Thursday, February 16, 2006

Economics of an Influenza Pandemic

The Australian Treasury have published a working paper 'A Primer on the Macroeconomic Effects of an Influenza Pandemic'. It is motivated by fears of Avian Influenza, SARS and so on.

The focus is admittedly on economic not health effects but the essential conclusion looks a bit odd to me. Using their TRYM model the paper finds that confidence effects and the large short-term withdrawal of labour promote most of the adverse economic effects. The suggestion: Adopt policies which restore confidence and which get people back to work.

Of course people would be staying home from work to avoid infection or to care for the sick so you would want to be careful about getting them back to work. In fact, staying at home might avoid high levels of overall infection and, after a transitional loss in income, might allow GDP to be restored more rapidly.

Deaths from the pandemic would raise GDP per capita and wages as claims on capital would accrue to survivors although (as the authors fail to note) there would be gains-from-trade losses. There would also be offsetting expenditures on health that would boost GDP and the possibility of a an expansionary fiscal response would further improve things. Thus long-run effects of the pandemic are likely to be milder than short-run effects. There would be big long-term effects on tourism and services.

This study contrasts with a much more pessimistic assessment of the effects of an epidemic by ANU researchers here.

'Up to 30 % cent of Australians could catch the deadly bird flu virus if there is a serious global outbreak this year, a landmark study by Australian experts predicts.The report forecasts that one per cent of the population, or 214,000 people, could die, along with 2.2 % of humanity – or 142 million people.

In addition to the potential human toll, the study says even a "mild" bird flu pandemic would have a major impact on the global economy.

The potential price tag on an extreme outbreak is around $4.4 trillion, with Australia forecast to lose 10.6 % of its economic output as the global economy shrinks by 12.6 %. '
(McKibbin & Sidorenko).

The McKibbin et al. paper makes specific assumptions about the parameters. The Treasury paper places emphasis on qualitative insights not numbers. Both papers supplement the health policy oriented discussion of pandemics prepared by the Department of Health and Aging.


Anonymous said...

There have been analyses of the Great Plague which have pointed out ameliorating (for the survivors). I think there may have been an NBER working paper recently (2004 or 2005) that made related points regarding the effects of AIDS on Africa

Mike Hart said...

The paper is interesting but essentially flawed in a number of areas. The mortality rate is selective, even so, other data about previous pandemics is incorrect. Over forty percent of the european population died in the Black Death (Tuchman 1978 and Westfall Thompson 1920, 1931) and probably about 50% or more of many Indian populations in the Americas.

The paper also attempts to model (inductively) a deduction hyptothesis that recent pandemics had low mortality rates therefore future ones will too. This deduction ignores empirical evidence from previous world pandemics, namely the bubonic plague episodes in europe in the 14th century and introduced western diseases into North and South America. There is also no shortage of good comparitive historical analysis on previous pandemics (Westfall Thompson 1920) which contradicts the mathematical model, therefore either the values ascribed or the functions are invalid. The bubonic plague came and went almost with a sine curve wavelength or interval of a decade before decreasing in amplitude and increasing over time until the pandemic was effectively over.

Taxation was increased substantially as revenues fell these fiscal impositions by various rulers in various countries gave rise to serious social instability and violence not a growth in GDP. Production slowed first, creating shortages of various goods, there were outbreaks of hyperinflation and rampant wage inflation (30% not uncommon). Politically the period was marked by repression and violence. In short the historical data suggests that the outcome of pandemics has been; economic chaos, social unrest, high prices, corruption and profiteering, breakdown of social constructs, serious decline in the production of goods and services, irrational expenditures, conspicuous consumption, and maladministration. This is not what the Treasury Paper model shows.

The paper also ignores the mortality rate of health workers irregardless of the worst case hypothesis for the entire population cohorts, because you work in the area does mean you don't get the diseases, simple fact. Biology not occupation determines immunity or resistance.

The treasury people should have done a little more research on previous pandemics and checked the data on previous pandemics.

Mike Hart said...

Final paragraph, a small correction, should have read...

just because you work in the area does not mean you don't get the disease...