S. Raihan, B. Khondker, G. Sugiyarto & S. Jha ...



Farming, fertility, food: why has Bangladesh done better than expected? Can it continue to do so?Michael Lipton, Sussex University, Brighton, UK1st Bangladesh Economic Forum, Dhaka, 21 June 2014 (revised 1 Aug 2014)(a) IntroductionBangladesh was born amid widespread pessimism. It has lived in an unhelpful natural and political environment. Yet, in the last quarter-century, it has achieved respectable economic growth and, on many social and economic dimensions, substantial mass advance. So have some other developing countries, with hugely different conditions, polities and problems. It is therefore likely (absent unmanaged, deep and wide environmental catastrophe, of which more later) that growth and mass advance can be achieved in many different ways, and that the necessary conditions for State conduct are rather simple. However, many other developing countries, especially in sub-Saharan Africa, have not achieved rapid growth or mass advance. What are the necessary conditions? How has Bangladesh met them? Can they - and hence mass advance - be spread and sustained?Early in his career, Adam Smith described apparently simple conditions for growth with mass advance:Little else is required to carry a state to the highest degree of opulence, but peace, easy taxes, and a tolerable administration of justice. He later added free trade; universal elementary education; and provision of some public goods. Malthus, in the first edition of his Essay on Population (1798), raised an objection: Smith's recipe might produce growth, but not mass advance. That would raise the capacity to buy food, and hence to support children. Given 'the passion between the sexes', this must stimulate earlier marriage and larger families. That would raise labour supply and cheapen it, raise demand for food and make it costlier, and so erode any gains for the working poor. If this did not happen fast, then - once spare farmland and agricultural improvement were exhausted - rising death-rates from famine, sickness or war must re-align population and food availability: prolonged and substantial mass advance was impossible. Later, however, Malthus's empirical studies transformed his pessimism into optimism. By 1826, he agreed with Smith that there were simple conditions for mass advance, but they were not Smith's: We are not however to relax our efforts in increasing the quantity of provisions, but to combine [the] effort ... of keeping the population ... at such a distance behind, as to ... unite the two grand desiderata,?a great actual population, and a state of society, in which abject poverty and dependence are ... little known; two objects... far from incompatible. [For] ... essentially and permanently bettering the condition of the poor, we must ... show them, that the withholding of the supplies of labour [by reducing fertility] is the only possible way of ... raising its price, and that they themselves, being the possessors of this commodity, have alone the power to do this ... [T]his country, with a proper direction of the national industry, might, in the course of some centuries, contain two or three times its present population, and yet every [person] be much better fed and clothed than ... at present.Thus 'Malthusian optimism' indicated, as conditions for mass advance - in addition to Smithian peace, 'tolerable' rule of law (including clear property rights), and education - State provision of laws, incentives, and public goods that increased food supply and stimulated voluntary reduction of fertility.Recent decades have seen "surprising" acceleration of growth and mass advance across many different polities, including Bangladesh. Are there simple common conditions that explain this? All agree that peace and 'tolerable' rule of law, including clear (but not unrestricted) property rights, are normally needed. Smithians add State-led liberalization, leading to rising trade/GDP ratios, specialization, and hence output. Malthusian optimists emphasises State action to stimulate farm productivity, food adequacy, and fertility reduction. This paper argues that more of Bangladesh's success is explained by post-Malthusian State action than by post-Smithian liberalization. Malthus, because of his brilliantly written but crude and unempirical first edition, is often seen as a pessimist. Yet State action on fertility, farming and food, as indicated by mature Malthusian optimism, goes far to explain Bangladesh's (and China's and Rwanda's) relatively successful trajectory, meeting the conditions for mass advance despite terrible handicaps. Smithian liberalization is relevant to growth - above all in farming, by removing the excesses of agro-"planning"; but rice research and the spread of irrigation mattered more, and their output effect was magnified by far-from-liberal input subsidies. Two other sources of Bangladesh's progress, growth of income and employment in ready-made garments and remittances, are sometimes credited to liberalization. As for RMGs, the GDP impact is reduced by the high proportion of RMG output that comprises not value-added but imported inputs (apparently about 75 per cent). Also - while RMG growth is due partly to the liberalizing removal of barriers to imported inputs (and of policy biases against exports) through tariff reduction and quota elimination - export-promoting policy interventions, some specific to RMG, also played a role. As for booming migrant remittances (now about two-thirds as much as exports), they can be seen as a Smithian response to free(ish) labour movement, or as a Malthusian response to population pressure. Anyway, both sources of progress seem fragile, sandwiched between new competitor countries, and strong and justified, but cost-increasing, democratic and moral pressures to improve labour safety and conditions, both of Gulf migrants and of RMG workers. These pressures arise despite the weakness in Bangladesh of any third, redistributive-radical, political-economy alternative to Smithian liberalization and post-Malthusian state action. That weakness has limited the poor's gains, leaving big gaps in mass advance. Also there are serious sustainability problems. Maintaining the post-1990 mass advance, including poverty reduction, depends, not only on continued progress on fertility, food and farming, but on open, flexible institutions, as the structure of output and employment changes. This in turn depends on two further f's: facts, and freedom to discuss and challenge. (b) Past grounds for pessimism but apparent success since 1990: the argument Around 1971, “Bangladesh pessimism” prevailed. Population was already exceptionally dense, yet in 1965-70 was growing by 3%/year, with total fertility rates of 6.9 - both far above India, and similar to West Africa. Perhaps 70% of the workforce received income mainly from agriculture, and there was very little high-quality unfarmed rural land. The rice grain yield in a typical year was only about 1.1 tonne/ha (about 1.6 t. paddy). The historical record was one of colonial extraction - first by Britain, then by Pakistan - punctuated by famine and disease and culminating in the War of Independence, in which intellectuals had been targeted for slaughter. Many thought that poverty and hunger would get worse. The background for improvement has been unpromising. Bangladesh led the world in per-person exposure to natural disasters, with 80 serious droughts, floods, storms or earthquakes in 1993-2002 and 55 in 2003-12 - more than in any other country