Milan Brahmbhatt, World Bank
Institut Pasteur – Paris, France
June 29, 2006
|> Presentation (166k pdf)- Follow Milan Brahmbhatt's speech with this presentation|
Let me start my presentation today with an obvious question. Why is an analysis of economic and social aspects relevant in what appears to be mainly an issue for animal and human public health policy? The more straightforward answer is of course that the animal and human illness and death caused by disease can have large economic and social costs. Till now, with the principal transmission of the H5N1 virus occurring between animals, and, to a very limited extent from animals to humans, the principal costs have been felt in the rural or commercial poultry sectors of affected economies. However, as these outbreaks continue and spread to new regions, they also increase the probability of a second stage, with human-to-human transmission and a global influenza pandemic, with enormously greater costs on a world scale.
The less obvious part of the answer is that disease prevention and control strategies themselves have various economic and social costs and impacts. The success of these strategies can sometimes be affected by their impact on the economic incentives facing people. There is thus a natural role for economic analysis of costs, benefits and incentives in helping frame effective health strategies.
Avian Flu Goes Global
Until recently highly pathogenic avian influenza H5N1 virus had been concentrated in East Asia, where some 10 countries had experienced outbreaks since late 2003, the most serious in Vietnam, Thailand, Indonesia and China. Then, in the last 6-9 months, the virus has gone global, spreading to over 40 more countries. In Western and Central Europe the majority of these new outbreaks have been among wild birds. But elsewhere, in South Asia, Central Asia, the Middle East and Africa, nearly all the new outbreaks have been among poultry. Among the many things not fully understood about the disease is its mode of transmission. However it now appears that both wild birds and domestic poultry are involved in transmission, the latter through poultry trade, both legal and illegal or informal. While wild birds seem to have been the main channel in Europe, in Africa trade seems to have an important factor.1 In Nigeria, for example, the virus seems to have arrived through import of chicks from abroad by a large industrial poultry farm.
The number of human infections and deaths reported to WHO has also accelerated in the last six months. [Slide 2] There were 41 deaths in all of 2005, but 54 in only the first half of 2006, more than twice the pace of last year. However, it is not clear if this pick up is due to acceleration in underlying numbers, or reflects an improvement in reporting to the WHO, or both. The number of countries in which deaths occur has also constantly risen, from one in 2003, to two in 2004, five in 2005 and 7 in the first half of 2006.
Till recently the economic impacts of the avian flu were concentrated in East Asia, but are now also spreading more widely. In most economies the impact at the macroeconomic level has been relatively limited so far, mainly because the poultry sector is a relatively small part of the world economy. As Slide 3 indicates, overall growth in many East Asian economies has in fact accelerated over the last three years, for example in Vietnam, Indonesia, Cambodia and Laos, or has continued to run at high rates, as in China. Among the AI affected economies, growth has slowed only in Thailand, in part because, as we will show, the economic impact of avian flu was bigger there than elsewhere, although a number of other factors have also contributed to the slowdown. In addition, even though the macro impact on most economies other than Thailand has been limited, the impact on specific regions and communities has been more severe.
Several types of economic impacts or costs of AI can be distinguished. First there are direct production costs because of losses of poultry, due to the disease and to control measures such as culling birds, with impacts extending not only to farmers but also to upstream and downstream sectors such as poultry traders, feed mills, breeding farms etc. According to FAO, over 200 million poultry have died or been culled since the end of 2003, mostly in East Asia. The largest declines have occurred in Vietnam and Thailand, where they were equal to 15-20% of the stock of poultry. Other relatively smaller losses of poultry have also occurred in other economies. The size of the poultry sector in the national economies of the region before the epidemic ranged from a little under 1 percent of GDP at the low end in countries like Vietnam, to a high of a little over 2 percent in the Philippines, with most countries centering a little over 1 percent of GDP. In an economy like Vietnam, where poultry output was down by around 20 percent, this part of economic loss was worth about 0.1-0.2 percent of GDP. Additional losses have occurred because of lower egg production and reduced activity in distribution channels.
