also the economy seems to have become less diversified over time. This is evident from Figures 2.10 and 2.11, which show the composition of exports and Mozambique’s export complexity, respectively. Figure 2.10 reveals that manufacturing exports (aluminium) plateaued by the later 2000s; but, starting roughly in 2010, natural resources exports (mainly coal) have quintupled in value and now account for around 50 per cent of all exports. Correspondingly, the relative shift towards lower value-added exports (which are lower in complexity) is evident in Figure 2.11.

Figure 2.10 Major exports in US$ millions (2000–2019)

Notes: Primary materials include energy exports; agricultural goods include fisheries.

Source: Authors’ estimates based on unpublished data from Central Bank of Mozambique.

Figure 2.11 Export complexity score (1991–2017)

Notes: Lower scores mean less complex exports.

Source: Smoothed estimates from The Growth Lab at Harvard University (2020) data.

These three conclusions are reinforced by trends in the final period, as shown in Table 2.1, which merit special attention. As discussed in Section II, in 2016 a series of hidden (illegal) debts taken out since 2013 from overseas commercial banks by publicly guaranteed enterprises came to light. The funds were premised on discoveries of large natural gas fields in the country, but (ex post) appear to have been squandered, various members of the ruling elite and overseas backers being implicated in corruption. Critically, this prompted a major freeze in foreign aid and, correspondingly, a sharp depreciation of the exchange rate. As plotted in Figure 2.8, growth rates simultaneously declined across multiple sectors and public sector investment approximately halved as a share of GDP (see Table 2.1) – that is the sudden lack of foreign inflows outside the natural resources sector, combined with significant macroeconomic uncertainty, had drastic growth consequences, revealing the fragile nature of the earlier growth and the economy’s dependence on foreign savings.

Another way of making sense of these trends is to note that Mozambique has been and remains a highly segmented economy. A large share of the population remains engaged in informal rural economic activities, predominantly smallholder agriculture. In comparative terms, productivity in these activities is low and is strongly associated with high rates of poverty (see below). The substantial FDI over the post-war period has focused on enclave-type industries, with few linkages via employment or provision of inputs to domestic productive sectors. Therefore, as Arndt et al. (Reference Arndt, Garcia, Tarp and Thurlow2012a) and others have formally quantified, inter-sectoral linkages in the economy have been weak and are highly compromised by poor infrastructure and large transport distances between areas of highest agricultural potential and population mass (rural centre and north) and areas of greatest effective demand (urban south). Furthermore, economy-wide benefits from natural resource development have proven elusive, especially as many investments have not reached the production (profitability) phase. In sum, the pattern of employment creation suggests that the infrastructure and business environment remain weak and skewed against labour-intensive manufacturing.

C Competent Macroeconomic Management

Despite the most recent period of crisis, the tail end of which has coincided with the COVID-19 pandemic, macroeconomic management has been competent. Broadly speaking, inflation has been well controlled and fiscal deficits have been almost fully covered by substantial foreign inflows (foreign aid) rather than domestic debt. Tax revenues have also steadily increased, from a low level, via a combination of sensible reforms and relatively rapid growth of specific formal segments of the economy (e.g., finance, natural resources). Table 2.2 supports this narrative, summarising a range of macroeconomic and public expenditure figures for selected years since 2002.Footnote ²⁰

Until around 2015, the combination of increased tax revenue and sustained (although relatively smaller) foreign aid allowed both government consumption and investment to expand. However, most recently we observe that consumption spending now accounts for almost 25 per cent of GDP, while public investment spending has fallen back below 10 per cent of GDP. This is explained by large cuts in foreign aid in reaction to the hidden debts and the ensuing rapid build-up of public debt, external and internal, entailing much higher recurrent servicing costs.

Thus, as compared with the mid- to late-2000s, when Mozambique benefitted from highly favourable debt write-offs, the government is now in a considerably more fragile fiscal position, with debt at distressed levels. Furthermore, the build-up of public debt has not been matched by any boost in investment spending that could be expected to yield future dividends.

D Steady but Slowing Social Progress

Table 2.2 confirms that the real pace of growth on a per capita basis is hardly phenomenal. Since 2000, the average real per capita growth rate equals just 3.4 per cent per annum, implying that it will require more than twenty years to double real incomes. Thus, Mozambique remains a very poor country by any measure. In 2017, Mozambique ranked 180 of 186 countries in terms of real GDP per capita (IMF 2018); and 180 of 189 on the UN’s human development index (UNDP 2018: 22–25).

The post-war period has shown progress on a range of social indicators, including consumption and multidimensional poverty. Data from a series of household surveys, undertaken approximately every five years since 1996, show particularly strong gains in poverty reduction in the immediate post-conflict period, but less impressive gains since then (see Table 2.3).

This is also consistent with the narrative, outlined above, of a weak connection between specific sources of aggregate demand (sectoral) growth and employment. Over the entire period, we estimate that the consumption poverty headcount fell by 2.2 per cent per year, well below the aggregate rate of growth. In turn, inequality has increased, again suggesting that growth has not been broadly shared,Footnote ²¹ and this tendency has worsened since 2015 (Egger et al. Reference Egger, Salvucci and Tarp2020).

Table 2.4 summarises progress in a range of other socio-economic dimensions, including household ownership of assets or capital (economic and human) (see MEF/DEEF 2016).

These are all dummy variables that take a value of one if the household head is literate, has access to electricity, has access to clean water, has clean sanitation, has a robust roof, owns some means of transport, has access to information technology (e.g., telephone), and owns some durable goods (e.g., bed). Throughout the period, all indicators have shown progress. Even so, absolute deprivations are widespread – only about half of adults are literate, half of all households have access to clean water, and less than a third have adequate sanitation. So, despite twenty-five years of steady growth in the post-war period, plus sustained access to large volumes of foreign savings, a lot remains to be done to ensure that all Mozambicans enjoy a decent standard of living.

