#251 Connecting the Dots

Election Crystal-gazing, China's Chip Troubles, Kahneman on Biases, and a Framework for Bengaluru's Water Woes

Programming Note: We will be on summer break for three weeks. The next edition will hit your inbox on April 28th. If you are yearning for public policy gyaan in the interim, consider applying for Takshashila’s GCPP and PGP programmes. The intake for the next cohorts of both these courses is open.

---

—RSJ

Before I jump into this edition, I should point out that my last post on the Apple-DoJ lawsuit received a fair bit of criticism. Some of it is available in the comments section of that post, and others came via personal messages. The upshot is either I got it mostly all wrong or that they disagree with me, especially on Apple’s closed-garden approach to things. Thank you for the feedback. I need to go deeper into this subject, and I must apprise myself better before commenting on it in the future. I will do that, and I will write about it soon. 

I have taken the liberty to paste Mihir Mahajan’s comment here in case readers who only follow the newsletter by mail missed it. It is quite illuminating.

RSJ makes several points regarding Apple and Amazon antitrust actions; I respond to those below.

Timing of case filing

Whether political signaling, happenstance, or otherwise, this is in the realm of speculation. An important reason to bring the case during a particular President’s term is to put forth the economic logic for the case that the administration agrees with. An antitrust case brought by a Democratic administration may have different contours (and chances of victory) than that by a Republican administration, even when both have a problem with the alleged monopolist. Investigations and putting together a case can take years. The DoJ’s investigation during the Biden years can only come to fruition if the case is brought during the term.

Apple’s monopoly

The gist of RSJ’s argument is that (a) customers have choice and yet, buy iPhones/iPads; (b) Apple’s “strange alchemy of design chops, innovation and brand image” makes it worth it; (c) customers can leave any time; (d) Apple is better at protecting customer data through tight control over ecosystem; (e) “go full tilt at breaking it up” is an extreme remedy.

Unfortunately, RSJ does not appear to know or appreciate the logic of American antitrust. Any antitrust case is made of 3 parts – liability (possession of monopoly in a properly defined product and geographic market); causation (that the monopoly was illegally acquired or maintained, shown with evidence of specific actions; and relief (whether fines or structural remedies).

Read the original DOJ complaint inline. The Table of Contents shows that the complaint is written in conformance with the antitrust textbook. The DOJ lays out contours of the case, includes anecdotes, and provides background on Apple and platforms (Sections I-III). It carefully defines the market as “performance smartphones in the US” (main position) or “smartphones in the US” as alternative positions (section VII). It lists specific actions by Apple by which Apple “unlawfully maintains its monopoly power” in the defined market (section IV). It shows anticompetitive effects (section V). It addresses alternative explanations (section VI), establishes jurisdiction and venue (section VIII) and lastly, alleges specific violations of law and requests appropriate relief (sections IX and X).

First, RSJ gets it wrong by saying DOJ wants to “break it up”– the “request for relief” on page 76 merely seeks that the Court adjudge Apple to have “acted unlawfully to monopolize/ attempted to monopolize the smartphone/performance smartphone market in the United States” and “enter relief as needed to cure any anticompetitive harm” and seeks to prevent Apple from continuing the challenged practices (control over app distribution, private APIs, and contractual terms”. Nowhere does the DOJ ask the court to break up Apple. In fact, while some in the DOJ prefer structural remedies like breakups, this article states that the DOJ has been careful to not address the exact remedies they will seek.

Second, RSJ is judging the case by saying “there’s no monopoly Apple has for any customer to choose an iPhone or iPad to begin with” before evidence from both sides is heard. The whole point of the DOJ's case is to have a court examine the market definition, and the pro- and anti-competitive arguments. It remains to be seen whether DOJ’s argument that Apple unlawfully maintains its monopoly power over smartphones/performance smartphones by (a) contractual restrictions, fees, and taxes on app creation and distribution; (b) uses APIs to control behavior and innovation of third-parties; and (c) Apple’s smartphone moat is wide and deep and extends to other products.

The DOJ has not alleged that Apple acquired the alleged monopoly illegally. It is a maintenance of monopoly case.