except China, India and the Philippines. Governance was marked by instability; by party rivalry driven by personalities rather than policies, and interrupted by suspension of democratic process; and by corruption. In each year from 2001 to 2005, businesspersons saw Bangladesh as the most corrupt country (i.e. the country that they had done most to corrupt) - though by 2013 it had improved to 136th of 186.Yet there is a widespread perception of disaster avoided, and since 1990 of success. Not only has growth been quite rapid, if slightly less than in India. On indicators of mass advance, Bangladesh has progressed faster than India. Poverty incidence and severity have declined. Education, at least as indicated by primary enrolments and literacy rates - and health and nutrition on many indicators - has advanced substantially. Key inequalities - female-male, urban-rural - have shrunk. How, why, with what gaps, how sustainably, and with what implications for other countries and for Bangladesh's future? Before tackling these issues, I summarize the main argument, and (with apologies for the superficiality of a foreigner) Bangladesh's experience.I argue that, if a low-income, initially farm-based country (e.g. Bangladesh 1971-2014) (1) gets a few things roughly right in three areas - farming, fertility, food; (2) limits war/violence; (3) publishes facts (contestable, verified data) and (4) has some freedom of policy debate, it can reach (1990-2014+) fast mass advance, despite otherwise rather unfavourable conditions in economy and polity. The conditions are that the government must either itself sectorally act “wisely” on farms, fertility, food, and/or allow, subsidise or incentivise “wise” NGO and/or private action. Class structure need not rule out “wisdom” - but progress leaves gaps and may be unsustainable. (c) Welfare - the record: production, distribution, poverty, well-being A main channel, from macro-economic performance to the "mass advance" of individual well-being, runs from (1) real gross national product per person and (2) foreign remittances; via (3) private consumption per person; via (4) distribution; to (5) the incidence and severity of absolute consumption poverty, here measured at the inter-national poverty line, $1.25/person/day at 2011 PPP; to (6) capabilities - nutrition and other poverty-related aspects of well-being. How did Bangladesh fare, in regard to improvement along this channel since about 1990? (1) Real per-person GDP fell by 24% from 1970 to 1972, due to war and its aftermath. In the next decade, apart from strong recovery in 1974 (8.2%), annual growth was around 1%; not until 1982 did Bangladesh regain the 1970 level of real per-person GDP. In 1982-90 growth was typically 1-2% (11.5% from 1982 to 1990). The rate rose further to around 3%/year, 40.6% in all, in 1990-2002, and to 4-5% thereafter - all at constant taka prices of 1980. A later base year is more suitable to measure recent growth, and purchasing-power parity (PPP) of 2011 for international comparisons; on this metric, real GDP per person rose at 3.7% per year from 1990 to 2012, i.e. from $1068 to $2364. This growth accelerated from a much lower level in 1971-89 to 2.8% between 1990/2 and 2000/2, and 4.65% between 2000/2 and 2010/12. GDP/person growth alone - with constant distribution and household-consumption/GDP ratios - would mean consumption poverty above pre-independence levels in 1973-82, but declining thereafter at an accelerating rate, sharply so after 2000. (2) Foreign remittances (not in GDP, but in GNP) rose from almost nil in 1976-7 to $1bn in 1994-5 and $11bn in 2010-11.Though they did not go mainly to the poor, from 1994 they increasingly accelerated poverty reduction. If in 2010-11 only $1bn reached 10mn (of the roughly 67mn) Bangladeshis in absolute consumption poverty (see (5) below), that is $100 each: over 20% of annual income at the international poverty line (begging the question of PPP vs. exchange-rate dollars). Even such crude arithmetic is suggestive. A causal model suggests 'that 1.7 of the 9 percentage points [fall in absolute poverty in] 2000-2005 was due to the growth in remittances'. (3) Household consumption per person in 1976-88 was 85-88% of GDP, but fell to 79% in 1998 and 75% by 2004, before levelling off at 74-75% in 2004-12. In constant prices, it grew slowly (0.5%/year) in the 1990s, but at 3.5%/year in 2000-12 (and since 2010 at over 4%), matching slower growth after 2000 of gross capital formation (9.7%/year in the 1990s, 7.7% since 2000) and of exports, both gross and net of imports. Rapidly increasing capital formation is often at the cost of current poverty reduction, but may accelerate it later.(4) As for distribution of per-person household consumption, its Gini coefficient rose from 25.9 (1984) to 33.5 (2000), before falling to 32.1 (2010). The poorest quintile's share was around 9.4-10% in four surveys 1984-92 (and the top decile's share around 22-24%), and around 8.8% in four surveys 1996-2010 (top decile's share, 27-29%). These levels and uptrends in inequality are low by Asian (and very low by African and Latin American) standards. Though they underestimate true inequality, they suggest that Bangladesh's rapid growth should have led to substantial falls in absolute poverty. However, the poorest two household quintiles were below the international poverty line in all eight household surveys 1984-2010, and the richest quintile above it: so these quintiles' consumption shares do not directly affect poverty incidence. Given changes in mean consumption, changes in incidence depend on how many poor people are "movers" across the poverty line - i.e. on the density of poor population near the line, and on levels and trends in proportions of consumption that such people enjoy. In 1992-2010, when poverty fell sharply (see below), "movers" were in the second-, third-, and (in 1992-6) fourth-poorest household quintiles. These middle quintiles' consumption shares - crucial for poverty reduction - changed in 1992-2010, respectively from 13.6% to 12.4%; from17.3% to 16.1%; and from 21.1% to 21.3%. This slightly "squeezed middle" (rising polarization) somewhat hampered poverty reduction. More important, however, these shares stayed rather above Asian norms, which helped poverty reduction. (5) Severe absolute "consumption poverty" (below $1.25/day at 2011 PPP) rose in 1984-92: incidence from 61% to 70% and severity from 18% to 24%. Slowly rising inequality, and from 1988 a falling consumption/GDP ratio, more than offset modest growth of GDP per person. From 1992, poverty fell, reflecting faster growth of mean GDP, levelling-off in consumption/GDP ratios, booming remittances, and only slight worsening of distribution: by 1996 poverty incidence had fallen to 61% (its 1984 level) and severity to 19%. Falls to 2010 were much faster: to 43% and 11% respectively. Trends in extreme poverty were worse than in India around 1984-93, but subsequently much better, even though Bangladesh's GDP was barely half India's, and grew somewhat more slowly.