There is a well-documented tension in systematic investing between what a rules-based strategy prescribes and what a human operator is willing to execute when the signal generates a trade that feels counterintuitive. This tension is not unique to Korean markets — but KOSPI 200 has structural features that make it particularly acute for Korean retail investors, and understanding those features matters for anyone assessing whether a back-tested rotation strategy will be implemented faithfully in real conditions.
This post examines the behavioral patterns visible in retail order flow around KOSPI 200 quarterly reconstitutions and rebalancing events, what a systematic back-test shows about the cost of discretionary overrides, and why rules-based rotation is not simply a performance claim but a discipline question.
Past-performance disclaimer: Back-test illustrations and return comparisons in this article use synthetic parameter examples. Backtest results are not a guarantee of future returns; this is research, not investment advice.
The KOSPI 200 Reconstitution Effect and Retail Herding
KRX announces KOSPI 200 quarterly reconstitutions (구성종목 변경) approximately five trading days before the effective date. The announcement generates a predictable retail order flow pattern: incoming constituents see buying pressure as retail investors anticipate index-replication demand, while outgoing constituents see selling pressure as retail investors extrapolate the removal as a fundamental signal. Both effects reverse partially after the effective date — a phenomenon documented across global index markets and observable in KRX data through the price behavior of affected names over the 10-day window straddling reconstitution.
The retail herding pattern around reconstitution is consistent with recency bias (최근 성과 편향): retail investors interpret recent price strength (which drove the constituent's inclusion) as forward-looking, and recent price weakness (which contributed to exclusion) as a reason to sell. The back-test evidence on this behavior is straightforward: strategies that buy incoming constituents at the announcement and sell at the effective date capture the reconstitution premium — but this is a different strategy entirely from a fundamental momentum rotation, and conflating the two generates spurious performance attribution.
Discretionary Override Costs in a Rotation Framework
Consider a synthetic scenario: an individual investor running a KOSPI 200 value rotation — quarterly rebalance, equal-weight top quintile by P/B — encounters a rebalance signal in Q3 2022 to buy a domestic steel company (코스피 철강주) trading at a deep discount to book value. The investor's narrative memory from 2021–2022 is dominated by 2차전지 (secondary battery) outperformance and the steel sector's near-complete exclusion from retail conversation. The signal says buy; the investor's pattern recognition says this sector has been dead for two years and the signal must be wrong.
If the investor overrides the signal and substitutes a position in a battery materials name already richly valued relative to the factor model, they have made a discretionary override. The cost of that override is not the single-trade P&L difference — it is the compounding of signal-following discipline across all future rebalances. Research into retail systematic strategies consistently shows that the gap between a back-tested rules-based strategy and the live discretionary variant of the same strategy grows with each override, primarily because overrides are not random: they are consistently pro-cyclical. Investors override signals to buy what has recently performed and avoid what has recently lagged, which is the precise opposite of what most value-based and mean-reversion rotation signals prescribe.
The Psychological Geometry of KOSPI Sector Narratives
Korean retail investor psychology is influenced by a sector narrative culture that is more concentrated than most other markets. The 2차전지 supercycle of 2020–2022 and the 반도체 (semiconductor) cycle that preceded it created a retail mental model where KOSPI sectors are categorized into "growth themes in play" and "value traps to avoid." This binary framing is actively unhelpful for systematic rotation, because rotation strategies by design shift capital from overvalued momentum names into undervalued laggards — precisely the names narrative culture labels as value traps.
We are not claiming that sector narratives are uniformly wrong — we are saying that when a retail investor's narrative conviction overrides a systematic signal, the resulting trade is no longer a test of the strategy but a test of the narrative. The back-test's Sharpe ratio is irrelevant to a strategy that is being discretionarily modified at every rebalance. What you are actually running at that point is a discretionary strategy that uses a systematic framework for signal generation but not for execution — a combination that typically produces neither the systematic strategy's discipline nor the discretionary manager's flexibility.
What a Rules-Based Back-Test Shows About Override Costs
One way to quantify override costs without requiring individual transaction data is to compare the simulated output of a fully systematic rotation against a "narrative-adjusted" variant where the strategy's prescribed trades are filtered by a simple narrative heuristic: skip any signal to buy a name that ranks in the bottom tercile of 12-month momentum (the "dead sector" filter) and substitute cash or a higher-momentum name instead.
In synthetic walk-forward tests on KOSPI 200 value rotation (2010–2023, quarterly rebalance, top quintile P/B, equal weight), applying a "no buy if bottom-tercile momentum" override reduces strategy performance by approximately 80–150 basis points annually. The degradation is not uniform: it is concentrated in years where value rotation was most rewarded — exactly the years where the overridden stocks recovered from depressed valuations and the "dead sector" assessment was most incorrect.
This concentration is the key insight. The override filter does not hurt the strategy in the years when narrative avoidance was coincidentally correct. It destroys value in the years when value rotation worked specifically because the narrative was wrong. Since those recoveries are the fat tail of the strategy's return distribution, removing them via systematic discretionary override truncates the upside without proportionally protecting the downside.
Reconstitution Timing and Systematic Discipline
Beyond individual stock selection overrides, the second common discretionary intervention is rebalance timing drift. A strategy specifying a first-trading-day-of-month rebalance may be executed on day 3, 5, or 7 as the investor monitors news flow and waits for "better conditions." Back-test results assume exact rebalance dates; execution drift creates a systematic deviation between the back-tested and realized strategy.
Rebalance timing drift is not neutral — it tends to be correlated with market conditions. Investors are more likely to delay rebalances during drawdown periods (when the signal is telling them to hold or add to positions that are declining) and more likely to execute promptly or early during momentum-chasing periods (when recent winners are being signaled for further purchase). This correlation means timing drift is systematically pro-cyclical in the same direction as stock-selection overrides, compounding the behavioral drag rather than averaging it out.
The Case for Removing the Override Decision Entirely
The argument for rules-based rotation is not that quantitative signals are always correct — it is that the alternative is discretion, and discretion in retail contexts has a documented bias toward narrative-driven pro-cyclicality that compounds negatively over time. A systematic back-test is not a promise of future performance; it is an honest measurement of what a rule would have done if followed exactly. The discipline question is whether an investor can tolerate the periods when following the rule feels wrong — which are typically the periods when the rule eventually proves most valuable.
Finology's rotation framework is configured around fixed, rule-defined rebalance dates aligned to the KRX trading calendar, with parameter sets that are locked before back-testing begins and not adjusted after review of results. The output includes the rebalance log showing each prescribed trade, so the gap between the back-test and any live discretionary deviation is visible and measurable. More on implementation mechanics on the Platform page.