There are also secondary or indirect impacts related to sharp shifts in market demand which result primarily from spontaneous efforts by consumers to reduce their subjective or perceived probability of becoming infected, as well as from trade restrictions on poultry trade imposed by governments seeking to prevent transmission of the disease. In so far as consumers switch from domestically produced poultry to other domestically produced foods, there ought not to be a large impact at the overall macroeconomic level, although there will be severe impacts on the poultry sector itself. In Romania, for example, which has suffered over 100 AI outbreaks over recent months, domestic poultry sales have fallen by 80%, bringing many producers to the verge of bankruptcy. In Iraq only 10% of semi-commercial farms remain operational, and there have also been large losses in Turkey. In France, Europe’s leading poultry producer, producers hit by sharply lower demand reportedly lost 40% of their income in the first quarter of 2006. The poultry feed sector in Europe, which accounts for a turnover of 42 billion USD, has been hit by the avian influenza crisis, with a 40% reduction in demand for poultry feed in some European Union countries. The FAO now projects an 8-9% fall in European poultry consumption this year, a huge number, which is contributing to sharply lower prices and poultry producer incomes worldwide. Thus even in Brazil, which has not experienced an outbreak of the disease, weakening world demand and lower prices have induced the main suppliers to reduce production by 15% this year.
Thailand, which is the only large net exporter of poultry in East Asia, had already experienced a sharp 40% fall in poultry exports in 2004 due to import restrictions in foreign markets on its uncooked ,poultry exports, [Slide 4] which likely doubled the overall macroeconomic costs it experienced in that year. However Thailand also provides an example of recovery based on strong policy responses. Together with Vietnam, strong control measures (e.g. culling of infected flocks) have resulted in no new outbreaks of AI for the last 6 months. Exporters have managed to switch from uncooked to cooked poultry exports, which are not affected by trade restrictions, as a result of which exports began rebounding last year. Domestic consumption has also been reviving due to increasing consumer confidence in the safety of cooked products. In fact, as we know, there is not much health hazard from eating chicken if it is properly cooked. Thus the sharp falls in market demand we are seeing are based to some extent on consumer misperceptions and panic. Hopefully, such effects should be short-lived, as consumers learn the true situation, a process which can be fostered by a transparent and credible public communications strategy.
We close this discussion of market demand effects, by taking a look at the impact of AI on tourism. [Slide 5] Avian flu could have a much bigger macroeconomic impact if there is a fall in international tourism to infected countries because of disease fears or travel restrictions, as happened during SARS. This does not appear to have occurred so far, however, with tourist arrivals in China, Vietnam and Thailand appearing to continue trending higher despite AI. Tourist arrivals to Indonesia have been without trend for several years, because of a series of adverse shocks such as September 11, the Bali bombing, SARS, the tsunami and the second Bali bombing. It is difficult to tell how far the current bad publicity on AI in Indonesia is having any additional adverse impact on tourism.
Finally, we note that the costs of prevention and control also need to be taken into account, including costs to the government of purchase of poultry vaccines, medications and other inputs, hiring workers for culling, cleanup, surveillance and diagnosis, and so on. Governments also face needs to at least partially compensate poultry owners, an important factor in persuading owners not to conceal outbreaks of bird flu.
It is also important to consider how the costs of avian flu will affect different social groups, such as poor rural households or small commercial poultry producers, a consideration closely related to the production structure of the poultry sector in different countries. As Slide 6 indicates, the proportion of poultry production undertaken by backyard and small commercial systems is much higher at lower levels of per capita income. In low income economies like Vietnam, where the bulk of poultry production is still by backyard producers, the impact has fallen mostly on individual rural households, and has only partly been offset by government compensation to farmers. Survey data show that in Vietnam the poorest quintile of households relies more than 3 times as much on poultry income than does the richest quintile, so there are also adverse distributional effects. Slide 7 shows an analysis developed by researchers at FAO, showing that income from poultry is much more equally distributed than overall income.2 Reductions in poultry income due to avian flu or to avian flu control strategies will thus tend to worsen income distribution in Vietnam. One possible strategy that has been discussed is simply to ban backyard poultry production and to concentrate poultry production in large industrial and commercial units. Slide 8 shows the likely impact of such a ban on the incomes of 600 representative Vietnamese households grouped from left to right according to income. Clearly such a policy would disproportionately reduce the incomes of the poor. On the other hand, in economies like Thailand, where more of production is undertaken by industrial and large commercial producers, the economic shock of avian influenza is likely to fall more on these large producers, with impacts which would be felt not on rural farmers’ incomes, but through greater unemployment of wage laborers, lost profits and corporate bankruptcies.