E Public Sector (in)Efficiency

The foregoing raises the important question: how efficient or cost-effective is the public sector in delivering core public services? As already noted, the relative size of the public sector in the economy has grown substantially. Table 2.1, for instance, shows that total government current expenditures have increased by almost 10 percentage points in relation to GDP, implying far more rapid expansion than in the private sector as a whole. However, according to Figures 2.5 and 2.6, the employment share of the public sector has increased only moderately, while the real value-added of public services (which includes privately provided health and education) has been almost flat over time. In this sense, there is no clear indication (in aggregate statistics) of any significant real economic impacts of the increased spending.

Table 2.5 compares budget allocations to outcomes in education and health for a range of sub-Saharan African countries. It shows that on a per capita basis, social sector spending in Mozambique is comparable to elsewhere (in the middle). Yet outcomes in Mozambique are considerably poorer – for example while more than 90 per cent of children attend primary school, the primary school completion rate is only around 40 per cent versus 80 per cent in neighbouring Tanzania, and under 5 mortality is 74.2 per 1,000 live births versus 53.0 in Tanzania.

Of course, Mozambique’s troubled history explains part of this gap, but other sub-Saharan African countries, such as Ethiopia, also entered the 1990s in dire circumstances and would seem to have made stronger progress and at a lower fiscal cost. Additionally, evidence in Mozambique points to extreme and persistent spatial disparities in access to public services (see Egger et al. Reference Egger, Salvucci and Tarp2020). The more urban south of the country tends to enjoy not just better quality services but also much higher effective public spending than elsewhere. Thus, overall, the public sector would seem to be not cost-effective and marred by inequities in allocations that fail to benefit the poorest.

F Demographic Dynamics

According to INE (2019), summarising the results from the 2017 Population Census, 39 per cent of the rapidly growing Mozambican population lives in Nampula and Zambézia as compared to 21.5 per cent for the provinces in the south. The remaining almost 40 per cent live in the provinces of Tete, Cabo Delgado, Sofala, Manica, and Niassa. The distribution between urban and rural areas shows that two thirds live in rural areas. The overall annual rate of growth of the population between 2007 and 2017 was 2.8 per cent, increasing from 1.7 to 2.5 per cent in the previous decades. Maputo province saw the largest relative increase in its population (60.7 per cent) between 2007 and 2017 while the population of the city of Maputo stagnated. Large increases of more than 40 per cent were experienced in Niassa, Tete, and Nampula, whereas increases of around 30 per cent were experienced in Manica, Sofala, Zambézia, and Cabo Delgado.

Reflecting the steady but slowing social progress discussed in Section D, life expectancy increased from 50.9 years in 2007 to 53.7 years in 2017. However, the population dynamics imply high dependency ratios that will continue at burdensome levels for well more than a generation (Jones and Tarp Reference Arndt, Hussain, Jones, Nhate, Tarp and Thurlow2012, Reference Jones and Tarp2013, Reference Jones and Tarp2016) as nearly half of the population is under fifteen years of age (INE 2019). A critical implication is that the working age population will grow rapidly over the next thirty years. This dynamic is essentially ‘locked in’ as it reflects past fertility decisions. Some 34 per cent of the total labour force is concentrated in the north, 41.9 per cent in the centre and 24.1 per cent in the south (MITSS, 2020), reflecting the regional distribution of the general population.

The majority (60.8 per cent) of the young labour force (15–34 years) is between fifteen and twenty-four years old and, according to MITSS (2020) based on the IV General Population Census 2017, it is estimated that between 2020 and 2024, the labour force will increase by 2.26 million, including 1.62 million young people (15–34 years). How can the labour market absorb this magnitude of new entrants? While it is impossible to make firm predictions, there is an internal logic to what is feasible, even in the best of policy environments. This derives from the existing structure of employment – sectors with large existing employment shares must be the principal sources of ‘new’ jobs. Put differently, even if sectors with small employment shares were to grow at an annual rate of 10 per cent or even 15 per cent, this would be insufficient to absorb all new entrants (Jones and Tarp Reference Jones and Tarp2012, Reference Jones and Tarp2013, Reference Jones and Tarp2016).

Informal non-farm enterprises, the second largest category of jobs, are both substantially smaller than agriculture and much bigger than formal employment. The latter constitutes a minor share – less than 15 per cent of all workers report receiving a wage, of which the large majority is in informal employment. This share configuration dictates that most new entrants into the labour market must seek work in agriculture or informal enterprises. This has very much been the situation for the past twenty-five years and will remain the case in the near future. Indeed, under a high-growth scenario of 8 per cent growth in formal sector jobs, the informal sector will continue to increase in absolute size until at least 2030 (Jones and Tarp Reference Jones and Tarp2012, Reference Jones and Tarp2013).

These facts inescapably imply that future dynamics in smallholder agriculture and the informal sector will be of fundamental importance to achieving inclusive growth. This is the core policy challenge facing the Mozambican government and its people over the next decades.

G Taking Stock

The evolving pattern and complexity of Mozambique’s growth trajectory, as well as the sheer scope of the challenges the country continues to face, limit the value of attempting to reduce this experience to a few well-defined underlying causes. We would argue that many of the conventional factors identified as constraints to growth may well be symptomatic of a deeper malaise, namely the lack of a clear (vision for a) domestic engine of growth and an excess reliance on external savings to finance demand.

While no formal comprehensive growth diagnostic in the spirit of Hausmann et al. (Reference Hausmann, Rodrik and Velasco2005) has been undertaken for Mozambique to date, a range of less formal ‘diagnostics’ point to weaknesses across multiple domains. These include low private returns to economic activity as well as the high costs of finance (the two distinct primary branches of the growth diagnostics framework). These exercises frequently end up documenting a long list of symptoms and do not identify specific ‘deep’ causes or reform priorities.

