Friday, April 30, 2021

TSMC graphics of the day

The Economist has an article on the rise and rise of TSMC as "the Hope Diamond of the semiconductor industry" or perhaps the most strategically important company in the world.

This is as much a story of the precipitous fall of Intel in slightly more than a decade, as it is of the spectacular rise of TSMC. 

In a world of declining investments, TSMC's investment appetite is massive,

In January it said it would raise its capital expenditure for 2021 to $25bn-28bn, up from $17bn in 2020. In April it raised the figure again, to $30bn. It plans to spend $100bn over the next three years. Fully 80% of this year’s spending will go on advanced technologies.

And this about its profitability,

Last year TSMC made an operating profit of $20bn on revenues of $48bn... It has also stopped cutting prices—which in chipmaking, where processing power has only ever got cheaper, is tantamount to raising them... IC Insights, a research firm, calculates that TSMC can charge between twice and three times as much per silicon wafer when selling its most advanced processes, compared with what the next-most-advanced technology will fetch. This creates a positive feedback loop. Developing the latest technology before anyone else allows TSMC to charge higher prices and earn more profit, which is ploughed back into the next generation of technology to continue the cycle. And the cycle is spinning ever faster. Four technological generations ago it took TSMC two years for those cutting-edge chips to make up 20% of revenues; the latest generation needed just six months to reach the same level (see chart 2). Operating income, which grew at an average rate of 8% year in the decade to 2012, has since risen at an average rate of 15%. Combined with revenues that chip-designers make from semiconductors ultimately forged by TSMC, the company and its customers account for 39% of the global market for microprocessors, according to VLSIresearch, up from 9% in 2000 and a third more than once-dominant Intel.

As the example of Intel shows, TSMC's fortunes can turn quickly. Besides, its biggest threat is what can happen across the Straits with China. And Taiwan's risk mitigation strategy makes TSMC even more vulnerable,

Taiwanese government encourages all its chipmakers, including TSMC, to keep their cutting-edge production on the island as a form of protection from foreign meddling... Reflecting this, 97% of TSMC’s $57bn-worth of long-term assets reside in Taiwan (see chart 3). That includes every one of its most advanced fabs. Some 90% of its 56,800 staff, of whom half have doctorates or masters degrees, are based in Taiwan. The firm has made soothing noises to both America and China and offered to invest more in production lines based in the two superpowers. But it is hard not to see this as diplomatic theatre. Its Chinese factory in Nanjing, opened in 2018, produces chips that are two or three generations behind the cutting edge. By the time its first American fab, designed to be more advanced than the one in Nanjing, is up and running in 2024, TSMC will be churning out even fancier circuits at home. By our estimates, based on disclosed investment plans, the net value of TSMC’s fabs and associated equipment will roughly double by 2025 but 86% will still be located in Taiwan.

This is another excellent read in FT about TSMC and its dominance of the cutting-edge of chip making. 

Update 1 (19.05.2021)

Nice Bloomberg story on why it's very hard to make semiconductor chips.
It takes years to build semiconductor fabrication facilities and billions of dollars—and even then the economics are so brutal that you can lose out if your manufacturing expertise is a fraction behind the competition... Manufacturing a chip typically takes more than three months and involves giant factories, dust-free rooms, multi-million-dollar machines, molten tin and lasers. The end goal is to transform wafers of silicon—an element extracted from plain sand—into a network of billions of tiny switches called transistors that form the basis of the circuitry that will eventually give a phone, computer, car, washing machine or satellite crucial capabilities. Most chips are groups of circuits that run software, manipulate data and control the functions of electronic devices.
Update 2 (03.08.2021)

Good primer on the global chip industry by Prosenjit Datta

Thursday, April 29, 2021

India's unicorns universe

Credit Suisse recently released this hugely informative report which generated considerable interest. It trawled the entire economic landscape to identify all the privately held companies with valuation of greater than a billion dollars. 

Against 336 listed companies with billion dollar-plus market capitalisation, there are now 100 unicorns in India with a combined market capitalisation of US$240 bn. Two-thirds of these firms started after 2005, whereas only 46 of the listed firms were founded this century, and as many as 112 started before 1975. The sectoral split is highly diversified: in addition to the largely expected e-commerce, FinTech, education technology, food delivery and mobility companies, there is a rapidly growing number of such firms in Software- as-a-Service (SaaS), gaming, new-age distribution and logistics, modern trade, bio-tech, pharmaceuticals, and even fast-growing consumer brands benefitting from accelerating penetration and formalisation. These are only at the top of a fast-growing pyramid of 80,000 start-ups in India, which are incrementally now nearly 10% of new companies formed every year; the number of firms is up 70% in 8 years. There is some geographical diversity in the cities where these firms started, though there is some concentration in Bengaluru, Mumbai and the National Capital Region (NCR), Delhi.

Some graphics from the report.

Private equity, mostly foreign, has outstripped public capital raising even in India in each year of the last decade

While the emergence of these private equity funded unicorns is a positive development, it should be a matter of concern that almost all this capital is foreign. Coupled with the fact that most of these unicorns are incorporated outside India, it is fair to say that the best ideas and businesses emerging from India are foreign owned. I had blogged here on the practice of flipping, whereby foreign investors force founders to register outside India, with the attendant wealth transfer from India. 

Are India's unicorns the conduits for the biggest systematic wealth transfer from India to foreigners after the British colonial period?

This on the age of the unicorns and the listed companies is striking
There has been an undoubted increase in the dynamism of Corporate India since the millennium as manifest in the form of new business incorporation. The number of active companies have grown from 700,000 to 1.3 million in just 8 years.
Even among these, creation of startups, those which are completely new entities, and not subsidiaries of existing businesses, too have risen sharply.
They now form over 10% of new company registrations, from below 7% in 2013.
Manufacturing continues to remain marginal as a destination for private equity 
Information technology, discretionary and healthcare continued to be the dominant sectors in 2020. The decline in volumes from peak in percentage terms has been the worst for industrials: from 113 in 2008 to just 56 in CY20. Staples saw a record high 106 deals in 2019, but the value per deal was just US$6 mn.

IT, and services like health care and education, and discretionary goods (personal care, spirits, packaged foods etc) remain the main destination for private equity. Aggregators (of people, finances, and goods), content delivery providers, and software firms, propelled by the advances in IT, are perhaps the biggest attractions for venture capital.   

This about the paucity of risk capital in India is important,
Low per-capita-GDP economies like India are generally short of equity capital. Not only do they have low wealth per capita, most household wealth is in hard assets like the land they own, the house they live in, the shop they own, the vehicle they use, or gold. The share of their wealth in financial assets is low. Further, even for financial wealth the first investment preference for households tends to be capital-assured asset classes like bank deposits. It is only beyond a certain quantum of weatlh, and usually after the purchase of a house that households begin to invest in equity of firms they do not run themselves. 

How did this happen in other economies when they were at the per capita income and wealth of India? First, the transition of currently developed markets out of their emerging market status occurred over a much longer period, with growth averaging 2-3% a year over a century or more. Wealth thus accumulated over a period of time, and was able to provide the risk capital necessary to finance this level of growth. To grow at 6-8% annually for a few decades though one needs significantly large amounts of risk capital. Some emerging markets (including the US in the early 19th century) also used debt as risk capital, and had several boom-bust cycles in debt. The Chinese model of growth has also relied on debt-funded growth, with state ownership reducing the risk of systemic instability when loans go bad. Given an aversion to boom-bust cycles of debt, and low availability of locally available risk capital, India has been short of risk capital. If a business needed Rs1 bn of equity capital, only a handful of business families could afford that.

