Rebalance on a Schedule or Only When Your Allocation Drifts? Two Different Rules, Tested
7 min read
Loading...
7 min read
Between January and June 2026, Indonesia's stock market fell about 35% while gold kept climbing. If your portfolio held both, that swing pulled your allocation badly out of shape, whether you noticed or not. Two different rebalancing rules handle a swing like that in opposite ways. Calendar rebalancing checks your portfolio once a year, on a fixed date. Threshold rebalancing only acts once your mix drifts past a set band, whenever that happens to be. Using real IHSG and gold prices from this exact stretch of 2026, here's what each rule would actually have done, and what each one costs.
Two rules, same goal: keep your portfolio close to the mix you originally chose. They just decide when to act in opposite ways.
Calendar rebalancing means picking a date, often once a year, and only checking your allocation on that date. If you're within range, you do nothing. If you've drifted, you trade back to target. The rest of the year, no matter what happens, you don't touch it.
Threshold rebalancing (also called band rebalancing) means you ignore the calendar and instead set a band around your target, commonly five percentage points on a major sleeve. A 70% stock target with a 5-point band triggers a trade once stocks hit 75% or 65%, and does nothing in between, whenever that happens to fall.
Both rules exist to solve the same problem: left alone, a portfolio drifts toward whatever has been moving the most, which quietly changes how much risk you're actually carrying. How Often Should You Actually Rebalance? covers the case for keeping the check simple and infrequent. This is about which trigger, a date or a drift, actually catches a real swing first.
To see the difference, take a simple, hypothetical Rp100,000,000 portfolio split 70% Indonesian stocks (tracked here with the IHSG) and 30% gold (tracked with Antam's own price), starting 31 December 2025. That day, the IHSG closed the year at 8,646.94, up 22.13% for 2025 (Investortrust.id), and Antam gold sold for Rp2,501,000 a gram (Tribun Tangerang). This is a hypothetical split, not a recommendation to hold exactly this mix. It exists so real, dated prices can be applied to round numbers and the two rules compared honestly.
Here's what happened to that split at three real, dated points in 2026:
| Date | IHSG | Antam gold (per gram) | Stock share | Gold share |
|---|---|---|---|---|
| 31 Dec 2025 (start) | 8,646.94 | Rp2,501,000 | 70.0% | 30.0% |
| 15 Jan 2026 (IHSG's 2026 peak) | 9,075.41 | Rp2,675,000 | 69.6% | 30.4% |
| 30 Jun 2026 (IHSG's 2026 low) | 5,643.19 | Rp2,630,000 | 59.2% | 40.8% |
| 21 Aug 2026 | 6,525.69 | Rp2,725,000 | 61.8% | 38.2% |
IHSG levels are the same peak, trough and current figures already verified for this window against Kompas, IDX Channel, Liputan6, Bisnis.com and Databoks (Databoks). Antam gold prices come from Kompas's own daily gold reports for 15 January, 30 June and 21 August.
At the January peak, the mix barely moved. Stocks and gold both rose, gold just slightly faster. By the June low, five months later, the stock side had fallen from 70% of the portfolio to just 59%, an 11-point drift. Even after stocks partly recovered by August, the split still sat nearly 8 points off target.
If your one check of the year fell on, say, 1 January, the next one wasn't due until 1 January 2027, well past every date in this table. Calendar rebalancing would not have looked at this portfolio once between the January peak and the August reading. It would have sat exactly as drifted as the last row shows, 61.8% stocks and 38.2% gold, all the way until its next scheduled date. No trade, no cost, but also no correction, for the better part of a year.
A 5-point band would have been breached sometime between the January reading, still inside the band at 69.6%, and the June low, already 11 points outside it at 59.2%. The exact day it crossed isn't in this data, but it happened well before the trough, while the sell-off was still underway.
That trigger forces a specific, uncomfortable trade: sell some of the gold that had been rising, and use it to buy more of the stock that had just fallen hard. Rebalanced at the June 30 point, moving the portfolio back to 70/30 meant shifting roughly Rp8,380,000 from gold into stock.
That's the trade nobody wants to make in the moment. It's also exactly the mechanism that makes rebalancing work: buying the loser and selling the winner, on a rule, not a feeling.
The two triggers don't just differ in timing. They pull the cost from different places.
Selling shares on the IDX carries a flat 0.1% final income tax on the gross sale value, charged whether the trade made money or not (Ortax), on top of a brokerage sell fee. Mandiri Sekuritas, as one named example, charges 0.28% to sell and 0.18% to buy (Mandiri Sekuritas); other brokers set their own rates. None of that final tax applies to buying, only selling, and the stock side of June's trigger was a buy, not a sell.
The cost instead falls on the gold leg. Selling roughly Rp8,380,000 of gold through a digital gold platform typically costs a 1% to 3% spread, the same range NetWort's comparison of Antam, digital gold and jewellery found for Pluang, so somewhere around Rp84,000 to Rp250,000, plus the roughly Rp15,000 buy-side brokerage fee on the stock purchase at Mandiri's rate. Call it Rp100,000 to Rp270,000 total for that one trade.
Run the drift the other way, stocks rallying past gold, and the cost structure flips: the sell leg lands on stocks, pulling in that 0.1% IDX tax on top of the brokerage sell fee. Which leg gets sold, and what it costs, depends entirely on which asset drifted up.
Calendar rebalancing, by contrast, cost nothing at all in this stretch, because it never traded. That's the whole trade-off in one line: threshold rebalancing costs real money exactly when it acts. Calendar rebalancing costs nothing until the one day a year it's allowed to.
Neither rule is strictly better. Threshold rebalancing reacts faster during exactly the moves that matter most, a real crash or a real rally, but only if you're actually checking often enough to catch the crossing, which takes more than a once-a-year glance. Calendar rebalancing costs nothing and takes no attention, but it can leave your risk badly out of position for months at a time, in either direction, purely by chance.
Combining both catches most of what each does well: check on a calendar date, and only trade if you're also past your band. How Often Should You Actually Rebalance? lays out that hybrid approach in more detail, alongside the case for keeping the trade count low overall.
Either way, the number that actually decides which rule would have helped you this year is your own current drift, not the example above. NetWort's Portfolio Health card tracks your portfolio's volatility, drawdown and other risk signals in real time, so you can see whether your own mix has quietly moved further from your target than you think. Check yours on your dashboard. You can also follow today's IDX and gold moves on NetWort's market page.