Dynamic Pricing of Tickets in Railways

Indian Railways is broke. A large number of their plans, particularly those relating to increase in capacity, are not moving forward because of lack of funds. Their operating ratio is over 90. That is, for each 100 rupees that they earn, more than 90 rupees are spent on the operating expenses alone. Obviously, no money is left for development work.

If you do not increase the passenger fares for 8 long years, and you have limited flexibility on the freight side of business, you are bound to land in the situation that Railways is in today.

With some increase in infrastructure, improvement in technology, higher efficiency, some more new trains can be introduced every year, but to really solve the problem of demand and supply, they would have to build infrastructure at a very fast pace. But for which they have no money.

In the last one month, the Minister and the officials have been talking about a middle-of-the-year increase in fares, which is good. The Minister has reportedly said that they don't want to get into the situation that Air India is in. They have already increase the freight rates, but given that they have a strong competition from trucks, they can't cover all their passenger traffic losses from freight. But I am sure no politician would want to increase the fares substantially in one go. And there is no doubt that the situation is so bad that a small tinkering with fares would not be of much help.

This is where dynamic pricing could help. And the good news is that the Minister is talking about it. (Read an article in Economic Times.) The Minister only needs to announce an increase in the normal fare, a fare that would be paid by only some of the passengers. Others pay the fare determined to a large extent by demand and supply. Just to give an example, if in the first 10 minutes of the reservation opening 90 days in advance, 10 percent of the train capacity is booked, the pricing engine would increase the fares. On the other hand 10 percent of the train capacity is not booked even 10 days after the reservation opens, the fare remains the same till an upsurge in booking is seen closer to the travel date.

The last minute tickets could be priced substantially higher than the minimum fare for the same on peak rush days.

Dynamic pricing has been completely accepted by public at large as applied by airline industry, except when the fares have become ultra-high, as it happened during Diwali time in 2010. Even in trains, the Tatkal rules of Laloo's times, when one had to pay the fare of end-to-end journey, and a premium on top of that with no cancellation refund, was accepted by public without much protest, though it would often be more than 2-3 times the normal fare, if one had to undertake the journey for only a part of the train route.

Dynamic pricing has lots of advantages, both for railways as well as passengers. Railways is able to get more funds, which can now be used for expanding the services. Passengers don't have to resort to bribing, finding loopholes, standing for several hours from early morning, 90 days in advance, so that they can get reservation on the day of opening, and so on. The VIP quota business can be thrown out, as with the new system, it would be possible to have some tickets available till the last day. Railways lose a lot of revenue, particularly in AC-1st and AC-2T because many people traveling on business do not want to book waiting list tickets. So for these classes waiting list is not long enough to take care of additional seats available when these VIP quotas are released for general public.

A system which allows some tickets at a relatively low fare on first come first served basis, and other tickets to be sold on the basis of market conditions, is also a fair and practical system. Selling all tickets at low fares 90 days in advance does not seem fair to those who cannot plan so much in advance. Selling all tickets through a lottery a few days before the date of journey is not fair to people who want to plan their vacations much in advance. Auctioning all tickets discriminate against economically weaker sections. But dynamic pricing takes care of all these concerns.

If one still has "fear of the unknown," it may be first introduced in upper classes, see the experience, and then take a decision to introduce it in non-AC classes.

Indian Railways could also introduce differential pricing for different seats in the same class. The pricing of lower inside berths could be higher than side upper berths.

Another pricing differentiation could be that the price of a short journey on a long distance train could be higher than the price of the same journey on a short-distance train. For example, Delhi-Kanpur ticket would be cheapest on Delhi-Kanpur trains, slightly costlier on Delhi-Allahabad trains, more costly on Delhi-Howrah trains, etc. Basically, encourage more long distance passengers on a long-distance trains, and you will find out that the long distance trains do not have many passengers from small stations, and therefore, those halts can be removed, and trains can be speeded up.

Another advantage of dynamic pricing (and the resultant availability of tickets closer to the day of journey) is that one can provide better service on super-fast trains. It may be worthwhile to recall that about a couple of decades ago, there were no unreserved compartments in many of the prestigious super-fast trains. They were added later on to enable those passengers to travel by these trains who have to reach their destination urgently because of some emergency, and they could not get reservation done in time. But now, if reservations are available closer to the date of journey, we could remove all unreserved coaches, replace them by reserved coaches, and make journey more comfortable for more people.

