The farm story repeats...

I have lamented here and here at the daunting challenge posed by India's agriculture. Indian Express has this article on the increased jeera cultivation acreage in Gujarat following record prices this year. Consider this,

Prices of jeera (cumin seed) hitting record Rs 21,000 per quintal levels at Gujarat’s Unjha market... In a year marked by low realisations for most agricultural commodities — be it cotton, groundnut, potato or tobacco —... jeera has been an exception... The above spike in prices has led farmers in Gujarat to plant 3.48 lakh hectares (lh) under jeera in the current sowing season, which extends from November to early December. This area, which is subject to upward revision, is significantly more than the 2.79 lh of last year. All the major growing districts have recorded higher coverage: Patan (from 40,700 to 68,100 hectares), Banaskantha (from 64,900 to 67,800 hectares) and Ahmedabad (from 12,400 to 19,800 hectares) in north/central Gujarat; Porbandar (from 12,100 to 23,200 hectares), Rajkot (from 4,900 to 22,000 hectares) and Devbhumi Dwarka (from 4,800 to 20,400 hectares) in Saurashtra; and Kutch (from 23,200 to 29,600 hectares)... Ram Patel has sown jeera in seven out of his 10-bigha holding this time. Last year, he grew the crop only on three bigha, while dedicating the remaining seven bigha for mustard (six bigha make a hectare). Jeera is extremely sensitive to weather fluctuations, making its cultivation riskier relative to wheat, chana (chickpea) and dhaniya (coriander), which are the other major rabi season crops in Gujarat. Overcast skies or dew at the time of maturity — jeera is harvested from February to March — can sometimes even lead to complete crop loss.

It is unlikely that policy actions can nudge farmers away from such herding based on previous year's prices.

The binding constraint against shifting in response to such trends is unlikely to be information. After all a farmer is likely to have seen several such cycles in his lifetime. It is more likely a deep behavioural bias that attaches excessive weight to the immediate experience over the less immediate experiences. Accordingly, the high incomes from the past season weighs so heavily on the mind of the farmers that they are willing to gloss over the memories of pain from earlier experiences of market cycles. What can be done to overcome this cognitive bias?

I would say precious little. Unless the farmer's state of poverty is alleviated, information supply will remain largely ineffective. This excessive preference most likely is a result of poverty and the desperation and insecurity arising from it. The farmer behaves like a poor person who fancies lottery tickets. It is the prospect of the windfall and the perceived favourable odds that make them take the bet.

This is a cautionary note for those who are bought into the story that providing (some even claim to want to sell) information to farmers or the poor can be a solution to many of their deep underlying reasons for distress. They are most likely to be disappointed.

Update 1 (06.01.2018)
A comment informs that the Agriculture Department of the Government of Andhra Pradesh succeeded with an information and awareness campaign to moderate the extent of cotton crop sown this year based on an assessment that cotton is not likely to do well this year.

Now there is a difference between good old public information campaigns (effectively done) and nudging  privately with information. The post was in the context of the latter, and did not make the distinction. Privately providing information as a nudge to farmers is qualitatively different from a massive public information campaign that targets the farmers as a collective. With the latter, mobilising a few influencers (or a platform like farmers groups) can swing the collective resolve towards the change.

So let me reiterate - privately nudging individual farmers with information to shift entrenched patterns or behaviours is unlikely to be effective especially given the (cognitively biased) perceptions of costs and benefits associated with status quo and change.

The problem with the standard information and awareness campaigns of government is that in systems with weak state capacity they are generally executed badly in a routine manner and to that extent becomes ineffectual. In the instant case, I have little hesitation that the outstanding individual heading the State's Agriculture Department made the difference.

The rote-learning trap

A very disturbing RISE study on the state of the high stakes high school assessment systems in India by Newman Burdett. Its verdict is damning, though not at all surprising,

Overall the CBSE examination papers are heavily biased to rote-learning, do not test higher- order skills, and actively discourage students who try and display them... they are almost exclusively assessing a large body of facts and intimate familiarity with text books that is of very questionable use... The papers from CBSE and Pakistan do not match in any meaningful way what is considered necessary or good education in any of the available benchmarks.

What is surprising is the magnitude of the problem and the fact that the quality is worse than in even countries like Uganda and Nigeria.  

The study classifies questions in each subject in the national-level standardised examinations based on recall (having to recall a fact or piece of knowledge), application (understanding that knowledge and applying it), and reasoning (critically analysing and evaluating facts and potentially putting those pieces of knowledge together in novel ways). It then seeks to identify the share of questions belonging to each category in the different subjects in each of the standardised examinations for all countries. 

While the PISA, TIMSS, PIRLS and other international assessment benchmarks focus more on application and reasoning, minimising or even avoiding recall, India's CBSE assessments for 10th and 12th Grades have no questions that test reasoning and are skewed towards recall.

Even those in the Level 2 category for Math is effectively reduced to a recall category given the repetition of questions (or at least their nature) over years. 