(5) Amartya Sen rightly insists that improved capabilities - a central aspect of well-being - are a goal of policy, and of efforts to reduce consumption poverty. However, to well-be one must first be. Bangladesh's good performance on life expectancy and child and maternal mortality owes much to better nutrition, due in part to reduced consumption-poverty. The proportion of under-fives stunted - more than two standard deviations below Harvard mean height-for-age - fell from 71% in 1993 to 37% in 2012: 45% in rural and 36% in urban areas (female-male gaps were negligible throughout; rural-urban gaps were significant but falling). Nutrition and health advance derives in part from private poverty reduction. However, for stunting and wasting to fall dramatically, better private access to food is only half the story; the other half is reduced risk of infection, and better general health. For that, non-private action - on water, sanitation, food safety, health-care provision and access, education, reduced disparities of residence and gender - has to complement consumption rises among the poor that raise their command over food. The question remains: how did Bangladesh manage this, and how can progress be sustained, given imperfect governance and high environmental risk? This needs to be explored, as there is still much to do: in 2013 only four countries not in sub-Saharan Africa (Afghanistan, Myanmar, Nepal and Yemen) had a Human Development Indicator below Bangladesh's.(d) Structure - the record: falling agricultural share, rising savings rate, or trade-led transformation? 1. Share of agriculture: A shift of output and workforce from agriculture to modern services and manufacturing is normally a structural enabler of development. During Bangladesh's accelerated growth, agriculture's share of the workforce dropped slowly, from 51.6% (1990) to 47.3% (2010). Its share of GDP dropped much faster, from 30% to 18%. So non-agricultural labour productivity was 2.5 times agriculture's in 1990, but 4.1 times in 2010. A little of this may reflect statistics - viz. rules that counted more women in the agricultural workforce (n. 23) - but most reflects (a) major capital formation and skilling in services and industry, including RMG, (b) an agricultural strategy focused (rightly) on raising land productivity faster than labour productivity. With land scarce, that is necessary if agriculture is to continue absorbing extra workforce. In Bangladesh, it should: though population growth has come down to just above 1%, workforce is still growing at 3%, and cannot yet be fully and affordably absorbed in non-farm activities, though their share of workforce should steadily rise. In early development, rapid industrialization of GDP, to be pro-poor, should accompany slower industrialization of workforce, and some continued absorption in farms with high and rising returns to land and capital. Since these are mainly small farms, agricultural progress and policy on landholding, as well as technology, were, and remain, critical to this key structural transformation. We return to this issue in sec. (e).2. Gross national saving: A sharp, sustained rise in the savings rate is often seen as the key structural change in the development process. Bangladesh rather more than doubled the ratio of GNS (gross national income minus consumption) to gross domestic product, from 18% in 1983-91 to 20-23% in 1992-8 and 27-28% in 2009-12. Growth of GDP per person equals the savings rate divided by the population growth rate, multiplied by the marginal capital/output ratio. The annual population growth rate slightly more than halved, from 2.64% in 1980-90 to 1.19% in 2010-15. Thus growth of GDP per person should have increased about 4.7-fold, if the marginal capital/output ratio and the depreciation/savings ratio stayed much the same. Both ratios will have risen, because as investment rose in 1983-2012 it shifted towards non-farm investment including infrastructure. This, plus (inevitably) very approximate data, explains why annual growth in GDP per head accelerated less than 4.7-fold: from 1.4% in 1982-90 to 4.65% from 2002, i.e. 3.3-fold, still a huge advance. This is merely a tautology, "attributing" rising growth in GDP per head to changes in savings and marginal capital/output ratios, and in population growth. It pinpoints the structural importance of Bangladesh's doubled savings ratio but says nothing about process. The sharp acceleration of savings-ratio growth since 2000 is partly due to remittances (included in gross national but not gross domestic product and saving). By fiscal year 2011 gross national saving (GDS) was almost 50% more than gross domestic saving. The remittance boom - while extremely welcome - may not represent a lasting, structural change; and policymakers cannot readily 'translate' international private intra-household remittances into higher investment. However, the savings ratio was rising, though more slowly, before the remittances boom; such a rise may well be a structural and sustainable source of economic growth. More disturbing is (a) the tension between a gross savings boom, worsening inequality, and poverty reduction: the more that extra incomes (and indeed remittances) go to the better-off, the more is the savings ratio, but the wider is inequality, and the slower poverty reduction; (b) the fact that all these numbers are for gross saving, disregarding depreciation, whereas it is net saving that is the fuel of growth. The only available adjustment shows gross national savings at 33.9% of GDP in 2008 (India 38.2%); consumption of fixed capital at 6.8% (8.5%); and net saving at 27.1% (29.7%), and then allows for changes in human and 'natural' capital.3. Trade: Was Bangladesh's progress 'led' by liberalization, especially of foreign trade? Exports of goods and services, around 6 per cent of GDP in 1980-91, rose between plateaux: 11-12% in 1995-8, about 14% in 1999-2004, 19% 2006-10, and 23% in 2011-13. Imports also plateaued: 18% in 1980, they fell to stabilise in 1984-94 at 12-13%, recovered to 18-20% in 1995-2003, and rose to 25-7% in 2006-10 and 31-32% in 2010-13. Certainly trade derestriction improves efficiency by allowing specialization, and eases the path to imports for diversification out of agriculture. Trade expansion, in a broad sense, accompanied growth in Bangladesh, but did it lead or cause? The expansion is overstated by counting re-exports, hugely booming, into both exports and imports (fn. 6). Nor is the trade-causes-growth reasoning supported by the surge-plateau pattern both of exports and of imports; the timing of surges vis-à-vis growth acceleration; the persistent and widening import surplus; and the concomitant increased financing of imports not by domestic exports, but initially by aid, and subsequently by remittances - by 2008-9 worth almost half imports (exports were worth two-thirds). My working hypothesis is still that Bangladesh's improved performance is based on rising annual output per hectare, mainly on small farms, together with falling human fertility; that this permitted both rapid mass advance, and structural transformation via gradual de-agriculturalisation and sharply higher savings; and that trade liberalization supported these trends (and was itself desirable) but did not substantially cause them. (e) What happened and why: farmingGross agricultural production rose 3.16-fold from 1961/4 to 2008/11, but around 1990 its growth rate roughly doubled: from 1.58% per year from 1961/4-1989/92 to 3.82% per year from 1989/92 to 2008-11. In this latter period there was little scope for rising farm area or cropping intensity (already around 1.8); the upsurge in agricultural growth came mainly from rising irrigated area (from 2.65m ha in 1990/1 to about 4m ha in 1996/97) and chemical fertilizer use (from 2m tonnes in 1990/91 to 3.02 million Mt in 1995/96), alongside faster adoption of a steady stream of improved boro and transplanted aman rice varieties, mainly from BRRI. Hence net paddy yields rose from 1990/2 