Chinese cities have become renowned for their extremely hazardous levels of air pollution. This is not surprising – many developing countries experience environmental degradation as they undergo industrialisation, and cities like Delhi, Cairo and Karachi are even more polluted than Chinese cities (WHO 2014). The initial phases of development typically involve rapid urbanisation with high pressure on metropolitan centres, which, coupled with little environmental regulation, can naturally lead to mounting pollution levels (Dasgupta et al 2002).
For the urban population of the People’s Republic of China, however, the problem of air pollution is particularly challenging, because the government exercises tight control over the economy and the media, and thus not only on pollution regulation, but also on public information regarding air quality. All major media outlets in China are owned by the government (Shleifer et al 2003) and the country is among the worst in terms of media freedom – ranking 176th out of 180 countries (World Press Freedom Index 2015). Moreover, information about air pollution is difficult to acquire individually – gathering data is costly and requires specialised technology (monitors) and knowledge (epidemiological research). Visibility alone is not sufficient to determine when hazardous pollution peaks occur, because there are many other confounding factors such as humidity or wind speed that contribute to changes in visibility. For this reason, the majority of the Chinese population exclusively relies upon the government for information regarding pollution. A much smaller fraction of the population may have access to other information via other (e.g. foreign) sources accessible over the internet. It is interesting to consider how these two parts of the population compare.
In a recent paper, we explore how a government with control over pollution levels and the media would bias information, in order to keep the population satisfied with working in polluted cities (Ravetti et al 2015). Other papers already found evidence that governments can strategically manipulate information to induce specific popular responses (Michalski and Stoltz 2013). In our work we argue that, if a large fraction of households relies upon the government’s signal to update their expectations about pollution, it may be optimal for the government to introduce a downward bias in information, declaring that the air is cleaner than it truly is. Further, the bias should be stronger when the signal is less noisy and when it can have a stronger effect on people’s beliefs. We explore these ideas in the context of information about air quality in Beijing, looking at the pattern of divergence between the official Chinese news, relative to that issuing from the US embassy in the same city.
Information about air quality
Recent research has shown that official reports of air pollution in China appear to be biased (Andrews 2008, Chen et al. 2012, Ghanem and Zhang 2014). In our study, we find further evidence that the public information signal in Beijing is being manipulated in order to influence people’s expectations, and that this has direct consequences on the behavioural response of households to pollution peaks. We compare the official daily information published by the Chinese Ministry of Environmental Protection and the one provided by the US Embassy in Beijing on its hourly Twitter account. The discrepancy between the two sources of information is quite stark (see Fig. 1 for a snapshot of a short time period) and we find that it is not coincidental – the Chinese signal is systematically lower the higher the pollution gets, and also appears to be biased specifically to influence how households perceive the health impacts of air pollution.
Figure 1. Chinese versus US index (daily minimum)
How is pollution information usually reported? Most countries – including China – adopt a simple classification, with colour-coding that maps the index of pollutants’ concentration into potential health consequences. Table 1 illustrates how this information was codified during our data interval (2008-2013). Ultimately, from the perspective of the general public, what matters is not a specific value of the Air Quality Index (AQI), but whether the pollution on a given day is ‘green’, ‘yellow’, and so forth.
Table 1.
These categories are crucial if the government is concerned about pollution perception. Indeed, we find that the downward bias becomes more prominent around the thresholds between different air quality categories. In particular, when air pollution crosses the threshold between green and yellow regions (i.e., at 100 points), and between the orange and red regions (i.e., 300 points), the negative bias in the official air quality announcement significantly increases. This manipulation generates an impact on people’s perception of pollution hazards that is much stronger than a similar change in intermediate values. This effect is significant and robust to different time series specifications, and thus it points to a specific reduction in public alarm and awareness of pollution exposure.
Other papers have highlighted that there may also be explicit political incentives for public officials to bias information away from the 100 threshold in the index, as cities receive special mention if they manage to achieve a large number of ‘Blue Sky Days’ (AQI<100) in a given year (Andrews 2008). Here we find that this might not be the only effect, because another threshold is also significant, the one that separates moderate/heavy pollution from a full health alert (AQI of 300). Thus, it seems that manipulating popular perception of air pollution may be an additional important determinant of the bias.
Households’ response to pollution and pollution information
So then, how do such information distortions impact the response of those households relying upon the government-controlled media? We conducted a household survey around different districts of Beijing, to examine the sources of information that households use and how these relate to their behaviour during pollution peaks. The costs associated with air pollution are quite substantial in our sample, with an average annual expenditure in medical costs, medicines and foregone wages of more than 3000 yuan, almost a month of an average salary. Families should therefore be quite concerned about protecting their health from air pollution. However, the majority of households still use traditional sources of information (radio, TV, newspapers) and many declare that these are sufficient for them to know about pollution peaks. We examine whether such people - who rely on government-controlled sources of information and perceive them as adequate - are less responsive to pollution peaks. We collect data on several different self-protective behaviours – cancelling outdoor activities, wearing a facial mask, changing means of transportation, undertaking preventive health checks, and using an air purifier. We find that, for several short term strategies such as the above, households relying on official information were less likely to adopt self-protective measures during pollution peaks. A key consequence of public misinformation lies in the failure of households to know when to protect themselves against health damages.
Final remarks
Dealing with public environmental goods in developing countries is always challenging, because the key priority is growth and poverty reduction, rather than environmental protection. China, however, faces a further difficulty – its government has the incentive to misinform the public about pollution hazards, rather than warning people about pollution peaks. Currently, the Chinese Communist Party seems to prefer to keep the majority of the population relatively desensitised to the scale of the pollution problem, at the cost of sacrificing some of the individual capacity to respond to pollution hazards. In the long run, this will result in unnecessary public health costs.
References
Andrews, S Q (2008) “Inconsistencies in air quality metrics: ‘Blue Sky’ days and PM10 concentrations in Beijing”, Environmental Research Letters, 3(3).
Chen, Y, G Z Jin, N Kumar, G Shi (2012) “Gaming in air pollution data? Lessons from China”, B.E. Journal Econ Anal. Policy (Adv), 13 (3).
Dasgupta, S, B Laplante, H Wang, and D Wheeler (2002) “Confronting the environmental Kuznets curve”, Journal of Economic Perspectives, 16(1): 147–168.
Ghanem, D and J Zhang (2014) “Effortless perfection: Do Chinese cities manipulate air pollution data?” Journal of Environmental Economics and Management, 68(2): 203 – 225.
Michalski, T and G Stoltz (2013) “Do countries falsify economic data strategically? Some evidence that they might”, The Review of Economics and Statistics, 95(2): 591–616.
Morris, S and H S Shin (2002) “Social value of public information”, American Economic Review, 92(5): 1521–1534.
Ravetti, C, Y Jin, M Quan, S Zhang and T Swanson (2015) “A dragon eating its own tail: Public information about pollution in China”, CIES Working Paper.
Shleifer, A, S Djankov, C McLeish and T Nenova (2003) “Who owns the media?” Journal of Law and Economics, 46(2): 341–81.
WHO (2014) Ambient (outdoor) air pollution database, by country and city. Public health, environmental and social determinants of health, World Health Organization.
World Press Freedom Index (2015) World Press Freedom Index 2015: decline on all fronts. Reporters Without Borders.
