Reading List – 12th of November

This week’s reading list is about discovering that several things you thought were separate turn out to be the same thing. The Reserve Bank raised the cash rate to 4.85 per cent on 3 November, on a six to three vote, and among its reasons was that rapid growth in AI related investment is pushing up the price of technology goods. Which means an Australian mortgage is now priced, in some small part, off the pace of data centre construction in Texas. Damodaran counts almost US$1.8 trillion of market value leaving AI stocks in a single session on 5 November, from a group that had grown to roughly 37 per cent of the entire US market. Jason Furman puts the build out at about two and a half per cent of American GDP, and reckons it has flattered growth by around 0.6 per cent. The twenty five largest US public pension funds hold between US$578 and US$707 billion of those same companies, close to two thirds of it in seven names. Deutsche Bank asks why the long promised crisis never arrived, and answers that the risk changed address rather than left town. The useful lesson for investors is that diversification is a claim about what you own, not about how many things you own. The best investors are not counting their holdings. They are checking how many of them are the same bet. 

📚 Statement by the Monetary Policy Board: Monetary Policy Decision: The Board lifted the cash rate 25 basis points to 4.85 per cent on 3 November, with six members in favour and three preferring to hold at 4.60. Inflation forecasts were revised up materially, and inflation is now not expected to reach the top of the 2 to 3 per cent band until late 2027, roughly three quarters later than the August forecast. Governor Bullock’s media conference is worth reading alongside it, including the observation that monetary policy cannot do anything about global oil prices. (Reserve Bank of Australia)

📚 The AI Trade Deconstructed: AI Builders, AI Suppliers and AI Product firms! Aswath Damodaran separates one crowded trade into three with genuinely different economics, which matters because owning the suppliers is not the same as owning the builders. By his own count on his own sample of listed companies, AI stocks shed close to US$1.8 trillion on 5 November alone, having peaked at about 37 per cent of US equity market capitalisation on 30 September. He estimates collective AI product and service revenues at roughly US$250 billion, a figure resting partly on leaked numbers, and he says so. (Musings on Markets)

📚 The AI Boom’s Two Engines: Jason Furman tries to answer the question everyone skips, which is how much of current American growth simply is the data centre build out. He puts data centre and computing equipment spending at about two and a half per cent of GDP, and calculates that the build out lifted GDP roughly 0.3 per cent above where it would otherwise have been, or about 0.6 per cent once the wealth effect from rising share prices is included. These are his own simulations on his own assumptions, not national accounts, and seven companies account for about two fifths of the rise in US public company value since ChatGPT launched. (PIIE RealTime Economics)

📚 When Tech Firms’ AI Exposure Increases, So Do Public Pensions’: Michael Katz reports research from the Equable Institute finding the twenty five largest US public pension systems hold between US$578 billion and US$707 billion in AI related companies, as at 30 June 2026. About 85 per cent sits in fifteen companies, with the largest seven alone accounting for 63.3 per cent; CalPERS holds the biggest dollar amount at US$65.2 billion, while Colorado PERA has the highest share of assets at 18.7 per cent. Equable describes its own estimates as a floor, since the work excludes private equity and private debt holdings. (Chief Investment Officer)

📚 Why hasn’t there been a major financial crisis in recent years? Henry Allen, Peter Sidorov and Camilla Siazon argue that crises have not been abolished so much as relocated, with the banking channel hardened after 2008 and the risk migrating into non banks. The bank share of financial assets peaked at 46 per cent in 2008 and had fallen to 38 per cent by 2024, while the non bank share rose from 43 to 51 per cent. Their candidates for the next one are AI capital spending, private credit and stretched sovereign finances; the “red zone” thresholds they use come from Greenwood, Hanson, Shleifer and Sorensen rather than from Deutsche Bank. (Deutsche Bank Research)

📚 How can Europe speed up its shift to prefunded pensions? Alexandre Mendonca wants fiscally capable European governments to issue debt worth 25 per cent of GDP and place the proceeds in individual retirement accounts, which is his own proposal rather than anybody’s policy. The interesting part for an Australian reader is the direction of travel: Denmark, the Netherlands and Sweden are held up as the models to copy, with higher replacement rates, lower old age poverty and deeper capital markets. Germany’s own committee proposed channelling less than one per cent of GDP into funded pillars in July 2026. (Bruegel)

📚 A Strange Quantum Theorem Has Killed Einstein’s Reality: Charlie Wood on the PBR theorem, published in 2012 by Pusey, Barrett and Rudolph, which leaves physicists choosing between a reality that is objective but deeply strange and no objective reality at all. Pusey’s own summary is admirably compact: if anything is real, the quantum state is real. Fourteen years after the proof, a Cambridge group led by Crispin Barnes finally ran an expanded version of the underlying card game on an IBM quantum computer. (Quanta Magazine)

📚 Elon Musk’s Grokipedia is littered with mistakes. But Wikipedia could still learn some things from it: Bill Adair looked up the Grokipedia entry about himself and found 5,600 words where Wikipedia manages about 700, which is either thoroughness or incontinence depending on your view. He counted 32 errors in it, seven of them outright fabrications, though that tally is his own and nobody has audited it. His more interesting concession is that the machine is genuinely good at dredging up true material from the far corners of the internet that no volunteer editor ever found. (Nieman Lab)

📚 The unlikely queen of the bagel business: Katherine Laidlaw profiles Beth George, 63, who has taught close to a hundred shops in a decade how to make bagels and, more to the point, how to make money from them. Clients pay a five figure fee plus two per cent of proceeds if they sell within ten years, and her own shop in Beacon, New York moves about a thousand bagels on a weekend. Her unit economics are folk wisdom rather than industry data, but they are the kind investors should recognise: rent above ten per cent of gross revenue kills you, goods above 35 per cent worries you. (The Hustle)

Sign Up to weekly newsletter to receive these in your inbox

* Indicates required field

Name(Required)