Everipedia Logo
Everipedia is now IQ.wiki - Join the IQ Brainlist and our Discord for early access to editing on the new platform and to participate in the beta testing.
Gauss–Markov theorem

Gauss–Markov theorem

In statistics, the Gauss–Markov theorem states that in a linear regression model in which the errors are uncorrelated, have equal variances and expectation value of zero, the best linear unbiased estimator (BLUE) of the coefficients is given by the ordinary least squares (OLS) estimator, provided it exists. Here "best" means giving the lowest variance of the estimate, as compared to other unbiased, linear estimators. The errors do not need to be normal, nor do they need to be independent and identically distributed (only uncorrelated with mean zero and homoscedastic with finite variance). The requirement that the estimator be unbiased cannot be dropped, since biased estimators exist with lower variance. See, for example, the James–Stein estimator (which also drops linearity) or ridge regression.

The theorem was named after Carl Friedrich Gauss and Andrey Markov.


Suppose we have in matrix notation,

expanding to,

whereare non-random but unobservable parameters,are non-random and observable (called the "explanatory variables"),are random, and soare random. The random variablesare called the "disturbance", "noise" or simply "error" (will be contrasted with "residual" later in the article; seeerrors and residuals in statistics). Note that to include a constant in the model above, one can choose to introduce the constant as a variablewith a newly introduced last column of X being unity i.e.,for all. Note that thoughas sample responses, are observable, the following statements and arguments including assumptions, proofs and the others assume under the only condition of knowingbut not
The Gauss–Markov assumptions concern the set of error random variables,:
  • They have mean zero:

  • They are homoscedastic, that is all have the same finite variance: for all and

  • Distinct error terms are uncorrelated:

A linear estimator ofis a linear combination
in which the coefficientsare not allowed to depend on the underlying coefficients, since those are not observable, but are allowed to depend on the values, since these data are observable. (The dependence of the coefficients on eachis typically nonlinear; the estimator is linear in eachand hence in each randomwhich is why this is"linear" regression.) The estimator is said to be unbiasedif and only if
regardless of the values of. Now, letbe some linear combination of the coefficients. Then the **mean squared error** of the corresponding estimation is
in other words it is the expectation of the square of the weighted sum (across parameters) of the differences between the estimators and the corresponding parameters to be estimated. (Since we are considering the case in which all the parameter estimates are unbiased, this mean squared error is the same as the variance of the linear combination.) The best linear unbiased estimator (BLUE) of the vectorof parametersis one with the smallest mean squared error for every vectorof linear combination parameters. This is equivalent to the condition that
is a positive semi-definite matrix for every other linear unbiased estimator.

The ordinary least squares estimator (OLS) is the function

ofand(wheredenotes thetransposeof) that minimizes the **sum of squares ofresiduals** (misprediction amounts):
The theorem now states that the OLS estimator is a BLUE. The main idea of the proof is that the least-squares estimator is uncorrelated with every linear unbiased estimator of zero, i.e., with every linear combinationwhose coefficients do not depend upon the unobservablebut whose expected value is always zero.


Letbe another linear estimator ofwithwhereis anon-zero matrix. As we're restricting to unbiased estimators, minimum mean squared error implies minimum variance. The goal is therefore to show that such an estimator has a variance no smaller than that ofthe OLS estimator. We calculate:
Therefore, sinceis unobservable,is unbiased if and only if. Then:
Since DD' is a positive semidefinite matrix,exceedsby a positive semidefinite matrix.

Remarks on the proof

As it has been stated before, the condition ofis equivalent to the property that the best linear unbiased estimator ofis(best in the sense that it has minimum variance). To see this, letanother linear unbiased estimator of.
Moreover, equality holds if and only if. We calculate
This proves that the equality holds if and only ifwhich gives the uniqueness of the OLS estimator as a BLUE.

Generalized least squares estimator

The generalized least squares (GLS), developed by Aitken,[1] extends the Gauss–Markov theorem to the case where the error vector has a non-scalar covariance matrix.[2] The Aitken estimator is also a BLUE.

Gauss–Markov theorem as stated in econometrics

In most treatments of OLS, the regressors (parameters of interest) in thedesign matrixare assumed to be fixed in repeated samples. This assumption is considered inappropriate for a predominantly nonexperimental science likeeconometrics.[3] Instead, the assumptions of the Gauss–Markov theorem are stated conditional on.


The dependent variable is assumed to be a linear function of the variables specified in the model. The specification must be linear in its parameters. This does not mean that there must be a linear relationship between the independent and dependent variables. The independent variables can take non-linear forms as long as the parameters are linear. The equationqualifies as linear whilecan be transformed to be linear by replacingby another parameter, say. An equation with a parameter dependent on an independent variable does not qualify as linear, for example, whereis a function of.

Data transformations are often used to convert an equation into a linear form. For example, the Cobb–Douglas function—often used in economics—is nonlinear:

But it can be expressed in linear form by taking the natural logarithm of both sides:[4]

This assumption also covers specification issues: assuming that the proper functional form has been selected and there are no omitted variables.

One should be aware, however, that the parameters that minimize the residuals of the transformed equation not necessarily minimize the residuals of the original equation.

Strict exogeneity

For allobservations, the expectation—conditional on the regressors—of the error term is zero:[5]
whereis the data vector of regressors for the ith observation, and consequentlyis the data matrix or design matrix.
Geometrically, this assumption implies thatandareorthogonalto each other, so that theirinner product(i.e., their cross moment) is zero.

