South Korea Services Production Index (SA)

KRSVCS is a monthly volume index of services-sector production that measures the real output of the service industries in a given month relative to a fixed base period set to 100, carried as a seasonally adjusted series at the level on which it is published and recorded with no further transformation.

The index expresses the current month's services output as an aggregate of many disaggregated activity relatives. Its conceptual roots lie in base-period-weighted aggregation and its current-period-weighted counterpart, between which the true volume measure is read (Laspeyres 1871; Paasche 1874; Fisher 1922). The modern volume-aggregation target that such a services index approximates is defined by the superlative aggregator, the log-change share-weighted index, and its continuous-time limit (Diewert 1976; Törnqvist 1936; Divisia 1925). The distinction between the chain and the fixed-base form is systematized by the general theory of index numbers, and the quantity-relative lineage by which an output index for an industry is built from underlying production data is separately established (Frisch 1936; Fabricant 1940).

The series sits within the double-entry social-accounting framework that defines the national-income aggregate and the boundary of economic activity and codifies them into consistent definitional rules, and it is one of the many comoving activity measures whose path traces the cyclical state of output (Kuznets 1941; Meade and Stone 1941; Stone 1947; Burns and Mitchell 1946). When the index stands above 100 the real output of the service industries lies above its base-period level, and a decline means the reverse.

KRED applies no transformation to KRSVCS and stores each monthly figure exactly as released, applying none of rescaling, deflating, smoothing, or annualizing. The only adjustment present in the series is the seasonal adjustment already embedded at the measurement stage. The within-year seasonal pattern is removed before the level is recorded, and KRED leaves that adjusted level untouched.

The measurement basis of the index that KRED records is the volume-index framework, in which a quantity index aggregates output relatives across service activities. The base-period-weighted form and the current-period-weighted form are each fixed, and the geometric mean of the two satisfies the time- and factor-reversal tests (Laspeyres 1871; Paasche 1874; Fisher 1922). This formula is exact for flexible aggregators and is therefore superlative (Diewert 1976).

The log-change share-weighted discrete-time aggregator, its continuous-time limit, and the annual chain-linking of volume relatives that builds a long index without a single fixed basket formalize this measurement (Törnqvist 1936; Divisia 1925; Hill 1988). This chain logic is already present in the true volume-measure theory and is systematized by the general theory of index numbers (Fisher 1922; Frisch 1936).

The construction and realification of volume relatives follow the tradition of the physical-output index for an industry (Fabricant 1940). Placing the resulting index inside a consistent system of production and expenditure accounts follows the definitional conventions of double-entry social accounting, and the national-income activity boundary determines which service outputs enter the aggregate (Meade and Stone 1941; Stone 1947; Kuznets 1941).

As the services component of aggregate output, KRSVCS measures the part of a modern economy that accounts for the largest share of activity within the national-income boundary, and within the double-entry accounts it maps to the production-side measure of value added by the service industries (Kuznets 1941; Meade and Stone 1941; Stone 1947).

Read month to month it is one of the comoving activity series whose clustered turning points define the reference cycle (Burns and Mitchell 1946). In the composite-indicator tradition that detrends, seasonally adjusts, and standardizes many activity series into a single barometer, a services-output index is a natural coincident gauge of current conditions (Persons 1923).

Because services output is published as a chain volume index, its growth rate is comparable across periods as a real-quantity change rather than a nominal one, the quantity-relative reading established for industrial output (Hill 1988; Frisch 1936; Fabricant 1940). The superlative theory, the ideal index, and the bounding pair of base-period and current-period weights make it economically interpretable as a measure of real activity (Diewert 1976; Fisher 1922; Laspeyres 1871; Paasche 1874).

Such an estimate carries measurement error, so its short-run movements are read as signals within a band rather than as exact quantities (Stone, Champernowne, and Meade 1942).

For market participants KRSVCS is a high-frequency real-activity gauge that arrives monthly and feeds the nowcasting of quarterly output. The services-output index is the part of the production accounts that dominates value added under the activity boundary and therefore moves the headline growth read (Kuznets 1941; Meade and Stone 1941; Stone 1947).

Treated as a coincident series in the comovement framework and the composite-barometer construction, its turning points inform positioning across the cycle (Burns and Mitchell 1946; Persons 1923). Because it is a chain volume index, its month-on-month and year-on-year changes are real-quantity changes comparable through time, the quantity-relative interpretation given to physical output (Hill 1988; Frisch 1936; Fabricant 1940).

The superlative-index reasoning, the ideal index, and the bounding pair of base-period and current-period weights let a desk treat the reported growth as a clean real-activity input to a relative-value or rates view rather than as a price-contaminated number (Diewert 1976; Törnqvist 1936; Fisher 1922; Laspeyres 1871; Paasche 1874). The estimate is a measured quantity with revision risk, so a position sizes around that risk (Stone, Champernowne, and Meade 1942).

KRSVCS is a seasonally-adjusted-at-source services-activity index, and its published level is the statistical object whose integration order the triplet assesses. The sample is 316 monthly observations spanning 2000-01 to 2026-04, and the tests are fitted with a constant and trend.

The augmented Dickey-Fuller regression of Dickey and Fuller (1979), in the lag-augmented form of Said and Dickey (1984), fails to reject the unit root at p = 0.133, and the stationarity-null test rejects trend stationarity at p = 0.031 (Kwiatkowski et al. 1992), so the augmented Dickey-Fuller and KPSS legs concur toward an I(1) reading. The nonparametric test of Phillips and Perron (1988) is the sole dissenter, rejecting the unit root at p = 0.029, which is the documented Phillips-Perron over-rejection toward spurious stationarity under near-unit moving-average errors (Ng and Perron 2001), so no clean order of integration is assigned. The GLS-detrended escalation of Elliott, Rothenberg, and Stock (1996), run only as confirmation of the augmented Dickey-Fuller leg, fails to reject at p = 0.066, marginally above the five percent level, and corroborates that leg without upgrading the ambiguous verdict.

Because the level is not a clean stationary object, the portmanteau and the break search are run on its first difference, since a level portmanteau reads mechanical near-unit autocorrelation and a level mean-break search spuriously segments the series (Perron 1989; Hamilton 2018; Bai and Perron 1998). The portmanteau statistic of Ljung and Box (1978), refining the form of Box and Pierce (1970), rejects the white-noise null on the differenced series at p = 0.000, while the multiple-break procedure computed by the dynamic-programming algorithm of Bai and Perron (2003) finds no break in the mean of the differenced series.

The seasonal battery is run with the interpretation that any residual seasonality reflects the source adjustment rather than an inherent stochastic cycle. The seasonal-unit-root test of Hylleberg et al. (1990), with the monthly mechanics of Beaulieu and Miron (1993), rejects unit roots at the seasonal frequencies at p = 0.000, so no stochastic seasonality is present, while the seasonal dummies are jointly significant at p = 0.023, so a residual deterministic seasonal component remains, which is the imprint of an imperfect source adjustment passed through unchanged rather than a seasonal cycle inherent in the data, consistent with the deterministic reading of the stationary-seasonality test of Canova and Hansen (1995) and the seasonal-adjustment literature (Ghysels and Osborn 2001).