Lledó (Reference Lledó, Clément and Peiris2007), for instance, notes that in both 2002 and 2006 firms identified the high cost of finance as a primary concern. However, on top of this, he notes that median returns to manufacturing firms are near zero, which reflects the ‘combined impact of low social returns and limited private appropriability’ (p. 344). In a more recent exercise, the World Bank (2016) echoes a similarly wide range of key challenges: ‘Inadequate physical and logistical infrastructure, excessive bureaucracy, weak public institutions, credit constraints and a complicated land-tenure system continue to discourage investment and narrow the range of economic opportunities available to domestic firms’ (p. 35).

The point is not that these diagnostics are necessarily wrong. Rather, it is that it is difficult to isolate unique growth constraints/opportunities or untangle causes from symptoms given the range of challenges and the segmented nature of the economy.

In the same spirit, it can be highlighted that previous diagnostic exercises for Mozambique have placed wildly different interpretations on the direction of public reform efforts. Ten years ago, improvements in the business environment were seen as impressive and indicative of a strong and coordinated government commitment:

Nonetheless, a 2017 diagnostic of the business environment suggests that these reforms have been cosmetic, that there is ‘a lack of political will to implement reforms because the status quo benefits politicians and the politically-connected [… and] business registration reform […] has resulted in little improvement of the process for entrepreneurs’ (Franco and Katiyo Reference Franco and Katiyo2017: 19–20).

Put differently, views change as to where priorities lie and where genuine progress has been made. And, as yet, there does not seem to be a strong politically independent domestic constituency for deeper reforms.

In our view, what this amounts to is not a specific set of distinct technical growth constraints per se but a more general failure to prioritise the upgrading of domestic competencies in either the public or the private sector beyond that of servicing international demands. A vision of a self-sustaining process of growth and development has been absent, and in its place, a fragmented and shifting external orientation has dominated, most recently degenerating into growth-damaging rent-seeking. Many of the growth challenges represent symptoms of this failed vision and corresponding growth engine.

Looking ahead, a key risk is that this lacuna will only expand as natural resource extraction (natural gas) plays a stronger role, cementing rentier elements in the public and private sectors and undermining competitiveness elsewhere in the economy. In the short-to-medium term, significant public finance gains from these industries are likely to be lacking, while exposure to commodity price volatility is likely to become more acute.

Avoiding an Angola-style scenario of deepening inequality and a bloated public sector that fails to deliver quality public services must be the priority. And this is especially important in Mozambique given political and spatial tensions that have already boiled over into military conflict, as discussed in Section II. Sadly, with public investment weakening and high debt obligations, the capacity of the government to support inclusive structural transformation has weakened since the late 2000s. Thus, decisive action must be taken.

A Overview

This section presents selected trends of the performance of Mozambique in different institutional indicators in comparison with two groups of selected countries: neighbouring – Tanzania, Malawi, and Zambia – and peer – Uganda, Ethiopia, Vietnam, and Lao People’s Democratic Republic (PDR). The criteria used in selecting the comparator countries were, on the one hand, geographic and, on the other hand, level of income and similarity in terms of historical and economic characteristics, either past or present.

The discussion of the indicators is not exhaustive, and it is important to keep in mind the limitations of some of these measures (e.g., see Kaufmann and Kraay Reference Kaufmann and Kraay2007; Gisselquist Reference Gisselquist2014; González et al. Reference González, Fleischer and Mira d’Ercole2017). However, they were selected based on data availability and relevance for the present study. The different indicators were obtained from a variety of well-established data sources, namely, Varieties of Democracy (Coppedge et al. Reference Coppedge, Gerring, Lindberg, Skaaning, Teorell, Altman, Andersson, Bernhard, Fish, Glynn, Hicken, Knutsen, Marquandt, McMann, Mechkova, Paxton, Pemstein, Saxer, Seim, Sigman and Staton2020; Pemstein et al. Reference Pemstein, Marquardt, Tzelgov, Wang, Medzihorsky, Krusell, Miri and von Römer2020), Afrobarometer (2020), World Economic Forum (2020), Worldwide Governance Indicators (WGIs; World Bank 2020), and Bertelsmann Stiftung (BTI 2020).Footnote ¹

First, we provide an overview of the scores for Mozambique in the six dimensions of governance suggested by the WGIs – voice and accountability, political violence, government effectiveness, regulatory quality, rule of law, and control of corruption – in comparison to neighbouring and peer countries in 2005 and in 2018 (Figure 3.1).

Figure 3.1 Worldwide Governance Indicators, 2005 and 2018

Note: Centre is at –2.5; lines further away from the centre correspond to better outcomes. The scores range from –2.5 to 2.5, with higher values representing better outcomes. Key: Voice and account, voice and accountability; Pol. violence, political violence; Gov. effect., government effectiveness; Reg. qual., regulatory quality; Rule law, rule of law; Control corr., control of corruption.

Source: Authors’ compilation based on Worldwide Governance Indicators (World Bank 2020).

In 2005, Mozambique was the best performer in terms of voice and accountability and control of corruption in both groups of countries. The graph for neighbouring countries shows similar performances for the countries considered, but there are noticeable differences when comparing Mozambique with its peers. While it has higher scores than Ethiopia and Lao PDR in all dimensions, Mozambique’s score in regulatory quality, for instance, is lower than those of Uganda and Vietnam.

It is tempting to compare the same indicators over time for 2005 and 2018 to get a sense of the institutional dynamics in specific countries. However, the norm of indicators in the WGI database changes from year-to-year. Bearing this in mind, the main point emerging from the graphs is that while Mozambique was doing better than the other countries in 2005, it did worse in 2018. Even though Mozambique remains the best performer in voice and accountability among the peer countries, its scores are lower than all the other four countries in terms of government effectiveness and rule of law. We return to some of these indicators in what follows.

