The reason is that you're taking the population of the USA where over 80% have licenses to drive (http://www.fhwa.dot.gov/policyinformation/pubs/hf/pl11028/ch...) and comparing with a country where most people don't own cars:
While nearly a third of people living in metros own personal transport, only 15 percent of the rural population own their own vehicle.
http://www.thehindubusinessline.com/catalyst/driving-indias-...
I think you would really want to compare serious accidents per vehicle mile driven to compare effectiveness of policy and driving behavior across countries.
Hope this clarifies my intent :)
https://en.wikipedia.org/wiki/List_of_countries_by_traffic-r...
That's using a lower number of deaths for the US (but it really is closer to 35,000 in recent years) and a larger number for deaths in India than you did.
Just for the sake of curiosity let us compare statistics of road accidents between USA and India. I'm taking the 2005 US road accident data as a reference: http://www.foxnews.com/story/2005/02/03/car-crashes-kill-400...
So according to the statistics from the above article, approximately 40,000 die every year in USA with 100 people dying every day due to road accidents. That is approximately 5 per hour. In India it is 100,000 per year. The population of USA is 318.9 million compared to 1.252 billion of India. If you put it in perspective, 0.013% of total population die every year due to road accidents in the US while only 0.00798% of total population die every year due to road accidents in India. So by looking at these calculations India "seems" to be a lot safer than the US (in-spite of horrible traffic rules and no proper emergency services). The issue is not the stats but proper implementation of emergency services which might help in saving a fraction of the total lives lost. But the real issue is regulating traffic (which is more erratic in the cities than in towns/villages in India).