RSJ’s points are essentially the case Apple will make:

(a) it will define a product market bigger than DOJ’s definition of “performance smartphones” and US as relevant geographic market;

(b) it will submit its innovation and privacy practices as the reason consumers pay a premium;

(c) it will likely argue technical and consumer benefit reasons for design choices it has made with respect to APIs, app store, and interoperability with other devices/ platforms;

(d) it will submit evidence of customer switching, and say no “lock-in”

Whether the DOJ has sufficient evidence, or whether Apple can persuade the court that its actions are legal will only be known through the litigation process.

PS: tablets are excluded from the DOJ’s market definition, so iPads are not even a subject of the case!

Lina Khan and Amazon

RSJ makes two points regarding FTC’s Amazon case – that a predatory pricing case presupposes future harms and that Lina Khan’s regret at not being able to “make the victims whole”. Additionally, RSJ objects to Lina Khan’s broad definition of the role of antitrust to “safeguard against concentrations of economic power”.

Khan has made these points throughout her career. Interested readers can refer to her very readable paper Amazon’s antitrust paradox.

I agree with RSJ that there are several problems with Khan’s arguments – at least in her paper, Lina Khan’s critique of Bork-era “consumer welfare” standard is not fully-formed and has lacunae. Preemptive regulatory action can cause harm and any preemptive action does have Orwellian aspects. Moreover, Amazon hasn’t always been successful and coming up with a coherent product market definition for Amazon is very difficult.

That said, Khan is correct on a lot of counts, especially the nature of platform economics. The problem of platform regulation stems not just from the behavior of platform companies, but also from insufficient progress on an economic theory of platforms.

Traditional microeconomics, with its “widget” factories, has described markets of “products” and “services” which for the most part are devoid of externalities/ scale economics. Externalities are added on later, as a complexity to be dealt with.

Common carrier/utility regulation attempts to address these for things like fax machines, telephone networks, etc. However, platforms are an odd beast where the demand-supply economics runs into limits. Multi-sided platforms combine externalities/scale economic with agents on multiple sides causing the conventional “price = marginal cost” in a competitive market to no longer be valid due to different amounts of market power and corresponding variance pricing decisions on different sides of the market. Simply put, it is far more difficult to determine abuse of power in such markets.

Economics literature and practice needs to evolve to address this, and however illogical “predatory pricing” may appear and however abhorrent preemptive action may seem, there is a lot of evidence that platform power in technology markets is hard to dislodge and Khan is right about such lasting concentration of power. “three decades” as RSJ argues may be short, but there can be plenty of harm within such a time period. Francis Fukuyama has argued that Amazon’s role in diminished downtowns across the US, while outside the strict consumer welfare standard and evidence of pricing, needs to be examined at a broader-than-Bork level in terms of its effects on society.

RSJ’s argument that “for going back in time and extracting money out of these enterprises for supposedly running a monopoly in the past when there was no case against them” is misguided and somewhat contradicts his own position against preemptive action.

DOJ/FTC bring cases post-facto, and if illegal acquisition/maintenance of monopoly is proven, both financial and structural remedies can be provided by courts. While consumers do not get “made whole” directly, indirect benefits do accrue. And in many cases, follow-on consumer class action cases from these victories (whether direct purchasers or indirect) can actually deter companies through financial awards to consumers (the US Government’s Microsoft cases were followed by consumer class actions in many states and did cost Microsoft several billion dollars), going back in time within permitted time windows.

As long as laws are in-place through the entire period, there can be no rational expectation for a firm to argue that they cannot take “ex-ante decisions” – firms hire expensive lawyers and economists precisely to estimate the risk when making product and pricing decisions!

On to other matters.

India Policy Watch #1: A ‘Point’-less Election

Policy issues relevant to India

— RSJ

The schedule for the Lok Sabha polls was announced a couple of weeks back - a 44-day, 7-phase affair that is more than twice the time the war in Kurukshetra lasted but almost half that of the war in Ramayana.

Just to put these things in the context of today’s times.

There are questions about the long, drawn-out schedule, the 7 phases in West Bengal, the arrest of Kejriwal, the freezing of Congress's bank accounts, or why Odisha is at the back end of the schedule. All are somewhat useful but besides the central mudda (point or issue) of these elections. And that is, what’s the mudda of these elections? To me, it appears like the most pointless LS election in living memory.

