U.S. Real Gross Domestic Product

USRGDP is the published level of real gross domestic product recorded without transformation, the chain-linked volume aggregate of final output produced within the economy, expressed in Billions of Chained 2017 Dollars at a seasonally adjusted annual rate. Because the printed figure is annualized, it states the yearly output level implied by one quarter of production rather than the volume produced during that quarter, and KRED carries that convention through unchanged.

The aggregate concept rests on the national-income definition whose production, income, and expenditure boundary fixes the range of measured output (Kuznets 1941), and double-entry social accounting renders the production and expenditure measures mutually consistent (Meade and Stone 1941; Stone 1947).

The qualifier real denotes a volume measure obtained by removing price change. That measure rests on index-number theory (Fisher 1922; Frisch 1936) and lies within the bounds set by base weighting and comparison weighting (Laspeyres 1871; Paasche 1874), superlative indices refine the volume index inside those bounds (Diewert 1976; Törnqvist 1936), the continuous-time limit supplies its benchmark (Divisia 1925), and annual chain-linking to a 2017 reference year fixes the shape of the published series (Hill 1988).

A rising level therefore means the physical quantity of final output has increased, consistent with the physical-output lineage behind quantity measurement (Fabricant 1940), and the level takes its place among the activity series whose comovement defines the reference cycle (Burns and Mitchell 1946).

KRED applies no transformation to USRGDP and stores the published level exactly as recorded, neither rescaling, deflating, smoothing, nor annualizing it. The annual-rate convention and the chained 2017 reference year are properties of the compilation at source, and the volume series is already seasonally adjusted there, so the seasonality treatment of the recorded level is inherited from that source adjustment rather than applied by any KRED step.

The measurement basis by which the figure comes to exist is the national-accounts production and expenditure framework, in which gross value added is summed across producers and reconciled with final expenditure (Kuznets 1941; Meade and Stone 1941; Stone 1947).

Conversion to a volume measure is the deflation of current-price aggregates through index-number procedures. The base-weighted form and the current-weighted form bound the true volume change (Laspeyres 1871; Paasche 1874), the ideal index takes their geometric mean (Fisher 1922), superlative indices approximate a flexible aggregator (Diewert 1976; Törnqvist 1936), and the framework organizing these choices together with the continuous-time benchmark supports them (Frisch 1936; Divisia 1925). Annual chain-linking of the volume relatives fixes the reference year in which the level is expressed (Hill 1988), and the physical-quantity deflation lineage completes the construction (Fabricant 1940).

Because annualization multiplies every quarter by the same factor, a period-over-period ratio computed from the published level returns a quarterly rate of change rather than an annualized one, a distinction the series carries but does not resolve. Revision and sampling error are inherent in the level, so the precision of early estimates is bounded (Stone, Champernowne, and Meade 1942).

USRGDP is the broadest quarterly reading of aggregate real activity, and its cyclical interpretation rests on the measurement tradition that defines expansions and contractions through the comovement of many activity series (Burns and Mitchell 1946; Persons 1923). Within that apparatus the output aggregate is the reference against which individual indicators are classified as leading, coincident, or lagging (Mitchell and Burns 1938; Moore 1961), turning points are dated by reproducible algorithm (Bry and Boschan 1971), and the measurement and timing properties of the indicator system have been catalogued (Zarnowitz 1992).

The boundary of the aggregate determines which production enters the cycle at all (Kuznets 1941), and the production-expenditure reconciliation allows the same movement to be read from the supply side or the demand side (Meade and Stone 1941; Stone 1947).

Because the object is a volume level, its economic content depends on the deflation and index-construction choices made at source (Fisher 1922; Hill 1988), so a change in the level isolates quantity movement from price movement. Latent-state representations treat the aggregate and the monthly indicators as noisy observations on one common cycle (Stock and Watson 1989; Stock and Watson 1991), while the measurement-without-theory caution recalls that such an apparatus acquires meaning only against an explicit model (Koopmans 1947).

Early estimates of the level are revised as more source data arrive, so the precision analysis marks the limit of any conclusion drawn from the most recent quarter (Stone, Champernowne, and Meade 1942).

Investors read USRGDP as the broadest coincident measure of the real cycle that drives earnings, default rates, and policy expectations, an interpretation organized by reference-cycle dating and the composite-indicator tradition (Burns and Mitchell 1946; Moore 1961; Zarnowitz 1992). Turning-point algorithms give that reading a reproducible form (Bry and Boschan 1971), monthly indicators fill the gap between quarterly prints (Shiskin 1961), and single-index representations translate both into one cycle variable (Stock and Watson 1991).

Because the published number is an annualized volume level, desks comparing successive quarters work with the ratio of two annualized levels, which is a quarterly growth rate, and the index structure governs how much of any surprise is genuine quantity information (Fisher 1922; Diewert 1976; Törnqvist 1936). The chained reference year and the continuous-time view frame how a growth rate is read off the level (Hill 1988; Divisia 1925).

The boundary of the aggregate matters for mapping the print to sectoral exposure (Kuznets 1941), and the measurement-error bounds bear directly on positioning, since the largest market reactions often accompany revisions to an already released quarter (Stone, Champernowne, and Meade 1942).

USRGDP is a seasonally-adjusted-at-source real output aggregate published at an annual rate, and its published level is the statistical object. The sample is 265 quarterly observations spanning 1960-01 to 2026-01, and the unit-root triplet is fitted on the 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 at p = 0.9587, the nonparametric test of Phillips and Perron (1988) concurs at p = 0.9639, and the stationarity-null test rejects trend stationarity at p < 0.01 (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) and the modified-criterion lag selection of Ng and Perron (2001) are reserved for ambiguous outcomes under the house protocol and are not needed on this clean reading.

Because the 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), does not reject the white-noise null on the differenced series at either horizon, with Q = 4.29146 at four lags and Q = 6.84398 at eight lags and p = 0.367995 and p = 0.553556, so the quarterly change carries no linear dependence detectable at these lags. The multiple-break procedure of Bai and Perron (1998), 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) rejects unit roots at the seasonal frequencies at p = 0.000, the seasonal dummies are not jointly significant with F = 1.0637 at p = 0.365012, and the seasonal-lag portmanteau on the first difference does not reject either, with a statistic of 6.844 at p = 0.5536, so no residual seasonality survives at the quarterly frequencies, 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).