Poverty in Mozambique: immobility and widespread vulnerability
Using innovative statistical methods, we provide new insights into Mozambique’s existing household data and reveal a more complex picture of poverty persistence and widespread vulnerability.
Until recently and following decades of civil war, poverty rates in Mozambique had been slowly falling, from 70% of the population in 1996/97 to 46% in 2014/15. Since then, the country has faced many social, environmental and economic shocks and stresses – including a national debt crisis, devastating cyclones and, along with the rest of the world, the COVID‑19 pandemic. Economic growth has slowed dramatically, threatening the gains made over the previous 20 years: even before the pandemic, research suggested that poverty rates could increase by 10%.
But to really understand the situation in Mozambique, and to develop effective poverty reduction policies and programmes, we need to know more than how many people were ‘poor’ or ‘not poor’ at a single point in time. By applying innovative statistical methods and modelling to existing national household survey data, we have generated new insights into poverty in Mozambique. These insights provide a more detailed picture of how people’s economic wellbeing might change over time – that is, the dynamics of poverty and vulnerability – and the factors that are associated with these changes (like the head of household’s age or level of education). In doing so we can suggest where the country might best focus its policies and programming, particularly in the context of COVID-19 response and recovery.
Our analysis, which looks at survey data from 1996/96, 2002/03, 2008/09 and 2014/15, shows that despite solid economic performance and moderate reductions in poverty rates nationally, a large proportion of households experienced no change in their situation. If a household was poor in the first of two survey years, they were more likely than not to remain poor in the second. And this probability of poverty persistence has increased steadily over time, from 55% between 1996/97 and 2002/03 to 70% between 2008/09 and 2014/15. For households that do move out of poverty during this period, they are extremely unlikely to find themselves in a position of economic security – a reality that top-level trends overlook. When we introduce ‘vulnerable’ as a distinct category for households that are above the poverty line but at risk of falling into it from shocks and stresses, our analysis of 2008/09 to 2014/15 transitions shows that the most likely movement out of poverty is into a position of vulnerability. Even then, the likelihood of this upward transition from ‘poor’ to ‘vulnerable’ (30%) is half as likely as the chance of remaining either poor (70%) or vulnerable (60%).
This picture of immobility is also reflected over a single year. Actual panel data from 2014/15 (information gathered from repeated interviews with the same households at different points in time) shows limited movement between different states of poverty and vulnerability between quarters. Instead, most people are either in poverty or out of poverty for the entire year. Poverty persistence is especially high during Mozambique’s rainy season (March to October), when stocks from the April harvest have been used up and heavy rains, floods and cyclones cause significant disruption. It’s also over this period that, for those households that do experience a change in their circumstances, a downward transition into poverty is more likely than an upward transition out of poverty. More worryingly, of these people that fall into poverty in the rainy season, there’s a relatively high chance (45%) that they won’t recover in the subsequent dry season but will remain ‘poor’ – which stresses the importance of making sure that temporary poverty as a result of shocks or stressors doesn’t become persistent poverty
Interestingly, individuals experiencing this particular intra-year transition (from non-poor in one quarter to poor in the next two) are slightly more likely to be in urban areas, have some access to basic services and own some assets. This is very different to the strong rural association with downward transitions and persistent poverty over multiple years (inter-year), and highlights two key findings. First, associations between certain household characteristics and poverty mobility are peculiar to the specific transition and its circumstances, and therefore policies that target poverty at a particular point in time are unlikely to ensure that households escaping poverty don’t fall back. Second, vulnerability in Mozambique is widespread and not necessarily confined to the poorest households – even in a relatively stable year. Given the shocks and stresses that the country has since faced (including two major cyclones – Idai and Kenneth – in November 2019), these patterns of poverty and vulnerability will have almost certainly worsened.
Based on this analysis, it will be critical to poverty strategies in terms of where they focus, who they reach and how they’re tailored to different household circumstances. For example, efforts to address longer-term persistent poverty might be most effective if focused on households in rural areas in the central and northern regions of the country, while additional support for households in urban areas may be needed in response to temporary or intra-year shocks. This shorter-term focus on urban households is especially urgent given the COVID-19 pandemic’s early economic impact appears to have hit urban areas hardest. Education is the single characteristic most strongly associated with poverty escapes and therefore continuing investment in education opportunities may decisively increase upward transitions.
Read the full paper here.