The first general election that I followed was the 1984 one that took place immediately after the assassination of the sitting PM, Indira Gandhi. The eight-year-old me with limited political understanding could also see the ‘wave’ in favour of INC thanks in no small measure to growing up then in a highly political milieu. But even in those lopsided elections whose results PM Modi wants to replicate in 2024, I remember the pulse of an active opposition as I accompanied an older cousin to the rallies of Chandrashekhar and George Fernandes. They were 24x7 politicians, and I guess they knew nothing else but to fight when the elections were on hand. I barely sense that today. It is true that most elections tend to be fought locally but any attempt to unseat an incumbent government requires a framing of the narrative and a conception of a core message that resonates with the electorate at large. It is this central theme that candidates can then ride on to make their case locally. That is completely missing so far. It is either the failure of the political imagination of the opposition leaders or the relative success of this government that we might not have any overarching theme that serves as the connective tissue to the multiple state or local issues. And that’s a pity because, more than anything else, elections serve as a disciplining tool in a democracy. In the absence of a real threat of losing power in elections, that discipline of being moderate, seeking consensus and thinking rigorously on policy issues can be lost. You may gain decisiveness and the ability to take bold calls, but the jury is out on how well political parties fare when choosing such issues correctly. In any case, while I can foresee lots of drama and rhetoric in campaigns at the state level, this will be a ‘no wave’ election, which means it will be impossible to dislodge the incumbent. 

While that may be true, the actions of the incumbent (the BJP) on the ground belie this optimism. The eagerness to stitch alliances across states and political spectrums - JD(S) in Karnataka, PMK in Tamil Nadu, LJP in Bihar - and the easy ceding of seats to these allies, the wait for BJD in Odisha, the proactive strikes on the electoral machinery of AAP and Congress and the decision to not give tickets to a large proportion of sitting MPs all suggest some distress signals from the ground that might be at odds with the general sense of optimism the party is projecting at the moment. I’m not sure what that could be because almost every opinion poll in the past month has shown a 350+ tally for the NDA combination, and going by past experience, this trend only consolidates itself further. So why the supposed paranoia? It is likely that the BJP has taken the 400+ seats target for its alliance to its heart and wants a record-setting mandate as a parting gift to PM Modi, who might not be leading the alliance come 2029. Or, it might be preparing itself for a post-Modi scenario of 2029 where anti-incumbency could be a real thing and a stable set of alliance partners could be helpful. Both these explanations seem flakey and far-fetched at this moment. But who knows?

---

Global Policy Watch: China’s Atmanirbhar Chips Undergo Agnipariksha

Global policy issues relevant to India

— Pranay Kotasthane

I characterise the Chinese State as a car with two accelerators and no brakes. Such a car outpaces others on its way uphill but will also get into major trouble downhill. Either way, it gets to the end before others. And so, observing Chinese economic policy is all the more important because it is a leading indicator of many trends—good and bad.

The current wave of de-coupling and securitisation of the economy began in China with the Made in China 2025 plan. A significant milestone in this downhill journey came to the fore a couple of weeks ago when Bloomberg reported that the Chinese government is asking its famed EV makers to increase their purchases from local chip makers.

Now, Tocqueville had written about the government, saying that it’s always hard to distinguish its counsels from its orders. That observation applies even more sharply to the Chinese party-state. Its seemingly innocuous ask is actually a local sourcing mandate. It might seem like a run-of-the-mill industrial policy with Chinese characteristics, but here’s the clincher: many of these chips are costlier to produce in China, and yet the government is imposing local sourcing on its EV makers. This is how Bloomberg describes it:

The ministry is now directly instructing firms to avoid foreign semiconductors if at all possible, said the people, who asked to remain anonymous discussing sensitive information. That means overseas chip firms effectively have to manufacture their silicon through a local foundry such as Semiconductor Manufacturing International Corp. or Hua Hong Semiconductor Ltd., one of the people said. During a recent tender conducted by a major domestic brand, one foreign bidder failed to secure a contract despite offering a price they estimated was 30% lower than the eventual winner, another person said.