This assumption is violated if the explanatory variables are stochastic, for instance when they are measured with error, or are endogenous.[6] Endogeneity can be the result of simultaneity, where causality flows back and forth between both the dependent and independent variable. Instrumental variable techniques are commonly used to address this problem.

Full rank

The sample data matrixmust have full columnrank.
Otherwiseis not invertible and the OLS estimator cannot be computed.

A violation of this assumption is perfect multicollinearity, i.e. some explanatory variables are linearly dependent. One scenario in which this will occur is called "dummy variable trap," when a base dummy variable is not omitted resulting in perfect correlation between the dummy variables and the constant term.[7]

Multicollinearity (as long as it is not "perfect") can be present resulting in a less efficient, but still unbiased estimate. The estimates will be less precise and highly sensitive to particular sets of data.[8] Multicollinearity can be detected from condition number or the variance inflation factor, among other tests.

Spherical errors

The outer product of the error vector must be spherical.

This implies the error term has uniform variance (homoscedasticity) and no serial dependence.[9] If this assumption is violated, OLS is still unbiased, but inefficient. The term "spherical errors" will describe the multivariate normal distribution: ifin the multivariate normal density, then the equationis the formula for aballcentered at μ with radius σ in n-dimensional space.[10]

Heteroskedasticity occurs when the amount of error is correlated with an independent variable. For example, in a regression on food expenditure and income, the error is correlated with income. Low income people generally spend a similar amount on food, while high income people may spend a very large amount or as little as low income people spend. Heteroskedastic can also be caused by changes in measurement practices. For example, as statistical offices improve their data, measurement error decreases, so the error term declines over time.

This assumption is violated when there is autocorrelation. Autocorrelation can be visualized on a data plot when a given observation is more likely to lie above a fitted line if adjacent observations also lie above the fitted regression line. Autocorrelation is common in time series data where a data series may experience "inertia."[11] If a dependent variable takes a while to fully absorb a shock. Spatial autocorrelation can also occur geographic areas are likely to have similar errors. Autocorrelation may be the result of misspecification such as choosing the wrong functional form. In these cases, correcting the specification is one possible way to deal with autocorrelation.

In the presence of spherical errors, the generalized least squares estimator can be shown to be BLUE.[12]

See also

Other unbiased statistics

  • Best linear unbiased prediction (BLUP)

  • Minimum-variance unbiased estimator (MVUE)


Citation Link//doi.org/10.1017%2FS0370164600014346Aitken, A. C. (1935). "On Least Squares and Linear Combinations of Observations". Proceedings of the Royal Society of Edinburgh. 55: 42–48. doi:10.1017/S0370164600014346.
Sep 25, 2019, 12:53 AM
Citation Linkbooks.google.comHuang, David S. (1970). Regression and Econometric Methods. New York: John Wiley & Sons. pp. 127–147. ISBN 0-471-41754-8.
Sep 25, 2019, 12:53 AM
Citation Linkbooks.google.comHayashi, Fumio (2000). Econometrics. Princeton University Press. p. 13. ISBN 0-691-01018-8.
Sep 25, 2019, 12:53 AM
Citation Linkopenlibrary.org, p. 110.
Sep 25, 2019, 12:53 AM
Citation Linkbooks.google.comHayashi, Fumio (2000). Econometrics. Princeton University Press. p. 7. ISBN 0-691-01018-8.
Sep 25, 2019, 12:53 AM
Citation Linkopenlibrary.orgJohnston, John (1972). Econometric Methods (Second ed.). New York: McGraw-Hill. pp. 267–291. ISBN 0-07-032679-7.
Sep 25, 2019, 12:53 AM
Citation Linkopenlibrary.orgWooldridge, Jeffrey (2012). Introductory Econometrics (Fifth international ed.). South-Western. p. 220. ISBN 978-1-111-53439-4.
Sep 25, 2019, 12:53 AM
Citation Linkopenlibrary.orgJohnston, John (1972). Econometric Methods (Second ed.). New York: McGraw-Hill. pp. 159–168. ISBN 0-07-032679-7.
Sep 25, 2019, 12:53 AM
Citation Linkbooks.google.comHayashi, Fumio (2000). Econometrics. Princeton University Press. p. 10. ISBN 0-691-01018-8.
Sep 25, 2019, 12:53 AM
Citation Linkopenlibrary.org, p. 23-note.
Sep 25, 2019, 12:53 AM
Citation Linkopenlibrary.org, p. 22.
Sep 25, 2019, 12:53 AM
Citation Linkopenlibrary.org, p. 135.
Sep 25, 2019, 12:53 AM
Citation Link//doi.org/10.1093%2Fbiomet%2F37.1-2.14910.1093/biomet/37.1-2.149
Sep 25, 2019, 12:53 AM
Citation Link//www.jstor.org/stable/23321582332158
Sep 25, 2019, 12:53 AM
Citation Link//www.ams.org/mathscinet-getitem?mr=00369800036980
Sep 25, 2019, 12:53 AM
Citation Linkcds.cern.ch"How to combine correlated estimates of a single physical quantity"
Sep 25, 2019, 12:53 AM
Citation Linkui.adsabs.harvard.edu1988NIMPA.270..110L
Sep 25, 2019, 12:53 AM
Citation Link//doi.org/10.1016%2F0168-9002%2888%2990018-610.1016/0168-9002(88)90018-6
Sep 25, 2019, 12:53 AM
Citation Linkbib-pubdb1.desy.de"On the determination of the b lifetime by combining the results of different experiments"
Sep 25, 2019, 12:53 AM
Citation Linkui.adsabs.harvard.edu1990PhRvD..41..982L
Sep 25, 2019, 12:53 AM