Government e Marketplace - A GeM of an Idea with seeds of becoming the ultimate disruptor

On 9th August 2016, the Government of India launched the Government e Marketplace (GeM) without much fanfare.
The platform, which allows online purchase of common office use Goods and Services by Government buyers,  was inaugurated by the Minister of Commerce and Industry, Smt Nirmala Sitharaman. As the blurb on the GeM site. says, "GeM represents our Government's firm commitment to bring greater transparency and efficiency in Public Procurement"

That it does, but the platform can and will do more, and has the potential of disrupting the "disruptors" of the information age (Think any platform - Amazon/Flipkart, Uber/Ola, TaskRabbit/UrbanClap). Here's why:

Why I expect GeM to Take-Off

1. GeM platform is inherently well placed to solve the chicken-or-egg problem faced by conventional platforms.
It makes procurement extremely easy for the Government buyers providing them the required incentive to procure through GeM. The Government rules were amended in May 2016 (GFR 141-A) to allow procurement through this portal. Orders can now directly be placed on any Supplier/Vendor on GeM  provided that the order amount is below Rs 50,000/- or if the Supplier/Vendor happens to be offering the lowest price among the available suppliers on GeM. There is no upper limit on the amount of order that can be placed in the latter case.
It (GeM Platform) welcomes Supplier/Vendor with an open arm with a Trust First Verify Later approach.  Suppliers can self register themselves - no waiting outside a Government procuring agency to get empaneled as a Government Supplier. The supplier provides relevant details about his/her firm which is validated online against Government databases  (MCA-21 Company Database, Udyog Aadhar number, PAN Database, TIN Number etc.). These would be subsequently verified through an offline process.
2. There is a strong support from the leadership team - The responsibility for running this platform has been placed on DGS&D, the Central Procuring Authority of India, whose leadership is fully committed to its success. Large number of trainings are being organised, both for Government buyers and sellers to make them understand the benefits of GeM (We at NIFM have conducted training for over 500 buyers and sellers in the last three months). The DGS&D, which is the central procuring agency, has decided to do away with the Rate Contracts (empanelment of suppliers) for all items that are made available for procurement on GeM. In a way, this is a commitment that there is no turning-back - the bridges are being burnt.

Why I think (and hope) that GeM can be the ultimate "Disruptor of Disruptors"

At the heart of GeM user authentication lies the Aadhar ID, the 12-digit unique identification number issued by the Indian government to every individual resident of India. Not only is Aadhar number unique, it is mapped to the biometric details of the citizen (fingerprint, iris scan).Few nations have their citizen's trust their government with this level of personal details. Every transaction on GeM, be it by the buyer or the seller, is e-Signed using the Aadhar number of the person initiating the transaction. With almost all interactions of the citizens/residents with the Government now becoming Aadhar enabled, and with the Aadhar enrolment in India at over 90 % , it has now reached the stage wherein it is ready to serve as the Trust Platform of the Government. The Government can embark on creating a trust score associated with each Aadhar ID for different types of transactions. The various Information age platforms, be it Uber/Ola or UrbanClap/TaskRabbit, or even Facebook ultimately derive value from the trust that they are able to provide to the transacting parties. If a trust platform is created by the Government using Aadhar, and this is made available as a Service to  a Government owned or Not-for-profit platforms, then the platform aggregators will stand disrupted. The sharing economy(a misnomer for the Information age disruptors) will truly live upto its name when it is the interacting and transacting parties which share all of the value generated by the exchange, and not the scraps doled out by the aggregator after taking its cut. We all know that Uber or Airbnb can charge 10 to 30 % as their commission for the matching service, and this when their algorithms are not open to public scrutiny!

As a first step, GeM can be expanded to allow residents and citizens, and not just Government buyers, to procure Goods and Services. The algorithms which match the buyer and the seller can have societal benefit as its primary goal, and not profit maximisation.  A government owned or backed platform can make its algorithms, both for trust scoring and buyer-seller matching available for audit- say by the Comptroller and Auditor General of India.

