by François Nielsen & Gary Gaddy

1.  Heterogeneity Bias

Techniques of pooled time series of cross sections are applicable in situations in which one has observations on N units (such as individuals, areal units, or countries) at T points in time (such as monthly, yearly, or every 5 years).
With data like these the standard linear regression model is written:
(1)  Yit = a + Xit'b + eit  with i = 1,...,N; t = 1,...,T
a is the intercept
vector Xit' contains K regressors for unit i at time t
vector b contains K regression coefficients to be estimated
by assumption E{eit} = 0 and Var{eit} = se2
There need not be the same number of time points for each unit of observation, but assume T is the same for all units in this presentation to keep the notation simple.

A main strength of longitudinal design is that it allows controlling for heterogeneity bias due to the confounding effect of time-invariant variables omitted from the regression model.

EXAMPLE:  A sample of N secondary school students are observed from the 7th to 12th grades (T = 1,..,6).
Suppose a researcher estimates the model

(2)  GPAit = a + b1SESit + eit
But the "true" model is
(3)  GPAit = a + b1SESitb2IQit + eit
Assume also that SES and IQ are correlated, that is r(SES, IQ) <> 0.
Then model (2) suffers from specification bias: the effect of SES is (typically) overestimated.

With longitudinal data, the effect of relatively time-invariant variables (like IQ in the previous example) will be similar to the effect of a unit-specific intercept, that varies across units but remains constant for a given unit over time.  If there is such a unit-specific intercept and it is not included in the regression model, the result is heterogeneity bias.  Heterogeneity bias may cause the OLS estimates of the parameters to be entirely different from what they are in the "true" model.  The mechanism is illustrated in the next exhibit.

Exhibit:  Mechanism of heterogeneity bias  (Hsiao 1986, Figures 1.1 to 1.5 p. 7)

Longitudinal data permit correcting for the effect of any combination of omitted variables, like IQ, that are stable over the period of observation.  This is done by "simulating" the combined effect of such time-invariant omitted variables by individual-specific intercepts ai.
Model (1) becomes:

(4)  Yit = ai + Xit'b + eit  with i = 1,...,N; t = 1,...,T
The individual-specific intercepts ai capture any combination of time-invariant variables that have been omitted, knowingly or not, from the regression model.
There are two approaches to estimation of model (4), the fixed effects model (FEM) and the random effects model (REM).

2.  Fixed Effects Model (FEM)

In the FEM, the ai (also called incidental parameters) are treated as fixed constants, as the regression coefficients ai in the equivalent model:
(5)  Yit = a1d1ita2d2it + ... + Xit'b + eit
where each djit is a unit-specific indicator (dummy) variable which is 1 when i = j and 0 otherwise.  There are N djit indicators, one for each unit is the analysis.  (5) does not include a general intercept a to avoid perfect collinearity with the set of N indicators djit .  For the obvious reason, (5) is often called the LSDV (Least Squares with Dummy Variables) model.

Rather than estimating (5) with N indicators, the LSDV estimate of b, bLSDV, can be obtained from an OLS regression of (Yit - Yi.) on (Xit - Xi.) with no constant term, where

Yi. is the unit-specific mean of Yit
Xi. is the vector of unit-specific means of the predictors Xit
In other words Model (5) is equivalent to an OLS regression using the deviations of all the variables from their unit-specific means.  This regression is sometimes called the within regression.
The unit-specific intercepts ai can then be estimated as
 ai = Yi. - Xi.'bLSDV

3.  Random Effects Model (REM)

The random effects model is
(6)   Yit = a + Xit'b + ui + eit
with assumptions
E {ui}= 0 and Var{ui} = su2
Cov{eit, ui} = 0
Var{eit + ui} =  se2 + su2  =  s2
Corr{eit + ui, eis + ui} = rsu2/(se2 + su2)
(Note that (6) includes a general intercept a.  Perfect collinearity is avoided by the assumption that the expectation of the unit-specific errors ui is zero.)
The unit-specific components are now denoted ui (instead of ai) to emphasize that they are now considered a stochastic (random) component of the same type as the error eit, with a certain distribution characterized by its mean and variance, rather than fixed parameters.