This, to me, is a leading indicator of the future. As the many planned semiconductor fabrication facilities come online to assuage national security concerns, we will end up with excess global capacity, even as the per unit cost of production of chips at the new facilities will be higher than what a comparative-advantage-led market would offer. The only way to bridge this gap would then be local sourcing mandates underwritten by taxpayers. JP Kleinhans and I had warned about this trend last year in a Nikkei Asia article:

Let us assume for a moment that the subsidies are wildly successful. New front-end fabs start coming up at various places as intended. But these chips will be costlier to produce outside countries where they enjoy a comparative advantage. TSMC estimates that the advanced chips coming out of their new Arizona fab will be roughly 30 per cent more expensive than the ones from Taiwan. The same is, most likely, true for other Asian foundries setting up fabs in the United States and Europe. The big question is, why would original equipment manufacturers (OEM) – companies building cars, smartphones, laptops, and so on – buy a costly chip from TSMC Arizona when they can buy the exact same chip at a lower price from TSMC Hsinchu?

If OEMs are hesitant to source domestically, governments might have to further intervene with local production requirements, just like the US government did with the Inflation Reduction Act. From a regulatory perspective, it would be relatively easy for a government to say that 20% of chip content in an electric vehicle, smartphone or data centre needs to be domestically sourced. However, implementing such a policy will add layers of complexity and costs to the already heavily intertwined global semiconductor value chain.

So, while it’s not surprising to see local sourcing of chips becoming a reality, what’s interesting to note is that it’s China which has gotten to this move before others. The much-famed ‘China cost advantage’ has withered away, and customers have to pay a national security premium for chips that were earlier easily importable a few years ago. This will hurt Chinese firms’ competitiveness going ahead, especially in products where the semiconductor Bill of Materials (BoM) is a substantial portion of the product cost. Looks like the car with two accelerators is going downhill for the moment.

Speaking of local sourcing mandates, I read another China-related news with keen interest. The Financial Times reported:

China has introduced new guidelines that will mean US microprocessors from Intel and AMD are phased out of government PCs and servers, as Beijing ramps up a campaign to replace foreign technology with homegrown solutions.  The stricter government procurement guidance also seeks to sideline Microsoft’s Windows operating system and foreign-made database software in favour of domestic options. It runs alongside a parallel localisation drive under way in state-owned enterprises. 

As it happens regularly nowadays, this sent observers into a tizzy. Those bedazzled by China posed this as proof that China’s industrial policy is working. To them, these guidelines indicate that Chinese chipmaking capabilities are closing the gap with those in the West.

My takeaway was exactly the opposite. Because China has splurged massive subsidies to uphold an inefficient and not-so-effective local semiconductor industry, the only way to recoup that investment is to issue local sourcing mandates like this one. To me, these mandates indicate that local chipmakers aren't able to produce chips competitively. Because if they were, there would be no reason to issue such mandates.

Moreover, many such mandates have failed in the recent past. A so-called "3-5-2" policy in 2019 had domestic substitution targets to replace foreign PCs and software in government offices. At that time, the goal was domestic substitution at a pace of 30 per cent in 2020, 50 per cent in 2021 and 20 per cent in 2022. Back then, we saw similar reports claiming that China would close the gap with the West. But none of those targets were achieved.

Another reason for my scepticism is that had the CCP truly replaced Intel/AMD PCs; it wouldn't have quietly released this procurement directive; there would have been fanfare and chest-thumping.

It is not that China has no domestic alternatives but that it faces an uncertain future after disrupting technology transfer streams that existed in the past. For instance, Hygon Information Technology can access an old version of x86 IP acquired through AMD. The AMD-China joint venture was added to the US Entity List in 2019, meaning that AMD could no longer work with its Chinese partners to develop newer x86 chips.

The other Chinese company, Zhaoxin, is a joint venture between VIA and the Shanghai Municipal Government. VIA is a Taiwanese fabless company that has an x86 license. These companies have been trying to substitute Intel and AMD for a long time. But with the joint ventures hobbled by US export controls, they will find it difficult to keep pace with the West going ahead.

Finally, even if they can succeed in design, Chinese fabless companies will have to rely on national manufacturing champion SMIC to produce these chips as the doors of TSMC are now closed to Chinese customers. So, SMIC's capabilities will remain a binding constraint.

