Tuesday, October 14, 2008

Krugman's Nobel Winning Theory and Software Industry

Paul Krugman won the Nobel in Economics today. Before you jump, I know its not a real Nobel prize. I have enjoyed his writings on New York Times along with other favorite economists like the one's behind Cafe Hayek - who funnily enough or sadly enough denounced Krugman's award today. In any case, how does this relate to software?

As it turns out, Krugman's award was a recognition for the "New Trade Theory". You can read a good explanation at Marginal Revolution.

Very simply put: unlike older models that tried to explain trade based only on comparative advantages - so if Australia has more copper ore reserves, it may produce more kitchen fittings made from copper, the New Trade Theory introduced the idea of economies of scale into the trade models - so rather than just natural comparative advantages increasing returns from scale and network effects - could result in advantage.

In other words, if you got big enough - that itself is an edge!


Leaders of software companies have repeatedly said - we are opening offices and hiring in India and China - not because its cheaper but that's the only place where we can find engineers.

With salaries rising in India every year (over 25% by many estimates), the differential is coming down. In other words, the comparative advantage derived from lower salaries, lower costs etc. is getting diminished every year. And yet, companies continue to flock to India. (I am using India as an example since I am most familiar with it but you can apply the model to other countries where scale is true).

I have often posed this question to my friends over last year or so: If you wanted to hire 5 software developers, where would you do that? And answers range from US (for we still have one of the best, if not the best), Ukraine, Russia, Israel, India to Ireland. The reasons for their particular pick range from existing economic ties, existing social ties, cost advantages, ability to find domain experts in particular fields - in other words, what may be called comparative advantages.

But then I ask the following: If you wanted to hire 5,000 software developers, where would you do that? And the answers are rather limited. If you wanted to get that many in a short period of time - you probably can only do that in India today. (Don't believe me - look for software companies with more than 5,000 developers in other countries hired over less than 5 years).

I didn't have a name for this - and thanks to Krugman's Nobel, perhaps now I do - I think its economies of scale and network effects kicking in. The New Trade Theory!

And India got there by a fortunate series of accidents - lack of employment opportunities outside of software for so-called elite engineers (due to regulation till mid-'90s), sudden spike in demand due to Y2K (same question - where else would you find 5,000 engineers? Or 100,000?), lower costs for nearly 15 years, excellent education for the select few that could obtain it and the Indian diaspora in US and other countries that had become successful in their new homelands. And external factors like cheap connectivity after the dot com bust (dark fiber). Due to a confluence of these factors, Indian software industry for outsourced services and development has achieved scale. And even as comparative advantages diminish, its likely to do very well.

So what shall I call it - (Anshu's) New Trade Theory of Software Outsourcing.

Here are questions for other industries and perhaps even companies:
  • Will the scale and network effects also play out for China in manufacturing, for example? (Or not?)
  • Will leading SaaS companies that have achieved scale be able to benefit from a similar effect?
  • If you are not yet at scale - what should you do and not do? Compete where comparative advantage (features) trump scale (network effects)?


Anonymous said...

I agree "Indian software industry for outsourced services and development has achieved scale." your informations are nice.

Robert Evans said...

I can understanding throwing large numbers of humans at testing and QA but on the creative side, where much of the challenge is harnessing the available IQ, do you think scaling is essential?

Seems to me that for green-field work, a small tight team of experts with a proven track record, ought to be sought where ever they may be located.


Anshu Sharma said...

In fact, I agree with you. You don't need thousands of developers unless you are trying to compete with (or are one of) the largest SI's or software companies in the world today. But if luck smiled on you and you became the next Google/Microsoft/Oracle/Cisco/Infosys etc. - its rather hard to hire enough people in the US.

I prefer hiring smart and hiring locally when I can.

Ashutosh said...

I would certainly go to India to hire 5000 engineers. Not only will I get them there, but I will also get HR managers who have hired 5000 engineers and a full working network that has delivered such numbers and knows how to do it!