Threat of a human influenza pandemic
As avian flu outbreaks continue and spread to other regions, infecting a growing number of people, they also increase the probability of a human influenza pandemic, with enormously greater human and economic costs on a world scale. This, after all, is why avian flu is attracting such intense concern. The WHO observes that “Best case scenarios, modeled on the mild pandemic of 1968, project global excess deaths in the range 2 million to 7.4 million. Other estimates that factor in a more virulent virus, similar to that responsible for the deadly 1918 pandemic, estimate much higher numbers of deaths. Both scenarios are scientifically valid. The differences arise from the assumptions about the inherent lethality of the virus, which past experience has shown to vary greatly.”3 Other experts go further to argue that clinical, epidemiologic, and laboratory evidence suggests that a pandemic caused by the current H5N1 strain would be more likely to mimic the 1918 pandemic than those that occurred more recently, resulting in far higher death totals.4
Given great uncertainties about the timing and potential lethality of a future human pandemic, the best that economists can do for now is develop various more or less detailed scenarios, based on assumptions about transmission, attack and fatality rates, as well as analysis of potential channels of economic transmission. Slide 9 shows three scenarios from a recent study by McKibbin and Sidorenko, with excess mortality (above the number of deaths in a normal season of 0.2-1.5 million) on one axis, and the associated loss of GDP on the other.5 One of these scenarios is modeled on a ‘mild’ scenario similar to the 1968 Hong Flu epidemic (resulting in a little over 1 million excess deaths and a loss of a little under 1 % of GDP. At the other extreme is a “Severe” scenario based on the 1918-1919 “Spanish Flu” (resulting in over 70 million excess deaths and a loss of 4-5% loss of GDP).
What are the main types of economic impacts one might expect to see in a pandemic scenario? Slide 10 shows the results of some World Bank modeling of a “severe” scenario, showing the size of three different types of impact, as well as results for different regions of the world . 6
Interestingly, the most immediate and largest economic impact of a pandemic might arise not from actual death or sickness but from the uncoordinated efforts of people to avoid becoming infected. This at least was the experience during SARS, when people tried to avoid infection by minimizing face-to-face interactions, resulting in a severe demand shock for services sectors such as tourism, mass transportation, retail sales, hotels and restaurants, as well as a supply shock due to workplace absenteeism, disruption of production processes and shifts to more costly procedures. To these results of private action could be added economic disruption and costs caused by emergency public policy measures such as quarantines and restrictions on domestic and international travel and trade, resulting in – among other things - a breakdown, at least temporarily, of international and domestic supply chains and logistics.
The World Bank study assumes 20% declines in demand for tourism, transportation and other key services. It estimates that these kinds of effects could lead to a loss of close to 2% of GDP for the world as a whole. This is very similar to the economic losses of this type associated with SARS, which were estimated at perhaps 2 percent of East Asian regional GDP in the second quarter of 2003, even though only about 800 people ultimately died from this disease. Note that a 2 percent loss of global GDP during an influenza pandemic would represent around $800 billion over a whole year.
In addition to these immediate costs of disruption, a serious global flu pandemic could also entail a sizeable loss of potential world output through a reduction in the size and productivity of the world labor force due to illness and death. The effect of disease on the size of the labor force would depend on the virulence and spread of the disease and on how it affected different age groups, among other factors. There would also be a general decline in labor productivity due to illness among the labor force at large, as well as costs of hospitalization and medical treatment.
