2.6 · Portfolio construction: the whole machine
Why this lesson
Section titled “Why this lesson”You now own parts: a global index core (1.5–1.6), a PH dividend basket and REIT pick (2.1–2.2), a peso ladder (2.3), a fenced speculative sleeve (2.5), and an inoculation against engineered yield (2.4). Parts are not a machine. A machine has proportions — how much of each, written down, with rules for what happens when markets move the proportions around. That’s this lesson: asset allocation, the decision research keeps confirming matters more than every security selection you’ll ever make, and the investment policy statement (IPS) — the one-page constitution that makes your future self obey your present, calmer self.
This is also where the course’s recurring currency thread gets tied off properly. You earn USD (agency receivables), spend PHP (life), and will someday borrow PHP (Level 3). That three-way split is not a complication to apologize for — it’s a structural edge most investors would envy, if the allocation is built to respect it. Growth in USD, income in PHP, liabilities matched to peso cash flows: the logic is three sentences, and most PH portfolios violate all three.
Rob Berger for the build, PensionCraft for the diagnosis, Berger again for the maintenance schedule.
Watch for:
- 02:20 — the whole core in three funds: total US + international + bonds. Your translation is even simpler: one all-world UCITS fund (VWRA holds both equity legs).
- 06:00 — diversification is about underlying holdings, not fund count: three funds holding 10,000+ securities at ~0.07% blended cost vs a 20-fund advisor portfolio charging 1–1.5% (lesson 1.5’s fee arithmetic, weaponized).
- 26:28 — “supercharging” satellites (small-cap value, EM tilts) — and his REIT caution for taxable accounts, which PH investors dodge entirely: our REIT dividends are 10% final, not ordinary-income (a rare PH tax advantage, lesson 2.2).
Translation notes: his tickers (VTSAX/VTIAX/VBTLX) and M1 Finance platform are US-only — swap the UCITS route from 1.6. And one figure to distrust: the video shows CAGR jumping from 8.64% to 16.12% merely by adding monthly contributions — that’s a quirk of how his tool blends contribution timing into “CAGR,” not a real return difference. Backtest tools flatter; more below.
Watch for:
- 01:22 — “clutter” diagnosis: portfolios accreted from years of tips and promos (the PH version: three UITFs, two VULs someone regrets, some GCash funds, a stock from a seminar) — and how to consolidate without losing diversification.
- 11:21 — concentration hides in styles, not just stocks: funds that returned 15–20%/yr for 17 years, then −27% and staying down. What looks like skill is often one bet repeated.
- 17:14 — “too complicated” as a portfolio red flag: if you can’t explain a holding in one sentence, it’s a candidate for removal. The IPS enforces this automatically.
Watch for:
- 01:30 — absolute vs relative bands: a “10% band” means wildly different things in each convention; your IPS must say which.
- 03:45 — the honest tradeoff: never-rebalancing won the full 29-year backtest (stocks outrun bonds), but was the worst performer through bear markets — rebalancing is risk control, not return enhancement.
- 11:35 — checking bands frequently but acting rarely beat calendar rebalancing in the study he reviews.
Core-satellite: the assembly
Section titled “Core-satellite: the assembly”Core-satellite is the architecture that ends the “index fund vs everything else” argument by giving both a home. The core — the largest single allocation — is the boring, evidence-backed compounder: your all-world UCITS fund at 0.22%, bought on schedule, never traded. The satellites are the deliberate, sized deviations, each of which must earn its seat with a written purpose: the PH dividend/REIT sleeve (purpose: peso income at 10% final tax — 2.1/2.2), the guaranteed layer (purpose: stability + known maturities — MP2/RTB/ladder, 2.3), and the speculative sleeve (purpose: asymmetric bets + tuition, capped — 2.5).
The industry map’s typical shape for your band (₱500k–2M) is the course’s starting template, not a prescription: ~15% cash/emergency · ~25% MP2/bonds · ~35–40% global index core · ~10–15% PH dividend/REIT · ~10% speculative/experiments. Two knobs matter far more than the exact numbers. First, the videos’ headline finding translated: your real “stock/bond split” is core+income vs guaranteed layer, and it should be set by risk capacity (agency income volatility, dependents, how soon Level 3 needs capital) and risk appetite (your honest 1.8 drawdown rehearsal), not by template. Second, correlation — the reason the machine holds different parts at all. Diversification only works when holdings move differently; the PSEi’s flat decade while the S&P compounded (lesson 1.5) is lived proof the two equity sleeves are genuinely different machines. But price the honest catch: in a global crash, equity correlations converge toward 1 — everything falls together, and only the guaranteed layer (MP2 literally cannot mark down; maturing rungs pay par) buys stability and the ready cash that makes rebalancing rules executable at the bottom.
A word on glide: the practice of shifting weights toward the guaranteed layer as the money’s use date approaches (retirement funds do this automatically — “target-date” funds are a glide path in a wrapper). At your stage the glide is mostly flat — decades of horizon, agency income covering life — but write the trigger into the IPS anyway: weights shift toward guaranteed as Level 3’s property purchase approaches, because money needed in two years doesn’t belong in equities (lesson 1.8’s liquidity spectrum, now with a calendar).
