Page 211 - AJWEP-v22i3

Effects of urbanization on air and health

estimates with IV results and showing first-stage Our results provide compelling evidence of a causal
diagnostics. chain linking urbanization to air pollution and
The first-stage results (column 2) confirm that our subsequently to public health outcomes in Vietnam’s
instrument – industrial output share – strongly predicts cities. The positive effect of urbanization on PM2.5 levels
PM2.5 levels. The coefficient of 0.378 indicates that aligns with extensive research on urban environmental
each percentage point increase in the industry’s GDP challenges in rapidly developing countries, while the
share raises PM2.5 by 0.378 μg/m on average. The first- magnitude we identified proves consistent with findings
3
stage F-statistics (117.83 and 92.47 in columns 3 and 4) from comparable Asian contexts.
substantially exceed the conventional threshold of 10, The urbanization coefficient of approximately
indicating our instrument is not weak. 0.36 μg/m per percentage point increase represents a
3
Comparing OLS and IV estimates reveals important substantial relationship. As Vietnamese cities continue
insights about potential bias in simple correlations. urbanizing at 3% annually, this implies PM2.5 increases
The OLS estimate (column 1) suggests that each of roughly 1 μg/m per year solely from urbanization
3
1 μg/m increase in PM2.5 associates with 2.98 additional processes without policy interventions. Over a decade,
3
respiratory disease cases per 1,000 population – a this cumulative effect could raise pollution levels by
3
substantial effect representing approximately 2% of the 10 – 12 μg/m , representing a 25 – 30% increase from
mean respiratory disease rate. present levels.
However, the IV estimates provide more credible This finding proves consistent with studies from
1
causal identification. In our preferred fixed-effect China, where Wang et al. reported that provinces
IV specification (column 4), the coefficient is 2.315, with 10% higher urbanization exhibited 5 – 8% higher
indicating that each 1 μg/m increase in PM2.5 causally PM2.5 concentrations. Our fixed-effect estimates suggest
3
generates approximately 2.3 additional respiratory comparable magnitudes: a 10% urbanization increase
disease cases per 1,000 population. This estimate is in Vietnamese cities correlates with approximately
3
somewhat lower than the OLS result, suggesting that 3.6 μg/m higher PM2.5, representing roughly 10% of
simple correlations may overstate pollution’s health baseline levels.
impact due to omitted variable bias. The causal pathway from urbanization to
The IV estimate implies substantial health impacts pollution operates through multiple mechanisms.
from present pollution levels. Considering that PM2.5 Urban concentration intensifies energy consumption
concentrations vary by about 25 μg/m between for residential, commercial, and industrial uses,
3
the cleanest and most polluted cities in our sample, predominantly from fossil fuel sources in Vietnam’s
this variation could account for approximately present energy mix. Simultaneously, urbanization
58 additional respiratory disease cases per 1,000 dramatically increases vehicle usage as cities expand
population – representing nearly 40% of the mean and rural populations migrate to urban areas. Vietnam’s
respiratory disease rate. rapid motorization, particularly motorcycle adoption,
To put this in perspective, if Vietnam’s most polluted has created dense traffic flows with limited emission
cities reduced PM2.5 concentrations by 10 μg/m controls.
3
(a substantial but achievable reduction), this would Industrial concentration represents another crucial
prevent approximately 23 respiratory disease cases per mechanism. As cities develop, they attract manufacturing
1,000 population annually – equivalent to roughly 15% and processing industries seeking infrastructure, labor,
fewer respiratory illnesses. and market access. Without stringent environmental
The policy implementation variable becomes regulations, these industries contribute substantially
statistically insignificant in the IV specifications, suggesting to urban pollution through emissions from production
that environmental policies primarily affect health through processes, power generation, and freight transportation.
pollution reduction rather than through direct pathways. Our health impact estimates prove substantial and
This supports our identification strategy and implies that policy-relevant. The IV coefficient of 2.315 additional
the health benefits of environmental policies operate respiratory cases per 1,000 population per μg/m of
3
primarily through their impact on air quality. PM2.5 implies that present pollution levels generate
significant health burdens across Vietnamese cities.
5. Discussion This estimate aligns with international evidence
from developing countries, where pollution-health
5.1. Interpretation of key findings relationships often prove more severe than in developed

Volume 22 Issue 3 (2025) 205 doi: 10.36922/AJWEP025130088