South Korea Official Foreign Reserve Assets

The Korean external-position stocks KRFXRESV, KREXTDEBT, KRFDIOUT, and KRFDIIN are raw stock series recorded as reported without any transformation, the cross-border asset and liability holdings of the Korean economy measured at a point in time. They span the official foreign reserve assets, the gross external debt outstanding, and the outward and inward direct-investment positions, so together they describe both sides of the country's balance sheet against the rest of the world.

Each stock is a national-balance-sheet magnitude in the lineage of the external-wealth accounts, where a country's foreign assets and liabilities are tabulated as an international investment position whose change reconciles cross-border flows with valuation effects (Lane and Milesi-Ferretti 2001; Obstfeld 2012). The gross positions carry information beyond the net figure, since offsetting claims of similar size can still imply very different exposure to valuation and rollover risk (Gourinchas and Rey 2007b; Lane and Milesi-Ferretti 2018).

The measurement sits within the financial-accounts framework of moneyflows accounting, where every claim is entered as an asset of one sector and a liability of another (Copeland 1949; Stone 1966). The sectoral tabulation follows the national balance sheets built to record financial assets and liabilities stock by stock (Goldsmith, Lipsey, and Mendelson 1963), under the compilation conventions for sectoring units and classifying instruments (Dawson 1996).

Because each series is a stock rather than a flow, it is read through the intertemporal accounting identity linking saving, investment, the current account, and the change in net foreign assets (Obstfeld and Rogoff 1995), and the reserve and debt aggregates connect to the long-run credit-and-liquidity record in which rapid external accumulation has preceded instability (Schularick and Taylor 2012; Jordà, Schularick, and Taylor 2017). These external-position stocks are original series carried without seasonal adjustment, so any seasonality they contain is a property of the data and not of a prior filter.

KRED applies no transformation to KRFXRESV, KREXTDEBT, KRFDIOUT, or KRFDIIN, neither deflating, re-scaling, smoothing, annualizing, nor converting currency, so each published number is the compiled stock exactly as measured in its source unit, whether official reserve assets in thousands of US dollars or the external-debt and direct-investment positions in millions. No seasonal adjustment is applied, since these are original series.

The measurement basis is the international-investment-position framework for external assets and liabilities, whose estimation methods build gross and net foreign positions economy by economy (Lane and Milesi-Ferretti 2001; Lane and Milesi-Ferretti 2007). The reserve, debt, and direct-investment components are compiled with component-by-component valuation, the practice established for measuring gross external positions over long spans (Gourinchas and Rey 2007a).

The underlying accounts are the financial accounts of the flow-of-funds system, first founded as moneyflows accounting (Copeland 1949) and later implemented as a full empirical construction (Copeland 1952), which records each instrument simultaneously as one sector's asset and another's liability under quadruple-entry bookkeeping. The stocks are read from the integrated social-accounting framework (Stone 1966) and tabulated on the national balance sheets (Goldsmith, Lipsey, and Mendelson 1963), following the handbook conventions for compiling financial accounts (Dawson 1996).

Aggregating heterogeneous cross-border claims into one currency total invokes index-number and aggregation theory, under which quantities combine through their weights into a consistent total (Barnett 1980; Diewert 1976). Each series is an end-of-period position rather than a modelled quantity, and its level is reconciled with cumulated flows through the valuation channel that separates position change from transactions (Gourinchas and Rey 2007b; Obstfeld and Rogoff 1995).

The external-position stocks are watched as gauges of a country's solvency and vulnerability to the rest of the world, since the balance of foreign assets against foreign liabilities conditions how shocks to capital flows transmit to the domestic economy. The international investment position that these stocks compose carries information beyond the current-account flow, because gross claims and obligations of similar magnitude can still leave an economy exposed through their currency and maturity composition (Obstfeld 2012; Obstfeld and Rogoff 1995).

Official reserve assets measure the buffer available to meet external obligations, while gross external debt measures the claims that must be serviced or rolled over, so the two together bound the country's external liquidity. The direct-investment positions record the more durable cross-border ownership links, whose valuation and return behaviour differ from those of debt and portfolio claims (Gourinchas and Rey 2007b; Curcuru, Dvorak, and Warnock 2008).