The share of financial wealth has remained stuck at this level for at least two decades.
India adds 3.5 million engineering graduates each year, and now employs 4.5 m people in the IT sector, though its growth has been declining sharply. It has fallen from 25-30% till 2007 to about 5% now.


The report is without doubt very informative and an important addition to the area. However, its inferences are questionable and grounded more on optimism than any realistic or historical basis. The point about valuations of a start-up and associated wealth creation which can be form the basis for significant contributions to increasing risk capital in India is questionable. 

A unicorn is a valuation. Instead we need companies to create wealth and generate risk capital. Unfortunately, there is a big difference between unicorn creation and risk capital creation. Most often unicorns are vapour ware. Only companies create hardware and software. 

Neelkanth Mishra has opeds on the report here, here, here, and here. Understandably, these are promotional pieces to market the Credit Suisse report. Much of the optimism is grounded on logic and theory, but have little real-world basis. 

For example, it remains to be seen how the much hyped Edtech unicorns will go beyond marginally improving the learning environment of a tiny sliver of students from middle-class families to helping improve the massive problem of poor learning levels that affect more than 90% of Indian students. Or whether Agtech firms will address any of India's several agriculture sector problems. Or the biotech and medtech companies will address the problem of access to affordable and good quality health care for more than 80% of Indians. Or whether the fintech companies will help address the problem of improving financing intermediation by increasing access to mass-market financial products and increasing India's financial savings, besides making formal finance mainstream for the 80% of the labour force working in the informal sector. Or ensuring access to finance simple for businesses in the informal sector. Or whether, like Alibaba's rural Taobao's, India's e-commerce sector has significantly improved market access and incomes in the aggregate to small manufacturers and traders. 

In fact, a simple smell-test is the answer to the question, which of these unicorns have had a meaningful impact on the incomes of at least some of 80% of Indians?

In all these cases, the respective IT solutions will undoubtedly improve the experience for the better-off consumers. But its impact on those really needing such disruptions is deeply questionable. At the least, it needs to be acknowledged that there is absolutely no evidence of significant improvements in any of these areas. 

Highlighting the point, Nivedita Mookerji strikes the note of caution here

Update 1 (12.01.2021)

Livemint has a data feature on the Indian unicorns universe in 2021. Its startups raised $39 bn over the year, compared to $13 bn in 2020, added more unicorns in the year (44) than all previous years combined, and now has the third largest number of unicorns after US and China. 

Social platforms formed a major share of the capital flows
If you discount for copycat innovations, as all of the above are, it's hard to not argue that there is little by way of genuine innovation happening in India's startup universe. 

Finally, Delhi is clearly breaking away as the startup capital of the country. 

Monday, April 26, 2021

Capitalism and European Soccer League

The problems created by the concentration of economic power, a trend that's now universal across business segments and geographies, has been a recurrent theme of this blog. Corporate interests invariably seek to change rules of the game and raise entry barriers, thereby lowering competition and blunting capitalism's productive forces. 

A recent reassuring example that not all is lost in the fight against such anti-competitive form of capitalism came from European football league

On April 18th a dozen of Europe’s top football clubs announced plans to disrupt the game with a breakaway “Super League”... The plan was for 20 clubs to compete in a Europe-wide league, kicking off in August. Fifteen “founding” clubs would be guaranteed a spot every year, with the remaining five places awarded competitively... Investors cheered. But fans revolted, broadcasters turned up their noses and governments vowed to block the plan. Within 48 hours half of its founding members dropped out. It was soon declared dead... The 12 clubs that broke cover comprised England’s “Big Six” (Arsenal, Chelsea, Liverpool, Manchester City, Manchester United and Tottenham), plus three from Spain (Barcelona, Atlético Madrid and Real Madrid) and three from Italy (ac Milan, Inter Milan and Juventus... JPMorgan Chase was to stump up €3.3bn ($4bn) of financing to get the league off the ground.

This interesting snippet explains the motivations,  

The venture’s stated aim was to give the world’s best clubs more chances to play each other than Europe’s main existing club competition, the Champions League. Barcelona and Bayern Munich have faced each other fewer than a dozen times in their history. Big clashes would bring in more viewers and more money: the Super League’s organisers had hoped that broadcasting rights might generate €4bn a year, nearly double the €2.4bn brought in by the Champions League in the 2018-19 season... 

Automatic qualification looked even more appealing. Unlike American teams, European sides play in open leagues, where poor performers get demoted to a lower tier, with stingier broadcasting and sponsorship deals. Club owners thus gamble on making it to the top, investing generously at the expense of profits. In closed contests like America’s National Football League (no relation to what Americans insist on calling soccer), clubs face no risk of relegation and so co-operate more. “Draft” systems allocate talent more equally and wages are often capped—something that the Super League hinted it might do, via an agreed “spending framework”. Clubs in closed leagues must worry only about economic competition from rival leagues, which require more upfront investment to start than an individual club. The combination of less risk and less competition for talent produces higher profits for owners. Forty-three of the world’s 50 most valuable sports teams are American, according to a ranking last year by Forbes magazine. By contrast, European sport is a dicey business: between 1992 and 2014 there were 45 insolvencies in the top three tiers of English football, 40 in France and 30 in Germany. “Football is essentially insolvent,” notes Stefan Szymanski, a sports economist at the University of Michigan. Without their deep-pocketed owners, most clubs would not be going concerns.

The backlash among fans was so much that JP Morgan apologised for backing the breakaway league. The FT has this long read about the episode.  

Some observations

1. The theory of Econ 101 inform us that competitive markets create fair and efficient outcomes. It is argued that business competition spurs dynamism and innovation, which in turn enhances quality and lowers prices. It's also argued that this competition keeps everyone honest and thereby prevent market abuse. 

However, in the real world, markets across sectors, when allowed to operate untrammelled, increasingly end up concentrating market share among a few companies. The dynamics of market evolution - aided by trends in technology, consumer tastes, marketing, and regulatory capture - leads to the emergence of dominant firms. The death-blow to competition comes from the emergence of entry barriers, often implicit or innate to the service or product. This results from the dynamics of the prevailing elite-captured political economy. 

Karl Marx was profound in claiming that there is an inexorable trend with the internal dynamics of capitalism which leads to exploitation and market abuse. The "competitive" world of European soccer league has resulted in this deeply anti-competitive denouement, one which US reached several decades back.

2. On a related note, despite its enduring image as the land of competitive capitalism, there is a case that American capitalism was always characterised by corporate power. It's success may have been because despite the dominant trend, there were always spaces available within the system which allowed for innovation and disruption. Anyways, in the world of multiple equilibriums, the US economy, for whatever reasons, had settled on to its own unique equilibrium which fortunately turned out to be productive and innovative. 

The US soccer league is a great example of such closed-club capitalism. The cabal which run NFL have effectively captured the market with its captive consumers, and nobody even realises it. The entry barriers are insurmountable. 

3. As the example of JP Morgan's support shows, Big Finance is never far away when Big Corporates make their moves. Thanks to financialisation, Big Finance and Big Corporates are all part of one Big Capital universe. With the scale of economic inequality, Big Capital has managed to seize control of the decision-making processes in the political system.  