B Rule of Law and Judicial Independence

The first institutional dimensions we address here relate to confidence in, and abidance by, the known rules to government actors and citizens as well as to the independence of the judicial system. In Figure 3.1, we observed that while in 2005 Mozambique’s score in rule of law was already low in comparison to its neighbouring countries, this positioning becomes more apparent in 2018 (even when considering confidence intervals). Compared with its peer countries, Mozambique’s score is only lower than that of Vietnam and similar to that of Uganda in 2005, but it is, together with Lao PDR, one of the lowest among the five countries in 2018.Footnote ²

Figure 3.2 represents the perception of respondents of how independent the judicial system is from influences of the government, individuals, or companies. Both graphs in the figure show that the level of independence of the judicial system is perceived as low in Mozambique, the lowest in comparison to the selected countries and across the period. It is also noticeable that there was a decrease from 2008 and a sharp fall after 2017.Footnote ³

Figure 3.2 Judicial independence, 2007–2019

Note: The scores range from 1 to 7, with 1 being ‘not independent at all’ and 7 ‘entirely independenť.

Source: Authors’ compilation based on the Executive Opinion Survey (World Economic Forum 2020).

C Voice, Participation, and Political Accountability

The second group of institutional dimensions recognises the need for citizens to be able to participate politically and hold the executive accountable, as well as the importance of freedom of expression and assembly. The scores for voice and accountability from the WGIs represented in Figure 3.1 pointed to the conclusion that the performance of Mozambique in this dimension seems to have weakened comparing the situation in 2005 with that in 2018, especially in comparison to neighbouring countries.

Figure 3.3 represents the overall level of political participation (represented at the top of the pentagon), which measures the extent to which the populace decides who rules and has other political freedoms. The level of political participation is derived from the BTI scores for the remaining four sub-components represented in the pentagon (clockwise), namely, free and fair elections, effective power to govern, association/assembly rights, and freedom of expression. The graphs for 2006 show that Mozambique was one of the best performers that year among the selected countries in all political participation dimensions. In contrast, according to the assessment in 2020, Mozambique had the worst performance on the overall indicator of political participation among its neighbouring countries, while its peers show worse scores in all sub-components except freedom of expression and assembly rights.

Figure 3.3 Political participation and sub-components, 2006 and 2020

Note: Centre is at 1; lines further away from the centre correspond to better outcomes. The scale ranges from 10 (best) to 1 (worst). Key: Pol. participation, overall political participation score; Elections, free and fair elections; Eff. power, effective power to govern; Ass. rights, association/assembly rights; Free. expression, freedom of expression.

Source: Authors’ compilation based on the transformation index of the Bertelsmann Stiftung (BTI 2020).

D Political Instability, Violence, and State Legitimacy

The third set of dimensions refers to the degree of recognition of the nation as a state, with adequate and differentiated power structures at national and subnational levels, and the likelihood of political instability and of politically motivated violence and terrorism. The scores on political stability and absence of violence motivated by political reasons, including terrorism – represented in Figure 3.1 as ‘Pol. violence’ – show that Mozambique’s relative position compared with neighbouring and peer countries deteriorated over time. It is interesting to note that while Mozambique’s score was similar to that of Malawi and Zambia in 2005, it was lower in 2018 (with more certainty in the case of Zambia).

Figure 3.4 represents a measure of political instability, defined as ‘stateness’, as well as the sub-components used to derive it: monopoly of the use of force, state identity, no interference of religious dogmas, and basic administration. In 2006, Mozambique’s performance was similar to that of its neighbours, and it was one of the best among its peer countries. Still, Mozambique’s score was relatively much weaker in 2020 compared with the comparator countries, especially concerning the monopoly of the use of force.

Figure 3.4 Stateness and sub-components, 2006 and 2018

Note: Centre is at 1; lines further away from the centre correspond to better outcomes. The scale ranges from 10 (best) to 1 (worst). Key: Stateness, overall score; Mon. use force, monopoly on the use of force; No rel. dogmas, no interference of religious dogmas; Basic admin., basic administration.

Source: Authors’ compilation based on the transformation index of the Bertelsmann Stiftung (BTI 2020).

E State Capacity and Autonomy from Private Interests

The fourth dimensions deal with the capacity of the state to fulfil the social contract and the separation of state power from private interests. Figure 3.1 included the scores for government effectiveness, a measure of perceptions about the quality of public services, civil service, and policy formulation and implementation, as well as the extent to which there is independence from political pressures. The scores for that indicator show that the relative position of Mozambique was quite good in 2005, but depreciated in 2018.

Figure 3.5 shows the level of pervasiveness of political corruption where a high score reflects a high level of corruption. The figure shows that Mozambique’s score remained consistently close to that of neighbouring Tanzania (up to 2016), below that of Malawi and above that of Zambia. Compared with its peer countries, the score for Mozambique stayed similar to that of Vietnam and Ethiopia (until 2017) and was lower than the score for Lao PDR and Uganda. Moreover, the hidden debt scandal is reflected in an increase in 2017.

Figure 3.5 Political corruption index, 2005–2019

Note: The scale ranges between 0 and 1; the index runs from less corrupt to more corrupt.

Source: Authors’ compilation based on Varieties of Democracy data (Coppedge et al. Reference Coppedge, Gerring, Lindberg, Skaaning, Teorell, Altman, Andersson, Bernhard, Fish, Glynn, Hicken, Knutsen, Marquandt, McMann, Mechkova, Paxton, Pemstein, Saxer, Seim, Sigman and Staton2020; Pemstein et al. Reference Pemstein, Marquardt, Tzelgov, Wang, Medzihorsky, Krusell, Miri and von Römer2020).