There are other ways in which Railways can get more revenue without having to increase the base fares by much. Currently, Indian Railways has a mechanism to book a ticket from a station different from the station where the passenger will board. This is, however, discouraged by Railways. If I want to go from Station 'B' to Station 'C', and if I try to book the ticket from Station 'A' to Station 'C', with boarding at Station 'B', the reservation engine will book me only against the quota of Station 'B', and therefore, it would be foolish to pay from Station 'A'. But if Railways allowed passengers to use quota of Station 'A', it is essentially increasing the fare of those passengers, while keeping the option to sell those seats between Station 'A' and Station 'B' and make additional money.

We need Railways for the economic growth of the country, and not just a defunct railways, but a vibrant railways, which is expanding to meet the growing needs of the nation. It has to be financially strong to invest in the future.

PS: On an unrelated note, I found this news item on Economic Times very interesting. It talks about a problem similar to Y2K problem in the context of railway reservation system.

Selecting a Dean

I had been thinking of writing about selecting the second rung of leadership in an academic institution for a long time, but this semester is not allowing me to write much. But when I saw Prof. Ram Mohan's blog on the same topic, I was encouraged to write my own views on it.

Typically, in Indian institutions, the top position (Director or Principal or Vice Chancellor) is often advertised, and the decision makers (whether promoters, including government, or the board) often do consider a broad search, including people from outside the institute. There are some lacunaes in the process. As Prof. Ram Mohan says, the board (assuming that this is the body responsible for recruiting the top leader) must first articulate what type of skills it considers more important at that stage. This is hardly ever done. A combination of decent academic record with experience in administration is all that matters. But at least the search is wider.

But, when it comes to the second rung of leadership, there is very little focus on the selection process. And the result is an obvious one - you get some good Deans, and with others, you just pray that they won't do enough damage in the 3 years that they would be occupying the office.

First of all, there is no articulation of the job description. Somehow we all are supposed to be aware of it. While one may generally be aware of what are the responsibilities of say, Dean of Students Affairs, a document specifying the role will great help the search process. Also, what are the specific focus for the next couple of years, or what are the immediate issues to be handled will also help potential candidates and those who are involved in nomination or selection. If the goal of the Institute is to privatize the messes of all hostels, and you bring in a Dean who is philosophically opposed to outsourcing, it is not going to help the goals of the Institute. But, if the job description only included one line about managing the hostel, then no one will ask about the philosophy of outsourcing from the potential candidates.

Second, there is a need to think about the desirable profile. Not that a person outside this profile cannot be a good Dean. But having a desirable profile makes it easier to think of names to nominate and for shortlisting, etc. For example, Dean of Research and Development whose office is expected to provide support to all project investigators should be one who has handled several projects himself/herself. Dean of Students Affairs should be one who has handled student interaction either as a warden or in some other capacity. Dean of Alumni Affairs will have to be one does not mind traveling and meeting a lot of people. And so on. The profile may include desirable past experience, age profile, interests, etc. Again, generally speaking, there is no profile that is made available during most searches.

Thirdly, unlike search for Director or VC which looks at external candidates, there is hardly any institute in India who will look for external candidates for a Dean's position. This must change. Normally, a good institute would not recruit its own fresh PhDs as faculty. The reasons are many. But one reason is that we want the person to have had a diverse experience, and bring in a different perspective to the institute. In the same way, a few leadership positions being filled by external persons would bring in a new way of doing things, a change of perspective. And it becomes a tool to attract talent as well.

Fourth, in most cases, the process is very secretive. Not having a job description and a desirable profile is part of that secretive culture. It is impossible to find out who all were nominated, who were shortlisted, how shortlisting was done, why someone was selected as a Dean, etc. No transparency at all. Having a more public process (even in private universities) will result in better selections. It should be known to stake holders who are being considered, and how shortlisting was done. The shortlisted candidates may be asked to give a presentation on their vision and plans for the job. Such presentations could either be open to all stake holders, or at least their videos be made available to stake holders later on. This would result in people who are really passionate about the job and has some clear vision and plans for the job to be selected.