4. The example of the European soccer league highlights the power of organised consumers, in this case fans. When consumers can overcome collective action problems and come to collectively exercise their preferences, that's a powerful check against Big Capital. It helps that European clubs are strongly tethered to their cities and communities of origin. 

Unfortunately, unlike club football and fans, collective action failures are the norm elsewhere. Machiavelli famously alluded to the "uncertain support of the powerless many (customers and community) compared to the strong enemity of the powerful few (the oligopolists)". New entrants face the same forces in modern capitalism.

Update 1 (30.04.2021)

FT reports that PE fund Silver Lake has taken a 12.5% stake (at USD 281 m) in the commercial rights of the entity that runs the New Zealand's All Blacks rugby team, despite opposition from players' association. 

The deal, if approved by the association, would initiate a battle for influence in one of the world’s most popular sports between Silver Lake and rival CVC Capital Partners at a time when the Covid-19 pandemic has smashed the rugby revenue stream. Luxembourg-based CVC already owns minority stakes in the English Premiership, Pro14 club competitions and the Six Nations. It has also held talks with South Africa, the reigning world champions, about buying a 15-20 per cent stake in the sport’s commercial arm in the country... New Zealand Rugby made a loss of NZ$34.6m in 2020 on a NZ$55m revenue fall caused by the coronavirus pandemic, which has disrupted fixtures.

Saturday, April 24, 2021

Weekend reading links

1. Harish Damodaran has an excellent oped about the real number of farmers in India, questioning the conventional wisdom of 100-150 million farming households,

The 2016-17 Input Survey report shows that out of the total 157.21 million hectares (mh) of farmland with 146.19 million holdings, only 140 mh was cultivated. And even out of this net sown area, a mere 50.48 mh was cropped two times or more, which includes 40.76 mh of irrigated and 9.72 mh of un-irrigated land. Taking the average holding size of 1.08 hectares for 2016-17, the number of “serious full-time farmers” cultivating a minimum of two crops a year — typically one in the post-monsoon kharif and the other in the winter-spring rabi seasons — would be hardly 47 million. Or, say, 50 million.

The above figure — less than half or even a third of what is usually quoted — is also consistent with other data from the Input Survey. These pertain to the number of cultivators planting certified/high yielding seeds (59.01 million), using own or hired tractors (72.29 million) and electric/diesel engine pumpsets (45.96 million), and availing institutional credit (57.08 million). Whichever metric one considers, the farmer population significantly engaged and dependent on agriculture as a primary source of income is well within 50-75 million.

This point about the farmer's worsened terms of trade is instructive,

In 1970-71, when the minimum support price (MSP) of wheat was Rs 76 per quintal, 10 grams of 24-carat gold cost about Rs 185 and the monthly starting pay for a government schoolteacher was roughly Rs 150. Today, the wheat MSP is at Rs 1,975/quintal, gold prices are Rs 45,000/10g and the minimum salary of government schoolteachers is Rs 40,000/month. Thus, if 2-2.5 quintals of wheat could purchase 10g gold and pay a government primary schoolteacher’s salary in 1970-71, the farmer has to now sell 20-23 quintals for the same. Fifty years ago, one kg of wheat could buy one litre of diesel at MSP. Today, that ratio is upwards of 4:1.

The real motivations and political economy in demanding MSPs,

The demand for making MSP a legal right is basically a demand for price parity that gives agricultural commodities sufficient purchasing power with respect to things bought by farmers. It is coming mainly from the 50-75 million “serious full-time farmers” who have surplus to sell and with real stakes in agriculture. They are the ones whom “agriculture policy” should target. Most government welfare schemes are aimed at poverty alleviation and uplifting those at the bottom of the pyramid. But there’s no policy for those in the “middle” and in danger of slipping to the bottom. 

An annual transfer of Rs 6,000 under PM-Kisan may not be small for the part-time farmer who earns more from non-agricultural activities. It is a pittance, though, for the full-time agriculturist who spends Rs 14,000-15,000 on cultivating just one acre of wheat and, likewise, Rs 24,000-25,000 on paddy, Rs 39,000-40,000 on onion and Rs 75,000-76,000 on sugarcane. When crop prices fail to keep pace with escalating costs — of not only inputs, but everything the farmer buys — the impact is on the 50-75 million surplus producers. They have seen better times, when yields were on the rise and the terms of trade weren’t as much against agriculture.

2. Mariana Mazzucato writes,

Consider the $40 billion the US government invests every year in the National Institutes of Health. The NIH (along with the US Department of Veterans Affairs) backed the Hepatitis C drug sofosbuvir with over 10 years of taxpayer-funded research. But when the private biotech company Gilead Sciences acquired the drug, it priced a 12-week course of pills at $84,000. Similarly, one of the first antiviral treatments for Covid-19, Remdesivir, received an estimated $70.5 million in public funding between 2002 and 2020. Now, Gilead charges $3,120 for a five-day course of it.

This speaks to a parasitic, rather than a symbiotic, partnership. The NIH must do more to ensure fair pricing and access to the innovations it funds, rather than chipping away at its own power, as it did in 1995 when it scrapped the Fair Pricing Clause from its cooperative research and development agreements.

3. Fascinating FT long read that describes the distortions engendered by the entry of private equity into the market for veterinary care, traditionally offered as small, locally owned enterprises. It narrates the example of UK veterinary care company IVC Evidensia which was bought over by Swedish PE firm EQT in 2016, and has since been on an acquisition spree snapping up local independent and small chains of vet care practices across Europe. It now is Europe's largest vet care provider with 1500 sites.

The IVC deal stands to be lucrative for its owners. It had been preparing to float in London this year until a last-minute change of course. Instead, in a February deal, Nestlé and California-based private equity group Silver Lake agreed to lead a €3.5bn investment at a valuation of €12.3bn. At the same time, the original EQT fund that bought the company sold most of its stake and a newer EQT fund bought in. The valuation is four times higher than just two years ago, and more than 32 times IVC’s earnings in the year to March — far higher than the levels at which buyout groups typically buy businesses, with the 2020 average in Europe being 12.6 times earnings according to a Bain & Co report. It is a “stratospheric valuation compared with those typically seen in leveraged finance”, according to a report published in April by data provider 9Fin. Such a valuation stands to make a big contribution to the pool of so-called “carried interest” bonus payments available to EQT executives, under standard private equity pay structures. On buying a practice, IVC centralises procurement and finance, and appoints business support managers to oversee practices. IVC sets financial targets for practices that several vets described as challenging. It recommends drug prices centrally, but says local practice management ultimately decides what to charge. However, the effort to meet targets can lead to steep price increases. One vet in southern England says the price of Metacam, a widely dispensed anti-inflammatory painkiller, rose at their practice by 28 per cent to £82.79 for a 100ml bottle after IVC took over. The price of ProZinc, an insulin for diabetic cats, rose 39 per cent to £99.34, the vet says, and Fortekor, a heart medication, rose more than 78 per cent to £76.85.