Figure 3.6 complements this analysis by asking how much respondents trust one element of the public service, the police. We use data from the two most recent rounds of the Afrobarometer (2020). In 2016, among the selected African countries, the responses for Mozambique show the highest percentage of respondents saying that they do not trust the police at all, and one of the lowest percentages of those who say they trust the police a lot. However, in the latest round the highest percentage corresponds to trusting the police a lot, whereas the response ‘Not at all’ received a lower percentage of responses.

Figure 3.6 Trust in police, Rounds 6 and 7 of the Afrobarometer

Source: Authors’ compilation based on Afrobarometer Rounds 6 and 7 (Afrobarometer 2020).

F Sovereignty and Independence

The final dimension considered relates to external factors and the degree to which Mozambique has a sovereign position in an international context and whether the political leadership in Mozambique is willing and able to cooperate with external supporters and organisations. Figure 3.7 represents BTI data on international cooperation (the top of the diagram) and its different sub-components, namely, effective use of support, credibility, and regional cooperation (represented clockwise). We note that while Mozambique did well in both 2006 and 2020 in terms of regional cooperation compared with all other countries considered, it is clear that the credibility and effective use of support scores, which were both at the top in 2006, were lower than everywhere else in 2020.Footnote ⁴

Figure 3.7 International cooperation, 2006 and 2020

Note: Centre is at 1; lines further away from the centre correspond to better outcomes. The scale ranges from 10 (best) to 1 (worst). Key: Int. coop., overall score; Use sup., effective use of support, which represents the extent to which the political leadership uses the support of international partners to implement a long-term strategy of development; Credibility, represents the extent to which the government acts as a credible and reliable partner in its relations with the international community; Reg. coop., regional cooperation, which represents the extent to which the political leadership is willing and able to cooperate with neighbouring countries.

Source: Authors’ compilation based on the transformation index of the Bertelsmann Stiftung (BTI 2020).

A Design, Implementation, and Sample of Respondents

B Results

1 The Most Important Constraints to Economic Development

This section describes the results obtained when we asked respondents to select and order the biggest constraints to economic development in Mozambique out of the fifteen areas referred in Section I. Figure 3.8 shows in part (a) the number of respondents that selected each of the fifteen areas, without considering their ordering in terms of perceived importance. The areas that were chosen more often among the respondents were human capital and poverty and inequality, followed by management of public administration and common vision of national strategy. Among the areas that were not identified as predominantly as main constraints, one finds foreign aid, gender equality, and autonomy in relation to the outside, followed by regulatory quality.

Figure 3.8 Choice of main constraints. (a) All ranking positions, number of occurrences. (b) Ranking = 1, number of occurrences

Source: Authors’ calculations based on quantitative survey.

This is confirmed when one looks at the areas that were selected by the respondents as the main constraint to development; that is, areas that received a ranking of one more often (Figure 3.8, part (b)). Common vision of national strategy was chosen more often than any of the other areas, in all likelihood reflecting the frequent change of course direction in economic strategy and policy, followed by poverty and inequality and human capital as reasons for the economy and society not working well. Insights from qualitative interviews (described in more detail in Section IV) confirm that Mozambique requires a coherent vision and plan to adapt to globalisation and that promoting equality is essential to maintaining a sovereign and united nation. Increasing inequality is likely to undermine social cohesion and stability. In contrast, and maybe somewhat surprisingly, the business environment did not come out at the very top as the main constraint to economic development in Mozambique, and, in line with Figure 3.8 part (a), only a few respondents selected autonomy in relation to the exterior, gender equality, and foreign aid as the main constraint.

Table 3.4 shows the number of occurrences for each area in each position of the ranking. Management of public administration, management of macroeconomic and sectoral policy, and human capital received the highest number of rankings of the second and third most important constraints. In fourth position, most of the respondents selected agriculture and access and use of land, management of public administration once more, and functioning of the legal sector. The last was also one of the most frequently chosen areas in fifth position of the ranking, together with business environment and management of natural resources. This likely reflects that the management of natural resources is understood as a public management/macroeconomic issue and that the constraints associated with the business environment are perceived as less important than agricultural development at this stage and often seen as reflecting the existing lack of human resources.

Table 3.4 Overview of ranking distribution

Areas	Number of occurrences
	1st	2nd	3rd	4th	5th	Total
Functioning of the legal sector	5	8	8	18	15	54
Decentralization of public power	8	16	10	8	10	52
Political participation	5	5	10	13	7	40
Common vision of national strategy	38	12	6	8	7	71
Agriculture and access to and use of land	8	6	13	20	13	60
Management of public administration	13	22	15	14	11	75
Management of macroeconomic and sectoral policy	7	18	18	13	10	66
Management of natural resources	5	10	12	13	15	55
Business environment	0	6	9	7	16	38
Regulatory quality	3	2	8	5	4	22
Human capital	24	22	18	13	7	84
Poverty and inequality	29	14	13	11	13	80
Gender equality	1	4	2	2	7	16
Foreign aid	2	1	3	2	7	15
Autonomy in relation to the exterior	1	3	4	2	7	17

Note: The three highest numbers of occurrences in each ranking position are highlighted in bold.

Source: Authors’ calculations based on quantitative survey.

When considering the difference in the choices by gender (Figure 3.9), one observes that female respondents (thirty-two respondents) have selected the management of macroeconomic and sectoral policy as one of the five main constraints more often than any of the other areas (Figure 3.9, part (a)). Still, in line with the male respondents, human capital and poverty and inequality were selected very often. In terms of the area selected as the main constraint (Figure 3.9, part (b)), the opinions are similar between male and female respondents, with the lack of a common vision of national strategy being identified as the main constraint more often than the other areas.

Figure 3.9 Choice of main constraints by gender (a) All ranking positions, number of occurrences (b) Ranking = 1, number of occurrences

Source: Authors’ calculations based on quantitative survey.

In Figures 3.10–3.12, we consider the responses given by respondents affiliated with business, academia, and public administration separately.