My colleagues tell me that secretive process is necessary since most faculty members in a good institute like IITs do not wish to become Deans. If they have to go through a transparent process, they will simply refuse. On the other hand, if they are selected after a secretive process, then they think it is their responsibility to take up any role assigned to them by the institute. So the secretive process helps the Institute in getting good Deans. Nothing could be farther from reality. While I don't deny that there are some faculty members who would agree to be a Dean after a secretive process and wouldn't have participated in a transparent process, but the number of such Deans is small. Most faculty who do not want to be an administrator will refuse to take up the role even after the secretive process. A lot of people who later turn out to be ineffective leaders as Deans, wouldn't have participated in the transparent process, but are happy to be Deans after a secretive process. And the transparent process does not mean that the Director or the board members cannot approach faculty members and encourage them to participate in the process. In fact, it is often necessary to do this in any selection process, if you want quality intake. We do this all the time for faculty selection, but somehow forget to do so when it comes to leadership positions.

Search for leaders often a long time. Hence it is important that the process starts six months in advance. The final decision should be announced about a month in advance of the incumbent's tenure coming to an end. It should be possible for the new Dean to spend some time as Dean-designate and observe how the previous dean carries out various tasks. This would enable the transition to be smooth.

I welcome readers, particularly those who are faculty members in Indian institutions, to write about any interesting mechanism that their institute follows to find leadership talent.

NRN Speech at PanIIT

A lot of people have commented on NRN's comments during Pan-IIT at New York last week. So, I am not going to add something new to the debate, but just putting my views across.

What I find very amusing is the exclusive focus on the numbers, 20 percent good, 80 percent bad. If media were to be trusted, Mr. Murthy made several points during his speech. He criticized the coaching culture and considered that as one of the reasons for decline of IITs. He said that in current form, JEE is not the best admission filter. He said that graduates of IITs have poor English language skills. He also said that IITs will have to focus a lot more on research if they want to be in the top league of universities in 10-20 years from now. All the statements that a lot of people have been saying in the recent past. You may agree or disagree with any one or all of these statements, but people have generally remained civil in their disagreements. The only statement which has inflamed passions is the 20-80 percent thing.

Is it fair to lose all sense of fairness and ignore the entire speech of NRN, and only focus on one statement of that speech. Isn't it obvious that Chetan Bhagat wants to be in the limelight just before the release of his book, and that is why he is bringing in completely irrelevant issues to the table like what Infosys has done or not done. (And, what does he know about Infosys anyway.) My friends who listened to him live in New York tell me that it was a great speech, where he talked about a lot of issues, with the only controversial statement being this one about 20-80 percent. Same thing had happened earlier when NRN had given a speech in IIT Gandhinagar in July. People listening to him there thought it was a great speech with deep analysis of what is right and what is wrong, and were totally shocked next day when media just harped on one particular statement during the speech.

I have had the good fortune of meeting NRN on several occasions. And every time I am impressed with his ideas to improve education. And he does not only make statements. Under his leadership, Infosys had initiated PhD fellowships long before IBM and Microsoft started doing so. Infosys support for conferences and other academic pursuits has been remarkable. The Infosys Prize has become the biggest honor that a scientist can receive in this country. Infosys has been supporting IT education in 100s of colleges around the country. (One may argue that it is only due to self-interest, but whether for self-interest or otherwise, you cannot ignore that they are serious about quality of education.) Of course, Mr. Murthy has also been supporting excellence in education from his personal funds to the tune of 10s of crores of rupees. Here is a person who is really passionate about education, has remarkable analysis, great ideas, and who has done a lot. And we ignore everything he says, but we can't ignore that one line.

Having said all this, let me also take this opportunity to reiterate some of the things I have been saying in this blog. First, the coaching issue. A lot of people have been saying that it is wrong to criticize coaching centers. They are servicing a need of the society. But has anyone said that it is not a legitimate business. Has anyone said that Bansals, PACE, FIIT JEE and others are not satisfying a need of the society. If they weren't, they won't be in business. The point being made is that the pedagogy employed in coaching is very different from the pedagogy that education experts tell us should be employed for teaching kids at that age. Kids coming from coaching classes demand the same pedagogy after they come to IIT. Also, the attitudes that coaching creates are different from what one would expect from a well rounded school education. They focus only on short term success. You can't blame coaching centers for any of this. The real reason is competition for a few good seats in a vast country like ours. But can there be no other admission process that allows a bright student to focus on school education and still get admission in IITs (or other top colleges). Can we not minimize the impact of coaching (not because coaching is illegitimate business but because of its side effects). I am convinced that IITs can come up with better admission strategies, and some of that I have talked about in this blog earlier.

