World Bank Document

6.3 Marginal Returns of Infrastructure Investment
Some studies show that infrastructure development has an inverted U-shape relationship 
with economic growth. For example, Bougheas et al. (2000) find an inverted U-shaped relationship 
between infrastructural development and economic growth for the United States. The threshold 
infrastructure development level above which infrastructure investment has negative marginal 
returns will vary across countries depending on several factors, including climate, land area, and 
spatial distribution of inhabitants across the country. Wylie (1996) illustrates this with a finding 
that Canada needs a much higher level of infrastructure as compared to the United States due to 
its colder climate and highly scattered inhabitants across the country. The nature of the threshold 
in turn, affects how the marginal return to infrastructure investment changes as infrastructure 
grows toward the threshold. However, more studies needed to understand the peak infrastructure 
level, after which the marginal impacts of infrastructure on economic growth is declining. Does 
the threshold level vary across countries? What does historical experience or data demonstrate? Is 
it true that the relationship between infrastructure and economic growth in today’s developed 
countries was stronger while they were achieving the ‘developed’ status? Is this true for newly 
emerging economies – China, Brazil? What about the Eastern European and Former Soviet 

---

31 
Republics where governments heavily invested in infrastructure before they transited to a market 
economy?
6.4 Distributional Impacts of Infrastructure  
Over the last decade or so, several studies have attempted to underscore the relationship 
between infrastructure and its impacts distributed across income groups or inequality (Démurger 
2001; Calderón & Chong 2004; Calderón & Servén 2004, 2010; Chatterjee & Turnovsky 2012; 
Chotia & Rao 2017a, 2017b; Hooper et al. 2018). The common findings from the existing literature 
are that infrastructure development has a negative relationship with inequality, meaning that 
infrastructure development helps reduce income inequality. However, what is still not known is 
how would infrastructure development helps the poor who use the infrastructure services 
disproportionally less compared to the rich. Some studies have reported that access to 
infrastructure mainly benefitted to the rich. Khandker et al. (2014) finds that the larger share of 
benefits of rural electrification in India accrues to wealthier rural households because poorer 
households use a limited amount of electricity, mostly for lighting only. One may raise a question 
– to what extent the government spends on infrastructure that disproportionally benefits low-
income households. It is possible that these studies might not have investigated indirect benefits 
that infrastructure investment ultimately brings to the poor. Construction of a rural road will 
provide farmers better market opportunities for their agricultural products. It is, therefore, 
necessary to investigate the impacts of infrastructure on the poor from a broader perspective than 
from the narrow perspective employed by the existing literature.