RupeeCase | Systematic Quantitative Investing Terminal by QC AlphaRupeeCase | Systematic Quantitative Investing Terminal by QC Alpha
Start free
A treemap of glass tiles in navy and teal, a few of them lit

RupeeCase notes . Topic

Portfolio construction

Diversification, correlation, allocation, rebalancing, position sizing and the risk profile that decides how much of the same portfolio you can hold.

  • 33notes in this topic
  • 6to read in order
  • 3questions answered

01Read in this order

Six notes that carry the idea.

  1. 01The mix beats the pick9 min .
  2. 02Twenty funds, one bet9 min .
  3. 03The lower number is your profile5 min .
  4. 04Risk profile fit: the drawdown you can actually sit through5 min .
  5. 05Capacity, tolerance, need: the three risks4 min .
  6. 06Size decides more than selection10 min .

02Every note

33 notes, newest first.

  1. Size decides more than selectionHolding research constant and moving the top three weights from 5 to 20 percent each quadruples a 40 percent shock's damage from 6 to 24 percent of the book. . 10 min
  2. Rebalancing is not free. And it is not always money.Two 10 percent sleeves at 25 percent volatility earn a 1.88 point bonus at minus 0.20 correlation and 0.16 at plus 0.90, while a trending decade charges 1.39 points a year. . 9 min
  3. Twenty funds, one betAt the 0.85 correlation Indian diversified equity funds show against each other, twenty funds buy 1.17 effective bets against 1.14 for five. . 9 min
  4. The mix beats the pickBrinson's 90 percent is about variability over time, not return; the same index at 100/0/0 and 40/40/20 trades 2.2 points of return for a worst drop of 28 instead of 52. . 9 min
  5. Who sells first: a household's real risk levelA shared portfolio is held by consensus and sold by veto, so the account's risk level is the lower partner's tolerance, not the average. . 4 min
  6. The one country betJapan was 45 percent of world equity in 1989 and took 34 years to regain its high; India is about 2.9 percent and most Indian portfolios sit near 100 percent India. . 4 min
  7. Home bias is rationed in India54 of 66 international funds were closed to fresh money under a 7 billion dollar industry cap set in 2008, so one country concentration is often found rather than chosen. . 4 min
  8. One profile, many goalsPointing one Moderate label at a two year flat deposit and a 25 year retirement pot loses at both ends; the deposit falls short and the far pot ends Rs 24.18 lakh lower. . 5 min
  9. Your biggest asset is not in your portfolioA 30 year old's future paycheques can be 15 to 20 times the brokerage account, so a portfolio should diversify away from the job, not toward it. . 5 min
  10. Allocation drift: the risk you didn't chooseAn untouched 60/40 drifts to about 71/29 in ten years at 12 and 6.5 percent, and falls 27.8 instead of 23.2 percent when equity drops 40. . 6 min
  11. Capacity, tolerance, need: the three risksA questionnaire scored one risk and pointed at 90 percent equity; capacity eight years from retirement and a need of about 9 percent a year both pointed near 60. . 4 min
  12. The risk you get paid forAbout 80 percent of a single stock's volatility is company specific and unrewarded, and an equal weight basket sheds it from 40 percent at one name to about 19 at thirty. . 5 min
  13. A risk profile has an expiry dateTwo years from a college goal, a 35 percent crash leaves a 90 percent equity saver Rs 10 lakh short while one who glided to 25 percent equity lands funded. . 6 min
  14. Single stock skew: own the haystackBessembinder found 57 percent of US stocks since 1926 did worse than a Treasury bill and four percent of names created all the net wealth. . 6 min
  15. The cash dial, not the stock listRisk appetite picks how much of the highest Sharpe portfolio to hold, not which stocks; swapping to a defensive basket instead costs about Rs 7.8 lakh on Rs 10 lakh over twenty years. . 5 min
  16. Volatility is half the story, correlation decidesAdding 20 percent gold at 15 percent volatility to an 18 percent volatility equity book brings the blend down to 15.0, because gold rarely falls on equity's bad days. . 4 min
  17. The calm day contractThe same investor scored Moderately Aggressive on a quiet day and Conservative three weeks into a wobble, so fill the profile on a boring day and treat it as a contract. . 5 min
  18. Capital weight is not risk weightA 60/40 portfolio's risk decomposes to about 95/5, and genuinely splitting risk down the middle takes roughly 22 percent in stocks and a 6 percent annual swing. . 4 min
