If the sea level rises considerably, the watery real estate of West Bengal will fall in value. Let’s say we knew that Calcutta would flood in fifty years’ time, how would the adjustment process work? Will people leave a dying city too rapidly or too slowly, as defined in economic terms?
Under one scenario, not everyone need leave the city. The city ought to shrink, but can survive at a less populated level. Furthermore then suppose that the stayers are better off, because they do not incur migration costs. Each person then will wish that others leave and he gets to stay. Migration will become a game of "chicken," and people will postpone leaving for as long as possible, hoping to be the lucky stayers. This is related to the reason why not all auto workers leave Detroit when the plant shuts down. They are hoping they will be rehired if/when a scaled plant reopens; everyone waits for the other guy to leave.
Alternatively, perhaps the flooding will be severe and everyone must leave Calcutta. Ideally the residents should coordinate on some new locale with urban increasing returns. Do you prefer to be the first one to leave or the last?
If the new locale is known, land rents might be bid up rapidly in anticipation of the forthcoming change. If you arrive first, you pay the higher rent without yet reaping the benefits from the new, still-forming city. You might rather wait for the others to come first. Markets will know this, and perhaps land prices won’t be bid up so much at first. The equilibrium likely involves mixed strategies and a partially successful solution. Of course many residents here in Calcutta don’t pay rent but rather sleep on the streets. They may be the ones who drive a satisfactory solution, since they can reap unpriced benefits from moving. Expect them to walk of course.
If the new urban locale is not known, it is hard for Calcutta residents to know where to go.
Note that if increasing returns are truly strong, having a long time to adjust may not help much. Once some of the city starts moving, the rest must follow. The period for adjustment becomes compressed, whether this happens right away or as the likelihood of flooding arrives.
I predict most residents will not form a new city to mimic the increasing returns of the old. Most likely, they will migrate to other very large Indian cites; when large numbers had to leave Pakistan during Partition, many of them settled in Delhi.
How costly will this additional concentration of population be? To Western observers, Indian cities appear impossibly overcrowded. But in income-adjusted terms, are Indian cities overcrowded at all? Might they be undercrowded and still capable of reaping additional economies of scale? In this case the real problem is that too many Indians keep a sentimental attachment to their rural areas, to the detriment of their potential urban neighbors. Migration out of Calcutta would again be too slow, relative to an optimum, but in absolute terms things would work out OK.
Adjusting for income, are American cities and suburbs more or less likely to be overcrowded than Indian cities?