Figure 3.10 Choice of main constraints from respondents affiliated with the business sector (a) All ranking positions, number of occurrences (b) Ranking = 1, number of occurrences

Source: Authors’ calculations based on quantitative survey.

Human capital is the constraint that was chosen more often among respondents linked to the business sector, followed by management of macroeconomic and sectoral policy (Figure 3.10, part (a)). When looking at the areas ranked as number one in terms of their importance, both common vision of national strategy and human capital were identified more frequently as the most important constraints (Figure 3.10, part (b)).

Common vision of national strategy is again one of the most selected constraints when we look at the answers from respondents affiliated with academia, followed by management of public administration, human capital, and poverty and inequality (Figure 3.11, part (a)). The last two categories together with common vision of national strategy are also the most commonly ranked as number one in terms of their importance (Figure 3.11, part (b)).

Figure 3.11 Choice of main constraints from respondents affiliated with academia (a) All ranking positions, number of occurrences (b) Ranking = 1, number of occurrences

Source: Authors’ calculations based on quantitative survey.

These same three dimensions – common vision of national strategy, human capital, and poverty and inequality – were chosen as important constraints by respondents affiliated with public administration, in both the overall number of occurrences (Figure 3.12, part (a)) and the number of times they were selected as the main constraint (Figure 3.12, part (b)).

Figure 3.12 Choice of main constraints from respondents affiliated with the public administration (a) All ranking positions, number of occurrences (b) Ranking = 1, number of occurrences

Source: Authors’ calculations based on quantitative survey.

2 Perceptions of the Quality of Institutions

We now discuss the perceptions of the quality of institutions according to the answers to the 136 statements referred to in Section A. As mentioned, they were grouped in eighteen thematic areas and all respondents expressed their degree of agreement with all of the statements. Figure 3.13 shows the distribution of statements across these eighteen areas.

Figure 3.13 Distribution of statements by institutional area

Source: Authors’ calculations based on quantitative survey.

Before highlighting the main insights from the analysis of the answers, we did a check on the number of respondents that selected ‘I don’t know’ in each statement to identify potential statements that received a lower rate of response. This could be because they were more specific and perhaps required a more in-depth knowledge about the area or because there was some lack of clarity in the formulation. The highest number of ‘I don’t know’ responses (thirty-three responses) referred to the statement ‘Auditing of formal firms results in tax adjustments when appropriate’. Three of the other statements with a number of ‘I don’t know’ responses higher than fifteen also belonged to the group of questions on regulatory quality and required a more detailed knowledge of this area. This might be related also to the fact that regulatory quality was among the areas that were not chosen that often as constraints to economic development. The only other statements with a number of ‘I don’t know’ responses greater than fifteen (eighteen and seventeen responses, respectively) were ‘Foreign companies frequently receive support from the state’ and ‘The unofficial costs (e.g., bribes) of starting a business are high’.

The overall results show an average response of 2.66 across all the statements. This is below the middle level of the scale of score 3, corresponding to ‘Don’t agree or disagree’, suggesting that respondents tend to disagree with positive statements about Mozambican institutions, in line with insights from the qualitative interviews of KIs. Figure 3.14 represents the full distribution of average scores across the statements.

Figure 3.14 Distribution of average scores across statements, according to the Likert scores

Source: Authors’ calculations based on quantitative survey.

We now describe in more detail the relative frequency of statements within the institutional areas considering their average score. The first group of bars in Figure 3.15 shows the percentage of statements with a score below 2.5 that are related to each institutional area out of all the statements with an average score of 2.5. The second group of bars results from applying the same analysis, but to the statements with an average score above 3. In general, one observes that the opinions regarding the different statements across the institutional areas do not tend to be unanimous, with rather low percentage values even in the areas with higher frequencies.

Figure 3.15 Frequency of statements under the institutional areas by average scores

Source: Authors’ calculations based on quantitative survey.

In terms of the areas where opinions seem to gather one of the more extreme scores, one can mention, for example, legal and constitutional matters, with 11.1 per cent of the 54 questions with a score below 2.5 and none of the 39 statements with a score above 3 being related to this institutional area. In contrast, 17.95 per cent of the statements with a score above 3 are related to foreign aid and only 5.56 per cent of the statements with a score below 2.5 are related to this institutional area.

The institutional areas with a higher frequency of statements with an average score below 2.5 are legal and constitutional matters, freedom and political participation, and international collaboration and autonomy. These are closely followed by public protection, land, public goods and services, and business environment. Freedom and political participation seem to divide the opinion of respondents, given that there is a high frequency of the statements linked to this institutional area that have an average score higher than 3. Also showing high frequencies for statements with average scores higher than 3 are state accounts and statistics, and foreign aid, as already mentioned.

Despite these institutional areas being more disaggregated than the institutional constraints analysed in the previous sections, one notices some consistency in the results. Issues related to public protection and to the provision of public goods and services can be linked to problems of poverty and inequality and human capital. A link can also be established between a lack of common vision of national strategy and weaknesses related to legal and constitutional matters, as well as international collaboration and autonomy.

Next, we focus on the statements that have the lowest and the highest average scores in order to identify what are perceived to be the main weaknesses and strengths, respectively. One should bear in mind that when statements were phrased as negative for the economic development of a country, the Likert scale was inverted so that higher scores reflected better functioning institutions.

Box 3.1 lists the top fifteen major institutional weaknesses according to the lowest average scores. They cover a broad range of institutional areas.