This will be a revolutionary development whose time has come. Information has to be recognised as a Public Good, and the Government has to step in to ensure that the digital exhaust of the citizens does not become the property of data aggregators be it Google or Facebook or Amazon, At the same time, any algorithm that works on this data has to be subject to public scrutiny lest the delicate balance between efficiency and fairness is lost for good.

Brexit - Analysis indicates Referendum may be the Wrong Way for Contentious issues

The world is still trying to come to grips with Brexit - Much has been said and much will be said on this emotive issue over the next few years. This is an event whose impact will slowly but inexorably play out impacting the lives of millions.
Here is an analysis of the referendum results - with some findings which merit serious discussion by all democracies who may think that Referendum is the way to go for all BIG decisions.
My conclusion - Referendum without mandatory voting by All is wrong. Had voting been mandatory - the Brexit referendum results could as easily have been different. A 4 % gap is no gap at all when 28% of the population chose to watch the show from the sidelines.
1. Results were close: Remain: 48.11 % Leave: 51.89 % - the gap between the two: 3.78 %.

The results are available for the 382 different local areas. In terms of sheer count of the local areas voting Leave - the gap appears wider.

2.Voter turnout was high- but anything below 100 % won't do for deciding on such a contentious issue 

While the overall voter turnout at 72.1 % was quite high as voting in democracies go (The last general elections in India had the highest ever voting % for India at 66.38%), it was far short of 100%. Let us understand why even a 72% voting may not be good enough to truly measure the will of the people. The plot below shows the voter turnout versus the absolute gap between the Leave and Remain vote %.

The graphic clearly shows that the Voter Turnout % going down as the absolute Gap between Leave and Remain increases. Thus the more uniform the views of a local area (with a clear majority in favour of Leave or Remain), the lesser is the voter turnout. Thus, the voting turnout is highest when the Leave and Remain voters were almost equally divided (the left hand side of the graphic), and lower where there was a clear gap in either direction. My explanation for this pattern of behaviour is the thinking embodied in all Democracies with "First Past the Post" system. Once a voter knows that her area is predominantly going to vote in a given direction, she may feel that her vote does not count and hence would not go out to vote. This kind of thinking does not work in Referendum when every single vote counts, and local area wins count for nothing individually - there is no MP or Congressman getting elected in the referendum process.
One may argue that a 72% turnout should be able to measure the overall sentiment accurately. This is backed by the Law of Large Numbers thinking - where a truly random sample, though small, can fairly accurately assess the % in favour or against a particular issue. This is why the poll results are reasonably accurate,
For this thinking to work, the 72 % who voted should be no different from the 28% who did not vote. But such was probably not the case in this Brexit referendum, Refer the graphic again, and you can see that the voter turnout % shows a steeper decline in case of Remain than in the case of Leave. These are two different groups, with different levels of commitment to the cause as measured by the voter turnout.

3, Voter turnout % differs for the Leave and Remain groups with consequent impact
The relationship between the voter turnout and the absolute gap (between Leave and Return %) is clear. But the data also shows that the Leave group was more committed and the local areas where Leave had a majority shows a higher voter turnout. This can be seen in the tabulation and graphic below. I have categorised Local Areas with an absolute gap of > 20 % as COMPLACENT, and those with gap of less than 20 % as ENGAGED.