Advantages of Aakash

So, we now have the world's cheapest tablet. Congratulations to all who made it happen.

I am curious about one thing. What is innovative in Aakash. I am yet to see a paper, or a reference to a patent, or any design document which describes the innovation part of Aakash. If any of my readers know something in this regard, please let me know.

The alternate theory of Aakash is the following.

Most hardware vendors (like most manufacturers in most industries) think of products for the richest 1 billion of the world. The margins are decent with this sector of population, and if you are sure that your product will sell a few million pieces, you go for it. This market segment would prefer good quality, good specs, multiple things in one box, and so on. Of course, once the vendors are convinced that the 2nd or 3rd or 4th billion also will buy their product if the price is right, they will bring the lower versions of the product, which is happening in the phone market today, where the growth market is 4th or even 5th billion, who can hardly afford anything fancy. But it has not happened in other markets. The lower-end laptops, PCs, and now tabs are at best being targeted at the 2nd billion today.

How do you get a tab whose price point is accessible to the 3rd billion or even 4th billion. You either do an innovation, which others haven't thought of. (And, by innovation, I include the possibility of a technical innovation as well as a financial innovation - may be some company will subsidize the device in exchange of some advertisement rights.) Or you convince a vendor that there is sufficient market for the product at the 3rd or 4th billion level. And one way to convince a vendor is to tell him that I am prepared to give a written order of 1 lakh pieces, and will further purchase a million devices from the market. And then you sit with him to figure out what is bare minimum configuration, which is useful for the purpose it is intended for. And if the bill of material cost is more than the price point you have announced publicly, then quietly wait for 3 months, 6 months, a year, or 2 years, for the prices to come down, and if they stubbornly refuse to come down to the level that you desire, throw in some subsidy as well.

But whatever may be the mechanism of getting this price point, I am quite sure that this will have a strong positive impact on higher education in India. And because of this reason, I must congratulate MHRD for thinking of such a device and taking the project to its logical conclusion.

Why am I sure that this will have a positive impact. Well, institutions in India are generally averse to bringing in technology in the teaching learning process. We don't know who is going to maintain those things. We don't want to be seen as elitist by insisting that students own some devices. And a lot of faculty isn't good at technology adoption anyway. All solvable problems, but who has the motivation. But institutions in India are also very aware of what MHRD wants, and would like to please the minister and his secretaries. It is obvious that if MHRD is putting in a huge amount of money and wants this to be a visible project, it is going to ask Directors to write reports on how Aakash is being used, how it is making a difference, etc. And a few Directors will want to make sure that they have something to tell MHRD. They will ensure that we pluck at the least the low hanging fruits. Can we do at least some sort of video recording of the lectures, and make them available on our servers for students to go through them later on. Can we ask each course instructor to prepare a playlist of youtube videos relevant to that course, which students can see on Aakash in their free time. Once a few institutes show that technology adoption is not a serious problem, and it really helps in the teaching learning process, I am sure others will join in, and we would have improved the quality of higher education in this country.

And, I am sure, if indeed we have a few million devices with students, private sector will bring in innovation so that the students can make use of these devices more effectively.

A similar experiment is underway in Tamilnadu where the state government is giving away laptops to all students. An excellent  blog article by K Satyanarayan on implication and details of the TN government scheme is a must read for those interested in this topic.

Update on 9th October, 2011:

In today's newspaper (Sunday Times, Kanpur edition), there is an interview Suneet Singh Tuli, CEO of Datawind, the manufacturers of Aakash. He talks about three "innovations" or reasons that have kept the price of Aakash low. The first one is that they shift the burden of processing from the client device to backend servers in the cloud, that reduces the cost of processor. (I read this as an admission that the device is not meant for any computation, but is just a web browser and a video player, and hence we have used a real low end processor and small amount of memory.) Second, he claims that they are a lot more vertically integrated than the average manufacturer who buys 50 parts and puts together the device. Datawind componentize it - they buy 800 parts. (It is not clear how much this can save.) And last point is interesting. Aakash connects to their app store, and they are expecting some users to buy some apps from them. This is likely to result in some profit for them, and help them keep the cost of Aakash low.