For reference, see the distribution for other major benchmarks...

... and for Alberta, Canada 

Achieving learning outcomes of desired levels will take time. A good place to start would be to signal intent and shape expectations by making the requisite changes to the assessment system itself. And it can start now.

I confess to having hitherto underestimated the depth to which our assessments have plunged and also the potential value of this as an important policy instrument in improving learning outcomes. This really is important and very useful practical lever to improve learning outcomes. 

Update 1 (02.01.2018)

Karthik Dinne has a very good suggestion,

The solution is to upgrade the quality of board exams but to avoid the political pressures during the transition, it would be good to initially conduct two versions (basic, advanced) of each subject, with the choice to the student to choose either. Gradually, one can work on bridging the gap between the two. The idea is that the advanced version will signal the higher expectations to schools and also creates demand from parents by providing them with an easily understandable metric gauge quality. Hopefully, this would lead to change in the status-quo. The splitting of the exam will also serve another problem of our board exam - the dual purpose of both certifying minimum standards and signaling the capability of students. The basic version can be taken by students who are interested only in getting the 10th certificate to demonstrate minimum competency.

Please also see his blogpost on this here.

Market participants game regulations, bureaucracy responds with more red-tape...

I had blogged just recently on India's private sector's collective propensity to game regulations wherever there is an opportunity and the associated bureaucratic response to err on the side of caution with restrictive regulations. 

Consider the latest example from the compensation scheme of the GST,

Goods and services tax (GST) tax returns filed for the July to September 2017 period by firms under the ‘composition’ scheme suggest there is massive tax theft by smaller taxpayers. The composition scheme is a special one to make GST filing easier for small firms; apart from simpler tax procedures, the returns have to be filed once a quarter. To that extent, the government’s plan to bring in the e-way bill and other ways to plug tax theft are quite justified. While there are roughly 15 lakh small firms registered under the composition scheme today, the number was around 10-11 lakh in September. Of these firms, around six lakh filed their returns for July-September by December 24. The total tax they paid was around Rs 250 crore. Assuming a 2% tax incidence on their turnover — it is 1% for traders, 2% for manufacturers and 5% for restaurants — this means these firms had an average turnover of Rs 2 lakh in that period, or Rs 8 lakh for the full year if you annualise the data. The problem, however, is that firms that have a turnover of less than Rs 20 lakh a year, don’t even need to pay GST or file returns. In other words, these firms are understating their returns in a big way.

Let's face this. This is not a problem at the margins - a small share of firms indulging in tax evasion. The sheer scale of evasion is staggering. It is clear that the major share of firms are complicit. In fact, not just in this case, corporate India has shown a consistent collective proclivity to game regulations to maximize illegitimate private gains. 

This story repeats. Government implements a reform. Private sector blames the reform regulations for being very restrictive. Government responds by relaxing the regulations. The private sector responds by collectively gaming the regulations to their benefit. Then why blame the government alone for excessive red tape, especially when at stake is a non-trivial share of the country's tax base?

If this is the scale of gaming, how can we blame the inspections and reporting and other requirements of the Department of Revenue from becoming more intrusive?

We forget that trust is the basis of well-functioning markets. Regulations and enforcement are built on trust, and are largely put in place as marginal deterrence - the vast majority comply on their own and only the small minority need the threat of enforcement to comply. However, when trust is consistently betrayed (or is deficient in the first place), then regulations and enforcement have to become the basis for ensuring compliance. 

In this repeat game with market participants, the bureaucrats have internalised a preference for excessive restrictions in anticipation of the inevitable deviation from businesses. 

New Year Greetings!

Wishing all readers a very happy year ahead in 2018!

Restoring the old glory of the IAS

I have a co-written oped with Dr D Subbarao in Indian Express which calls on reinventing the IAS - transform itself so as to be able to lead the social transformation, lead the movement to discard the chalta hai attitude and embrace a badal sakta hai attitude. 

Thoughts on NCLT and GST

The two biggest economic reform events in India in 2017 have been the roll-out of the Goods and Services Tax (GST) and the operationalisation of the Insolvency and Bankruptcy Code (IBC) through the National Company Law Tribunal (NCLT). A few observations on both. 

On GST, the government faced two challenges. One, achieve consensus among States through protracted negotiations involving give-and-take so as to get the Parliamentary approval. Two, operationalise GST among millions of enterprises through a very intrusive IT solution, the vast majority of whom had never used such systems.  

Both these meant that the best GST policy design and implementation plan was off the table. The challenge was to negotiate and ensure the most satisfactory second-best solution. 

In a raucous democracy with high-stakes negotiations conducted in full public glare, that too in real-time and at frenetic pace to meet an imminent deadline, and given the confrontational collective attitudes and behaviours of all parties concerned, the government cannot be faulted for prioritising the achievement of consensus to get legislative nod over the purity of the GST design. After all, nothing about the design is set in stone. 