As can be seen above, the Voter Turnout % is overall higher when the voters know that there is a tough fight ahead (Engaged Category - Gap of less than 20 %) than when the voter is complacent knowing that the results in her area is clearly going one way. This behaviour of increased voter turnout is seen in both the Local Areas which voted to Remain and those which voted to Leave. However, the Local Areas where the majority voted to Leave shows a higher voter turnout in each category.This is the key to the final conclusion. The 72 % voters may not be truly representative (in statistical terms) of the 100 % population, and the results could have been different if there was 100 % turnout.
It is standard thinking based on the work done by behavioural economists that losses loom twice as large as gains. In a Zero-Sum game with the population equally divided between those who perceive the outcome as gains and those who perceive it as a loss, it is the group which perceives the outcome as a loss which would be more motivated in putting up a fight. In such a case, the people who actually come out to vote would be skewed in favour of those who perceived the outcome as a loss, and hence not be representative of the population.This pattern would not be so easily discerned in a pre voting poll, but the self-selection bias would manifest itself on the day of the actual voting.
The Graphic and tabulation above shows that the voters in favour of "Leave" ensured a higher turnout. More than the Status-Quo bias, it was the asymmetry of Gains and Losses which finally decided the issue.
If the Leave and Remain groups were equally divided, and the Leave group was 5 % more motivated than the Remain group to go out and vote,(ie, if Remain Group has 70% Voter turnout rate, then Leave group at 5% higher rate would be 73.5% turnout)  then any voter turnout less than 96% of the total electorate would lead to the Leave group winning by 51.2 % against 48.8 %. Only a 100% voter turnout could reveal that the two groups were equal.
This is the problem with Referendum, where the two opposing sides, though having clear views on which way they stand on the issue may have different levels of motivation to go out and vote. Even a 5 % difference in the motivation level and consequent voter turnout would lead to results different from what the population as a whole wants.
Referendums should be used only when voting is made mandatory and near 100 % turnout is assured. I hope that we do not see a spate of referendums in the EU with an outcome which does not truly represent the views of a fairly equally divided population.


Election results data available at the UK Electoral Commission website and can also be downloaded here.

Civil Services Examination - Game of Thrones or Game of Chance?

The Civil Services Examination of India is one of the most selective examinations in the world with probably the lowest success rate for any competitive examination. In 2015, a total of 465,882 candidates appeared for the first stage of the Exam – Prelims, of which 15,008 (3.2%) were selected to appear for the next stage of written examinations – Mains. Around 20 % of these candidates, specifically, 2797 were called for third and final stage – the Personalty Test or Interview of which a total of 1078 candidates were finally selected and recommended for appointment to one of the many Services of the Government of India, notably, the IAS, IFS, IPS, IA&AS, IRS, ICAS, IDAS, IRAS, IP&TAFS etc. The success rate is thus a measly 0.23%. 

I took the Civil Services Examination in 1994  and again in 1995, cleared on both occasions, and  based on my 1995 exam rank of 136, joined the Indian Audit & Accounts Service (IA&AS) in 1996, and have been happily working with the Audit Department since then.

In those days, the marks secured by the successful candidates were not available in the public domain. Recently, when  I came to know that this data is available on the UPSC website, I decided to analyse the same. Here is my analysis, based on the presentation that I made to the Civil Services Officer Trainees (Probationers) of the various Accounts and Finance Services  undergoing training at NIFM (where I am presently on deputation as Professor).

The dataset in MS Excel and pdf format can be downloaded here. (MS Excel / PDF)

The above visualisation shows the stratification of final Rank (bin size of 85 ranks) with relative share of the four different categories in different rank bin. This is fairly stable till about Rank 600, which has been taken as the cut-off rank for subsequent analysis.

The above visualisation (Histogram) shows the very narrow band in which the written scores of most candidates lie - between 40 to 45 % in both Civil Services Exam 2015 and 2014. (Each bin of width 1 %)
The Personality Test (Interview) however shows a very different behaviour - the spread here is far wider. A very interesting feature of the interview scores distribution is the "lumpiness" of the data, with prominent spikes at round number scores of 55 %, 60 %, 65 % and 70%. Based on this, one can safely infer that a) The UPSC Interview board is assessing the candidates on a percentage scale and then converting it into marks (out of 275) and b) Some of the Interview Boards are not attempting to be very precise in their assessment, and are willing to grade candidates in round number scores (say 65% instead of 64 % or 66 %).

What could be seen visually in the histogram is now clear in the above tabulation of the Standard Deviations of the Written and Personality Test (PT - Interview) Scores.