Overall, these two developments tell me that the Chinese semiconductor industry is not in good shape, despite oft-repeated perceptions that the US export controls have been ineffective in slowing China’s progress on this front.

---

India Policy Watch #1: A Wicksellian Disconnection

Policy issues relevant to India

— Pranay Kotasthane

The Wicksellian Connection is a public policy framework that can help us understand the root cause of Bengaluru’s water problems. This term, coined by Albert Breton, describes an intuitive idea—the taxes you pay should be closely linked with the services you receive.

Think of a Venn Diagram with three partly overlapping circles representing “those who decide,” “those who benefit,” and “those who pay.” These are called the three circles of budgetary policy. The intuition is that the higher the overlap between the three circles, the better the accountability and, hence, delivery. The stronger the Wicksellian Connections, the better the result.

With this framework at hand, we can describe Bengaluru’s water system as a Venn Diagram in which the overlap between these three circles barely exists. The disconnections exist not at one but several layers. Consider a few.

First, half of the city’s water supply comes from Kaveri, which is 90 km away and 350m lower. A WELL Labs report estimates that it costs the city’s water supply agency Rs 3 crores per day merely in electricity bills to lift this water. Before independence, Bengaluru, due to its elevation, relied largely on the rainwater stored in nearby lakes. But as the city population grew, the engineering marvel of pumping water up from the Kaveri was dreamed and realised. Many of its subsequent stages were financed through international aid by the Japan International Cooperation Agency (JICA). The lakes, now having little practical use, fell into disuse, turned into breeding grounds for mosquitoes, and became susceptible to encroachment. While the water pumped from Kaveri alleviated Bengaluru’s problems for a while, it created a Wicksellian disconnection — the state-level politicians decided, the old areas of Bengaluru benefited, and the state residents paid out the JICA loans through their taxes.

Now consider another loop of the water system. The other half of the city’s water needs are met through groundwater. These are the new parts that have been subsumed in the city. These parts have apartment complexes and technology parks that made Bengaluru the global city it is today. These parts account for the lion’s share of the city’s economic activity and, hence, taxes, including property taxes. But they aren’t the ones that benefit. In fact, apartments must compulsorily do water treatment, operate rain harvesting systems, and refrain from letting any treated water flow outside their premises. Despite bearing all these costs, they aren’t the ones that get any water from the city government. In fact, the state government uses the tax money collected largely from these areas to subsidise the Kaveri-pumped water for residents of older areas of Bengaluru! My colleague Anupam Manur estimates that while a household of four pays Rs 172 as the monthly water bill on average, the cost to the city’s water pumping agency is Rs 1248-1560, a subsidy of over 80 per cent. Another stark case of the Wicksellian disconnection.

This framework has a prescriptive dimension. The only way out of this water problem over the long term is to increase the overlap between the three circles, which implies:

1. Increase user fees for water usage. The heavy subsidies that few residents of Bengaluru enjoy must be phased out quickly.

2. The water drawn from Kaveri for Bengaluru’s use should be lowered over two decades. Extracting water from the Kaveri externalises Bengaluru’s problem and disincentivises the city from creating, storing, and utilising its own water resources. While completely forgoing the Kaveri water might not be politically feasible, the government must stabilise the current extraction levels at the very least.

3. The city must not incorporate any new areas under its administration without the water supply agency having the capacity to provide water through groundwater, lakes, and tankers.

These three measures will increase the overlap between the three budgetary circles and result in a sustainable solution for a chronic crisis.

---

Global Policy Watch: Kahneman On Expertise & Biases

Global policy issues relevant to India

— RSJ

Daniel Kahneman passed away this week. Kahneman was not a trained economist, yet he employed his training as a psychologist to question the fundamental assumption of humans as rational actors that forms the basis for modern economic theory. Along with Richard Thaler, he founded the field of behavioural economics, and his books and papers have helped us to question our biases and engage with the fallacies in our thinking while making decisions. In terms of impact and followership, he was among the most important economic thinkers of the past half a century. Over the years, those who have followed Kahneman have stretched the discipline of behaviour economics a bit too thin and used it as a convenient hammer to beat down every aspect of rational-actor economics. That’s wrong, of course. However, some of this adventurism led to a more critical appraisal of Kahneman and some undeserved scrutiny of his theories. That shouldn’t take away from his seminal contributions, his willingness to listen to the other side (evidenced in numerous interviews and podcasts), and his ability to demystify concepts in simple and clear language. 

