Saturday, October 13, 2012

Reliance GSM data plans


  • The default plan charges Rs 10 per MB upto 50MB and Rs 1 per MB after the 50MB limit. 
  • The other plan is Rs 99 for 2.5GB per month. 

Basically reliance sells bandwidth at following rates:

  • Rs  10.00  per MB 
  • Rs    1.00  per MB 
  • Rs    0.04  per MB 
The default plan costs 250 times the cheapest plan. That is best use of behavioral economics I have ever seen. 

Sunday, October 07, 2012

Prepaid Traffic Fine Card

The problem we are going to solve today is the corruption in the traffic police. It is obvious that given the circumstances corruption is the best strategy for both police and people.

  • Let's say traffic light jump is fined at Rs 200. If a corrupt policemen charges Rs 199, it is good deal for both the driver and the policemen. Because people are not just rational but emotional too, a fair deal will probably happen around Rs 100. 
  • This means that if fines are 0, their will be no corruption. The reason fine is not 0 is that our objective as a society is not just reduction of corruption, but:
    • As less travel time as possible 
    • Less and less accidents 
The problem is it is was hard to measure these things when traffic rules were invented. Accidents are not too hard to count, travel time across all people shouldn't be so difficult given mobile phones and GPS.  Traffic police has no incentive to solve these problems. So lets leave these things aside and just concentrate on the corruption part. 

Lets define the Prepaid Traffic Fine card. Basically drivers can buy such a card with various denominations. It could be a coupon booklet like sodexo. Lets say everyone knows their own driving style, so they know what kind of "crimes" they are likely to commit and how much they end up paying to corrupt cops. Now when I say crime, I don't mean running over people. Just the soft variety like may be 10% over the speed limit or 20% over the speed limit. Jumping a red-light by 2-3 seconds etc. Or may be 10%-20% over the alcohol limit. 

Whenever a driver is fined, he can pay his fine using this booklet. The other condition is that this booklet just like sodexo expires after a year. Now:
  • Driver should always pay using the booklet because he already paid for it. 
  • This money is already with the government, so nothing goes to the corrupt police guy. 
Few more cases arise.
People either underestimate or overestimate their "fine budget". If they overestimate, they have incentive to always pay using the booklet because any money they give to the traffic police guy will overshoot their "budget". If people underestimate or don't care to buy the booklet, the usual scenario applies which means government gets 0 and corrupt police guy and the driver share 50% each of the fine amount. Now lets assume that we give some discount on this booklet. 10% should be easy because government can earn 10% interest in one year with this amount. The more "interesting" or "value for money" government can make this "offer", the incentive for corruption reduces by the similar amount. For example, 50% discount means that the driver has no incentive not to buy it because that is the maximum he gets by sharing the profit with the corrupt police guy. Which means the corrupt police guy now has to ask for 25%, giving 75% to the driver. 

I think at some point, the convenience of the time saved in bargaining, social pressure, some sense of dignity, investing in the future would make this converge at much lower discount levels. So here we have the solution to the corruption by traffic police. In the super rational humans case government need to give 100% discount. In case humans act like humans, government gets money, humans get peace of mind and corrupt traffic police 0.  This fund can be used to give education to children of traffic police, bonuses, construction of roads, better technology to monitor and manage traffic, something sensible. Using average of Rs 500 per two wheeler and Rs 2000 per 4 wheeler, Bangalore traffic police will have about Rs 370 crores to work with per year.

Saturday, October 06, 2012

People Futures

I don't know if the title does justice to what I am about to say. The other two I considered were "Investing in people" and "Partial Slavery".  "FDI in people" was another contender. I am not going to talk about future of people in the sense we usually do, but instead the "future" here refers to ones used in financial markets.

Labor doesn't have a future. Again I mean in the financial sense of the term to avoid any Marxist connotation. The question is what if it did? Let's say any person could sell upto "49%" stake in himself subject to investment. Just to simplify implementation, lets say income tax of the person will be used by government to pay dividends to the "share holders" and in case of death of the person, they automatically get assets of the person (bank balance, property, etc) in proportion to their shares. Further assume that these shares are traded on a stock market so that other people/corporations can trade them.

Now that we have defined the model, lets try to see if it helps us in any ways.

Obviously anyone who made investment in Amitabh Bachchan or Sachin Tendulkar or Shah Rukh Khan would be making tons of money. In theory at least some investments could be handsome.

Consider education. Let's say the retirement fund of a teacher is exclusively or at least heavily influenced by the future income tax of the students he teaches. May be even his salary. Does this gives a teacher an incentive to educate the children to best of his abilities?

Consider education again. Does an industry benefits by investing in students which will be part of its labor force tomorrow? May be industry doesn't realizes this, someone else with money does.

Consider sports other than cricket in India. If their is any possibility of someone winning a medal in Olympics and possibly signing up for brand endorsements, would someone invest in players of these games?

I don't know what would this all really mean. But I do know that people are the future of any possible future we have. May be if that future is invest-able,  some people might invest in it. If nothing else, it gives us some measure of "human capital" vs "other capital" so that at least we are clear on what we as a generation believe is the "future".  May be this brings some reality (funny that speculation is what I propose as real?)  to the over simplified model of people as commodity.

Thursday, October 04, 2012

Simple solution for unemployment

There is lot of literature on the subject with many ifs and buts.  Most of it points to stimulus to revive the economy. Because I am not an economist, I don't have much say on their effectiveness.

But I do have a very simple solution to the problem.  Declare Friday as a holiday

How does this works? We first need to understand the reasons for unemployment.
  • Companies decide to reduce production and hence need less labor.
  • Companies don't need more people. So the students coming out of college don't get any jobs.
  • New methods of productions increase efficiency and require less people. 
Now lets assume government  declares Friday as a  holiday.

Good things that can happen:
  • Companies only have 4 days to produce whatever they were producing in five days. 
  • They need to employ more people (about 20%) to meet their weekly targets. 
  • They might need to invest more in infrastructure to ensure that they can have working space/ machines etc for 20% more people.
  • They might even decide to pay less per person from existing salaries, but they do need to invest in infrastructure and hire more people. Less paying jobs is infinitely more better than no job.
Bad things that can happen:
  • They might ask people to work more hours per day. But labor regulations and bonus for extra work might not have linear cost structure. 
  • They might reduce production further and increase prices. 
I can't completely predict what will happen, but if unemployment is caused by "excess labor", then this could be simple and cost effective way for governments to artificially reduce "excess labor" from the market. BTW instead of Friday, Monday will work too. Or if possible to enforce, 20% less working hours per week per person would also yield the same result. Actually government doesn't even needs to implement it, just declare it. Assuming 40 working hours per person, government can just announce that every 2.5% increase in unemployment will result in 1 hour less working hours per week. I am sure companies will optimize everything to find out what gives them optimal results. Who knows within a decade or two we will all work couple of hours a week ;)


Wednesday, October 03, 2012

Opportunity for github

Sell maintenance for open source projects. It could be a yearly payment or per problem based.

  • They know the people who contribute
  • They also know which areas they contribute 
  • They can easily know who downloads
  • Pay developers for fixing the bugs or adding features.