There is a lack of detailed studies of what these mortality and morbidity costs of a flu pandemic might amount to at a global level. However one 1999 study of the United States calculated that a relatively mild epidemic similar to the 1968 epidemic could lead to between 100000 and 200000 deaths in the US, together with 700000 or more hospitalizations, up to 40 million outpatient visits and 50 million additional illnesses.7 The study by McKibbin and Sidorenko tries to model how mortality might vary around the world according to two sets of factors. The first is how easily the virus might reach and travel around a country, based on information about the country’s participation in international air travel, and on its population density. The second is an index for the quantity and quality of health services available in a country, which are obviously much poorer in developing countries than in developed ones.
In the “Severe” scenario, these assumptions lead to an average attack rate of around 35% for the world as a whole, as well as a case fatality rate of around 3%, resulting in a 1% mortality rate, or about 70 million people, resulting in a 0.4% decline in world GDP. Mortality rates are much higher in developing economies though, resulting in an economic loss twice as high in those economies than in developed economies. Finally, there are economic losses due to the enormous morbidity typically associated with influenza, leading to absenteeism, school closing, declining productivity and crowded hospital emergency rooms. The short run costs of morbidity are expected to be more than twice as large as those for mortality, around 0.9% of GDP. Putting it all together, the World Bank study arrives at a total loss of 3.1% of world GDP in a severe scenario. All told, given a world GDP of around $40 trillion, the World Bank and McKibbin and Sidorenko studies suggest that global economic losses in a “Severe” scenario could be in the range of $1.25-2 trillion. (It goes almost without saying that these broad scenarios are not meant to be forecasts; they are only exercises to help think through the various channels of impact and possible orders of magnitude).
Looking at impacts at the level of the individual firm, it seems clear that businesses that are the most resilient to extreme volatility are the ones more likely to survive a pandemic, or to survive it with relatively less financial damage. Thus it would be firms with strong balance sheets and capitalization that are more likely to survive steep downturns in demand and cash flow that could last from 6 months up to 2 years, especially in the sorts of services sectors noted above. Bankruptcies would likely surge among highly leveraged firms, and in particularly exposed sectors, such as airlines. Governments may need to plan for policies to facilitate corporate restructuring and getting industries relaunched promptly after the crisis. Firms also need to plan for all sort of unexpected disruptions in supply chains, logistics and the availability of key business services, for example business travel. It would make sense for firms to work with their existing suppliers on tactics to cope with potential supply chain disruptions, while also investigating alternative sources of supply and alternative means of doing business. Firms could also consider building up buffer stocks of critical inputs. That may eat into short term profit margins and runs against current “just-in-time” management philosophy, but could help ensure the very survival of the firm during this type of catastrophic event. Finally, firms obviously also need to plan for mitigating the impact of the epidemic on their own employees, working in close concert with the public health authorities
Let me turn finally to policy issues, to policy actions that are being undertaken to combat avian and human pandemic influenza, to some emerging lessons as well as questions that need further study. The preceding discussion establishes that the benefits of preventing or even mitigating or delaying a global influenza pandemic are likely to be large indeed, which has motivated an unprecedented collaborative effort between governments, multilateral development banks and key technical agencies such as the World Health Organization, Food and Agriculture Organization, and World Organization for Animal Health. A global meeting in Geneva in November 2005 identified key components of a global action plan to control avian influenza in animals and simultaneously limit the threat of a human influenza pandemic. In January 2006 an International Ministerial Pledging Conference in Beijing secured pledges of some $1.9 billion in grants and credits from bilateral donors, the European Commission and multilateral development banks, including a new World Bank funding facility of $500 million, the Global Program for Avian Influenza Control and Human Pandemic Preparedness and Response (GPAI).
These resources are made available for the following purposes:
• Developing countries, for integrated country programs;
• Regional organizations, for a range of technical assistance, stockpiling, and coordination activities;
• International technical agencies at the global level; and
• The AHI Facility, a multidonor trust fund facility based at the World Bank, for provision of grants to country, regional, or global recipients.
By the end of April 2006, some $1,150 million had been committed, of which $331 million had been disbursed.