Box 3.1 Top fifteen major institutional weaknesses

• 98. Corruption distorts the business environment. (Yes; 1.66; 99%)

• 101. Corruption in public administration is prevalent. (Yes; 1.72; 100%)

• 74. Public services are of an adequate quality. (1.80; 99%)

• 37. The economic and political influence of the elite class in Mozambique is very strong. (Yes; 1.81; 97%)

• 76. Coverage of public goods and services is unequal across the country. (Yes; 1.83; 99%)

• 125. Donors and international organisations are influential with respect to national economic policy. (Yes; 1.86; 99%)

• 67. Land related conflicts are frequent. (Yes; 1.88; 99%)

• 3. The judiciary is independent with respect to the state. (1.89; 99%)

• 1. All citizens are treated equally before the law. (1.90; 99%)

• 126. Aid is a major contributor to the national budget. (Yes; 1.92; 99%)

• 2. There is a clear separation of judicial, executive, and legislative powers. (1.95; 99%)

• 79. Public procurement of goods and services is transparent. (1.96; 98%)

• 58. Theft is widespread in Mozambique. (Yes; 1.99; 99%)

• 118. In the past, the country has been affected negatively by political events in neighbouring countries. (Yes; 2.00; 99%)

• 61. The use of child (below fifteen years old) labour is common in rural areas. (Yes; 2.02; 99%)

Note: The parentheses indicate (i) ‘Yes’ if the scale was inverted because the statement was phrased in the negative; (ii) the average score obtained according to the following Likert scale: 1 = ‘Completely disagree’, 2 = ‘Disagree’, 3 = ‘Don’t agree or disagree’, 4 = ‘Agree’, and 5 = ‘Completely agree’; (iii) the share of the sample giving an opinion on the statement (in percentage).

Source: Authors’ calculations based on quantitative survey.

• Corruption seems to affect both the business environment and public administration.

• Three of the statements relate to legal and constitutional matters, highlighting a lack of judicial independence and no separation of judicial, executive, and legislative powers, as well as inequality in the treatment of citizens before the law.

• On average, the respondents agree that there is a very strong economic and political influence of the elite class in Mozambique.

• The scores also indicate some weakness in the provision of public goods and services, which are not of adequate quality and not equally distributed across the country.

• Additionally, respondents agree, on average, that there are frequent conflicts related to land issues.

• There is some suggestion of weak public protection, with the scores pointing to theft being widespread in Mozambique and child labour being common in rural areas.

• Finally, the scores highlight the negative influence of political events in neighbouring countries in the past as well as aid dependence and influence of donors and international organisations in national economic policy.

We now turn to the major institutional strengths, identified based on the highest average scores, which mean agreement with statements that account for positive aspects and disagreement with statements that refer to negative aspects.

Box 3.2 represents the fifteen statements with the highest average scores. Again, they cover a wide range of aspects.

Box 3.2 Top fifteen major institutional strengths

• 27. The media is politically pluralist, representing a diversity of viewpoints. (3.25; 98%)

• 13. The elected political authorities have control over the public bureaucracy. (3.25; 98%)

• 47. Discrimination and/or segregation by the society on the grounds of religion is prevalent. (Yes; 3.28; 99%)

• 35. Economic policy is actively debated within civil society. (3.29; 98%)

• 131. Aid significantly reduces the accountability of government. (Yes; 3.29; 94%)

• 39. There is a strong sense of national identity in Mozambique. (3.30; 98%)

• 130. Aid significantly improves the quality of governance institutions. (3.37; 97%)

• 129. Aid significantly improves the quality of economic policy. (3.38; 98%)

• 51. Traditional solidarity links (e.g., through family, neighbours, associations, religious groups, etc.) are effective in providing support to those in need in rural areas. (3.47; 94%)

• 19. Civil society participates in politics, for example, through official commissions, opinion polls, public debates, and op-eds. (3.51; 100%)

• 88. The state has autonomy in determining and implementing socio-economic policy and reforms independent of religion and traditional norms and values. (3.64; 99%)

• 120. There is a significant degree of collaboration with neighbouring countries. (3.75; 99%)

• 25. People are free to form associations to collectively express, promote, pursue, and defend common interests (e.g., religious, ethnic, occupational, political). (3.78; 99%)

• 127. Aid significantly improves infrastructure development. (3.83; 99%)

• 128. Aid significantly improves health and education. (3.95; 99%)

Note: The parentheses indicate (i) ‘Yes’ if the scale was inverted because the statement was phrased in the negative; (ii) the average score obtained according to the following Likert scale: 1 = ‘Completely disagree’, 2 = ‘Disagree’, 3 = ‘Don’t agree or disagree’, 4 = ‘Agree’, and 5 = ‘Completely agree’; (iii) the share of the sample giving an opinion on the statement (in percentage).

Source: Authors’ calculations based on quantitative survey.

• It is striking to observe that there is agreement in the perception of foreign aid as a strength in Mozambique, given that five of the statements refer to it.

• Three of the statements refer to freedom and political participation. They suggest that people are free to form associations, the media is politically pluralist, and the civil society is able to participate in politics. Related to the latter is the fact that economic policy is actively debated within civil society.

• They also highlight that there is no discrimination and/or segregation on the basis of religion and that traditional solidarity links provide effective support in rural areas.

• Two of the statements relate to public power and autonomy, suggesting that there is political control over the public bureaucracy in Mozambique and that the state is autonomous from religion and traditional norms and values in the formulation and implementation of socio-economic policy and reforms.

• Finally, there is agreement with a strong sense of national identity in Mozambique, together with a significant degree of cooperation with neighbouring countries.

Lastly, we explore the heterogeneity in respondents’ answers according to their gender by listing, in Table 3.5, the nine statements with the largest statistical difference (i.e. largest t-statistic derived from two-sample t-tests of mean comparisons) between female and male respondents. The main differences in the perceptions of the two groups are related to the independence of the Central Bank in managing the monetary policy and the social support network in rural areas. On average, men tend to agree that aid increases corruption to a stronger degree than women do. The remaining differences are not as big in terms of absolute values, despite being statistically significant.