This is a good summary of the standard PE operating model,

IVC’s race for growth is the latest large example of a model used widely by the private equity industry. Known as a “roll-up”, it involves buying large numbers of smaller or independent businesses using debt, and merging them into a large group that can cut costs with economies of scale. Rival buyout groups have snapped up dental surgeries and petrol stations in a similar way. The model rests, in part, on a piece of financial engineering known as “multiple arbitrage”, in which buyout groups calculate that the price they pay for a small business, per dollar of its earnings, is lower than the price they can later receive for it as part of a bigger group. In the process, the companies typically amass large debts. IVC’s junk-rated net debts and leases total £2bn, or 6.2 times the £322m it earned before interest, tax, depreciation and amortisation in the year to March, according to figures that the company shared with lenders.

4. I find it inexcusable that IIFCL and NIIF, whose collective contribution to crowding in private capital (which would otherwise have not come) to infrastructure financing is negligible, are so miserly with sharing information about their projects in their website as well as their annual accounts. The starting point for good governance is information disclosure. Its absence is there to see in both institutions. See this blog post on NIIF.

5. The case for a "single price to be paid to vaccine makers for all the doses that they supply". This is a strong case against the just announced vaccine policy of the government of India. 

I am not sure about the single price argument. But the case for canalising vaccines through public systems and even having it delivered only through government system is very strong. 

It is being reported on SII's price,

Even the Rs 400 procurement price — applicable to both State and new Central procurement orders — is higher than the price at which governments in countries such as the US, UK and in the European Union are sourcing directly from AstraZeneca. It is also higher than the price agreed by countries such as Bangladesh, Saudi Arabia and South Africa for supplies of the vaccine from SII. In most of these countries, the shots are being administered for free, with the governments absorbing the costs... In terms of per dose pricing, the 27-nation EU is paying $2.15-$3.50 for a shot of the vaccine across locations in Europe, a high-cost manufacturing destination. The EU had, incidentally, invested $399 million “at risk” in AstraZeneca way back in August 2020, in return for 400 million doses of its vaccine. The UK, which had a smaller investment commitment to AZ, is paying about $3 per dose and the US has been offered the vaccine at $4 per dose, according to data compiled by British Medical Journal. Both the US and the UK are paying the amount directly to AstraZeneca. Meanwhile, Brazil is reported to be paying $3.15 per dose for the AZ vaccine through state-owned Oswaldo Cruz Foundation (Fiocruz), another licenced producer.
Bangladesh, according to Reuters, is paying an average of $4 per dose supplied by SII, with the BBC having quoted a health ministry official in Dhaka citing that it entails a total cost at $5 per dose, incorporating the margin charged by Beximco, the vaccine’s Bangladesh distributor. Both South Africa and Saudi Arabia had paid over $5.25 per dose from SII, according to UNICEF’s Covid Vaccine Market Dashboard, which collects publicly reported price information. This is higher than the price at which Indians will get vaccinated at state government hospitals, without subsidy.

This is unlikely to have a nice ending for SII.   

6. FT has a good article on the impact of vaccines. From UK

From France.

From Chile.

It appears to be working.

It is also reported from a new UK study from more than 350,000 people between December 2020 and April 2021 that a single dose of either Oxford/AstraZeneca or Pfizer/BioNTech vaccines cut the rate of infections by around 65%, a second dose reduced by 70-77%, protected older and more vulnerable people almost as well as younger and healthier individuals, it took around 21 days after the first jab for the immune system to mount a decent response, and that vaccinated people could still re-infect and spread asymptomatic spread of the virus. 

7. The conventional wisdom is that generous unemployment benefits discourage work and distorts the labour market. The Economist points to the contrast of Denmark which has the most generous unemployment benefits program and the best functioning labour market.

Danish benefits are worth more than 80% of previous earnings after six months out of work, compared with 60% across the rich world and less than 50% in Britain (America is even stingier). For Danish parents who lose their jobs, replacement rates can approach 100%. The generosity of Denmark’s unemployment system is the flipside of its liberal regulation of employment contracts—a combination called “flexicurity”. Danish employers can hire and fire workers pretty much as they please. Jobs therefore come and go, but people’s incomes are stable. Yet the state’s munificence has not produced a class of feckless drifters. Denmark’s unemployment rate is lower than the rich-world average and its working-age employment rate is higher. Long-term unemployment is low. When Danish people lose a job, they find a new one faster than almost anyone else in the world, according to the OECD.

This generosity comes with tough love

Denmark makes it hard for people to live off welfare. Recipients must submit a CV to a coach within two weeks of becoming unemployed. They can be struck off for not trying hard enough to search for work or to keep up with adult-education programmes. As a share of GDP Denmark spends four times as much as the average OECD country, and more than any single one, on “active labour-market policies” to make people more employable.

8. The work of the newest John Bates Clark medal winner, Isaiah Andrews, points to a much needed methodological correction to social sciences research. 

In work with Maximilian Kasy, now of Oxford University, Mr Andrews cast light on the fact that economists are too eager to pursue striking results, and journal editors too prone to approve them. Studies that find that minimum wages have no significant effect on employment, for example, are only a third as likely to be published as those finding a stronger negative effect. The researchers developed a way to adjust reported results for this publication bias—and found that the average effect of a higher minimum wage on employment fell by half.

Mr Andrews, with others, also explored what is called the winner’s curse when it comes to choosing between policies: the policy that performs best in a trial may owe its top rank to chance, and will later be doomed to disappoint. To illustrate this the researchers turn to a trial that assesses the most effective ways of encouraging people to donate to charity, by combining requests for specific donations with promises to match the initial contribution. The researchers find that if the charity chooses the method that does best in a trial, it will always overestimate its donations. They suggest ways to make a more realistic estimate that takes account of the role of chance.

Andrews' work reveals the deficiencies in the methods of economists and provides tools to correct them.

9. Finally, very informative interviews on the impact of the farm laws on agriculture processing.

There is a huge asymmetry in information, when you invisibilise transactions and players, businesses will have more information on the farmers through credit scores, data stats, but the farmers won’t have that much information on these businesses and that puts them at a disadvantage

Wednesday, April 21, 2021

Aadhaar in eligibility identification - a reality check

There is no doubt that Aadhaar has been a game-changer and a market catalyst in many areas of the Indian economy. The India stack has been truly transformational in areas like fintech. But what about its role in welfare services delivery?

When conceived, one of the primary motivators of Aadhaar was the need for a unique identifier to enable more effective targeting welfare benefits. Then, state and central governments in India were grappling with the problem of egregious inclusion errors - duplicate and non-existent beneficiaries for various schemes, especially the Public Distribution System (PDS). Aadhaar was thought to be the solution to this problem.

And it cannot be denied that Aadhaar has nearly eliminated the issue of duplicate or bogus (the person itself not alive or existent) through enrolment and physical screening. However, sceptics argued that in the process of Aadhaar screening, significant numbers of eligible poor would be screened out, as a collateral damage. But it was thought that over time - as technology improved, databases got cleaned up, stakeholders got used to it, and redressal mechanisms were streamlined - this problem of exclusion errors would disappear. Unfortunately, that has not proved to be reality.

The real problem lies in the use of Aadhaar in the targeting of beneficiaries, especially through screening them out by linking multiple databases. It was all along known, though not explicitly acknowledged and clarified by all concerned (including the government), that beyond elimination of duplicates and non-existent beneficiaries, in the Indian context Aadhaar's role in targeting was limited.