to 2010/11 by 2.3% per year (from 2.6 tonnes/ha to 4.2), and the crop-mix shifted towards higher-value crops. Does this agricultural upsurge cause, or merely coincide with, Bangladesh's sharply better economic growth and "mass advance" after 1990? Growth (and therefore acceleration) in any large sector contributes arithmetically to GDP growth. Farm growth - unless distribution of farm income worsens - contributes even more to poverty reduction, since over 80 per cent of Bangladesh's $1.25-poor depend mainly on agriculture, around double the proportion of the non-poor. However, this arithmetic merely shows that Bangladesh's agricultural acceleration might have helped cause its faster and/or more pro-poor growth after 1990. Two "theoretical" considerations suggest that this did happen. First, agricultural growth is much more labour-intensive than growth in the rest of the economy, i.e. it bids up the real wage and/or the level of employment. This tends to make farm growth relatively pro-poor (unless it induces anti-poor structural change, such as major eviction). This is fortified by the strong rural "growth linkages" of extra farm to extra non-farm employment and income. Second, about half of consumption at the poverty line comprises staples (and over two-thirds is food); where agricultural output is staples-dominated as in Bangladesh, its growth moderates local staples prices, and stabilises availability, of the poor's main need. As for evidence, cross-country regressions and country studies (e.g. South Africa, China, India, Indonesia, Philippines) show that growth is more poverty-reducing if agricultural. A recent review concludes that typically an extra $1mn of agricultural output reduces poverty incidence over 3.5 times more than similar non-farm growth; only in Latin America, with much more unequal land than most other regions, are exceptions found, confirming that exclusion, by land inequality, of the poor from farm income is a big cause of poverty. In this context, and contrary to oft-reiterated conventional wisdom, Bangladesh's generally very small farms are a source not only of well-distributed mass advance, but of more intensive land use than larger farms, and hence of faster economic growth. Farmers operating above 2ha (7.3% of farmers, on 31.1% of arable land) produced gross annual output worth 76583tk/ha, as against 102914 taka/ha for farmers operating below 0.2ha in 2011 (36% of farmers, on 9.6% of arable land). For net output (value added) the respective annual figures were 37,800 and 58,400 taka/ha - over 55 per cent more The smaller the farm, the greater the use of fertilizer, and the proportion of area irrigated, due in part to "sweat capital" in installing and maintaining irrigation, especially tubewells (and other wells). Rice yields did not significantly differ by farm size; tiny farmers' greater land productivity derives from their low labour-supervision cost, and hence greater eagerness to use more labour-per-hectare to generate a higher-value crop-mix. This "inverse relationship" in labour-plentiful, land-scarce developing countries is confirmed by numerous studies, and by the tendency, revealed by successive agricultural censuses in developing countries for increasing proportions of farmland to be found to the smallest size-groups. Clearly, this means that land grab in Bangladesh - if, as recent research suggests, it tends to displace the smallest farmers, who may lack the political muscle to insist on full compensation - is hostile to agricultural efficiency as well as equity. The inverse relationship also makes a case for land redistribution in Bangladesh, but how? First, farmers owning 2-5ha are difficult targets for redistributive reform, being less wealthy, much more numerous, and no less vociferous than urban groups seeking to reform them. The lowest feasible land ceiling may be 5ha, and it is not clear how much land that would release for the landless and near-landless. Second, as a recent survey shows (fn. 39), inequality of operated farmland is much reduced by sharecropping, mainly from larger owners to the land-poor or landless - whose welfare might be enhanced by tenancy reform, but not if the owners respond to reform by resuming personal cultivation, as often happens. Tenancy reform works if accompanied by ceilings on ownership that can be implemented, but can be counter-productive otherwise. Might the shift of larger landowners into urban business provide an opportunity for politically manageable land reform in Bangladesh? Third, experience in Indonesia, Nepal and elsewhere has shown that, even in very densely populated rural areas, distribution of home gardens to the landless and near-landless is feasible, even where wider reform is not.Anyway, concentration in very small, labour-intensive farms is one reason why Bangladesh's agriculture, while accelerating its growth rate after 1990, has continued to provide a large share of productive employment. Rapid agricultural growth has helped to release savings and labour for non-farm expansion, and to permit better nutrition without greatly expanded food imports. In all this - given land scarcity, and a still fast-rising supply of workers - rapid technical progress in agriculture is crucial (for equity as well as for efficiency) - but so is the form of that progress. For all major groups of poor - rural near-landless, urban poor, tiny farmers - to benefit from agro-technical progress, it has to raise labour productivity, but to raise land productivity by a higher proportion; and it has to restrain stales prices, but raise agriculture's total factor productivity by a higher proportion. The Green Revolution, being based on improved seeds that made heavier fertilization and more irrigation profitable, met both these conditions, but globally the gains from "conventional" plant breeding are declining in each decade. This is especially the case because soil-water constraints, land loss, and climate change create new requirements, if yield growth is to be maintained. Hybrid rice and other innovations can help, but will not maintain the rates of advance Bangladesh enjoyed in the 1990s - while workforce, seeking productive employment and needing extra food, will. In this context, for preserving agriculture's post-1990 contribution to pro-poor growth, transgenics is "the only kid on the block". To steer its gains to the poor cannot be the top priority for private companies and their researchers, but can be for Bangladesh, co-operating with the International Rice Research Institute and other public agencies. Bangladesh is well placed here, having a strong science and research base, and apparently less handicapped by the ill-informed pressures and emotional hangups that have denied even the available transgenics (except Bt cotton ) to China, India, and indeed Europe. (f) What happened and why: fertility and demographic transitionWe have seen that Bangladesh experienced an upsurge - a sharp acceleration - around 1990 in income-per-person growth, "mass advance" (e.g. consumption-poverty reduction), and structural change (e.g. falling GDP shares in agriculture and rising savings ratios). Absent radical institutional improvement, three sorts of cause are plausible: greater openness to trade, especially internationally; agricultural change; and population change. As for international variables from 1990, there was a sharp acceleration of remittances, but neither of trade liberalization - powerful and desirable, but in Bangladesh more gradual than in India - nor, probably, of