I will close with one of my favourite Kahneman articles that he co-authored with Gary Klein, titled ‘Conditions for Intuitive Expertise: A Failure to Disagree”. At the heart of the article is that old question - does the intuitive judgement of experts work well, or do they tend to be overconfident in their abilities? How reliant should we be on experts and their intuitions as against using a ‘heuristics and biases’ (HB) approach? As their summary states - 

“Starting from the obvious fact that professional intuition is sometimes marvellous and sometimes flawed, the authors attempt to map the boundary conditions that separate true intuitive skill from overconfident and biased impressions.”

So, what do they conclude? Coming from different starting positions (Klein favours expert intuition while Kahneman prefers the HB approach), they find quite a bit of convergence in their views. I have reproduced the conclusions below:

In an effort that spanned several years, we attempted to answer one basic question: Under what conditions are the intuitions of professionals worthy of trust? We do not claim that the conclusions we reached are surprising, but we believe that they add up to a coherent view of expert intuition, which is more than we expected to achieve when we began.

1. Our starting point is that intuitive judgments can arise from genuine skill—the focus of the NDM approach— but that they can also arise from inappropriate application of the heuristic processes on which students of the HB tradition have focused.

2. Skilled judges are often unaware of the cues that guide them, and individuals whose intuitions are not skilled are even less likely to know where their judgments come from.

3. True experts, it is said, know when they don’t know. However, nonexperts (whether or not they think they are) certainly do not know when they don’t know. Subjective confidence is therefore an unreliable indication of the validity of intuitive judgments and decisions.

4. The determination of whether intuitive judgments can be trusted requires an examination of the environment in which the judgment is made and of the opportunity that the judge has had to learn the regularities of that environment.

5. We describe task environments as “high-validity” if there are stable relationships between objectively identifiable cues and subsequent events or between cues and the outcomes of possible actions. Medicine and firefighting are practiced in environments of fairly high validity. In contrast, outcomes are effectively unpredictable in zero-validity environments. To a good approximation, predictions of the future value of individual stocks and long-term forecasts of political events are made in a zero-validity environment.

6. Validity and uncertainty are not incompatible. Some environments are both highly valid and substantially uncertain. Poker and warfare are examples. The best moves in such situations reliably increase the potential for success.

7. An environment of high validity is a necessary condition for the development of skilled intuitions. Other necessary conditions include adequate opportunities for learning the environment (prolonged practice and feedback that is both rapid and unequivocal). If an environment provides valid cues and good feedback, skill and expert intuition will eventually develop in individuals of sufficient talent.

8. Although true skill cannot develop in irregular or unpredictable environments, individuals will sometimes make judgments and decisions that are successful by chance. These “lucky” individuals will be susceptible to an illusion of skill and to overconfidence (Arkes, 2001). The financial industry is a rich source of examples.

9. The situation that we have labelled fractionation of skill is another source of overconfidence. Professionals who have expertise in some tasks are sometimes called upon to make judgments in areas in which they have no real skill. (For example, financial analysts may be skilled at evaluating the likely commercial success of a firm, but this skill does not extend to the judgement of whether the stock of that firm is underpriced.) It is difficult both for the professionals and for those who observe them to determine the boundaries of their true expertise.

10. We agree that the weak regularities available in low-validity situations can sometimes support the development of algorithms that do better than chance. These algorithms only achieve limited accuracy, but they outperform humans because of their advantage of consistency. However, the introduction of algorithms to replace human judgement is likely to evoke substantial resistance and sometimes has undesirable side effects.

---

HomeWork

Reading and listening recommendations on public policy matters

1. [Paper] Here’s a must-read, foundational paper on an Innovation Policy for India by RA Mashelkar, Ajay Shah and Susan Thomas.

2. [Article] A neat backgrounder on the Indian government’s aggressive stance at the WTO’s thirteenth ministerial conference.

3. [Podcast] Our Puliyabaazi on ‘one nation one election’.

---