Sectors and Economic Structure
Jan 2024 - Alex Alejandre

Louis-Vincent Gave called Hong Kong a “Tax Paradise” Timestamp at 20:45 because with a blanket 16% rate, corporate and private, there’s no need to optimize structures. Bill Gates roommate in Harvard, Andy Braiterman, intimidated Gates away from math into CS. What became of him? Tax Attorney! The convoluted US tax system (17,000 pages) has bred a large cottage industry (6th largest employer in aggregate) of professionals catering to optimizations other systems don’t need. Their economies have better primitives, abstracting away complexity. (On the other hand, a large cottage industry of professionals increases economic complexity. Is this tradeoff (burden on other enterprises) accretive to growth? I think no.) Our system misallocates its best human capital in tax and finance.

What can we do with this?

Industry Classification Systems

Once upon a time three sector models primary: raw materialssecondary: manufacturingtertiary: services quarternary: R&D, knowledge quinary: lol were enough, nowadays unifying the myriad taxonomies in use is impossible. These are the most relevant today:

  • Market Based
    • GISC Global Industry Classification Standard
      • 11 sectors
      • 25 industry groups
      • 74 industries
      • 163 subindustries
    • ICB - Industry Classification Benchmark
      • 11 industries
      • 20 supersectors
      • 45 sectors
      • 173 subsectors
    • HSICS - Hang Seng Industry Classification System
      • 12 industries
      • 31 sectors
      • 102 subsectors
    • TRBC - Reuters' used in the Refinitiv Indices
      • 13 sectors
      • 33 business sectors
      • 62 industry groups
      • 154 industries
      • 898 activities
    • MGECS - Morningstar Global Equity Classification Structure
      • 3 super sectors (cyclical, defensive, sensitive)
      • 11 sectors
      • 55 industry groups
      • 145 industries
  • Establishment based
    • NAICS - North American Industry Classification System
      • Here each establishment is no larger than a single physical location though different administrative groups in the same location can be separate establishments. The decision is made on its largest output.
    • SIC - Standard Industrial Classification
      • Here an establishment is a US headquartered business, based on its largest product lines
    • NACE

The difficulties of wrangling with their inconsistent mappings This kills naive factor models. distract us from more important comparative questions. (Although I’ll use index weightings like everyone else, there are many problems from mislabelling (CLF has been mislabelled an iron miner for years, stuck in mining ETFs etc.) and poor mapping and I personally hand label everything I analyze (which lead to my own problems, thankfully uncorrelated!)) They are inconsistent because they don’t take the ecosystem into account, what role companies play in the economy or in other industries.

What Role do these Sectors Play?

In the US, healthcare makes up ~20% of the US economy, compared to 5% in 1960. In China, France, the UK etc. most healthcare is paid through the government and taxes (and at lower GDP) with much less activity in privately traded companies.

The Chinese government reports GDP differently, not including state-provided services (healthcare, (non-cram) education). Charles Gave Louis-Vincent’s father once called GDP “misleading if not outright criminal” for measuring private and public sector contributions equally although they’re judged by different theories of value (marginal vs. labor/cost theories of value). When university costs the German government $9000 instead of an American consumer $40,000…

We have 3 possibilities which ruin econometric research:

  • healthcare in GDP and equities market (US) It’s in all equities markets, but see below.
  • healthcare in GDP (EU)
  • healthcare not shown (China)

We can find many cases where GDP lowers while people’s wealth or well-being grows. Lowering the cost of health inputs (through large single payers, be it Singapore or UK style) materially lowers GDP while helping individuals and the overall economy as they can consume more total goods and services. France and Spain have recently grown their GDP through public debt-fueled spending (even debt servicing adds to GDP!)

Back to weightings, we see MSCI Europe’s energy peak at 11.5 in 2002, shrinking to 5% in 2001. Financials went from 25% to 15%, Utilities from 12% to 3% in the same time frame. MSCI USA has energy peak in 2007 at 12%, crashing to 2% in 2001, which is closer to MSCI World. But if that’s so easy, why did I wax wordy?

MSCI Europe has a (slightly) higher healthcare weight (15%) than MSCI USA. Analyzed more specifically, the composition is very different (Pharmaceuticals, Equipment etc.) but more importantly: world tech monopolies based in the US shrink other industries' proportion/weighting dramatically. Why is it that healthcare makes up 20% of GDP and tech only 10%, yet healthcare is half tech’s size on the stock market, yet tech profits are more than double healthcare’s? Because it’s a poor measurement on both sides. Healthcare does not include health insurance, which is in “financials”.

With perfect data, we could investigate whether reducing total spending on necessities (health, education, food) results in growth and well being, how it differs in kind etc. but confined to our worldly data the typical analysis constrains itself to cost theory of value although with health specifically there are (politically motivated) indices of outcomes.

We know multiples differ by industry and country. We often misappreciate precisely what this means and how the differences multiply. I will start with misleading data: EV/EBITDA I don’t like EBITDA, but easy data. in the US: 3 for coal, 20 for med tech or diversified REITs vs. 3.5 and 22 in Europe. Why is this misleading? Because market wide EV/EBITDA in the US is 14.3 vs. 8.2 in Europe (personal data with sample bias of what interests me. This source disagrees with my personal data, showing recent divergence of 15.4 US to 11.2 EU but overall similarity.) A dollar of profit for a European energy company translates to a higher valuation than for an American one (surprising given the poor regulatory environment with windfall taxes and courts ordered decarbonization) but an American tech dollar in profit increases valuations way higher than a Euro in a consumer defensive. Such distortions lead to beautiful value plays (but also result in structural value traps.)


P.S. Local investor habits and tradition play similar roles:

In China, real estate is the primary investment instrument, resulting in cheaper stocks (ADRs have further issues due to political fears.)

In Germany, investors focus on ETFs (A2PKXG is the current “holy grail” replacing A1JX52.) (Generally Germans are unsophisticated investors, many just use a Sparbuch or follow bank advisors' advice to buy products.)

India and the Middle East love gold!

Brazil offers fascinating fixed income options (resulting in Yen carry trade) and literal free money.

Please send me further examples!