Wendell Cox writes (pointer via Planetizen):

Most suburban growth is not the result of declining core city populations, but is rather a consequence of people moving from rural areas and small towns to the major metropolitan areas. It is the appeal of large metropolitan places that drives suburban growth.

Larger metropolitan areas have more lucrative employment opportunities and generally have higher incomes than smaller metropolitan areas. This is particularly the case in developing countries. As a result, the big urban areas attract people seeking to escape what are often the stagnant or even declining economies in smaller areas.

A very Jane Jacobs insight, and one I find compelling.

In The Economy of Cities, Jacobs controversially argued that virtually all of economic growth traces back to cities: In her view, cities actually precede agriculture. Early cities, according to Jacobs, spurred agricultural development by providing trading centers for agricultural products.

While it's common to think of suburbs as draining off city assets, today's metropolitan areas with their urban cores and suburban and ex-urban rings, are really expanded cities. Up until the early-to-mid 20^th century, cities were able to capture peripheral growth by annexing new development, until suburbs figured out they could prosper by becoming independent municipal entities - thus the now famous concentric-ring, or in some cases the hole-in-the-donut, pattern of our metro regions. The growth of gargantuan mega-regions like the Boston-New York-Washington corridor is essentially the next phase of this process of geographic development.

It's important to understand how these two interrelated geographic processes - outward geographic expansion and the more intensive use of existing urban space - combine to shape  economic progress.