Why this is our appraisal

The fear cycle has turned; four independent measures support the conclusion. The gauge is hovering over ‘Neutral’. Foreign selling peaked in late July, at about IDR 79tn cumulative, and has since reverted to IDR 68tn, with August a net buying month and foreigners picking up IDR 2.3tn in the week to 4 September. The Rupiah has firmed to 17,629, with reserves recovered to a five-month high of USD146.5bn, and Bank Indonesia holding in August, after 100bp of hikes. Note how Indonesia is no longer Asia's least-preferred market in a BofA fund manager survey - India grabbed that dubious place in August, while Indonesia's “Underweight” narrowed from 32% to 27%. Four different instruments, one direction.

Earnings are better than the headline, while the headline is bad. Bottom-up across 65 names, covering 88% of the large-cap free float, reported how first-half profit fell 6.6%. Strip out six names carrying identified one-off items and the same basket grew 11.7%. The median names grew 17.0%. A negative aggregate against a median of that size tells you the weakness is concentrated in a handful of large companies – not spread across the market.

The consumer is split, and ‘s is now visible in earnings rather than surveys. Mid-market retail grew first-half profit hard - ERAA +38.0%, MAPI +26.5% - while staples went backwards: ICBP -33.1%, INDF -19.1%, KLBF -2.8%, AMRT -0.2%. The K-shape that Bank Indonesia's consumer survey showed through September 2025 is still running, and the listed companies are now reporting it.

The valuation is cheap and the mechanical problem is why. The index trades at 14.72x trailing earnings against a four-year mean of 16.12x, the 13th percentile of its own window. But MSCI Indonesia is down to 11 constituents and USD58.9bn of float-adjusted market cap - smaller than MSCI Vietnam, which is categorized as a “Frontier Index”. A discount only becomes a return when someone is structurally able to buy it.

The range held. What broke was the floor, not the ceiling

The July note said the index would grind sideways, shifting between 5,800 and 6,400 in a high cost-of-capital regime, and most of the period since it has done exactly that. But the first half of the year saw something the band was never asked to describe: a 42% drawdown from 9,135 in January to 5,342 in June, driven by policy noise rather than by earnings. The index reclaimed 6,000 on 13 July and has not been below that figure since. At 6,662 it is 24% down on the year, and 25% up off its low.

That combination - a violent de-rating followed by a durable reclaim of a round number - is what the end of a “fear phase” looks like. To test that rather than assert it, we built a sentiment gauge for the JCI on the CNN Fear & Greed template: six components, equal weight, each ranked against its own trailing 500 sessions. Three measure the level of the market (price against its 125-day average, drawdown from the 52-week high, recovery off the 52-week low) and three its rate of change (20-day return, 120-day return, and 30-day volatility against its own 50-day average, inverted).

The mix is the design, and it is disclosed because it decides the answer. Level components alone read 45 - still Fear, because after a 42% drawdown they stay depressed for a year. Rate-of-change components alone read 65 - already Greed, because they turn the moment a bounce becomes evident. Neither alone is informative. The blend reads 54.7 (Exhibit 1).

Riset BCA Virtue

MEDC - Riding the Tailwind

UNTR - The Good, the Bad, and the Rest

DMAS - The default address for data center wave

Semua Riset

Judul

Analis

Industri

Perusahaan

Tanggal Terbit