foreign trade net of re-exports. As for agricultural variables, there was a sharp acceleration around, or soon after, 1990. On population change, I first review which demographic variables, in what order, benefit income-per-person growth and "mass advance". I then show that in Bangladesh these indeed improved sharply, either in the early 1990s or with appropriate time-lags, and thus, like agricultural variables, may well have helped explain the upsurge.Demographic transitions in 1920-2015 usually started with reductions in disease, especially dysentery and malaria, and/or improved nutrition. That led to big falls in death-rates, especially among under-fives, with static or even slightly rising birth-rates. The swelling child numbers meant sharp rises in population (rising to 2-3%/year) and, at the young end, in dependency ratios. This process was followed, some 10-40 years later, by falls in total fertility rates and, with a lag due to population momentum, in birth-rates: the crucial causal role of falling under-5 mortality in inducing TFR falls, and the supportive role of education and earnings, especially for women - raising the cost of motherhood, and allowing parents to substitute child quality for quantity - are widely agreed, but there is controversy about whether the role of contraceptive supply is major or minor. Anyway, falling birth-rates came to outweigh the effects of (still-falling) death-rates, so population growth declined (though still haltingly and slowly in Africa): after 40-60 years population growth falls to near-zero or negative in many countries. Dependency ratios meanwhile improved, both as growing numbers of children survived to enter the workforce, and, later, as falling fertility reduced child numbers. Finally an ageing population and fertility decline led to a rising dependency ratio and, in many cases, to some population decline.What are the effects of this long sequence upon economic growth and "mass advance"? Initially this was investigated by cross-country regressions. One-third of international growth-rate differences within East and South Asia around 1960-90 was attributed to fertility reduction, mainly via higher savings, lower dilution of capital, and lower dependency ratios. International differences in poverty reduction rates were linked to the extent of earlier fertility reductions, with the effect channelled almost equally via faster growth and less-unequal income distribution. Such regressions, despite increasing sophistication, were widely thought to face insuperable problems in sorting out cause and effect. More recently, however, country studies - and simulation effects at national level of proven micro-effects of fertility change on well-being - have supported the initial findings. However, all this relates only to the current phase of demographic transition, the demographic gift, when fertility and proportions of under-15s are falling, thus reducing the dependency ratio. This is a "demographic window" between two phases of demographic theft: the phase of rising mortality, near-static TFRs and growing young-end dependency; and the phase (long reached in Japan and Europe) of very low death-rates and even lower birth-rates, with a large and growing old-age dependency burden. In Bangladesh, demographic progress speeded up in 1990, helping explain improved growth and mass advance: Population growth, similar from 1975/80 (2.65%) to 1985/90 (2.60), fell to 2.20% in 1990/9, 1.99% in 1995/ 2000 - and 1.56% in 2000/05 and 1.09% in 2005/10, tracking a further rise in GDP-per-person growth. Under-five mortality between 1950/5 and 1965/70 fell from 284 per thousand births to a still-shocking 228, i.e. by 0.99%/year. The rate of decline in 1965/70-1975/80 was similar (0.91%.year). Under-five mortality then fell more and more sharply: 2.94%/year in 1975/80-1980/5, 3.58% in 1980/5-1990/5 and 5.11%/year in 1990/95-2000/05 and 2010/15, reaching 42 - one-fifth of the rate in 1965/70. Did the slow improvement from 1950/5 to 1975-80 gradually, and with a time-lag, make couples ready to have fewer children, confident of survival prospects; and was the time-lag greatly shortened, and the proportion of couples ready to reduce fertility increased, by the rapid acceleration of child mortality decline from 1975/80 to 2010/15?Total fertility rate is consistent with this. From 1950/5 to 1970/5 it rose (from 6.36 to 6.91 children/woman), a pattern familiar in Africa, reflecting lower stillbirth rates; low confidence in reduced child mortality (or female education, etc); and, perhaps, reduced "unmet demand" for children. TFR then fell, slowly at first (by 0.28 children/woman in 1970/5-1975/80 and 0.65 in 1975/80-1980/5), but accelerating, to decline by 1.87 children - 3.68% per year - from 1980/5 to 1990/5, when TFR was 4.11. It reached 2.87 by 2000/5 and 2.20 by 2010/5. Birth-rates (live) changed by less than likely measurement error from 1950/5 (48.5 per 1000) to 1970/5 (46.4), but then fell at a rising rate: 0.92%/year from 1970/5 to 1975/80, 1.25% to 1980/5, 2.11% from 1980/5 to 1985/90, 2.57%/year from 1985/90 to 2000/5, and 2.85%/year from 2000/5 to 2005/10, by when the birth-rate had fallen to 20.3 per 1,000. Birth-rates tend to fall at a slower annual rate than TFRs - and to fall to the same level as death-rates, so population stops growing - well after TFRs have reached 'normal' replacement levels of 2.1-2.3 (as they have in Bangladesh) due to population momentum - the high and rising proportion of women in child-bearing age-groups. Nevertheless, the sequence supports an accelerated transition from about 1990.Direct dilution of capital, natural and human-made, fell due to slowing population growth, causing a switch, speeding up from 1990, from capital widening to deepening: higher true net savings and so real per-head GDP.Dependency ratios moved as suggested at fn. 44 for the region, though Bangladesh's "gift" arrived rather later than India's. Over-65s have remained 6-7% of Bangladesh's population from 1950 to date: changes in the dependency ratio therefore stemmed entirely from the share of under-15s. From 1950 to 1980 the dependency ratio rose from 82 to 92, because birth-rates changed little, while child survival improved somewhat. The DR then fell: by 7 (0.79/year) to 85 during 1980-90, but faster, by 15 (1.92%/year) to 70, during 1990-2000 and faster still, by 13 (2.94%/year) to 57, during 2000-2010. In the two decades 1990-2010, the number of dependants, supported by the average person aged 15-64, fell by just one-third, from 85 to 57. This accelerating process must be a big part of the explanation for Bangladesh's accelerating growth and poverty reduction, especially since, on top of the effects of age-structure, workforce growth - absolutely and relative to dependants - was further accelerated by substantial rises in female participation rates. Two caveats are needed. First, these extra workers-per-dependant are keys to growth and poverty reduction only if they find employment at rising labour-productivity. Workplaces at affordable investment cost are therefore critical, and quite late in the development process this continues to make the case for labour-intensive small farming with rising labour-productivity but (given land scarcity) faster-rising annual value-added per hectare. Second, the demographic gift - Phase 2 of the transition, with falling fertility, population growth, dependency ratios and capital dilution - is a brief respite between two demographic thefts: Phase 1, with high mortality, very high