The implementation plan, while largely a bureaucratic exercise, is critically dependent on the design. When the design features were subject to negotiations till the last minute, the challenge of incorporating the changes, and testing and validating them in very aggressive timelines becomes formidable. In any case, as countless examples of such massive project roll-outs show (recollect the chaotic roll-out of the Affordable Health Care Act insurance exchanges in the US), the challenge is not so much to get the best solution out, but to respond swiftly to emergent challenges and rectify them at the earliest. The initial few months of adoption chaos is par for the course. 

A very active GST Council has been consistently refining the GST design elements since its roll-out. Given that these changes have to be incorporated into the IT system, almost in real-time, the implementation challenge looks even more daunting. 

I am not well-placed to comment on the basic software design features (not related to emergent GST design changes), including ease of use for the largely IT illiterate users, of the GST solution or on the implementation processes. On the former, given that one of India's leading software companies has been in charge of its development for a long-enough time, I would be inclined to hold them responsible for such technical failings. On the latter, we need to examine the robustness of decision-making and process protocols, within the implementation Secretariat, as well as engagement with State Commercial Taxes departments. Was the commitments made for incorporation of GST Council decisions into the IT solution based on judgements that accounted for technical considerations? The bureaucracy cannot absolve itself off the blame on failings in these areas. 

It would be very useful to study, compare, and learn from the decision-making protocols and processes associated with Aadhaar and GST, the two big IT solution roll-outs in the country's history.

More changes on the GST design features are on the anvil. In any case, a year-long iterative process and an implementation stabilisation period of 12 months is not at all unreasonable for such roll-outs. The verdict on the whether GST roll-out was done well or not will have to wait till then. 

My initial reaction after the disastrous first case resolved under the IBC was one of dismay. But the government reacted swiftly to make changes to limit promoters gaming the resolution process and gaining control through the backdoor, including prohibiting promoters with non-performing loans for more than a year from bidding with resolution plans. One can argue that those changes could have been easily anticipated while framing the original regulation and that the response in terms of completely shutting out promoters, and not just "wilful defaulters", from even placing proposals before creditors is too restrictive. 

On the former, I am sympathetic. But on the latter, I am not sure whether India's corporate eco-system is mature enough to responsibly engage with such flexibility and not abuse it. Any regulation is a trade-off involving flexibility that can be embraced or abused. A consistent feature of India's business landscape, across sectors, has been the egregious gaming of regulations, not so much at the margins but by large and responsible sections of corporate sector. In the circumstances, the standard bureaucratic response has been to err on the side of caution with restrictive regulations. The universal voucher reconciliation requirement or the Anti-Profiteering Agency in case of GST are examples. I feel that while reasonable people can agree or disagree on the specific set of restrictions in each case, it is important to appreciate the bureaucratic line of thinking in such cases. 

As Livemint reports, the IBC appears to have already triggered some changes in the behaviours of errant promoters. That is welcome news. And I feel this deterrence taking root and shaping debtor attitudes, could become the biggest achievement of this reform. But on the substantive part of resolution, I foresee challenges ahead with the capacity of the system to execute the resolution process with fidelity and that of the market to absorb these transactions. I foresee challenges and struggles. 

China facts of the day

The Economist has these figures on the Chinese appetite for lifts,

In 2000 some 40,000 new lifts were installed in the country. By 2016 the number was 600,000—almost three quarters of the 825,000 sold worldwide. China not only wanted more skyscrapers; it wanted taller ones. More than 100 buildings round the world are over 300 metres; almost all of them were built this century, and nearly half of them in China. The country is home to two-thirds of the 128 buildings over 200 metres completed in 2016. Other countries may content themselves with a few show-off pinnacles. China buys them by the dozen.

This is a fascinating (and disturbing) account of how Chinese digital companies are using personal data to score a person's social credit and offer the full spectrum of services covering the daily lives of Chinese. Sample this,

Alipay knows that at 1 pm on the afternoon of August 26, I rented an Ofo brand bike outside Shanghai’s former French Concession and rode north, parking it across from Jing’an Temple. It knows that at 1:24 pm I bought a snack in the mall next to the temple. It knows that afterward I got in a Didi car bound for a neighborhood to the northwest. It knows that at 3:11 pm I disembarked and entered a supermarket, and it knows (because Alibaba owns the supermarket, which accepts only Alipay at checkout) that at 3:36 pm I bought bananas, cheese, and crackers. It knows that I then got in a taxi, and that I arrived at my destination at 4:01 pm. It knows the identification number of the taxi that drove me there. It knows that at 4:19 pm I paid $8 for an Amazon delivery. For three sweet hours—one of which I spent in the swimming pool—it does not know my whereabouts. Then it knows that I rented another Ofo bike outside a hotel in central Shanghai, cycled 10 minutes, and at 7:11 pm parked it outside a popular restaurant. Because Ant Financial is a strategic investor in Ofo, Alipay might know the route I took.