The supporters of Aadhaar argued that, apart from the duplicates and the non-existent, it could provide the accurate unique identifier that can help link multiple databases on employment, vehicle registration, tax payments, electricity consumption, bank loans and so on. This, it was claimed, could be a game changer in targeting social benefits. 

The problem with this argument/claim lies in the quality of the validating databases. Apart from 10-20% of Indians who are either employed, or own vehicles, or have taken loans, or purchased property, the vast majority have no formal signatures of their existence. In other words, at least 80% of Indians have no formal database on which they exist reliably so as to be used for Aadhaar-based validation. 

In this scenario, the government ends up using various questionable datasets for Aadhaar screening. Governments also force people into various formal database inclusion requirements so as to generate digital trails for subsequent Aadhaar validation. 

Most importantly, since their primary objective is to eliminate inclusion errors, none of the various software applications of state and central governments to enrol and process welfare schemes do not have  the requisite safeguards to prevent egregious exclusion errors. This is a manifestation of the inclusion errors elimination bias within governments. I blogged about it here

Finally, and surprisingly, technology itself has not held up well and has been a major contributor to the problems with Aadhaar implementation. The apparently simple activity of biometric validation - biometric capture, validation with Aadhaar database, and the data communication - has its struggles, especially but not only in the rural areas of the country.

The consequences of all these efforts to use Aadhaar in welfare services delivery end up doing great damage, often more than offsetting their beneficial effects. 

A recent article in the Indian Express points to the problems with Aadhaar-linked validation on accessing PDS and other welfare schemes benefits. It reproduces the salient findings of a Lokniti-CSDS Survey conducted during the 2019 LS elections.

And the denials rates were very bad in the Bimaru states,
The survey findings speak for themselves and presents strong evidence to support the concerns about the unqualified use of Aadhaar in welfare services delivery, especially in targeting. I've blogged earlier on Aadhaar's targeting challenge here

The point here is not to advocate abandoning the use of Aadhaar in welfare services delivery. Instead it is required to acknowledge the limitations of Aadhaar and the potential damage it can cause, use it where it serves the purpose, and also put in place mechanisms to address them while using Aadhaar. 

I have been surprised about the absence of good research which points to the collateral costs of Aadhaar-linked targeting. After all, it only required some surveys. It's also perhaps a testament to the misallocation of social sciences research priorities in and about India that issues of first-order importance not only get little or no attention but also get marginalised by bandwagon studies conducted by foreign researchers (which unfortunately only ends up fuelling the dominant but misleading narratives). 

It's interesting that these findings have come as a collateral benefit from an election survey conducted by an opinion polling agency, and not from mainstream research. 

See also this oped by Jean Dreze on the absence of a mechanism for poor for grievance redressal if their benefits are terminated due to Aadhaar validation problems. 

Monday, April 19, 2021

The problems with explaining development and planning implementation

Human beings are caught up with the urge to have logically consistent explanation for every thing and an operationally detailed plan for the implementation of any idea. Unfortunately, in reality, with many things and ideas, especially all-encompassing phenomena and public policy ideas, there are no such explanations or plans. Worse still, even their (explanations and plans) pursuit is a signal of having lost the way. 

Francis Fukuyama (I am only using his speech as an opportunity to highlight them, for these insights have been made by others too) illustrates both these points in a brilliant and educative lecture here.

Consider some of the biggest problems in development. Why are some nations poor and some rich? What does it take to establish stable democratic institutions? How does civil society emerge? How do countries develop capable states? How can markets be created? 

In the context of development, Fukuyama talks about the challenge of "getting to Denmark" - achievement of rule of law, strong state, accountable government, and efficient markets. It's really difficult. As he says, even the Danes don't know how they got there! Every country will have to chart its own unique path to Denmark.

But this has not deterred social scientists from trying to formulate theories that can explain these questions with logically consistent explanations. Leave aside popular story tellers like Jared Diamond or Yuval Noah Harari, even serious academic researchers have struggled to avoid this trap. 

As Fukuyama says, institutions develop organically in ways that are dependent on the particular histories and cultures of each country. Successful examples of countries nurturing modern institutions have invariably involved countries taking into account their own traditions while trying out these social, political and economic ideas. This has involved adapting ideas to their contexts and experimenting and iterating over long periods to develop acceptable and stable models. No two trajectories have been the same. 

This post by Dietz Vollrath nicely captures the point being made - differences in development are about incredibly persistent antecedent differences in populations, their cultures and institutions, and not countries. 

In a recent book, Radical Uncertainty, John Kay and Mervyn King talk about the difference between puzzles and mysteries. You can try to solve puzzles, whereas mysteries are inherently unsolvable. The questions raised above are not puzzles. They are mysteries and will remain so. 

Explanations are relevant only to the extent of being able to draw from history and inform about the challenges, pitfalls, and opportunities. It provides a perspective and an analytical framework to make better sense of the underlying processes in the unavoidable, uncertain, hard, and long journey. It cannot tell us what needs to be done in the form of actionable steps.

The book Narrow Corridor is a good example of an "explanation" of this kind so as to better comprehend such phenomena (related to development and economic growth). 

On the second point about plans to implement public policy ideas, Fukuyama again refers to how China implements policies. 

He describes what we all know about China, the Communist Party has complete control over rule of law. It can take over property and do whatever it wants. It can dispossess people and hand it over to developers who can construct malls and condos. It can also build roads and high speed rails. It's just so that most of those dispossessed unjustly are the poor and voiceless, and rarely the elites. But it does serve the purpose of facilitating economic growth. 

This is a teachable example of the limitations of "theoretical explanations" approach. The classic institutional theory and its explanation of economic growth claims that if you don't have western-type inalienable property rights - the rule of law and legal economic system that protect property rights - then you cannot realise modern economic growth. But we know that while this applied to the west, it cannot explain the growth in East Asia and China, which had none of these institutional requirements

In the context of comparing China and the US, Francis Fukuyama makes a very important remark about how both look at policy making. He says, 

The real challenge to western economic theory is this... If you don't have western property rights but have mechanisms to skew property rights in favour of those who can use it more productively, then you may actually get superior growth compared to a neutral system... That's the secret to Chinese growth than any (theoretical explanations)... What's most impressive is that they will roll out an experimental program in one province. If it doesn't work they pull the plug on it, and if it works they scale it up and move it elsewhere.
And in that sense, they are more pragmatic than Americans. I think one of our problems is that we are full of all these ideological theories of how the economy works like if you lower taxes then you'll get more growth and there are certain parts of our society that takes it as gospel and no amount of empirical information will make them not believe this anymore. Whereas the Chinese, if they try something and if it works they do it and if does not, they stop doing it.

There are no grand nation-wide scaling plans, in the sense we imagine of plans (in case of projects or the private sector). This is "crossing the river by feeling the stones". It is also similar to the framework that Horst Rittel and Melvin Webber formulated to make the distinction between wicked and tame problems. Or the idea of Minimum Viable Product and iterative execution that is often used in management thinking. 

A few observations in this context:

1. History is important. It informs with knowledge and illuminates with perspectives. It guides on the direction, and exposes to challenges and possibilities.

2. Theory is useful, but only unto a point. Theorising beyond that point is not only not useful but can be damaging. 

3. All that matters is whether the idea gets implemented or not. If this is the objective, then the only focus should be to support with implementation.

Sunday, April 18, 2021

Weekend reading links

1. In the context of the new Production Linked Incentive (PLI) scheme of the government, the Business Standard examines the air conditioners market in India. 

