Skip to main content

Quantifying the costs of drought: new evidence from life satisfaction data

Abstract

We estimate the cost of droughts by matching rainfall data with individual life satisfaction. Our context is Australia over the period 2001 to 2004, which included a particularly severe drought. Using fixed-effect models, we find that a drought in spring has a detrimental effect on life satisfaction equivalent to an annual reduction in income of A$18,000. This effect, however, is only found for individuals living in rural areas. Using our estimates, we calculate that the predicted doubling of the frequency of spring droughts will lead to the equivalent loss in life satisfaction of just over 1% of GDP annually.

This is a preview of subscription content, access via your institution.

Notes

  1. 1.

    Blomquist et al. (1988) used variation in wages and housing across regions in the USA to ascertain the compensating wage differential for many socioeconomic and climate conditions, which, in principle, means one gets an estimate of the long-run willingness to pay or accept conditions. In that type of analysis, however, one needs to fully account for differences in economic opportunities across regions, to separate the compensating variation in wages from unobserved individual heterogeneity.

  2. 2.

    The Australian Bureau of Meteorology (http://www.bom.gov.au/climate/glossary/climate.shtml) states that: “the weather of any place refers to the atmospheric variables for a brief period of time. Climate, however, represents the atmospheric conditions for a long period of time and generally refers to the normal or mean course of the weather.” We refer to drought as a climate event, rather than a weather event, because we are using measured rainfall over a period of time (most commonly seasonal rainfall) in our regressions below.

  3. 3.

    Australian Prime Minister John Howard, media release, 9th December 2002 (http://www.pm.gov.au/news/ media_releases/2002/media_release2025.htm). The Australian Bureau of Meteorology described the 2002-03 drought as “one of the worst on record” (see: http://www.bom.gov.au/climate/drought/).

  4. 4.

    Horridge et al. (2005) also estimated that multiplier effects, which could have occurred in urban or rural areas from the drop in agricultural production, would have resulted in GDP being 0.6% further below what it would have otherwise been. However, as our primary interest is in the direct effects on rural communities, the comparison figure we use is the direct effect of drought on GDP.

  5. 5.

    An important point to note here is that although droughts are costly to rural communities, this does not directly imply that individuals should be compensated for this cost. Climate phenomena are natural events that, to some extent, are part and parcel of agricultural production and part of the normal ‘business-risk’ of an agricultural enterprise for which the usual equilibrium argument of Blomquist et al. (1988) applies. Whilst one can make the case that capital market imperfections should mean a government takes over an insurance role, the case for compensating farmers for droughts seems weak from an economic point of view.

  6. 6.

    More details about the data can be found at http://acqol.deakin.edu.au/index_wellbeing/index.htm.

  7. 7.

    The income measure used in this paper is derived from a question where individuals were required to give their total household income gross of tax in the previous 12 months. The income measure was deflated back to the May 2001 level using the Australian Bureau of Statistic’s CPI. One limitation is that in our survey, like in many such surveys, income levels were given in bands. Subsequently, we simply took the midpoint of each band to calculate our log household income measures. The original income bands were: <$15,000, $15,000–<$30,000, $30,000–<$60,000, $60,000–<$90,000 and $90,000+. We have also examined the robustness of our main result by using a dummy variable for each income band to the find the effect of drought on life satisfaction (holding income constant) and found the result unchanged. The results are also robust to the amounts used to define the lowest and highest bands.

  8. 8.

    A postcode is a four-digit number used by Australia Post to assist with mail delivery. There are 2,435 postcodes with valid street delivery area (giving a mean number of people per postcode of approximately 8,000). The size of the sample in each postcode in each sample is usually quite small. The range of the sample varies from 0 for some postcodes in some quarters to 21 with a median of two people (where the postcode has people in the sample).

  9. 9.

    Where there was missing data for the rainfall collection station, the next nearest rainfall collection station to the centroid of the postcode was used.

  10. 10.

    The Australian Bureau of Meteorology recognises that it is the State Government of Australia’s responsibility to declare drought. Therefore, in Australia, the official definition of drought is an administrative measure that considers the meteorological, hydrological and pastoral drought concepts.

  11. 11.

    The key issue here is that very detailed geographical information is required to link the drought data to the life satisfaction data. As the level of detail that rainfall is collected in Australia, it is possible to map rainfall data to the postcode level. To the authors’ knowledge, however, administrative, agricultural and hydrological drought measures cannot be matched to the same level of geographical detail.

  12. 12.

    It should be noted that northern Australia has quite different rainfall patterns to the rest of Australia, and therefore, we would expect quite different seasonal results there than from the rest of Australia. As most of the Australian population is located in southern Australia, the results presented in this paper will be primarily influenced by the rainfall patterns there.

  13. 13.

    An anonymous referee pointed out that the marginal rate of substitution (MRS) of income for other variables (such as climate) is an ordinal concept. That is, if true latent life satisfaction is a monotonically increasing function of the discrete observed variables (as is assumed in ordered choice models), the MRS should be unaffected by the details of the function.

  14. 14.

    The annual rainfall specification is not shown, but is available from the authors upon request.

  15. 15.

