KPSS test¶

A time-series test whose null hypothesis is level or trend stationarity and whose alternative is a unit-root process.

Version: v2 · 2026-09-19 · History
Domain-specific #: 5223
Domain group: Social Sciences
Origin domain: Economics & Finance
Subdomain: Econometrics → Economics & Finance
Aliases: Kwiatkowski–Phillips–Schmidt–Shin test

Core Idea¶

The deterministic component, long-run variance estimator, kernel and bandwidth determine critical values and finite-sample behavior; failing to reject does not prove every form of stationarity. The series is regressed on a constant or trend, partial sums of residuals measure accumulated departure and their scaled quadratic magnitude is normalized by an estimate of long-run variance. The abstraction is therefore identified by a declared carrier, a transformation or constraint over that carrier, and an invariant that tells an analyst whether the named structure is genuinely present.

How would you explain it like I'm…

Steady-or-Wandering Check

Imagine watching a bouncing ball's height over many days. You start by guessing that it keeps wiggling around the same middle height. The KPSS test adds up how far the ball keeps drifting away from that middle. If the drift piles up too much, you stop believing your guess. If it doesn't, you keep the guess, but you haven't proven it.

Steady-or-Drifting Check

Economists track numbers over time, like prices, and want to know whether a series stays around a steady level or trend, or wanders off with no anchor. The KPSS test begins by assuming it is steady. It fits a flat line or straight trend, then adds up the leftover gaps one after another. If that running total grows too large compared with how much the series normally jiggles, the test rejects 'steady'. Not rejecting doesn't prove the series is steady in every way, so people often use it together with other tests.

Stationarity-Null Test

The KPSS test checks a time series for stationarity around a level or a linear trend. Unlike unit-root tests such as Dickey–Fuller, its null hypothesis is stationarity and its alternative is a unit root. You regress the series on a constant, or on a constant and trend. Then you form running sums of the residuals, which measure how much departure accumulates. The test statistic is the scaled sum of the squared running sums, divided by an estimate of the long-run variance. Choices like the kernel and bandwidth in that variance estimate affect the results. Failing to reject does not prove the series is stationary in every sense, which is why KPSS is often used together with complementary unit-root tests.

The KPSS test reverses the usual unit-root setup. The null is level- or trend-stationarity and the alternative is a unit root. For an observed series over a stated sample period, regress y_t on a constant or on a constant plus linear trend and obtain residuals e_t. Form partial sums S_t = Σ_{s≤t} e_s, which accumulate departures from the deterministic component. The statistic is the scaled sum of squared partial sums, T^{-2} Σ S_t², divided by a long-run variance estimate of e_t. That estimate is typically a kernel (e.g., Bartlett) estimator with a chosen bandwidth. The deterministic specification, kernel and bandwidth all affect the critical values and finite-sample size and power. Failure to reject does not establish every form of stationarity. KPSS is commonly read together with complementary tests whose null is a unit root, such as ADF.

Scope of Application¶

KPSS test belongs to econometrics and is useful where the analyst can specify the typed econometrics carrier, including objects, relations, parameters, conventions, evidence, boundaries, and comparison targets, then evaluate the observed series and sample period, level- or trend-stationary null, unit-root alternative, deterministic regression, residual partial sums, long-run variance kernel and bandwidth, statistic, critical values and interpretation with complementary tests are explicit. The scope is broad within that domain but bounded by the need for the observed series and sample period, level- or trend-stationary null, unit-root alternative, deterministic regression, residual partial sums, long-run variance kernel and bandwidth, statistic, critical values and interpretation with complementary tests are explicit.

Clarity¶

The abstraction clarifies a crowded vocabulary by making the observed series and sample period, level- or trend-stationary null, unit-root alternative, deterministic regression, residual partial sums, long-run variance kernel and bandwidth, statistic, critical values and interpretation with complementary tests are explicit the center of the account. A claim should name the carrier, the governing operation or relation, the applicable assumptions, and the recognition test.

Manages Complexity¶

Without the abstraction, an analyst must reason directly over many local details: the carrier roles, admissibility assumptions, competing conventions, derived invariants, boundary cases, and proof or validation obligations specific to KPSS test. KPSS test compresses them into the roles in the structural signature. That compression permits comparison across instances without erasing the variables that determine validity. It also exposes which details may be varied safely and which are constitutive.

Abstract Reasoning¶

Identify the carrier. State what the elements, states, objects, or observations are: the typed econometrics carrier, including objects, relations, parameters, conventions, evidence, boundaries, and comparison targets. Reject examples whose alleged carrier belongs to a different problem.
Lock the constitutive rule. Express the observed series and sample period, level- or trend-stationary null, unit-root alternative, deterministic regression, residual partial sums, long-run variance kernel and bandwidth, statistic, critical values and interpretation with complementary tests are explicit independently of one notation or implementation.

Knowledge Transfer¶

Knowledge transfers strongly among subfields of econometrics because they reuse the typed econometrics carrier, including objects, relations, parameters, conventions, evidence, boundaries, and comparison targets, The series is regressed on a constant or trend, partial sums of residuals measure accumulated departure and their scaled quadratic magnitude is normalized by an estimate of long-run variance., and type the carrier, state every parameter and convention in the definition, test that the observed series and sample period, level- or trend-stationary null, unit-root alternative, deterministic regression, residual partial sums, long-run variance kernel and bandwidth, statistic, critical values and interpretation with complementary tests are explicit, compare the nearest accepted identity, and report counterexamples, uncertainty, and limiting cases.

Relationships to Other Abstractions¶

Current abstraction KPSS test Domain-specific

Parents (1) — outgoing typed relations

KPSS test is a kind of Hypothesis Testing (Null vs. Alternative) Prime

It remains its own entry because its identity is fixed by the observed series and sample period, level- or trend-stationary null, unit-root alternative, deterministic regression, residual partial sums, long-run variance kernel and bandwidth, statistic, critical values and interpretation with complementary tests.

Hierarchy paths (4) — routes to 4 parentless roots

KPSS test → is a kind of Hypothesis Testing (Null vs. Alternative) → is a kind of Statistical Inference → is a kind of Inductive Reasoning

Show alternative paths (3)

Neighborhood in Abstraction Space¶

KPSS test sits in a crowded region of the domain-specific corpus (23^rd percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Economic & Financial Indicators (42 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25