fertility, and fast-rising young-end dependency; and Phase 3, when population stabilises with a growing burden of old-age dependants like me. The dependency ratio in Bangladesh is projected to fall to 44 by 2035, but it then rises to 50 by 2050, 69 by 2070 and 84 (the same as in 1955) by 2100. These long-distance projections should not be taken too seriously, but the trends are clear, and suggest that Bangladesh (and other countries) need to make the best of their brief demographic window. (g) What happened and why: food and nutritional outcomesProgress against undernutrition among under-fives is indicated by falling proportions stunted or wasted. Stunting depends partly on feeding and infection ever since conception; wasting shows "recent and severe weight loss" due to starvation or disease. So reduced stunting tends to lag behind improved food intake and reduced infection. Some 71% of under-fives in Bangladesh were stunted through 1983-94, but the incidence fell to the low 60s in 1996-9, the low 50s by 2001-4, the low 40s by 2007-11, and 37% by 2012 (fn. 21). This is consistent with a sharp decline in family underfeeding and infection, starting around 1989-90 and continued at least to 2009-10 - and probably due in part to faster poverty reduction. Wasting depends largely on food or infection "shocks" to the most at-risk. Incidence stayed around 17% in 1983-91; fluctuated within 15.6-20.6% in 1991-8, 12.4%-14.6% in 1999-2004, but 16-17% in 2007 and 2011. This suggests that, though poverty fell and average well-being improved sharply from the early 1990s, risk to well-being in bad times for the most vulnerable 20% may not have fallen. This at-risk fifth probably comprises mainly very poor families, mostly rural; many remote; and most with several young children and one working adult, farming no or little land. Together, (1) changes in frequency or severity of shocks (rice price surges as in 2008; bad harvests; unemployment; local surges of dysentery or malaria), and (2) changes in resilience against shocks (e.g. via social protection), left the "at-risk fifth" too poor to reduce risks of child wasting.What accounted for much better performance against stunting from the 1990s? Urbanisation, and the slightly lower incidence of urban undernutrition, imply lower total undernutrition. This is tautological rather than causal, but new earning chances for poor women in urban RMG probably helped to reduce child stunting (known to be much more closely related to maternal than paternal income). Improved average nutrition, including that of the poor, is related to the upsurge in rural non-farm activity, to apparently better outcomes on the smallest (<0.2-ha) farms, and - perhaps most strongly - to the post-1990 surge in rice yields. That surge casts a shadow: better rice yields led (as in India) to reduced cultivation of pulse crops and other sources of micronutrients. This appears to be especially serious in slum areas, for zinc (49% rural, 50% slum, 29% other urban), and for iron (37% rural, 23% slum and other urban). The long-run answer is crop and diet diversification as incomes rise. But it will be several decades before all, or almost all, Bangladeshis can meet their micronutrient needs that way. Meanwhile, biofortification of rice - breeding and popularizing the best varieties with added, bio-available micronutrients - is a cost-effective, increasingly available solution, with promising and currently researched varieties tested in Bangladesh for effective bio-availability and absorption of zinc and iron. Bodily absorption of adequate nutrients depends not only on food intake, but also on infection, and therefore on sanitation, water, and treatment of diarrhoea. The proportion of households with sanitation remains small, but surged in 1990-2008, from 26% to 36%. The proportion with piped water, already 78% by 1990, changed little to 2008 (80%). The proportion of children with diarrhoea given oral rehydration surged from 48% in 1995 to almost 80% from 2008, twice the proportion in India. In nutrition, and probably elsewhere, Bangladesh's surge in capabilities from the early 1990s depends not on growth alone, nor even wholly on less poverty reduction and more, but also on resources devoted to social actions - by both government and NGOs - against disease, dirt, ignorance, and their underlying causes. These resource uses are probably complementary: ten billion taka (if well spent) will improve nutrition (and well-being) more if divided among stimulating growth, reducing poverty, raising rice yields, and advancing preventive and curative health, than if concentrated on just one or two of these. (g) Challenges: gaps in mass advanceAs shown, despite major progress in under-fives' nutritional status since 1990 by reduced stunting - explained mainly by better access to staple foods, and to oral rehydration - child wasting remains as alarmingly common as in the 1980s; and zinc and iron deficiencies remain widespread. More surprisingly, while child obesity is still rare, the proportion of adults overweight or obese climbed from 7% in 1980 to 17% in 2013. Linked to this are Type 2 diabetes and hypertension incidence: diabetes soared from 4% in 1995 to 10% (among over-35s) by 2013, with pre-diabetes affecting a further 22%; adult hypertension incidence rose from 11% to 15% in 1995-2010. It is commonplace to speak of the "double health burden": poor countries' need to cope with the high cost of treating non-communicable adult diseases linked affluence and lifestyle, while still burdened with the communicable diseases linked to child poverty. Those suffering from (often avoidable) lifestyle diseases are usually better off - and more urban and articulate - than those suffering from dysentery and malaria. They are thus likely to capture health funding that could, perhaps, be more cost-effectively spent on the inexpensive measures needed to improve the nutritional status of the poor.However, the double burden" approach is not the best way to approach the problem. Though overweight and obesity, and the resulting increased incidence of Type 2 diabetes, first appear in developing countries among the highest income deciles, over time they work their way down the income scale, and soon come to be concentrated among the poor. Furthermore, foetal and childhood food deprivation and stunting appear to predict higher risk of cardiovascular disease in old age - and of its frequent precursor, repeated infections in middle age. Attacking childhood (and foetal) undernutrition and micronutrient deficiency - and the poverty that helps to cause them - may in the long run be the most cost-effective way to reduce incidence of diseases of old age. As for poverty itself, while Bangladesh has accelerated the rate of reduction of $1.25/person/day consumption poverty, it still affected 43% of the population in 2010. And a large majority - 86%, as against 95% in 1990 - were below the higher poverty threshold, $2.50/person/day. The case for getting more resources, including land control, to these people remains extremely strong. I hypothesis that regions and groups experiencing slower falls in total fertility rates, slower growth in rice yields, and/or more unequal distribution of farmland did worse in cutting poverty, and that causal relationships can be verified. If so, that makes a strong case for maintaining, speeding, and concentrating on less-affected groups and regions, efforts to reduce child mortality - the precursor of falling TFR - and to finance research leading to rapid