﻿

# Unit root

In time series models in econometrics, a linear stochastic process has a unit root if 1 is a root of the process's characteristic equation. The process will be non-stationary. If the other roots of the characteristic equation lie inside the unit circle, then the first difference of the process will be stationary.Consider a discrete time stochastic process {$y_\left\{t\right\},t=1,infty$}, and suppose that it can be written as an autoregressive process of order p:

:$y_\left\{t\right\}=a_\left\{1\right\}y_\left\{t-1\right\}+a_\left\{2\right\}y_\left\{t-2\right\} + cdots + a_\left\{p\right\}y_\left\{t-p\right\}+varepsilon_\left\{t\right\}$

Here, {$varepsilon_\left\{t\right\},t=0,infty$} is a serially uncorrelated, mean zero stochastic process with constant variance $sigma^2$. For convenience, assume $y_\left\{0\right\} = 0$. If $m=1$ is a root of the characteristic equation:

:$m^\left\{p\right\} - m^\left\{p-1\right\}a_\left\{1\right\} - m^\left\{p-2\right\}a_\left\{2\right\} - cdots - a_\left\{p\right\} = 0$

then the stochastic process has a unit root or, alternatively, is integrated of order one, denoted $I\left(1\right)$. If "m" = 1 is a root of multiplicity "r", then the stochastic process is integrated of order "r", denoted "I"("r").

For example, the first order autoregressive model, $y_\left\{t\right\}=a_\left\{1\right\}y_\left\{t-1\right\}+varepsilon_\left\{t\right\}$, has a unit root when $a_\left\{1\right\}=1$. In this example, the characteristic equation is $m - a_\left\{1\right\} = m - 1$. The root of the equation is $m = 1$.

If the process has a unit root, then it is a non-stationary time series. That is, the moments of the stochastic process depend on $t$. To illustrate the effect of a unit root, we can consider the first order case:

:$y_\left\{t\right\}= y_\left\{t-1\right\}+varepsilon_\left\{t\right\}$

By repeated substitution, we can write $y_\left\{t\right\} = sum_\left\{j=1\right\}^t varepsilon_\left\{j\right\}$. Then the variance of $y_\left\{t\right\}$ is given by:

: $Var\left(y_\left\{t\right\}\right) = sum_\left\{j=1\right\}^t sigma^2=t sigma^2$

The variance depends on t since $Var\left(y_\left\{1\right\}\right) = sigma^2$, while $Var\left(y_\left\{2\right\}\right) = 2sigma^2$. Note that the variance of the series is diverging to infinity with t.

Estimation in the presence of a unit root

Often, ordinary least squares (OLS) is used to estimate the slope coefficients of the autoregressive model. Use of OLS relies on the stochastic process being stationary. When the stochastic process is non-stationary, the use of OLS can produce invalid estimates. Granger and Newbold (1974) called such estimates 'spurious regression' results: high R2 values and high t-ratios yielding results with no economic meaning.

To estimate the slope coefficients, we can
* assume the process is stationary (has no unit roots) and use OLS, or
* assume that the process has a unit root, and apply the difference operator to the series. OLS can then be applied to the resulting (stationary) series to estimate the remaining slope coefficients.

For example, in the AR(1) case, $Delta y_\left\{t\right\} = y_\left\{t\right\} - y_\left\{t-1\right\} = varepsilon_\left\{t\right\}$ is stationary.

In the AR(2) case, $y_\left\{t\right\} = a_\left\{1\right\}y_\left\{t-1\right\} + a_\left\{2\right\}y_\left\{t-2\right\} + varepsilon_\left\{t\right\}$ can be written as $\left(1 -lambda_\left\{1\right\}L\right)\left(1 - lambda_\left\{2\right\}L\right)y_\left\{t\right\} = varepsilon_\left\{t\right\}$ where L is a lag operator that decreases the time index of a variable by one period, $Ly_\left\{t\right\} = y_\left\{t-1\right\}$. If $lambda_\left\{2\right\} = 1$, we can define $z_\left\{t\right\} = Delta y_\left\{t\right\}$ and then : $z_\left\{t\right\} = lambda_\left\{1\right\}z_\left\{t-1\right\} + varepsilon_\left\{t\right\}$is stationary. OLS can be used to estimate the remaining slope coefficient, $lambda_\left\{1\right\}$.

If the process has multiple unit roots, the difference operator can be applied multiple times.

Properties and characteristics of unit-root processes

* Shocks to a unit root process have permanent effects which do not decay as they would if the process were stationary
* As noted above, a unit root process has a variance that depends on t, and diverges to infinity
* If it is known that if a series has a unit root, the series can be differenced to render it stationary. For example, if a series $Y_t$ is I(1), the series $Delta Y_t=Y_t-Y_\left\{t-1\right\}$ is I(0) (stationary). It is hence called a "difference stationary" series.

ee also

* Dickey-Fuller test
* Augmented Dickey-Fuller test
* Unit root test
* Phillips-Perron test (PP)
* Weighted symmetric unit root test (WS)
* Kwiatkowski, Phillips, Schmidt, Shin test, known as KPSS tests

Wikimedia Foundation. 2010.

### Look at other dictionaries:

• Unit root test — A unit root test tests whether a time series variable is non stationary using an autoregressive model. The most famous test is the Augmented Dickey Fuller test. Another test is the Phillips Perron test. Both these tests use the existence of a… …   Wikipedia

• Root Sports Northwest — Launched 1987 (as Northwest Cable Sports) Network Fox Sports Net Owned by DirecTV Sports Networks Country …   Wikipedia

• Unit — may refer to:In mathematics: * Unit vector, a vector with length equal to 1 * Unit circle, the circle with radius equal to 1, centered at the origin * Unit interval, the interval of all real numbers between 0 and 1 * Imaginary unit, i , whose… …   Wikipedia

• Unit 88 — распавшаяся неонацистская группировка, основанная в Веллингтоне Коллином Уилсоном и также действовала в Окленде, Новая Зеландия. Была наиболее активна в течение 1997 1998 годов. Лидером скинхедов был Джей Бар. Группировка Unit 88 действовала …   Википедия

• Root (linguistics) — The root word is the primary lexical unit of a word, and of a word family (root is then called base word), which carries the most significant aspects of semantic content and cannot be reduced into smaller constituents. Content words in nearly all …   Wikipedia

• Root locus — In control theory, the root locus is the locus of the poles and zeros of a transfer function as the system gain K is varied on some interval. The root locus is a useful tool for analyzing single input single output (SISO) linear dynamic systems.… …   Wikipedia

• Unit (ring theory) — In mathematics, a unit in a (unital) ring R is an invertible element of R , i.e. an element u such that there is a v in R with : uv = vu = 1 R , where 1 R is the multiplicative identity element.That is, u is an invertible element of the… …   Wikipedia

• Root of unity — The 5th roots of unity in the complex plane In mathematics, a root of unity, or de Moivre number, is any complex number that equals 1 when raised to some integer power n. Roots of unity are used in many branches of mathematics, and are especially …   Wikipedia

• root — root1 rootlike, adj. /rooht, root/, n. 1. a part of the body of a plant that develops, typically, from the radicle and grows downward into the soil, anchoring the plant and absorbing nutriment and moisture. 2. a similar organ developed from some… …   Universalium

• Root cause — A root cause is an initiating cause of a causal chain which leads to an outcome or effect of interest. Commonly, root cause is used to describe the depth in the causal chain where an intervention could reasonably be implemented to change… …   Wikipedia