The importance of each Mark in determining the final rank of the candidates can be seen in the visualisation above, which shows the number of candidates with the same Marks - right from the highest score of 1063 with Rank 1 at the left to a score of around 877 with Rank of 600 at the right . While there is sufficient gap between the candidate's marks in the top 10 or so ranks, we have 4-5 candidates per mark around the 50th Rank and 9-10 candidates per Mark around the 100th Rank, going as high as 40 candidates with the same score at around Rank 600 !!
Each mark matters - and with Interview Marks showing a round number bias and using a % scale (with each % equal to 2.75 marks) added to the low reliability, an element of randomness/chance is introduced in the selection process of the candidates.

The final rank of the selected candidates is based on the total of the Written and Personality Test (Interview) scores. If UPSC was looking for a certain attribute, lets call it "IAS-ness" for want of any better name, which was present in both Written Exam and Personality Test, then these two scores would have shown some relation (correlation) with each other. For example,  the oft-quoted spurious correlation seen between ice cream sales and shark attacks occurs because both are related to the common variable - Temperature. We however do not see any correlation between the Written and Interview Scores - actually , a negative correlation is seen. This happens because the successful candidate with a poor Interview score necessarily has to have a higher Written Score and vice-versa; else she would have not made the cut-off.

The above definitions and examples of Reliability and Validity are taken from the excellent book by Richard Nisbett - Mindware-Tools for Smart Thinking. (This book, along with Daniel Kahneman's Thinking Fast & Slow are mandatory reading for all).
Which brings us to a very important Question:

A serendipitous natural experiment dataset was available within the Civil Services results for the year 2013,2014,2015. As a fairly large number of candidates clear the exam in successive years, their Interview scores in the two years could be examined to see the Reliability of UPSC Interview.
The above visualisation- scatter plot, which shows the correlation between the Interview scores of the same set of candidates in two successive years was the most surprising (and disturbing) finding. One can very clearly see that there is little if any relation between how a candidate may fare in the Personality Test in two successive years. With an R-Squared of around 0.1, the Interview process can be said to have limited Reliability(and as a consequence also of limited Validity.).
Whatever the Personality Test tries to assess (the "IAS-ness") lack of reliability leads to a doubt on the validity of the measurement.
Daniel Kahneman speaks about the "Illusion of Validity" as he recounts his experience of evaluating candidates for officer training as part of a group of evaluators. In his words "Our impression of each candidate's character was as direct and compelling as the color of the sky.... A single score usually came to mind and we rarely experienced doubts or formed conflicted impressions." However, the assessment of the candidate was in variance with the actual performance of the candidates in the officer-training school. As Kahneman goes on to say-"our ability to predict performance at the school was negligible. Our forecasts were better than blind guesses, but not by much". 
Richard Nisbett calls it the "Interview Illusion", and says that "predictions based on the half-hour interview have been shown to correlate less than 0.10 with performance ratings of undergraduate and graduate students, as well as with performance ratings for army officers, businesspeople, medical students, Peace Corps volunteers, and every other category of people that has ever been examined".

While the Personality Test (Interview) is seen to show limited Reliability, the Written scores of the same candidates across successive years shows a slightly higher consistency (Greater R-squared value), but not as high as a typical measure of pure ability.

The final rank of the candidate is based on the total of her Written and Interview Scores. The higher the score, the better the rank. Since it is not the absolute but the relative score that determines your rank, a regression model was built to predict the rank in percentile terms based on the candidate's percentile on her Written score and the percentile on her  Interview score in relation to the other candidates. The model was developed for the top 600 ranks for Civil Services Exam 2014 and 2015, and is shown below:

The model shows a good fit with the data as measured by the high R-Squared value, and based on the relative weights to the two parameters, it is seen that the Personality test carries a 40% weight in determining the final Rank!

Interview or the Personality Test has a high 40% impact weight, but at the same time it shows low reliability (year to year score consistency). The Written scores too can vary across years.
So what should the Civil Services aspirant do?