These magnitudes sit within the external-wealth accounts, whose construction across many economies made the net foreign asset position a comparable object of analysis (Lane and Milesi-Ferretti 2001; Lane and Milesi-Ferretti 2018). Read against the saving-investment identity, the change in the net position is the counterpart of the current account (Obstfeld and Rogoff 1995), while the underlying financial accounts locate each claim within the circular flow of payments (Copeland 1952; Stone 1966).

Rapid accumulation of external claims and liabilities has repeatedly preceded financial stress in the long-run macrofinancial record, so the level and growth of these stocks inform assessments of external and systemic risk, and the sectoral balance sheets make that accumulation an observable structural indicator (Schularick and Taylor 2012; Jordà, Schularick, and Taylor 2017; Goldsmith, Lipsey, and Mendelson 1963).

For market participants the external-position stocks are inputs to sovereign and country risk, since the size of the reserve buffer relative to short-term external debt shapes the probability of a disorderly external adjustment and feeds into sovereign spreads. Gross positions must be read with attention to valuation, because measured returns and levels on cross-border claims are sensitive to the estimation method (Gourinchas and Rey 2007a; Curcuru, Dvorak, and Warnock 2008).

Reserve adequacy is assessed against external liabilities, and the direct-investment positions matter for the stability of external financing, since equity-like ownership stakes tend to be more durable than debt under stress (Obstfeld 2012; Lane and Milesi-Ferretti 2018). The net foreign asset position, of which these stocks are components, is the balance-sheet counterpart of cumulated current accounts and a determinant of the sustainable external path (Obstfeld and Rogoff 1995; Lane and Milesi-Ferretti 2001).

Because rapid growth in external and credit aggregates has been linked to crisis frequency, desks use the level and momentum of these stocks to calibrate tail scenarios for currency and rates exposures (Schularick and Taylor 2012; Jordà, Schularick, and Taylor 2017). The valuation channel through which positions move independently of flows is itself a source of gains and losses on the external balance sheet (Gourinchas and Rey 2007b).

The financial accounts and national balance sheets that frame these positions are the same statements lenders, reserve managers, and macroprudential overseers consult when sizing external and funding risk (Goldsmith, Lipsey, and Mendelson 1963; Copeland 1949; Dawson 1996).

KRFXRESV is the official foreign reserve assets recorded without seasonal adjustment, and its published log-level is the statistical object, expected to be integrated of order one because these are positive reserve stocks rather than a signed flow. The sample is 438 monthly observations spanning 1990-01 to 2026-06, and the unit-root triplet is fitted on the log-level 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 with a statistic of -0.33 at p = 0.989, the nonparametric test of Phillips and Perron (1988) concurs with a statistic of -0.39 at p = 0.987, and the stationarity-null test rejects trend stationarity at p < 0.01 with a statistic of 0.82 (Kwiatkowski et al. 1992), so all three procedures agree on a clean I(1) classification. The GLS-detrended escalation of Elliott, Rothenberg, and Stock (1996) is reserved for ambiguous outcomes under the house protocol and is not invoked on this clean reading.

Because the log-level is I(1), the portmanteau and the break search are run on its first difference, since a level portmanteau reads the near-unit autocorrelation of a stochastic trend and a level mean-break search spuriously segments it (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, with Q = 105.17 at lag 12 and Q = 112.51 at lag 24, both at p = 0.000, and the multiple-break procedure computed by the dynamic-programming algorithm of Bai and Perron (2003) places two breaks in the mean of the differenced series, at 1998-01 and 2005-01.

Because these reserve data are not seasonally adjusted, the seasonal battery describes genuine seasonality rather than the residue of a prior filter. 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 the series carries no stochastic seasonal unit root, and the seasonal dummies are not jointly significant, with an F statistic of 0.40 at p = 0.958, so no deterministic monthly pattern is present. The seasonal portmanteau nonetheless rejects on the differenced series, with a statistic of 131.88 at p = 0.000, which reflects residual serial dependence at the seasonal lags rather than a seasonal unit root. This reading of stationary residual seasonality without a seasonal unit root is consistent with the seasonal-frequency framework of Canova and Hansen (1995) and with the seasonal-adjustment literature (Ghysels and Osborn 2001).