and labour-intensive agricultural growth.Bangladesh is exemplary in terms of documents integrating nutrition and poverty reduction with policies for agriculture and health. I don't know how effectively this has been co-ordinated, associated with authority to act, and implemented. A handful of countries (Canada, Norway, Sri Lanka) have addressed these issues. Mine, Britain, has a great record of official talk and a terrible record of action.Turning to structural transformation, how can the gaps be filed, and the process maintained and accelerated, especially by shifts of workforce into rural and urban manufacturing and service activities? Faster growth in agriculture's total factor productivity, and in labour and small-farm incomes, is the key to accelerating rural non-farm growth (fn. 36): the latter has been much supported by several NGOs (BRAC, Grameen, Proshika). Another gap is the persistent concentration of non-farm activities on RMG and migration for remittances - both susceptible, at least in some sub-sectors, to sluggish worker rewards, unsafe and even non-voluntary working conditions, and high sensitivity to a 'race to the bottom'.(h) Challenges: sustaining mass advanceWith well over 40% of workforce still mainly engaged in agriculture, continued growth in its productivity remains essential for labour and savings to flow into the non-farm economy. Also, continued rapid workforce growth - to be a source for GDP growth rather than rising unemployment - requires that the extra workers find productive workplaces. To be affordable, many of these (though a falling proportion) will continue to be in agriculture. However, keeping farm yields rising carries risks to sustainability. It is often said that cultivable land is shrinking at 1% per year. The causes of shrinkage are many, natural (soil erosion, nutrient loss, sea incursion, salinity) and unnatural (mining of inadequately priced or controlled groundwater, inadequately compensated and sometimes corrupt land grab for private non-farm development). A survey of 24 villages in 2001-8 showed 0.56% of land per year converted from farm to non-farm uses, reducing rice output by 0.86 to1.16% (fn. 42). Climate change is likely to worsen: sea incursion, as global levels rise; river-water uncertainty and flooding, as glacier melt accelerates; and yield, due to greater heat and evapotranspiration. Investment in structures, including for water management, can mitigate these effects, but is costly. There is very little unused, economically arable land; the scope for further increase in irrigated area or cropping intensity is small; and, while average NPK/hectare (rather above 150kg) is far below China's level, it is enough for worry about side-effects of further increases, including nitrate/nitrite pollution of drinking water. Further breeding for salt, heat and submergence tolerance can help. Breeding may even turn too advantage rises in atmospheric carbon dioxide. BRRI's world-class capacities, together with international inputs, are well placed for innovation in these areas. Is national and international public spending on agriculture, especially research and investment, enough to meet these threats, and thus to continue increases in food supply and employment sufficient to sustain recent rates of mass advance? Agriculture in 2013-14 was budgeted Tk Cr 12,275, and water resources (many agriculture-related) 2593 Cr, together 8.8% of total expenditure (excluding interest and other fixed liabilities) of Tk 168,652 Cr. Of this sum, about Tk Cr 9000 was allocated to agricultural subsidies. I know far too little about the local detail of Bangladesh's complex agriculture to suggest that this is too much, too little, or just right. (Agriculture employs around 45% of the workforce and produces around 18% of GDP.) However, the 73% of the agricultural budget devoted to subsidies seems high from the standpoint of agricultural sustainability. Some of this subsidises labour-displacing equipment; much of it subsidises fertilizer and irrigation, whose over-use has external costs, especially downstream and over many years. Such spending is to some extent at the cost of agricultural research and investment needed, among other things, to mitigate the effects of climate change.Mass advance based on agriculture, where the large majority of the very poor work, can come only from growth or distribution (e.g. less land grab or more land reform). There is no 'soft third option' in cutting household or merchant post-harvest losses of rice, which are quite small, contrary to a widespread belief. Agricultural growth will come mainly from improved varieties. As the global 'innovation shadow' from the dwarfing and other breakthroughs of the 1960s continues to shorten, major gains will increasingly require transgenics. As for demand, rice, dominates calorie intake (60% urban, 70% rural) as well as farm output and value-added (70-75%), but Bangladesh is about self-sufficient. As population growth slows, further expansion of demand for farm products requires removal of foodgrain export bans and/or diversification of demand - inducing diversification of cropland - out of rice; but into what? The only substantial current plantings are potatoes, jute, and all vegetables, all fruit and all oilcrops. Each occupies only 3-4% of gross area under rice. Vegetables - with high value-added, labour-intensity, income-elasticity of demand, and dietary advantages - are suitable for micro-farmers with water control, especially in urban peripheries and home gardens. (i) How did Bangladesh do it? Two more Fs: facts and freedomWe have some idea in Bangladesh which economic sectors, places, and groups are performing well, normally and badly, or better or worse than before; and, often, to test explanations. Data are plentiful and usually fairly reliable, or at least with, say, (roughly) known 5% confidence intervals. An economist who has recently worked with the 'poor numbers' available for, say, smallholder food output or national income in most countries of sub-Saharan Africa is dazzled by the wealth of serious data in Bangladesh. Gatherers, processors and users of data are usually trained, paid, numerous, and supervised. Where the data are in doubt, as with the last population census, a lively and open debate gives hope of correction: data are useful only when seriously, and freely, contestable.Some basic freedoms of speech and expression, then, are necessary for this essential access to basic economic facts. The use of data before, during and after policy experiments, the scaling-up of such experiments, and the freedom of NGOs to take part in all this, are crucial features of Bangladesh's accelerating mass advance. The role of Matlab, subsequently ICDDR - first in testing and promoting contraceptives, later in pioneering and implementing widespread use of oral rehydration - is a striking example. So is the Bangladesh Academy for Rural Development, Comilla, and the Grameen Bank that, in part, grew out of it, as a major source of experiments in rural poverty reduction, including non-farm micro-credit based on peer monitoring of lending. The Bangladesh Rural Advancement Corporation - now the world's largest NGO - has played a huge role in experimenting with, and then bringing to scale, new forms of basic education (it runs over 6 per cent of schools and has led in education for girls, who have now overtaken boys in numbers and performance) and micro-business management. More radical is Proshika, focusing on training and support for co-operative