This is a good summary of India's manufacturing challenge, the much smaller size of local market than expected,

Currently, in the Rs 18,000-crore local AC market, 70 per cent of the cost material used in assembly are imported. Key parts like compressors, variable speed motors in indoor units, and high quality copper pipes, among others, are imported. But with no incentive for incremental assembly, manufacturers are now hoping that large global component makers set up shop here. Since key components that are being imported require huge investments to manufacture locally, setting up such facilities will not be a viable business proposition for entities in India, clarified companies. To turn such investments profitable, the kind of scale that is required does not exist in the local market. At 6 million units a year, India’s AC market is much smaller, compared to leading global markets like China (50 million units), the US (17 million), and Japan (12 million).

2. On the importance of manufacturing to developed economies, Rana Faroohar writes,

In the US, for example, although manufacturing represents just 11 per cent of gross domestic product and 8 per cent of direct employment, it drives 20 per cent of the country’s capital investment, 30 per cent of productivity growth, 60 per cent of exports and 70 per cent of business R&D, according to figures from the McKinsey Global Institute... A fascinating study by MGI, to be released on April 15, examines 30 main manufacturing sectors in the US. It finds that 16 of them stand out for their economic and strategic value, as measured by their contribution to national productivity and economic growth, job and income creation, innovation and national resilience. Apparel is not on the list. But semiconductors, medical devices, communications equipment, electronics, autos and auto parts, and precision tools are.

And this very interesting snippet about China

Chinese producers exported 71 per cent of finished apparel goods in 2005. By 2018, it was just 29 per cent.

3. I've never understood the case for lowering corporate tax rates in India. This is a good set of graphics.

4. Larry Summers makes a very persuasive critique of the post-Covid fiscal policy in the US,

If you look at the economy at the beginning of this year, prevailing forecasts were that Covid would reduce wages and salaries to American households by $20bn-$30bn a month, with that figure declining over the year. So, that would be a $250bn-$300bn hole in wages and salaries over the course of the year. So, I look at this hole and then I see $900bn of stimulus in the December package, $1.9tn of stimulus in the recently passed package and $2tn in the savings overhang, which is also likely to be spent. I see the Fed with its foot on the accelerator as hard as any Fed has ever done... That could manifest itself, as a much smaller period of excess did during the Vietnam war, in rising inflation and a ratcheting-up of inflation expectations. It could, as has often happened, manifest itself in the Federal Reserve feeling a need for a sharp and surprising increase in interest rates, and the subsequent deceleration of the economy into recession. It could manifest itself in a period of euphoric boom and optimism that leads to unsustainable bubbles, or it could all work out well... 

There’s not much argument that the 2009 stimulus, in retrospect, was too small. It was 4 to 5 per cent of GDP over a couple of years, so it was 2.5 per cent of GDP in the first year, against a gap that was 6 or 7 per cent of GDP and growing, so it was perhaps a third or half of that gap. Today’s stimulus is above 10 per cent of GDP in the face of a gap that is 3 or 4 per cent of GDP. Relative to the gap, this stimulus is already of the order of five or six times as large as in 2009... I could have been comfortable with a headline figure well in excess of $1.9tn if it had been a large-scale, multiyear programme of public investment responding to our deepest societal concerns. But that’s not what this is. It transfers to state and local governments that don’t have any new budget problem, according to the latest figures. It’s paying people, who have been unemployed, more in unemployment insurance than they earned when they were working. It’s giving cheques to families in the 90th percentile of income distribution. It doesn’t seem prudent on resource allocation grounds, as well as being problematic on macroeconomic grounds.

5. Mahesh Vyas points to the informal market distress,

As people lost jobs and jobs became scarce in 2020-21, labour that lost jobs moved from one kind of occupation to another. Large numbers eventually migrated to agriculture, apparently, when all other possible occupations failed. As a result, employment in agriculture in March 2021 was nearly 9 million higher than it was in 2019-20. This implies an eight per cent increase in labour in agriculture. Agricultural output is estimated to have increased by 2-3 per cent in almost each of the four quarters of 2020-21. The 8 per cent increase in labour implies a sharp fall in labour productivity. We believe that this huge influx of labour into agriculture is largely disguised unemployment. It hides the greater employment challenge in March 2021 than the 5.4 million net jobs lost. The biggest loss of employment in 2020-21 was among the salaried employees. As of March 2021, there were 76.2 million salaried employees. This was 9.8 million less than the 85.9 million salaried jobs observed in 2019-20.

Salaried jobs are mostly in urban India. Urban India accounted for 58 per cent of all salaried jobs in 2019-20. But it accounted for only 38 per cent of the 9.8 million salaried jobs lost. Over 6 million salaried jobs were lost in rural India. Most of these are likely to have migrated to farming. Rural India also saw nearly 3 million business persons being rendered unemployed. These could also have migrated to farming. Farming saw an increase of 9 million jobs in rural India. So, the churn in rural India seems to have been people losing salaried jobs and losing their business and these unemployed people moving into agriculture for unproductive employment. The increase in agricultural jobs in March 2021 was essentially a migration of people who lost non-farm jobs in rural India into farming. This was not an urban to rural migration.

6. More disturbing news, which points to outright fraud, about Sanjeev Gupta's business activities,

Last week the FT reported that several loans to Liberty Commodities, part of GFG, were based on suspect invoices and that Credit Suisse executives were becoming increasingly concerned that their clients were victims of fraud. Several European metals businesses told the FT last week that they had not carried out any business with Gupta’s groups, despite invoices linked to them being repackaged as notes by Greensill and sold to Credit Suisse investors.

7. The Government of India has approved the Russian vaccine Sputnik V for use in India. It has also accepted a recommendation by the National Expert Group that vaccines approved by health regulators in the EU, US, Japan, and the UK and by WHO should be granted emergency-use approval in India. Till now India had insisted that these vaccines had to still conduct additional 'bridging' trials in India before use.

This decision, which ought to have been taken much earlier and is now precipitated by the acute shortage of vaccine stocks, is a precedent for several other areas. Regulators in other sectors could explore the possibility of using the regulatory approvals in other countries to allow use in India. 

8. Interesting contrast between the employee attrition rates of TCS and Infosys. The rates for the last quarter of 2020-21 was 7.2% and 15.2% respectively for the two companies, both being all-time lows and highs. Is there something about their respective business models which explains this big differential? Or does it tell us something about work cultures in the two companies?

9. Noushad Forbes points to an interesting fact,

In January 2021, India approved its first vaccines for use. The Ken Nutgraf tells us that between July 2020 (before any vaccines had passed testing) and January 2021, the US ordered (and paid for) over 600 million doses. That’s for a total population of 300 million. In the same period, India ordered 11 million doses, for a country of 1,300 million.

Indian Express has an article which points to how India did not put forward at-risk capital (or advance market commitment to purchase vaccines) to promote vaccine manufacturing. 

From all available accounts, India did not invest “at-risk” in SII and its first commercial agreement on vaccine offtake only came in mid-January 2021. And the pricing of Covishield is a factor in SII’s struggles to keep up with demand as the private, unlisted firm has committed to deliveries under AZ’s deals and through multilateral arrangements such as COVAX. SII has now sought “roughly” Rs 3,000 crore from the government to expand its “very stressed” capacity, SII CEO Adar Poonawalla told NDTV. “The globe needs this vaccine and we are prioritizing the needs of India…we’re still short of being able to supply to every Indian that needs it,” he said. “At the moment, the price (Rs 150 per dose) that is set is profitable. However, it is not profitable enough to re-invest substantially in building capacity, innovating new vaccines — including the new variants that we may need to develop and make and go into clinical trials and other things,” he added.

