(Almost) Daily Musings

Evening all

Northern Germany: Cold. Wintery. All year round. Just as I arrive for the winter break expecting Michael Buble, Marzipan and Rotkohl, I am greeted with a heat wave (at least for local standards): 11 degrees and sunshine.

While I swap Glühwein with Margaritas, I am sharing today’s musings.

Cheers
Phil

P.S. With the holiday season in full swing, the (Almost) in this newsletter’s title may live up to its name. Potentially expect more sporadic updates over the next two weeks before we return to regular programming.

Chart Art

Class Clownery

Education spending is up, results are down, and somewhere Paul Graham’s graph on the U.S. education system is sighing loudly in the corner. Despite mountains of cash being funnelled into schools, OECD scores in literacy and numeracy continue to drop, while the gap between the most and least skilled workers widens. Legacy systems are failing those who need help the most, and investors in EdTech are conflicted, to say the least.

Let’s break it down:

• AI is democratizing education: Platforms like ChatGPT and Perplexity are offering access to learning that doesn’t require a second mortgage.

• Post-pandemic reality check: Remote learning tools thrived during lockdowns but struggled to stay relevant when schools returned in-person.

• BYJU’s cautionary tale: Once a unicorn darling, BYJU’s valuation collapse and dry exit markets have scared investors into staying put.

Still, optimism persists. For all the funding inefficiencies, there are bright spots—EdTech players carving out wins in specific niches:

• Physics Wallah — Raised $210M in September at a $2.8B valuation, leading the charge in India’s exam prep space.

• Eruditus — Bagged $150M in October, pushing forward in professional education with backing from TPG’s The Rise Fund.

• SchoolLinks — Secured $80M in October to help students prepare for college and careers after K-12.

Capital deployment? Messy.

Opportunity? Massive.

Legacy systems are failing, the skills gap is widening, and EdTech—done right—can transform outcomes. The real potential lies in democratizing access: platforms that empower the underprivileged and focus on vocational training to help those left behind catch up later in their careers.

It’s not just about fixing education—it’s about leveling the playing field.

Some context on the funding data:

• Not just deal volume and count has decreased since the height of 2021 but so have valuations relative to revenue — with multiples down and the opportunity this large, potential entry points may be attractive to investors that can spot opportunities that address the real underlying challenges.

• Seed stage: Median pre-money valuation is $3.2M, with capital invested at just $0.6M—a 26.3x valuation/revenue multiple. Not bad for day one.

• Series A: Median pre-money valuation jumps to $11M with a 2.1x step-up and 1.5 years between rounds. Valuation/revenue multiples are still strong at 12.5x.

• Across all series: Pre-money valuations hit $6.9M, and the valuation step-up sits at 4.0x—a testament to how capital is flowing despite volatility.

Regressing GDP

Bubbles are usually transparent — this one differs. The mother of all bubbles, as FT's Ruchir Sharma coined it, seems visible to all. The share of US equities as a percentage of global markets has grown to excessive proportions. Shiller PE, Put/Call Ratios, Greed vs. Fear Indices — all signals are screaming the hymn of West Ham United: Forever Blowing Bubbles.

Almost as apparent, however, is the view that no deflation of said bubble is in sight. At least not any time soon. Wall Street’s crystal ball is unanimous: US stocks will dominate again in 2025. Economists, not to be outdone, argue that with household and corporate balance sheets looking bulletproof, the economic boom isn’t going anywhere.

So where could the pop come from?

Perhaps the fuel to the economic machine may be the very downfall of it as well: Debt. Per the FT: $2 of new government debt are needed to generate $1 of GDP growth — a halving of where it stood about five years ago.

I decided to look at the data myself since 1980:

• Looking at the two line charts in Matplotlib, it’s apparent the relationship is not clear cut albeit an initial trend clearly emerges — downward sloping. Yet the initial blanco R-Squared is not strong enough to warrant strong conclusions, so I continued to dig into the stats and clean some data. More below.

I cleaned for some outliers and calculated an efficiency ratio between the Debt to GDP and GDP Growth variables.

The analysis, with a p-value at 0.000 (every statistic prof’s dream), indicates a statistically significant trend that is clearly downward sloping over time. But the R-squared was just around 25%. OK, but definitely not great.

Clearly, growth should be lagging debt. Including a lagging effect (t-2 years), the results are stellar (at least in my statistical naiveté):

• 45% R-squared

• P-values at 0.000

I wouldn’t agree with the FT that it’s half the rate it was 5 years ago. Alas, the trend remains clear: Debt is not as effective to fuel growth as it used to be.

Consider, of course, that including net transfers, the current account deficit in the US is already back over 4% of GDP as well.

That being said, the consequence will be increasingly wary investors looking elsewhere—either for a road out of deficit or for ever-increasing yields to compensate for the additional risk. As of now, it sure doesn’t seem that Trump or the Fed are looking to lower rates. It’s a double-sided coin that poses risk to the economy: weakening US growth relative to strengthening global economies, albeit the results of measures such as China's stimulus spending to prop up its economy are yet to show.