We use new and unique nationally-representative panel ISA data for Malawi, Tanzania and Uganda to explore the degree of consumption insurance in Sub-Saharan Africa. Partly, our contribution is to construct accurate and consistent measures of consumption and income across time and space for these countries at the household level from the LSMS-ISA Surveys. Our main result is strong evidence of income inequality leading to consumption inequality, in particular, in rural areas. First, our full-risk sharing tests suggest complete markets allocations are strongly rejected (though less so in urban areas). Urban households can insure better despite facing (as we discuss below) more sizeable permanent shocks, while rural households insure poorly even though they are likely to get hit mostly by transitory shocks. We use the richness of our data in its full extent to explore the robustness of this result to a large set of idiosincratic risks beyond income such as population risk, wealth risk (livestock, land), health risk (illnesses/injuries/hospitalizations), marital risk, migration risk, refugee risk, weather risk, and intermediate inputs risk. Our results are robust to the type of consumption item (food, clothing, etc.) and income source (agricultural, labor, etc.). Geographic determinants such as distance to roads, city, coastal areas also support our results. Further, we find that the degree of insurance decreases with the level of aggregation, that is, individuals within enumeration areas/districts tend to be more insured than across them. Second, we show that the raw and residual variances of income and consumption over the life-cycle are highly correlated in both rural and urban areas suggesting that income shocks do transmit into consumption in both areas. Further, we find that in rural areas residual inequality does not grow over the lifecycle, suggesting a prominent role of transitory shocks in these settings (not so in urban areas where residual inequality accumulates). These two cohesive observations, the fact that (i) consumption in rural areas (where the vast majority of households in these poor countries live) largely responds to income shocks and that (ii) these shocks are of transitory nature, cast doubt on the permanent income hypothesis for rural areas. Then, we explore market structures that can explain this finding: transitory shocks have strong implications for consumption growth in rural areas. To do so, we use an agricultural model with subsistence consumption that incorporates potential credit and savings constraints. To asses the effects of saving constraints we use information on substistence consumption, food security, storage technology (harvest lost and infrastructures) and crime/theft/violance indicators. To assess the effect of credit constraints on insurance we use complementary data on the ability to borrow that includes detailed information on loan applications and its outcome (this includes self-selection reports into loan application as those who did not apply but wanted a loan are also examined). Further, our decompositions of insurance tests across wealth (separately for livestock and land) groups further substantiate the evidence on credit constraints. The associated empirical tests find strong evidence of the presence of both constraints and their signicant eect on consumption insurance. Finally, we explore the dichotomies between self- and mutual insurance, and between private and public transfers. Using self-reported data on the use of insurance mechanisms we nd that those individuals that claim to be only mutually insured (e.g. resort to family, friends, etc. to cope with shocks) are able to insure their consumption better than those who are only self-insured (through savings, labor supply, etc.). In this context, the fact that relatively few individuals (about 20% only) rely on mutually insurance mechanisms suggests the presence of limitted commitment economies where default is highly present. Last, we nd little evidence of crowding out using the fertilizer policy implemented in Malawi, the most important public transfer conducted in these three country, as individuals receiving fertilizers also tend to be more privately insured. We conclude that low partial insurance is the norm in the rural areas of SSA and that credit constraints, savings constraints, and imperfect enforceability with a high risk of default are good reasons for it. The one lesson that we are learning is that theories that incorporate credit constraints, savings constraints, and some degree of lack of commitment are the ones more suitable to understand the degree of consupmtion insurance in SSA and possibly growth in these contexts. We are currently pursuing a quantitative theory of that short.