10. Ed Luce on tax avoidance by US companies,

Last year, 55 of America’s largest companies, including Nike and FedEx, paid nothing in corporate taxes in spite of collectively making about $40bn in profits. The headline US corporate income tax rate is 21 per cent, which Biden wants to lift to 28 per cent. However, the official rate is not the point. The effective US corporate tax rate is just 11.2 per cent, which is below that of Ireland. The US Chamber of Commerce and the Business Round Table complain that the nation’s corporate taxes are higher than the western average. In practice, they end up close to the lowest. US tax collections amount to 1 per cent of gross domestic product, compared with a 3.1 per cent OECD average. All such avoidance is entirely legal.

11. As Covid relapses and schools start to shut down, a good report on where India stands with respect to schooling.

Learning and future prospects of a cohort of children may be the biggest long-term casualty from Covid 19 lockdowns.

12. Interesting change in the IBC, for MSMEs, which now allows promoters to remain in control during the restructuring negotiations with creditors.It may be an appropriate response for the Covid 19 induced business stress, but given India's business environment, it remains to be seen how this will work out. 

13. Progress in a graphic

 

Thursday, April 15, 2021

The rise in share buybacks in India

Business Standard reports that Infosys has announced a share buyback of Rs 9200 Cr.

The company board approved share buyback programme worth up to Rs 9,200 crore priced at Rs 1,750 per share, a premium of 25.12 per cent over the stock's closing price on the BSE, of Rs 1,398.60 per share, as on Tuesday.

It also has another article pointing to the rise in share buybacks in corporate India


The article argues that the recent changes in taxation has prompted this rise. From April 1, 2020, dividends in the hands of shareholders were brought under the tax net and that too at the marginal tax rate, and dividend distribution tax on the company was removed. Further, capital gains on buyback are taxed at 20%. The arbitrage opportunity is clear and irresistible.  

“Dividends impact promoters and significant shareholders with an additional tax burden of 20-30 per cent over a buy back. No reason why shareholders would want to bear that burden, especially companies with large promoter or investor shareholding,” said Praveen Raju, Partner, Spice Route Legal.

I am not sure why such regulatory arbitrage should be allowed. 

Shareholders are by definition putting forth risk capital. Technically, for them, it should not make any difference as to how they realise returns - either as dividends or buybacks. Both are realised returns on their investment. I don't know a reason why one form should get a preference over the other in taxation in the hands of shareholders.

Of course, for a company, there are important implications on which route they take to share gains with shareholders. And the corporate tax treatment should take that into account. 

The rise of buybacks also raises an intriguing point. On the one hand, corporate India has done well to emerge out of Covid in good health (though the outbreak of the second round raises questions). It's well placed to invest and grow. But, on the other hand, this rise in buybacks among the biggest companies points to less confidence on economic prospects. Buybacks are after all displacement of capital from investments. 

Much has been written on the corrosive effects of share buybacks in the US, including in The Rise of Finance

Wednesday, April 14, 2021

Domestication of fire, animals, and grains and state formation

James C Scott is arguably the most perceptive of social science scholars, belonging to the highest standards of cross-disciplinary enquiry. He's been a trenchant critique of the modern state for over half a century. 

Seeing Like a State is only one among a pile of classic works. I have recently been listening to his lectures on how the domestication of fire, grain, and animals created the conditions for emergence of the modern state (this and this). They form the basis of his last work chronicling the role of grain in creating the state. It contains fascinating historical and anthropological portraits and stories.

The sequence goes something like this. Homo sapiens can be traced back to 200,000 years, and 50,000 years outside of Africa. The first signs of sedentary communities can be traced back to 11000 years ago, roughly the same time as evidence of domestication of animals and plants. But there is no  evidence of villages with domesticated animals and plans and settled agriculture till about 7000 years back. The first form of a state appeared in the Mesopotamian lowlands (Uruk etc) about 6000 years ago, which was the last 3% of human history on the planet. The first forms of modern impersonal bureaucratic states can be traced to the Qin China of 3rd century BC. In other words, states are a very recent part of human history. 

This sequence raises questions on the conventional narrative of civilisation, which goes something like this. The domestication of plants allowed us to settle in one place, make small villages and then towns, and ultimately create civilisations. If civilisation is an achievement of the state, and if early civilisation was about sedantism, farming, irrigation, and towns, then there is something odd with this historical sequence. All these existed before anything like a state emerged on the horizon. 

A reconstructed story on civilisational development would go something like this. Human beings domesticated fire, animals, and grains. Fire ended up vastly reducing the radius of a meal. By allowing for cooking and storage, it enabled people to increase the types of foods they could consume. All this, over a long time, engendered spatial, societal, cultural, and organisational norms and practices that were appropriate for sedentary lives. This, in turn, led to the emergence of large and dense human settlements like villages and towns, which, in turn, created the conditions for state formation. 

The impact of domestication of grain and animals was transformational on the lives of human beings. It meant organising our lives around the requirements of often a single plant. Routines around preparing the field and sowing, managing the crop, harvesting and post-harvest activities, and preparing them for cooking, and finally cooking transformed human lives. It impacted even the shapes of our physical bodies, created their patterns of cooperation and coordination, and organised our work life, settlement patterns, social structures, our built environment of domus, and even our rituals. 

Moving from hunting and foraging to settled agriculture, and that too involving just one or two crops, meant a substantial narrowing of focus and simplification of tasks. In this context, Scott points out that the differences in complexity at all levels associated with the life of foraging and even slash and burn cultivation compared to sedentary agriculture based communities are of an order similar to the difference between early and Taylorised manufacturing. He argues that the transition to sedentary agriculture as similar to that of a skilled worker being reduced to the role of an assembly line worker. Alexis Tocqueville made the perceptive observation in the context of Adam Smith's pin factory example, "What can one expect from a man who has spent the last 20 years putting heads on pins?"

Fixed field agriculture is very labour intensive. Plough agriculture was avoided till population pressure and scarcity of land forced people into it. People tend to avoid it if they could. Settled agriculture was an extremely drudgery filled task, especially so for the homo sapiens who spent no more than 10-15% of their time on gathering food.

This also meant that there is nothing natural in terms of the progress from hunting and foraging to slash and burn cultivation and then to settled agriculture. He points to the physical drudgery and risks associated with domestication and settled agriculture. Unlike hunting, foraging and slash and burn cultivation, fixed field agriculture involved ploughing and other field preparatory works which entailed intense physical activity. Also, unlike foraging with its diversity of food sources, domestication of a few woodgrains concentrated risks and exposed people to risks of plant diseases and vagaries of the weather. For these reasons, people had a strong incentive to avoid domestication and sedentary agriculture. 

Accordingly, Scott points to the long time of nearly four millennia it took for people to adopt agriculture and assume settled lives. He also points to how episodes of large population destructions were generally followed by reversion away from settled agriculture towards either slash and burn cultivation or increased focus on animal husbandry. Even the Great Plague of Europe in the fourteenth century was accompanied by such reversals from settled agriculture. 

