There are two paths to nice cities: to take a messy old city and fix it, or to start from scratch.
The process of solving the problems of an old city is hard. Some cities started over owing to the destruction caused by war, an earthquake, or a great fire. We wouldn't wish that on any city. So there is going to be no opportunity to start over.
The other strategy consists of building new cities from scratch. This has traditionally been a troublesome problem. Cities like Chandigarh or Brasilia haven't worked out too well. The designers who dreamed up these cities did not understand the complexities of the genuine urban life.
I don't know if Bandra-Kurla Complex was intended to be an attempt at urban planning. But if it was, the designers surely knew nothing about what makes a financial centre. All they have there is a few towers housing large corporations. The entire ecosystem of finance - myriad small firms, service providers, even coffee shops - is missing.
There are people who think differently. There is increased confidence these days that it's possible to start from scratch and build good new cities. In India, the most important experiment of this nature is GIFT, which is near Ahmedabad/Gandhinagar. It will be interesting to see how this works out. Read Greg Lindsay's article in fastcompany.com on these radical ideas and opportunities.
Another experiment which is taking place in India, with the establishment of a new high-quality city, is Lavasa.
The real challenge is that of going beyond a first sensible urban plan and some sensible urban infrastructure. Once the machine is set in motion, how will it continue to work? It will inevitably run into problems of urban governance. How can we be sure that the problems of governance and elections, where Bombay has failed so badly, will not bedevil Lavasa?
Here again, there is a slow long path and there is an attempt at a short solution. The slow, long path is to undertake deeper reforms. The 73rd and 74th Amendments to the Indian Constitution began that process but a lot more needs to be done.
In China, the biggest cities (Beijing, Tianjin, Shanghai and Chongquing) have the status of provinces. This idea, of a `direct-controlled municipality' is equivalent to cities like Bombay, Delhi, Calcutta, Madras, Bangalore being states who directly elect a state government and directly get Finance Commission funds from the central tax sharing. This would be a big step forward in improving urban governance in these giant cities.
In some countries, governance is so bad that such reforms are inconceivable. Some countries are just not able to figure out how to do urban governance. Paul Romer's 2nd big idea is: Would it help if a country like Sweden or Canada was given a 50-year lease on a 100 square kilometre block of land, where it would build a high quality city with first world urban governance? He calls this charter cities. I have a blog post which engages in an alternative history on such a theme.