Table 3.5 Top issues with significant differences between men and women

Statement	Average score female	Average score male	t-statistic	p-value
110. The Central Bank is managing monetary policy independent of government.	3.55	2.96	2.72	0.007
51. Traditional solidarity links (e.g., through family, neighbours, associations, religious groups, etc.) are effective in providing support to those in need in rural areas.	3.87	3.36	2.68	0.008
132. Aid significantly increases corruption.	2.56	3.06	–2.49	0.014
47. Discrimination and/or segregation by the society on the grounds of religion is prevalent.	2.41	2.80	–2.24	0.027
20. Political corruption, such as vote buying, illegal campaign financing, etc., is prevalent.	3.43	3.90	–2.15	0.033
66. Land rights are secure in urban areas.	2.48	2.93	–2.12	0.035
80. Most public policies are guided by a long-term strategic vision.	2.88	2.43	2.07	0.040
79. Public procurement of goods and services is transparent.	2.26	1.88	2.00	0.047
109. Auditing of formal firms results in tax adjustments when appropriate.	2.68	3.10	–1.98	0.050

Note: For all the statements, the Likert scale is 1 = ‘Completely disagree’, 2 = ‘Disagree’, 3 = ‘Don’t agree or disagree’, 4 = ‘Agree’, and 5 = ‘Completely agree’.

Source: Authors’ calculations based on quantitative survey.

We repeated the same analysis considering the affiliation of the respondents (Appendix A). Those affiliated with the business sector tend to agree that poverty reduction is a priority for the government, whereas the average response of the remaining respondents is below score 3 (Table A3.1). Additionally, on average, they agree that it is easy for private foreign investors to invest in Mozambique, whereas the average of the remaining responses points to disagreement with this statement (Table A3.1). Academics, on average, seem to be more pessimistic than the remaining respondents in terms of the debate of economic policy within government and parliament, the enforcement of judicial decisions, and the security of land rights in urban areas (Table A3.2). Finally, we find that respondents affiliated with public administration, on average, are more pessimistic than the rest of the respondents in terms of how reactive the public bureaucracy is to changes in the economic and social context (Table A3.3). The same happens for the ability of the opposition to influence political decisions and the predictability of the delivery of government services despite corruption (Table A3.3).

C Final Remarks

The overall results of the analysis of the quantitative survey point to the importance of the institutional areas linked to management of public administration and of macroeconomic and sectoral policy, as well as a lack of common vision for a national strategy, as important constraints to economic development. It also seems to be consensual among respondents that problems related to human capital together with poverty and inequality have been impediments to development in Mozambique.

A more detailed analysis of a broad range of institutional areas shows that the answers are spread into different aspects and do not reveal a consensual view on a particular group of constraints. However, there is some indication that there are constraints in the areas of legal and constitutional matters, public protection, and the provision of public goods and services. More specifically, the analysis of the respondents’ opinions indicates some weaknesses also in terms of resolution of land issues and of corruption in the business environment and public administration. Some results echo the view that Mozambique has been negatively affected in the past by events in neighbouring countries. However, at the same time, the degree of collaboration with neighbouring countries is highlighted as an institutional strength. Additionally, even though the answers show agreement with a high level of aid dependence, they also suggest that respondents regard development aid as a strength to the economic development process.

A Introduction

This section summarises the views of KIs on major institutional constraints for economic development and their thoughts on possible solutions for the following five fundamental themes:

• rule of law and judicial independence;

• voice, participation, and political accountability;

• political instability, violence, and state legitimacy;

• state capacity and autonomy from private interests;

• sovereignty and independence.

Interviews with 22 KIs – consisting of politicians, business people, academics, and liberal professionals – took place from mid-April to mid-June 2019. During more than 50 hours of interviews, KIs were requested to identify important institutional issues and their perceptions were used to verify the relevance of the five themes and as a basis for further analysis.

Within the framework of the EDI project, ‘institutions’ were broadly defined as the ‘formal or informal rules of the game, which political, social and economic actors are expected to follow, individually and collectively’ (EDI 2020).

KIs recognised significant advances in institutional development since Mozambique’s independence in 1975, although there were phases when the party in power and government made choices that represented an institutional regression. The country went through periods when institutions were considered as improving and other periods when institutional capacity weakened, and their functions were distorted. Some islands of excellence have been created that seem to persevere, but generally KIs argued that during the past decade some institutions degraded, considering the democratic model as reference and recognising the value of public goods.

As a note on methodology, what is reported in the following sections is a brief set of institutional constraints and proposals for solutions, based on lively KI testimony. Accordingly, this summary is focused more on constraints and less on main institutional advances. Given that many KIs commented on agriculture and education, their views on these sectors are summarised separately. Finally, the conclusion points to a selection of constraints and proposed solutions, which were common to most KI views and seen as relevant for understanding the connections with the economic development.

B Rule of Law and Judicial Independence

C Voice, Participation, and Political Accountability

D Political Instability, Violence, and State Legitimacy

E State Capacity and Autonomy from Private Interests

F Sovereignty and Independence

G Views on Agriculture

H Views on Education

I Summary of KI Observations

The KIs highlighted the following:

• The judicial system remains weak and not independent of the executive power.

• The rule of law needs reinforcement to prevent/discourage crime and enforce contracts.

• Democratic institutions and regulations need reform to ensure a more inclusive society and bolster it against elite capture and – ultimately – foreign dominance.

• The state apparatus requires better-qualified personnel, downsizing, and a more stable structure, manned by professionals selected and promoted using objective and merit criteria.

• Legislation and institutions should be adapted to further general economic efficiency in the markets, including in public enterprises.

• Decentralisation, including of the public finance system, and a serious reform in agricultural and fisheries policies are required for increased productivity and income levels mostly for smallholder producers.

• There is a need for investment in education quality at all levels, and improvement of technical and professional education.

• Mozambique is a low-income country, and it could make better and more careful use of foreign investments and aid. It is under foreign pressure for its riches, natural resources, and strategic location. It should have its own coherent plan and good quality leadership with integrity, working towards an inclusive Mozambican society to stay independent and peaceful.