While the above is a strange (and funny) trend with lower roll number more likely to lead to success in the exam (Which way does the Causal arrow point?), the candidates need to rethink their Civil Services strategy.

So, any Civil Services aspirant should be prepared for the long haul.
The median number of attempts of the candidates who finally join the service is likely to continue increasing.
And what a waste of precious years of the talented  young Indian boys and girls slogging away to clear a highly unpredictable exam.
I hope that the expert committee constituted by the Government under the chairmanship of Sh BS Baswan to examine the various issues regarding the Civil Services Examination is able to fix some of the problems in the current system, and more importantly, prevent the colossal waste of time and effort of the youth of India.

Bounties for the Well-Off?

In one stroke of his pen, Arvind Subramanian – Chief Economic Adviser to the Government of India, has made me and most of you become the new Indian economic elite. We have been kicked hard - Upstairs - and have vaulted from our cosy conception of being part of the Indian "middle class" to the Piketty reviled elites forming the 1 %. (ET ran this article on 27th Feb 2016)

Be warned, for the Occupy Dalal Street movement that you were probably contemplating of leading or joining, is now a protest against you- the economic elite of India.

NO, the economic adviser did not promise any Bounties to make us all well-off, he  did the math and said that anybody earning more than 2 lakh rupees in India is Well Off (thats you and me, and anybody you know). And the meager tax breaks that he/she gets, are the bounties that needs to be clawed back by the Government of India.

His simplistic argument can be simplified further.

Anybody in the top 2 % of the population is a part of the economic elite, top 5-6% are definitely Well-Off and not a part of the vaunted middle class of India.

Based on the Income Tax data, any person in India having an income of more 2 lakh is in the top 6 % of the population, earning more than 5 lakh puts her in the top 1. 5% and more than 10 lakh places her in the envious top 0.5 % of the population.
A key assumption that is made here, which only the very naive would accept, is that this actually represents the true income distribution in India. The number of tax payers has remained near stagnant at around 3 crores for at least a decade now; and not because the number of people with taxable income has not increased.

And this cake shaped graphic takes the cake:

The cake above conceals more than it reveals. Lets try to understand it better.

The graphic is drawn to show the percentiles. (the axis on the left). But the markings on the right, which shows the income slabs are not drawn to scale.
The gap between 5 and 10 lakh appears to be less than a third of the gap between 2 and 5 lakh. And the upper slab of taxable income, which spans Rs 10 lakh to Rs 160 Crore (Yes, 1600 times more than the entry income of 10 lakh) is shown as a narrow sliver. Had this been drawn to the scale of income, the "cake" visualisation would have looked more like burj Khalifa, probably taller.
In this narrow looking band of 0.5 % of the population is where the real income variation is seen, and the true "Rich" and "Well-Off" reside.
To say that the meagre investment made by people in PPF or GPF Accounts, where the maximum investment (with tax saving incentive) is capped at Rs 1.5 lakh per annum, leads to a subsidy for the "well-off" adds insult to injury. This is no "bounty" to the well-off - for most, this is the only investment that they make from their sub-10 lakh salary.
The rich, who inhabit the broad earning range from 10 lakh to 160 Crores, are not investing their surplus in PPF accounts. Their investments lie in financial assets where capital gains are taxed far more favourably.
Instead of targeting the small savings (Yes EA, these are small savings) made by people with income between 4 lakh and 10 lakh, the Economic Advisor could have raised tougher questions - like why capital gains should not be taxed at a higher rate, and why the peak income tax slab starts at Rs 10 lakh and stops at 30 %, when we have people in India with income a thousand times greater than this? Even the US, whose free market ideology we seem to be aping, has higher income tax rates of 33%, 35% and 39.6 % for the top income slabs.
Hope the government did not take the Economic Adviser's advice seriously, and ignored chapter 6 of the Economic Survey (Bounties for the Well-Off) while finalising the Budget proposal.