Many Stocks, One Market Risk: What Beta Reveals in an IHSG Selloff

Owning more stocks does not necessarily give a portfolio more places to hide. Different businesses can respond to the same withdrawal of risk appetite, leaving a long holdings list exposed to one dominant force: the Indonesian equity market.

Trade Inc’s study of stock beta against the IHSG, Indonesia’s composite equity index, makes that distinction tangible. In the report’s historical sample through 19 October 2026, lower-beta stocks generally moved less aggressively during market declines. They still suffered substantial losses. Relative protection was not capital protection.

This editorial interprets the supplied Trade Inc research rather than presenting a new backtest. Trade Inc publishes both this article and the underlying report. All research figures below belong to the report’s dated sample, not to a live portfolio or current trading recommendation.

The essential distinction

• Beta describes estimated sensitivity to the market, not the probability of a profit.
• More holdings can reduce company-specific risk without removing a shared market exposure.
• A negative result in a short factor study does not establish that a strategy can never work.
• Cross-asset comparisons need consistent treatment of currencies, dividends, closing times and data limitations.

A long holdings list can still depend on one outcome

The report records declines in 319 of 323 stocks meeting its liquidity criteria on 16 October. That denominator matters. These were not all IDX listings: eligibility required median trading value of at least Rp1 billion over 60 sessions and trading activity on at least 80% of those days.

Company selection does not become irrelevant in such conditions. Debt, cash generation and the price paid still influence long-term investment outcomes. But a sound business assessment does not guarantee that buyers will be available when many investors want to reduce exposure together.

A portfolio can span several industries while remaining dependent on the same funding conditions, currency and appetite for Indonesian risk. Diversification by company name is therefore different from diversification by economic exposure.

The first reduces the damage from a problem at one company. The second asks whether the portfolio has anything that behaves differently when its main market is under pressure. The two jobs should not be confused.

Beta is a sensitivity estimate, not a price forecast

Beta summarizes a historical relationship between stock returns and market returns. A value above one indicates greater estimated sensitivity; a value below one indicates less. Neither provides a loss limit or an instruction to buy.

The report combines volatility and correlation estimates and shrinks extreme results toward one. Its beta figures are therefore not simply the slope of a line through the last few trading days. Lookback length, data completeness and estimation choices all affect the result.

One illustration assigns 40% to ICBP, 30% to BBCA and 30% to PTRO, using reported betas of 0.62, 0.94 and 1.33. Their weighted average is 0.929, rounded to 0.93.

For a hypothetical Rp100 million portfolio, multiplying that estimate by a scenario in which the IHSG falls 10% gives a market-sensitivity component of approximately minus Rp9.29 million. It is a linear approximation, not a forecast. Company news, changing correlations, execution constraints and unstable beta estimates can produce a very different realized loss. The holdings are an arithmetic example, not a proposed allocation.

Defensive does not mean untouched

For groups formed using beta on 30 December 2025, the report shows a 21.8% decline in the lowest-beta quintile through 19 October 2026, compared with 42.2% for the IHSG. A quintile is one fifth of the sample, not a list of five stocks.

That difference is economically meaningful. It is also a poor reason to describe the lower-beta group as safe. The report says the same group fell as much as 38.9% below its starting value at its deepest point during the period. An investor needing cash before a recovery cannot rely on the more comfortable end-date figure.

Nor did beta order the final outcomes perfectly. The highest-beta group did not deliver the worst 2026 return among all five groups. Market sensitivity captures one relationship; it does not exhaust the forces affecting an investment.

The useful decision is therefore not to search for an immune stock. It is to ask whether a plausible loss remains compatible with the investor’s cash needs and capacity to hold through uncertainty.

The BAB result needs a narrower conclusion than a slogan

Frazzini and Pedersen’s Betting Against Beta research studies a factor that takes leveraged exposure to lower-beta assets and opposing exposure to higher-beta assets. That is not the same as simply owning defensive stocks without borrowing or short positions.

Trade Inc’s reported implementation produced a compounded annual return of minus 6.1% between October 2023 and September 2026. Its reported CAPM alpha was minus 7.2%, with an alpha t-statistic of minus 0.82. The base sample contained only 36 monthly returns, with a median of 199 stocks per month, before transaction costs.

The proportionate conclusion is that this sample did not provide strong evidence of a positive BAB premium. It also did not establish a reliable advantage for the opposite strategy. A statistically insignificant result is not proof that an effect cannot exist.

Readers will see minus 5.1% in the report’s base-case sensitivity table. That table annualizes the arithmetic average monthly return by multiplying it by 12; the minus 6.1% headline uses compounding. They should not be compared as if they were identical annual-return measures.

Implementation creates another gap between a research factor and an investable portfolio. Financing, short-position availability, costs and achievable execution prices need separate examination. A model result is not evidence that an investor could have realized it.

Correlation explains why diversification feels weaker in stress

Beta connects a stock to the market. Pairwise correlation describes how stocks move together. They are related but different measures.

The report records average pairwise correlation of 0.34 among liquid stocks on 19 October, using a 60-session window. Its median for periods classified as calm was 0.06. A correlation of 0.34 does not make the stocks identical. It does suggest less room for their movements to offset one another than in the study’s calm periods.

The diversification exercise illustrates the consequence. Of 2,000 randomly constructed 20-stock portfolios, 96% lost money during 2026 through the observation date. All of the simulated 20-stock portfolios declined on 16 October.

Those percentages are not forecasts for next year. The eligible sample required prices throughout the 2026 observation period, creating a selection limitation tied to data availability. The exercise describes this episode under its stated rules, not a universal probability of loss.

Different risks matter more than historical winners

The cross-asset section changes the question from how many stocks to how many sources of risk. Assets outside Indonesian equities may respond differently, but introduce currency, price and investment-access risks of their own.

The historical findings should not become a claim that dollars, gold or money-market funds always protect capital. The crisis sample is limited. Gold and US equities did not help in every episode studied. All comparisons are in rupiah, so exchange-rate movements contribute to the results.

The proxies also differ. The S&P 500 series reinvests dividends, whereas the IHSG series is a price index. Money-market performance is represented by one fund; observations after 5 October 2026 are projected using its preceding 60-day rate. The mixed-asset results are therefore neither a perfectly like-for-like product comparison nor the actual record of an investor’s portfolio.

The durable lesson is not to buy the historical winner. It is to examine whether everything owned depends on the same conditions—and whether money needed soon is relying on an equity recovery that may not arrive in time.

Use the research as a risk map

Start with the largest exposures, then assess market sensitivity and liquidity needs. Consider losses worse than the beta approximation, not just its central implication. A closing price may not be an available exit when queues and trading limits impede transactions.

This research supports greater discipline about market exposure, not the abandonment of company analysis. Longer samples, more complete corporate-action adjustments and results after implementation frictions would strengthen the evidence. Changes in market composition and correlations could alter the conclusions.

Explore the Stock Beta and IHSG report in Trade Shop for the charts, quintile comparisons, sensitivity tests and full methodology. This article focuses on the investment implications; the report supplies the context needed to inspect the numbers.

For adjacent questions, see our guides to stock-return correlation and maximum drawdown.

Primary source: Trade Inc report dated 20 October 2026, with data through 19 October. Research figures are attributed to the supplied report, not independently recomputed from its complete raw dataset. Its methodology removes 204 daily returns outside ±35% as possible unadjusted corporate actions; this is not a substitute for comprehensive adjustment. No new backtest was run for this article. This is not personalized investment advice.

This article is for research and information only. It is not personalized financial advice or an instruction to trade.