Further, unlike foraging and hunting, settled cultivation was also amenable for greater participation by women. Most of the tasks associated with settled cultivation, including cooking, were tasks that are historically associated with women. Therefore societies that took to settled cultivation were associated with greater involvement of women in production and also greater drudgery of their lives. 

Given the extent of their influence on human civilisation, Scott also draws attention to the epistemological point about who is getting domesticated. He quotes Michael Pollan and Evans Pritchard to say that while it is argued that human beings domesticated plants and animals respectively, it can just as well be argued that human beings ended up being domesticated or enslaved by the forces unleashed by domestication of animals and grains. Their personal behaviours and practices, livelihoods, daily routines, familial relationships, social organisations, and so on were all deeply impacted and transformed. Indeed the entire human life, including even their biological processes, was transformed by the process of domestication of fire, grains, and animals. 

At another level, there is also the role of slavery in motivating the earliest wars so as to get state residents who can be captive cultivators and thereby tax payers. 

Scott characterises domesticated animals, like domesticated fire, as expanding our scope of food. They go far and wide, eat up different kinds of food, metabolise them, and provide us with valuable proteins and fats. They therefore become the dedicated foragers of human beings. 

Scott describes a "domus complex", whereby stable groups of humans gathered themselves and an assortment of nonhuman animals, to highlight the emergence of densely populated communities.

Density, in turn, brings with it a proliferation of diseases, as populations provide the necessary concentration of hosts for viruses and bacteria to survive. Combined with domestication of animals, it also provided the perfect opportunity for transmission of zoonotic diseases contracted from animals. The examples of how whole populations got exterminated in the New World from diseases brought by invaders from the Old World highlights the point. 

On state formation, sedentary agriculture was central to the project. Consider a few snippets. One, all early states were slave states. Most early wars were fought to conquer populations who could then be brought home and made to undertake agriculture which could then be taxed to support the state. Besides, these slaves also could fight more wars. In fact, most early trade was trade in slaves. Two, the big walls, like Great Wall of China, were built just as much to keep farmers inside the state as it was to prevent barbarians from invading.  

Scott points to the domestication of three grains - rice, wheat, and maize - as critical  in the process. Incidentally, these three grains make up nearly half the calories consumed by human beings even today. These three grains had pre-defined production cycles, were standardised, had to be harvested and the production could be easily assessed through routinised processes (unlike tubers which could be left inside the ground or millets whose production cycles were less certain). It became possible to assess production and levy taxes, which was a critical requirement to support the state. In fact, only these cereal grains can form the basis for taxation. 

Root crops are a form of state resisting crops, since they are easy to grow, and have no specific harvest requirement and can be left underneath for a long time. Scott describes growing them as agriculture of evasion.

In fact, even among grains, there is nothing like wet rice cultivation in concentrating state making. Rice grows above the soil, is grown the same times of the year, ripens at the same time, can be easily confiscated after threshing, stores quite well, high value per unit weight, and can be transported easily over long distance.  

Scott points out that sedentary agriculture by itself did not lead to the formation of states. But he's not clear (using his narrative on domestication of grain) about what forced state formation in the Mesopotamian lowlands. It is likely, as others have pointed out, that when faced with climate related pressures, those residing in those areas doing settled agriculture consolidated to form states. 

States in turn fought with each other, which necessitated more taxes, and organised armies to fight those wars. This, in turn, called for bureaucracies to collect taxes and manage the logistics of recruiting and managing armies. The modern state gradually evolved. 

In sum, as Scott says, states need alluvium or loose soils (to grow grains), navigable waters, and places where you can cram large numbers of people into small areas (grain-manpower module). Sedentary agriculture is an essential requirement for state formation. 

The geography of state depended on flat land and navigability over water to concentrate populations and undertake trade and conquests respectively. It also highlights Scott reinforces the point about how historically water was far easier for social contact and integration through trade and conquest than land. For example, as late as 1800, it took less time to travel from Southampton to Cape of Good Hope by sea than from London to Edinburgh by stagecoach. 

To highlight this, he puts forth the fascinating idea of "friction of distance maps" or maps of the world that measures sea-based and land-based distances separately in terms of their respective speeds of travel (with ships and carriage). In these maps, the the unit could be a distance travelled in a day. It would vastly reduce the distances where there was easy waters and vast increase distances where the topography was rugged mountains and rivers. This would give a much better grasp of social contact and integration. 

On the resistance to state building projects, Scott points to the more or less contiguous hilly periphery of South East Asia, encompassing the hill regions of eight countries, which are largely populated by non-state communities. In contrast the valley and alluvial low lands are populated by the mainstream populations of these countries. He describes these hill populations as having emerged in response to flight from state-building projects with its taxation, conscription, diseases, and dissent including religious persecution (eg. dissident Buddhist sects in hills of Myanmar). 

In fact, using the example of S E Asia, Scott argues that far from being primitive tribes totally unexposed to modern civilisation, the hill people are those who have fled the Han, Burmese, India, and Thai civilisations over a long period of nearly 2000 years. He argues that this is deeply reflected in several signatures across their social and personal lives and livelihood activities. 

While such non-state communities occupy the hills in South East Asia, they are also known to occupy the valleys in South America or swamps in Middle East and other places in Africa. But in general, state formation projects have struggled except in plain and alluvial lands because of the problems with conquering and concentrating populations. Scott's point is not that civilisations can't climb hills, but that for the last 2000 years, peoples have been climbing hills to escape the control of the state. These areas are regions of refuge. As Ernest Gellner has written, marginal tribalism is the type of tribal societies exists on the margins of non-tribal society as refuge from state building projects. 

He gives the example of Cossacks, who are considered the most soldieristic ethinic group in Russia and were used by Tzars in their armies. They were originally runaway serfs from Russian serfdom to the peripheral non-state areas. And their identities were formed at the margins as they fled the oppression. 

In summary, Scott characterises state making as being associated with several undesirable features for its populations - forced agriculture, slavery, diseases, wars, weakening of kinship ties etc. In fact, he calls state a "late Neolithic multi species resettlement camp"! Therefore, populations most often preferred to remain outside state limits. Given the central role of settled agriculture in state formation, populations made strategic choices to remain as nomads and foragers. In fact, the term barbarians is used to refer those residing outside state limits. They were nomadic, pursued foraging, and were governed by kinship norms, and did not live under states.

Reading Scott, one gets the impression of a historical narrative which is so tightly fitted into a prior hypothesis. It appears like a mega history. What makes him by orders of magnitude more formidable than the likes of other faux mega historians (Yuval Harari, Jared Diamond, Peter Turchin etc) is the rigour of his research and the orthodoxy of his conceptual frameworks. In many respects, I feel only another equally formidable and granular scholarship can refute Scott. 

At a higher level, Samuel Moyn  has a very good critique of Scott's works, 

Scott rarely mentions the forms of social justice that only modernity and its states have permitted and put into practice, however faulty and outweighed by state crime and excess they are. Instead, he has sought to project an immemorial dialectic between the state and its enemies onto the whole of human history. This has made him one of the greatest teachers of how costly modernity has been; yet it has also caused him to obscure the fact that modern states could strive not simply for civilizational splendor, but also for the freedom and equality of all.
This is a broader alternative point, which however cannot refute the descriptive interpretative parts of Scott's work.