South Korea Business Survey Index (Outlook)

The Business Survey Index (Outlook) is a raw series recorded without transformation, holding as reported the forward-looking component of a qualitative business-tendency survey in which firms answer whether they expect business conditions in the coming month to improve, stay the same, or deteriorate. The index is a net-response, diffusion-index statistic that subtracts the share of firms expecting deterioration from the share expecting improvement and then standardizes the result about a neutral benchmark, so that a reading above the benchmark means optimistic firms outnumber pessimistic ones.

The qualitative up-same-down design originates in the theoretical link that the trichotomous reply bears to realized output (Anderson 1952). The balance statistic that converts the trichotomous replies into a single continuous reading is an established quantification (Theil 1952), an approach later systematized through probability thresholds and critically reviewed (Carlson and Parkin 1975; Nardo 2003).

The diffusion-index family to which the statistic belongs is a formalized tool of cyclical measurement (Moore 1961). The reference-cycle definition that an expectational business indicator is meant to track was fixed within the cyclical-indicator program (Mitchell and Burns 1938; Burns and Mitchell 1946). Such monthly diffusion readings have been treated as real-time signals of cyclical turning points and as one series within the coincident-leading-diffusion taxonomy (Shiskin 1961; Zarnowitz 1992). The underlying object is recast as the latent state of the economy that the survey proxies, a measurement program whose atheoretical character was closely scrutinized early on (Stock and Watson 1989; Koopmans 1947).

KRED applies no transformation and stores the published outlook index exactly as released, describing here only the measurement basis by which the business-tendency diffusion statistic comes to exist. The measurement begins with the qualitative survey instrument, in which a panel of firms reports its expected business conditions as an ordinal up-same-down judgment rather than a cardinal magnitude (Anderson 1952).

The balance of replies, the share answering favorably minus the share answering unfavorably, summarizes these ordinal responses in a single number (Theil 1952). The probability thresholds under which the trichotomous answers map to a quantitative series were formalized, and the competing balance, probability, and regression quantifications were catalogued (Carlson and Parkin 1975; Nardo 2003). The resulting statistic is presented on the diffusion-index scale about a neutral value, so that its sign carries the cyclical signal used for monthly turning-point detection (Moore 1961; Shiskin 1961).

The indicator-selection logic that justifies treating an expectational survey as a cyclical gauge derives from the cyclical-indicator program, and the reproducible turning-point dating was established as an algorithmic procedure (Mitchell and Burns 1938; Burns and Mitchell 1946; Bry and Boschan 1971). The dynamic-factor formalization of the common cyclical signal and the general design principles of an economic-status survey are likewise in place (Stock and Watson 1991; Hussmanns, Mehran, and Verma 1990). KRED performs no rescaling, deflating, smoothing, annualizing, or seasonal adjustment.

As a forward-looking business-tendency reading, the outlook index is used to gauge the near-term state of the economy ahead of the official output statistics it anticipates. This is the role documented in the early work that linked the balance of survey replies to realized production (Anderson 1952).

Within the cyclical-indicator program, an expectational survey functions as a leading-to-coincident gauge whose comovement with the reference cycle is the object of interest (Mitchell and Burns 1938; Burns and Mitchell 1946). This framing was evaluated across the coincident, leading, and diffusion families and operationalized for real-time recession-and-recovery signaling (Zarnowitz 1992; Shiskin 1961).

The cyclical signal embedded in such monthly series is reinterpreted as the latent common factor of the economy, alongside a diffusion-index reading of the share of firms turning more optimistic (Stock and Watson 1989; Stock and Watson 1991; Moore 1961).

Interpretive caution is required, since a measured indicator is not itself a structural model (Koopmans 1947). The evidence that survey-based attitudes and expectations carry information about subsequent real activity is broadly established, and how far such sentiment readings independently anticipate spending has been assessed (Katona 1951; Carroll, Fuhrer, and Wilcox 1994; Ludvigson 2004). Labor-market flow measurement complements this reading of cyclical conditions (Clark and Summers 1979).

For market participants the outlook index is a timely, high-frequency read on the business cycle that arrives before quarterly output data, so its level relative to the neutral benchmark and its turning points are watched as early signals of activity.

The turning-point logic that gives such a series its forward content was proposed in the cyclical-indicator program and made algorithmic (Mitchell and Burns 1938; Burns and Mitchell 1946; Bry and Boschan 1971). Monthly diffusion readings were treated as real-time recession-and-recovery signals and placed within the leading-indicator toolkit used to anticipate cyclical reversals (Shiskin 1961; Moore 1961; Zarnowitz 1992).

Whether a qualitative sentiment reading carries independent predictive content for spending and output beyond what hard data already reveal has been tested and surveyed, resting on the behavioral premise that expectations measurably shape demand (Carroll, Fuhrer, and Wilcox 1994; Ludvigson 2004; Katona 1951).

The factor-model basis for extracting the common cyclical signal that markets price is in place, and the link from the survey balance to the output it foreshadows is fixed as well (Stock and Watson 1989; Stock and Watson 1991; Anderson 1952).

KRBSI is a not-seasonally-adjusted monthly business-sentiment index, tested over 281 observations from 2003-02 to 2026-06 under a constant-plus-trend specification. Its survey scale is bounded, but the index stays far from those bounds over the sample, so the standard unit-root null nearly coincides with the bounded-support null and no bounded correction is required (Cavaliere and Xu 2014).

The augmented Dickey and Fuller (1979) test with the lag augmentation of Said and Dickey (1984) rejects the unit root at p = 0.0001, the nonparametric Phillips and Perron (1988) test rejects at p = 0.0036, and the KPSS test does not reject stationarity at p > 0.10 (Kwiatkowski et al. 1992), so the three procedures agree without dissent on a clean I(0) classification. The DF-GLS escalation of Elliott, Rothenberg, and Stock (1996) is reserved for ambiguous readings and is not required here.

Because the three procedures concur on a clean I(0) verdict, the level is the stationary object and its mean-break dates are citable. The Bai and Perron (1998) procedure, computed by the algorithm of Bai and Perron (2003), locates two mean shifts, at 2012-07 and 2023-01, read as level movements of a stationary series rather than as stochastic-trend breaks (Perron 1989).

The index is not seasonally adjusted, and the seasonal battery is the designed diagnostic here. The HEGY seasonal-unit-root test rejects seasonal unit roots at p = 0.000 (Hylleberg et al. 1990), the seasonal-dummy F test rejects at p = 0.000, and the QS statistic on the differenced series rejects at p = 0.000, so the seasonality present is deterministic rather than stochastic, consistent with the stationary-seasonality null framework of Canova and Hansen (1995) and the seasonal-time-series analysis of Ghysels and Osborn (2001). On the stationary level the Ljung and Box (1978) portmanteau rejects the no-autocorrelation null at p = 0.000 at both the lag-12 and lag-24 monthly rungs, indicating the strong serial dependence expected of a persistent sentiment level.