production by groups of about 20. BARD, Grameen and Proshika each have many million participants. These NGOs share focus on, and much control by, women; a largely rural (but increasingly international) sphere of action; and a participatory ethos. All are really "semi-non-governmental", with extensive independence, but some funding (perhaps 10-20%) from government, and sometimes more from international organizations, to permit more expansion and experiment.As important, Government has to some extent allowed its own technical specialists and experienced laypersons to develop and implement policies in rural financial services, agricultural research, family planning, and nutrition. NGOs cannot substitute for government and should not try; but they can be complements, and sometimes policy leaders and experimenters. NGOs are only part of the answer to the puzzle of how - with a terrible inheritance, in a uniquely difficult and risky environment, and with often problematic governance - Bangladesh has achieved its accelerated mass advance. Solving that puzzle involves explaining why things got so much better from the early 1990s. This timing suggests that, while liberalization played a part, more important was focus on, resources for, and a certain degree of NGO-assisted depoliticization of, the food-fertility-farming nexus. Further pro-poor structural transformation and growth continue to require policies for that nexus. Bangladesh commenced improvement in all these fields in the 1980s, and has sharply accelerated it since 1990. Yet poverty incidence and child nutritional wasting remain among the most serious in the world for a large population; so probably do the land-water challenges to agriculture. Climate change and pollution present special problems in a very densely populated and urbanizing country. Alongside industrialization and urbanization, these challenges may need transformed or new institutions. Certainly they need continued freedom of enquiry and experiment.(j) Must progress on food, farms and fertility now give way to different sources of mass advance? I have argued that Bangladesh's sharply improving performance - given an open policy debate and decent, improving data - is mainly due, not to liberalization (though aspects of that were very helpful), but to State-led and NGO-supported actions on food, fertility and farming. This is persuasive until about 2000-10, but, many would claim, not subsequently: developmental success has changed both priorities and paths to achieve them. As for priorities, agricultural proportions of GDP and employment have fallen, and the easier paths to growth have been exhausted; total fertility rates had fallen to 2.2, about replacement level, by 2010-15, leaving hardly any scope for further reductions; and, while bad nutrition remained serious and widespread, its incidence and severity had fallen and the means of further improvement had become costlier. As for methods, the most promising State and NGO actions on physical and social infrastructure had been taken by c.2000, when emphasis shifted towards liberalizing, i.e. relaxing or removing, State-imposed constraints and distortions; that shift is still needed.Nevertheless, there remains much scope for State and NGO actions affecting the "three f's"; and the scope for liberalization, while real (especially for agriculture), remains limited, and its outcome - especially for income distribution - contested. Farming is still the main occupation for around half the workforce, a substantial secondary occupation for many more, and a promising source of affordable and productive workplaces for a population aged 15-64 still growing at 2% yearly. In particular, agricultural growth is needed to accommodate the welcome rise in female participation, especially informal, and hence reliant, for decent wages as well as working times, on growth in agriculture (and rural processing). While much farm growth will come from diversification, large parts of rice area are still sown to traditional varieties, with almost no transgenics; with important recent developments in submergence tolerance, scope for higher rice yields surely remains.As for fertility, when TFR was 6.9 around 1970 the scope for policy was clearly much more than now, when it is (supposedly) 2.2. However, the poor, the rural, the remote, and the uneducated have much higher TFRs; reducing these (for which lower child mortality is a precondition) will not only attack poverty and malnutrition, but will also prevent subsequent gradual rises in national TFR and hence, later, in the birth-rate and the rate of population growth. Moreover, while TFR of 2.2 leaves population only self-replacing in Europe, that is not so in Bangladesh: population growth in 2010-15 is 1.2%, and is not projected to fall to 0.5% until 2035-40, with TFR at 1.72. Accelerated falls in TFR would therefore not mean population decline, while they would further improve dependency rates and female participation. A more radical approach might enquire whether it would be worrying even if population eventually fell: some (neurotic?) Europeans worry about this prospect, but Bangladeshis, given their dense population and their sparse, threatened environment and resources, perhaps need not. In brief, further sharp lowering of TFR among the rural, the remote and the poor still represents a 'bargain sector' for Bangladesh. Clearly the proportions of children substantially stunted and wasted (see above), though reduced, still demand policy attention; and that reduction has exposed other serious nutritional issues (micronutrient deficiencies, obesity). But can policy make progress on these issues, as it has started to do with child undernutrition? That improvement was not a low-hanging fruit: progress has been fairly recent, and North India still struggles to improve on very slow progress, similar to Bangladesh's before 1990. Bangladesh has some moderately successful micronutrient interventions, but their limitations suggest casting the net more widely, perhaps to biofortification. The argument about the extent to which Bangladesh has diversified its routes of mass advance away from the "three fs", and should do so in future, will continue. Nobody should denigrate alternative routes, most obviously RMGs and remittance incomes, which since 1990 (and increasingly since 2000) have taken many Bangladeshis out of poverty. However, neither problems (notably worker conditions), nor threats from newcomer competitors, in these areas will be simply resolved. Obviously, the "flying geese" of economic development will, as time goes by, fly to new pastures. The role of the State includes helping the private sector to identify these; making the flight less bumpy, and more accessible to the poor; and probably building some appropriate infrastructure (analogous to "airports"). Almost no economist believes that only "getting the prices right" and minimizing regulations will work the magic for Bangladesh: - it certainly was not the path pursued by the Asian tigers, early and medium-sized like China,Taiwan and South Korea, or late and gigantic like China. But nor do many economists defend some of the remaining illiberalisms in Bangladesh (and India), such as the huge proportions of State agricultural spending devoted to farm input subsidies, at the inevitable expense of investment in physical and market infrastructure and in research. Michael Lipton, Sussex University, Brighton, EnglandRevised draft 1 August 2014 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download