There are several key principles and guiding ideas that are important for the global effort on influenza. One very clear lesson is the need for governments to establish a track record of credibility through honest, accurate and timely disclosure of information to their own people and to the outside world. There is evidence that during SARS the costs arising from panic and disruption were magnified by an initial lack of public information in some countries. In the 1918 pandemic governments tried to hide the truth so as to maintain morale during wartime, although this contributed ultimately to the worldwide spread of the disease, loss of trust among the public and social chaos. We need to understand this problem better, and to consider what kinds of institutional mechanisms or incentives could help over come it.
Second, given the close links between animal and human health, and their ties to economic and social considerations, country level responses need to be based on integrated cross-sectoral plans that draw on the best available knowledge and guidelines from the international technical agencies such as OIE, FAO and WHO. Preparation and implementation of these integrated plans needs to bring together the several concerned ministries in a country, such as agriculture, animal heath, human health, finance, local governments and communities, all under strong political leadership from the highest level. The emphasis to be placed on specific elements of the country programs needs to be decided according to specific country circumstances.
Third, there needs to be careful evaluation of results from various strategies that are currently being implemented, to judge if actual outcomes are meeting expectations. One such question might be to gain more information on the relative efficacy and implications of different control strategies, for example culling and vaccination of birds. A linked question is that of the appropriate levels of compensation needed to make culling successful. Too little and farmers have an economic disincentive to comply with culling; too much and they may have an incentive to deliberately infect their flocks. Different countries have different compensation schemes. A cross-country evaluation of the results so far could throw a lot of light on this question. Obviously this sort of policy research needs to be complemented with scientific research to fill in more of the gaps in our scientific knowledge.
Fourth, it is also important to strike a balance between short and long term measures. Avian flu is becoming endemic in parts of East Asia and will require a long effort to suppress, which should however have a high pay-off, especially for the incomes of poor farmers and availability of low-cost protein for poor people in developing countries. In addition, a human pandemic may emerge not now but later, and not from H5N1 but from some different strain of flu virus. Other zoonoses and pathogens also continue to emerge. The general problem of global infectious diseases is going to be with us for a long time, and may get worse. Thus it makes sense to also undertake broader long term measures to strengthen the early detection, surveillance, institutional, regulatory and technical capacity of the animal health, human health and other relevant sectors. These will be valuable investments both in the short and long run.
Lastly, there are a critical set of questions partially grounded in economics that need to be solved at the global level, including how best to foster R&D and innovation of new types of vaccines and anti-virals, how to foster rapid growth in manufacturing capacity for anti-virals and vaccines (the latter once a human pandemic strain has been identified) and how to establish adequate stockpiles of vaccines and anti-virals that would allow ‘rapid response’ to a pandemic outbreak. In recent years economists have been developing many creative new ideas about how to accomplish these objectives more effectively and cheaply than before. Some of these ideas now need to be taken off the drawing board and considered for implementation, at least in prototype forms.
1FAO AIDE News. Issue No. 40. June 19, 2006.
2 Joachim Otte, FAO Pro-Poor Livestock Policy Initiative (PPLPI); David Roland-Holst, UC Berkeley &
Dirk Pfeiffer, RVC London. “HPAI Control Measures and Household Incomes in Viet Nam”.
3World Health Organization. “Avian Influenza: Assessing the Pandemic Threat.” January, 2005. WHO/CDS/2005.29.
4Michael T. Osterholm. “Preparing for the Next Pandemic.” New England Journal of Medicine. 352: 1839-42. May 5, 2005.
5Warrick McKibbin and Alexandra Sidorenko. “Global Macroeconomic Consequences of Pandemic Influenza”. Lowy Institute. February 2006.
6World Bank DEC Prospects Group: Global Development Finance 2006.
7Martin I. Meltzer, Nancy J. Cox and Keiji Fukuda. (1999). “The Economic Impact of Pandemic Influenza in the United States: Priorities for Intervention.” Emerging Infectious Diseases. Volume 5, No. 5. Sept.-Oct. http://www.cdc.gov/ncidod/EID/vol5no5/meltzer.htmxford Economic Forecasting.