South Korea All-Industry Production Index (SA)

This is a seasonally-adjusted series that KRED records without transformation, namely the monthly all-industry production index that aggregates the real output of every producing sector, including mining and manufacturing, construction, market and non-market services, utilities, and public administration, into a single volume index normalized so that a chosen reference period equals 100. A rise in the index signifies an expansion in aggregate output volume, and a fall a contraction.

The measurement object is an output volume index that belongs to the lineage of industry-level physical-output indices. This volume index stands in the tradition of the physical-output indices that Fabricant (1940) built industry by industry. The accounting frame that holds it is the production-boundary and aggregate-output system of national accounting, defined for that purpose by Kuznets (1941) and fixed in double-entry social-accounting tables by Stone (1947) and Meade and Stone (1941).

Combining heterogeneous sectoral quantities into a single number is an index-number problem, organized through statistical, test, and functional approaches (Frisch 1936). The aggregation rests on the base-period weighting rule of Laspeyres (1871) and its current-period counterpart in Paasche (1874), whose geometric mean yields the ideal index of Fisher (1922).

Modern compilation chain-links these building blocks year by year. The theoretical grounding is the superlative index theory of Diewert (1976) for flexible aggregator functions, the continuous-time aggregation of Divisia (1925), and the annual chain-linking practice that Hill (1988) surveyed. The share-weighted form of Törnqvist (1936) supplies the discrete-time aggregator.

Read as a co-movement summary of many activity series, the index continues the reference-cycle interpretation of Burns and Mitchell (1946) and the composite-of-many-indicators construction of Persons (1923).

KRED applies no transformation to this series and stores the index exactly as compiled, neither rescaling nor deflating nor smoothing nor annualizing it. The only construction the series embodies is therefore the measurement basis by which an output index comes to exist, not any KRED-side recomputation.

That basis is the volume-index framework in which physical quantities across industries are converted to a common real basis and aggregated. It is the method that Fabricant (1940) established for measuring manufacturing output, and the way in which the national-accounting framework of Kuznets (1941), Stone (1947), and Meade and Stone (1941) embeds that output in a consistent set of production and expenditure totals.

The aggregation weights follow the base-period formula of Laspeyres (1871) and the current-period formula of Paasche (1874), bracketed by the ideal index of Fisher (1922). Chain-linking across successive weight periods rests on the superlative index theory of Diewert (1976) for flexible aggregator functions, the continuous-time limit of Divisia (1925), the share-weighted aggregator of Törnqvist (1936), and the annual chain-linking practice that Hill (1988) sets out. The choice among these formulas is itself the index-number problem that Frisch (1936) posed.

One step outside KRED is already present in the input, since the series arrives carrying monthly seasonal adjustment performed at source. The within-year calendar and seasonal pattern has been removed before the data reach KRED, which is consistent with KRED adding nothing further.

Because the compiled figure is a measured quantity rather than an exact total, it inherits the precision-of-estimate considerations that Stone, Champernowne, and Meade (1942) formalized. As a composite of many co-moving activity series, the index follows the standardization-and-combination procedure of Persons (1923) and the comovement reference-cycle logic of Burns and Mitchell (1946).

As a broad monthly read on aggregate real activity, the all-industry production index is used to date and gauge the business cycle. It serves as a roughly coincident indicator of the economy's expansion and contraction phases, in the reference-cycle tradition of Burns and Mitchell (1946) and the composite business-conditions reading of Persons (1923).

The index proxies, at monthly frequency, the real output that quarterly national-accounts aggregates measure only with a lag. Because it covers the whole production boundary that Kuznets (1941) defined and that Stone (1947) and Meade and Stone (1941) embedded in the accounts, it is a workhorse for nowcasting current-quarter growth and for locating the output gap.

The sectoral detail lets analysts trace which producing sectors drive a given turn, continuing the physical-output lineage of Fabricant (1940). The index foundations and the chain-linking govern how movements over time are read as real volume changes rather than artifacts of weights or prices (Frisch 1936; Laspeyres 1871; Fisher 1922; Diewert 1976; Divisia 1925; Hill 1988).

Because the published figure is an estimate subject to revision, early readings of the cyclical signal can shift as source data accrue. The reliability bounds that Stone, Champernowne, and Meade (1942) formalized underpin this caveat.

In markets the all-industry production index functions as a high-frequency activity surprise. The gap between the released volume index and its expected path moves government-bond yields, the policy-rate path priced in money markets, and the currency, since a stronger real-output print raises the odds of tighter policy.

Traders use it to confirm or contradict the turning points that leading indicators imply. It reads as a roughly coincident cyclical gauge and as a composite-of-activity barometer for assessing current conditions (Burns and Mitchell 1946; Persons 1923).

The index lets cross-asset desks gauge whether a slowdown is broad or confined to specific sectors. It spans the full production boundary of Kuznets (1941) and the accounting structure of Stone (1947) and Meade and Stone (1941), while the physical-output detail of Fabricant (1940) guides rotation between cyclical and defensive equity sectors.

Quantitative signal extraction depends on comparability across time, which the volume-index machinery and the chain-linking underpin (Laspeyres 1871; Fisher 1922; Diewert 1976; Divisia 1925; Törnqvist 1936; Hill 1988). Revision risk warns that positions taken on a first print can be unwound when the estimate is updated (Stone, Champernowne, and Meade 1942).

KRAIP is a seasonally-adjusted-at-source all-industry 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 = 0.162, and the stationarity-null test rejects trend stationarity at p < 0.01 (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 = 0.0035, 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 = 0.205 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 absence of seasonality reflects the source adjustment rather than an inherent property. 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, and the seasonal dummies are not jointly significant at p = 0.40, so no deterministic seasonality survives, which reflects the seasonal adjustment already applied at source and not an inherent absence of a seasonal pattern, consistent with the stationary-seasonality reading of Canova and Hansen (1995) and the seasonal-adjustment literature (Ghysels and Osborn 2001).