Economies of scale are not well understood by economists. Standard economic theory assumes diminishing economies of scale, i.e. more of the same leads to less output. When a company expands, it runs into coordination problems sooner or later. We frequently hear stories of corporate slimming for that reason. At an early growth stage, coordination at the company level is not a problem, while the advantages of shared resources are significant. Hence, a company exhibits increasing economies of scale during its early development and diminishing economies of scale when it matures. As economic theory deals with state of equilibrium in economic activities, it's correct to assume diminishing economies of scale.

Network economies of scale are similar to those at the company level, but with more intensity. When a network construction project is just beginning, its economic benefit is probably linear or proportionate to the investment scale. When a network is about to be completed, its economic benefit grows exponentially. When a completed network expands, the benefit diminishes relative to investment, i.e. every additional dollar of investment gets less economic benefit.

Has China's infrastructure investment reached the point of diminishing return? It's hard to say. But investment's effectiveness is much less than in the past because networks have been completed. As there will be fewer productivity gains from infrastructure, productivity gains are needed from elsewhere to maintain the same economic growth momentum.

China's macro policy is now being dominated by the pumping of liquidity into business and government sectors. The argument is that spending more will bring the economy back, regardless of efficiency, and that what's needed are the same results with less efficient investment. Of course, if there are no negative consequences, printing money would solve any economic problem: If something's not working, just print more. However, boosting liquidity and printing money lead to inflation. The less efficient the investment, the earlier inflation arrives. As I have elaborated before, inflation could infect the system more easily now that the benefits of globalization and IT have been absorbed.

China has a lot of room to maneuver: The government's debt level is low, the banking system's loan deposit ratio is low, and household leverage is low. A high leverage level in the corporate sector is a worry. But the country has excellent spending power overall reflected in a large foreign exchange reserve and persistent trade surplus.

Today's options reflect past successes in achieving productivity gains and export penetration. Both stories are running into barriers. Hence, China should spend money carefully to nurse existing golden gooses into the future, and not steer money into unproductive, white elephant projects to pump up GDP today. China should spend carefully even though it can spend more, because money won't come as easily in the future. The days of easy productivity gains and export penetration are over.

This is why China should tighten its monetary policy as soon as possible. Between December 2008 and April 2009, China's bank lending rose by 20 percent. If the current pace is maintained, the annual loan growth would reach 50 percent. Lending hasn't risen this rapidly since 1992. The result that year was rampant inflation. If inflation becomes a problem again, tackling it will be more difficult. In the mid-1990s, China tightened policy to contain inflation while exports rose rapidly to maintain economic growth. The same lucky scenario is unlikely to be repeated this time.

If inflation surfaces while the economy remains sluggish, the government may not be willing to tighten down sufficiently to cool inflation. It may lead to a prolonged period of high inflation rates and sluggish growth. This risk of stagflation is quite significant for China, the United States, and the world. While the initial burst in lending was needed to stabilize the economy, its continuation may carry more negatives than positives. It is not worth risking virulent inflation merely for a gain of one or two more percentage points in this year's GDP growth rate.