    There are several reasons why this should not be our preferred definition of drought. In particular, the level of the 10th percentile will vary widely between northern Australia and the rest of the country. An absolute measure of rainfall is more important because Australia’s primary production tends to be reliant on a relatively narrow range of production (such as pastoral farming, and certain types of crops). This means that a similar level of rainfall and water stocks will be required.

  16. 16.

    We also investigated whether the influence of weather differed by region by using an interaction between the weather variables and whether the region concerned was in Queensland and the Northern Territories (which are in the north of Australia where a monsoon climate prevails). We found no significance of such interaction terms. The limited additional explanatory power of additional variables probably shows that our data can only pick up the large climate effects and are not suitable for a more disaggregated analysis.

  17. 17.

    We are grateful to two anonymous referees for the suggestion to undertake these robustness checks.

  18. 18.

    Its worth pointing out that in the case of a truly exogenous regressor, the mere act of averaging that regressor within categories does not bias its coefficient in a linear regression as long as the mean differences between those categories correspond to the true mean differences. Indeed, averaging would increase the coefficient if there is measurement error in the original series, as averaging overcomes measurement error bias.

  19. 19.

    A wide 95% confidence interval was calculated for this effect. It was calculated using the 95% confidence intervals on income and drought in spring separately. Using this method, the 95% confidence interval on the drought quantification ranges from $8,300 to $25,000.

  20. 20.

    The expected well-being cost of climate change is an upper estimate. If there is indeed a doubling of dry springs, we would expect that that there would be out-migration from these regions, and, within the regions, that the industry composition would change and that alternative production methods would be employed (for example, there may be a drop in the harvesting of water-intensive crops, such as rice and cotton). In addition, part of the large costs associated with dry springs is likely to be transitional as people’s expectations change about their futures.

References

  1. Bin O, Polasky S (2004) Effect of flood hazards on property values: Evidence before and after Hurricane Floyd. Land Econ 80:490–500

    Article  Google Scholar 

  2. Blomquist G, Berger M, Hoehn J (1988) New estimates of the quality of life in urban areas. Am Econ Rev 78:89–107

    Google Scholar 

  3. CSIRO (2001a) Climate change: impacts for Australia. CSIRO, Canberra

  4. CSIRO (2001b) Climate change: projections for Australia. CSIRO, Canberra

  5. Ferrer-i-Carbonell A, Frijters P (2004) How important is methodology for the estimates of the determinants of happiness? Econ J 114:641–659

    Article  Google Scholar 

  6. Frey B, Stutzer A (2002) What can economists learn from happiness research? J Econ Lit XL:402–435

    Article  Google Scholar 

  7. Frijters P, Van Praag B (1998) The effects of climate on welfare and well-being in Russia. Clim Change 39:61–81

    Article  Google Scholar 

  8. Greene W (2003) Econometric analysis: 5th edn. Prentice Hall, New Jersey

    Google Scholar 

  9. Horridge M, Madden J, Wittwer G (2005) The impact of the 2002–2003 drought on Australia. J Policy Model 27:285–308

    Article  Google Scholar 

  10. Intergovernmental Panel on Climate Change (IPCC) (2001) Climate change: the scientific basis. IPCC, http://www.unep.ch/ipcc

  11. Kahneman D, Diener E, Schwarz N (eds) (1999) Foundations of hedonic psychology: scientific perspectives on enjoyment and suffering. Russel Sage Foundation, New York

  12. Maddison D, Bigano A (2003) The amenity value of the Italian climate. J Environ Econ Manage 45:319–332

    Article  Google Scholar 

  13. Managing Climate Variability (MCV) Program (2004) The adequacy of research on climate variability and its impacts on agriculture. A submission to the Australian Drought Review Panel

  14. Rehdanz K, Maddison D (2005) Climate and happiness. Ecol Econ 52:111–125

    Article  Google Scholar 

  15. Troy A, Romm J (2004) Assessing the price effects of flood hazard disclosure under the California Natural Hazard Disclosure Law (AB 1195). J Environ Plan Manag 47:137–162

    Article  Google Scholar 

  16. Van Praag B (1988) Climate equivalence scales, an application of a general method. Eur Econ Rev 32:1019–1024

    Article  Google Scholar 

  17. Van Praag B, Baarsma B (2005) Using happiness surveys to value amenities: the case of airport noise. Econ J 115:224–246

    Article  Google Scholar 

  18. Welsch H (2002) Preferences over prosperity and pollution: Environmental valuation based on happiness surveys. Kyklos 55:473–494

    Article  Google Scholar 

  19. Welsch H (2006) Environment and happiness: Valuation of air pollution using life satisfaction data. Ecol Econ 58:811–813

    Article  Google Scholar 

Download references

Acknowledgements

The authors would like to thank Alison Booth, John Freebairn, Andrew Oswald and two anonymous referees for helpful comments.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Nick Carroll.

Additional information

Responsible editor: Junsen Zhang

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Carroll, N., Frijters, P. & Shields, M.A. Quantifying the costs of drought: new evidence from life satisfaction data. J Popul Econ 22, 445–461 (2009). https://doi.org/10.1007/s00148-007-0174-3

Download citation

Keywords

  • Drought
  • Life Satisfaction
  • Income-equivalence

JEL Classification

  • I31
  • Q10
  • R10