The currency layer
Section titled “The currency layer”Three sentences, then the rules. Your growth engine is USD-denominated (all-world fund — really a basket of the world’s currencies, but USD-quoted and dollar-dominant), because that’s where the compounding evidence lives and it hedges the peso’s long slide (₱58–62 now; ~2–2.5%/yr average depreciation over the last decade — a tailwind on USD assets, with genuine reversal risk in any given year). Your income sleeve is PHP-denominated (dividends, REITs, ladder), because your expenses are PHP and income that must cross an exchange rate to buy groceries isn’t reliable income. Your future liabilities are PHP (Level 3 mortgages), so the assets that will service them stay PHP too — never fund peso amortizations from an asset that can fall 20% in peso terms because the dollar sneezed.
Your structural edge, stated once more as a rule: earn USD, invest USD, spend PHP. Agency receivables landing in USD can fund the UCITS core without ever touching pesos — no spread paid, no conversion timed. Convert to PHP only what the PHP side of the machine (expenses, income sleeve, ladder) actually needs. Currency exposure is then not a risk you’re passively suffering but an allocation you’ve chosen: roughly, the core’s share of the portfolio is your USD weight. Write it in the IPS like any other weight.
The IPS: one page, ten minutes, decades of service
Section titled “The IPS: one page, ten minutes, decades of service”The investment policy statement is the document Bernstein and every advisor worth their fee insist on, and it fits on one page. Its power is entirely in when it’s written: now, calm, with no crash on the screen. Sections:
- Purpose & horizon — what this portfolio is for (the 0.3 ladder: income floor → Level 3 capital → terminal wealth) and when each part is needed.
- Target weights — the five sleeves with a number each, plus the USD/PHP split.
- Contribution rule — the monthly amount (1.5’s Do-it) and where it lands (into whatever’s most underweight — self-rebalancing by default).
- Rebalancing rule — the course default: check quarterly; act only when a sleeve drifts more than 5 percentage points (absolute) from target; rebalance with new contributions first, sales last — sales cost STT and spreads on the PH side, and selling winners has tax consequences the contribution route never triggers.
- The never list — products this portfolio does not buy, pre-written: VULs (0.4), anything failing the 0.4 armor or the 2.4 decomposition, yield above the 2.2 spread benchmark without a written decomposition, US-situs holdings beyond the estate line (1.6), speculation above the sleeve cap (2.5).
- Trigger rules — what you do at −30% (pre-committed; lesson 2.8 writes these properly).
Everything in it should be one sentence long. If a rule needs a paragraph, it’s a discussion, not a rule — and discussions lose to fear in real time.
The lab, and its honest limits
Section titled “The lab, and its honest limits”Backtesting sites (testfol.io, portfoliovisualizer.com) let you run your target allocation against history — the closest thing to a flight simulator this course has. Use them for exactly two outputs: the max drawdown your mix would have suffered (rehearse that number against your 1.8 tolerance, in pesos, out loud) and the relative behavior of mixes (how much smoothing each 10% of guaranteed layer buys, what the income sleeve does to the ride). Distrust them for: predicted returns (past decades don’t repeat on schedule — the survivorship lesson from 1.5 applies to time periods too), anything PH-specific (US proxies stand in badly for PSEi sleeves and can’t model the 10%-vs-20% tax seam at all), and contribution-inflated “CAGR” figures like Berger’s 16.12% artifact. The simulator teaches the shape of your machine’s behavior; the hurdle math from 1.1 still prices its parts.
Assemble the machine on paper, then in the tools:
- Write the target-weight table. Five sleeves + USD/PHP split, starting from the band template, adjusted by your written risk capacity and appetite (one sentence of justification per deviation from template — the discipline is the sentence, not the number).
- Run the lab. Backtest your equity/guaranteed split (US proxies are fine for shape): record max drawdown, worst year, and the same for ±10% guaranteed-layer variants. Write the peso value of your mix’s max drawdown at today’s portfolio size, and your honest reaction to it. Resize if the reaction resizes you.
- Draft the one-page IPS — all six sections, every rule one sentence. Date it. This document is the Level 2 capstone’s backbone; 2.7 adds the tax map and 2.8 adds the trigger rules.
- Build the dashboard — one sheet, updated monthly: current value per sleeve · current vs target weight (drift) · trailing-12-month income per income sleeve, net of tax · net portfolio yield vs the two hurdles (MP2 7.12% tax-free; global index expected total return) · one cell for “action required per rules: yes/no.” Ten minutes a month; the point is that drift becomes a number, not a feeling.
Check yourself
Which single decision drives portfolio outcomes most, per the evidence in Berger's video?
In core-satellite architecture, what must every satellite have?
Risk capacity vs risk appetite — the distinction PensionCraft insists on — is:
The course's currency logic in three clauses is:
Your IPS sets a 5-percentage-point absolute band and your index core drifts from 40% to 47%. The rule-following action is:
The rebalancing backtest Berger reviews found never-rebalancing won over 29 years but was worst in bear markets. The honest conclusion:
What are backtesting tools like testfol.io trustworthy for, per this lesson?
You can move on when… your target-weight table exists with a one-sentence justification per deviation from template, the lab’s max-drawdown number has been rehearsed in pesos without resizing you (or you resized), the one-page IPS is drafted and dated, and the monthly dashboard tracks drift and net yield against both hurdles.
Go deeper
Section titled “Go deeper”Next: 2.7 · The tax layer — the machine is assembled; now the plumbing that keeps the state’s share correct and your Level 3 bankability intact: every instrument’s tax, the 8%-vs-graduated choice, and the filing calendar.