In recent decades, macro-level studies in economics beginning with David Aschauer in the late 1980s have reported a strong positive association between higher levels of investment in infrastructure and economic growth. A larger stock of infrastructure is thought to fuel economic growth by reducing the cost of production and transportation of goods and services; by increasing the productivity of input factors; and by creating indirect positive externalities.
The Chinese planners have paid considerable attention to these orthodox claims by macro-economists. Zeng Peiyan, the then minister in charge of the State Development Planning Commission, told the New York Times, (24 September 1998): “Only if overall fixed-asset investment (e.g., highways, bridges and power grid) grows by 15 to 18 percent, can we reach 8 percent economic growth”. Academic studies corroborate this narrative of more infrastructure is better for China and argue that investment in and proximity to transportation infrastructure have had a positive effect on economic growth in Chinese cities and provinces.
In contrast, drawing on micro-level evidence from multiple cases of investments in transportation infrastructure in China, we unsettle the orthodox narrative that heavy investment in infrastructure leads to economic growth. Here we show that even in China, the purported infrastructure prodigy, the actual costs of building infrastructure are systematically underestimated. The effect of cost overruns is exacerbated by shortfalls in forecasted stream of future benefits—direct or indirect. Consequently, a typical infrastructure investment yields a negative net present value—i.e. a net drag on the economy. The pattern of poor project-level outcomes and inaccurate cost and benefit forecasts in China is consistent with the notion of “lying with numbers” or “strategic misrepresentation” and “over-optimism” put forward in recent literature of major infrastructure programme management.
Our bottom-up evidence on the costs and benefits of transportation investments in China thus indicates that the high rates of return from infrastructure investment at the aggregate level calculated by David Aschauer and others working in his tradition are likely incorrect. Instead, what the macroeconomists appear to be measuring is correlation not causation between higher infrastructure spending and economic growth.
In interpreting our results, we argue that economists have tended to over-stress the need for infrastructure in the economy by dwelling on the link between infrastructure investment and short-term economic growth. It is a given that increased infrastructure spending will increase the Gross Domestic Product (GDP) in the short run as a natural accounting consequence of piling investments (productive or not) into fixed capital. Moreover, arguments related to positive externalities and spillover effects reinforce the tendency of “lying with numbers” by cooking forecasts of substantial wider economic benefits that fail to materialise.
The implications of our research are two-fold: First, GDP growth and obfuscatory notions of spillover effects are the wrong metrics to measure whether an infrastructure investment is beneficial. We propose that a simpler approach that focuses on the lifetime cash flow profile of a proposed infrastructure project ought to be adopted. Only infrastructures that yield net present value in terms of demonstrable user fees or a direct increase in tax revenues should be built. Second, we suggest that when it comes to infrastructure less is more: we propose ways in which policy makers can enhance the efficiency and productivity with which fewer infrastructure assets are used to create resilient economic prosperity.