  19. 60/40 is a 94/6 bet with better brandingAt 18 percent equity volatility and 5 percent debt volatility, a 60/40 book carries 94 percent of its risk in equity, and a true risk split needs about 22/78. . 4 min
  20. The optimizer is an error maximizerOne point of forecast error moved eighty points of portfolio weight in a three stock demo, which is why DeMiguel, Garlappi and Uppal found equal weight hard to beat out of sample. . 5 min
  21. Survivorship bias: the backtest told by the survivorsRunning a strategy on today's fifty index names flatters the curve by a couple of points a year, because the twenty five deletions left after they had already hurt. . 5 min
  22. The lower number is your profileA real risk profile scores willingness and capacity separately out of 50 and takes the lower one, which is how a self declared daredevil with an eight month horizon ends up conservative. . 5 min
  23. The split decides more than the pickBrinson's finding is about variability, not return, and holding the same index at 100/0/0 versus 40/40/20 trades two points of return for 34 points of drawdown. . 5 min
  24. The repo number is not the number that moves your moneyA 25 basis point cut lifts a liquid fund 0.03 percent and a long gilt 2.25 percent, and the yield curve usually priced the move weeks before the MPC met. . 5 min
  25. Position sizing: what Kelly actually saysAt a 60/40 win rate and even money, Kelly says 20 percent of the book and half Kelly says 10, because doubling the bet roughly quadruples the drawdown. . 5 min
  26. Rebalance cadence: the latency nobody readsThe same thirty names rebalanced quarterly and fortnightly landed at 30.2 and 33.8 percent CAGR over five years, because drift is a silent overweight. . 6 min
  27. The conservative starter is a drawdown number, not a nameMost investors tick the box for a 40 percent drawdown and panic at 15 to 20, which is why the first SIP should be sized to the worst window you would keep adding into. . 5 min
  28. Why adding gold can improve an all equity bookA 20 percent gold sleeve at 0.15 correlation takes a 13/19 equity book to 12.0 percent return and 15.9 percent volatility, because the variance cross term shrinks. . 4 min
  29. The diversification cliffIn March 2008 and March 2020 the textbook diversifiers fell together, because correlation is conditional and jumps from 0.10 to 0.55 for equity and gold in a liquidity crunch. . 5 min
  30. Rebalancing harvests volatilityAn equal weight five fund SIP drifted to 38/17/16/15/14 in two years, and a two asset napkin shows how rebalancing at the swing turns a flat round trip into a 10 percent gain. . 6 min
  31. How many stocks do you actually ownSeven mutual funds held 92 unique names, but the top ten carried 47 percent of the money and the long tail was too small to matter either way. . 5 min
  32. Risk profile fit: the drawdown you can actually sit throughMax drawdowns across the RupeeCase shelf ran from 16.92 to 39.48 percent when this was written, and SEBI's two axis rule makes your honest profile the lower of capacity and tolerance. . 5 min
  33. The Nifty 50 is 10 stocks plus 40 footnotesThree stocks are 30 percent of the Nifty 50 and the top ten cross 58 percent, so a cap weighted index fund is far more concentrated than the word diversified suggests. . 6 min

03Go deeper

From the notes to the course, and to the cards.

Learn path 2The free Learn course on the same ground, with a certificate at the end. 18 index algosEvery RupeeCase strategy, model back tested, in your own demat. All eight topicsBack to the hub: reading paths, search and the latest notes.

04Questions, answered with figures

Three questions readers ask about portfolio construction.

How many funds do I need to be diversified?

Fewer than you think. At the roughly 0.85 correlation Indian diversified equity funds show against each other, twenty funds buy 1.17 effective bets against 1.14 for five. A second sleeve with a genuinely different driver takes you to about 2. Read the note: Twenty funds, one bet

Is a 60/40 portfolio balanced?

In capital, yes; in risk, no. With equity swinging about 18 percent a year and debt about 5, a 60/40 book carries about 94 percent of its risk in equity. A true risk split needs roughly 22 percent in equity. Read the note: 60/40 is a 94/6 bet with better branding

What should a risk profile measure?

Two axes, not one: tolerance (what you would do when the screen is red) and capacity (horizon, emergency fund, debt). Score both out of 50 and your profile is the lower of the two. SEBI's adviser rules ask for both. Read the note: The lower number is your profile

Explore next

Where to go next.