Note that the assumptions about Model (6) imply that the variance-covariance matrix of the composite error term (ui + eit) is not scalar, as assumed for OLS, so that OLS is not the best estimator.  Model (6) can be estimated by Generalized Least Squares (GLS).  Assuming that the variance-covariance matrix of the error term is known, say s2W, the GLS estimator becomes

bGLS = (X'W-1X)-1X'W-1Y
where W-1 is the inverse of the matrix W.
It can be shown that the GLS estimate associated with the REM boils down to an OLS regression of
Yit - qYi.
(1 - q) and (Xit - qXi.)
where (1 - q) corresponds to the constant term and q is between 0 and 1.  In other words, the original data are transformed by removing a fraction q of the unit-specific means Yi. and Xi., instead of removing all of the unit-specific means, as the LSDV transformation does.  (In fact the FEM implemented as LSDV can be viewed as a limiting case of GLS, where q = 1.)

q is calculated as

q = 1 - se/s2
s22 = se + Tsu2
To estimate q one must estimate se  and su2.  There are several ways of doing this.  (One of the ways is to use the residuals from the LSDV regression; another uses the OLS residuals.)

4.  FEM versus REM

There are several considerations involved in choosing between FEM and REM:

1.  The fixed effects and random effects approaches can be contrasted by comparing the data transformations with which they are equivalent.

Therefore, the REM transformation may be seen as preserving more of the information (between units variation) in the data than the FEM transformation.  The GLS transformation is more efficient than the LSDV transformation, when REM assumptions are satisfied.

2.  The consistency of REM, however, depends on assumption that the ui are uncorrelated with regressors in the model.  If they are correlated, the estimates are inconsistent.  FEM does not require the assumption that the ai are uncorrelated with the other regressors, since the ai are treated as the coefficients of ordinary indicator (dummy) variables that are allowed to covary with other regressors.  While the assumption of non-correlation for REM may seem restrictive, it is often no more implausible than the usual assumption that the error term is uncorrelated with the regressors in ordinary regression models.

In small samples the net result of the trade-off of efficiency versus consistency is not easy to derive analytically, so that some of the literature on this topic has used Monte Carlo simulation to examine the small sample properties of the alternative estimators.  The GLS approach is often found to perform better overall.

In some situations, such as models with the lagged value of the dependent variable, the ui are necessarily correlated with one of the regressors.  In such cases REM is not justified.

3.  FEM uses up all between units variation and therefore does not allow including time-invariant variables in the model, as these are collinear with the (explicit or implicit) set of unit-specific indicators representing the fixed effects.  The REM model permits the use of time-invariant variables.

4.  Two statistical can be used in the context of panel regresion.


The previous models can be extended by allowing for a time-specific component in addition to the unit-specific component.
The FEM version of the time-specific component model is
(7)   Yit = a + ai + lt + Xit'b + eit
In model (7) the ai and the lt are constrained to sum up to 0.
The REM version of the model is
(8)   Yit = a + Xit'b + ui + wt + eit
The estimation methods are derived in similar ways.
It is also possible to mix FEM and REM by using explicit indicators for the time component, say, and the REM for the unit-specific component, or vice-versa.


1.  Example - Income Inequality and Economic Development

Exhibit:  Title page of Nielsen & Alderson (1995) with Kuznets curve (Figure 1)

Exhibit:  Depiction of between versus within country inequality trends (Figure 2)

Exhibit:  LIMDEP program

Exhibit:  LIMDEP output for model of income inequality

Exhibit:  Published table with model of income inequality  (Table 2a)

Exhibit:  Joint testing of groups of variables using OLS  (Appendix A)

2.  Example - Infant Mortality in European Countries

<This example is not available at this time>

3.  Example - Dynamic Model of Educational Enrollments

<This example is not available at this time>


We don't know of any "easy" introduction to pooled time series of cross sections analysis.  You may find that Rosenfeld and Nielsen (1984) is the closest thing to it.  We find Chapter 29 "Fixed and Random Effects Linear Models" in the LIMDEP 6.0 manual very helpful (Greene, 1992).  A more detailed theoretical discussion of the statistical issues involved can be found in the text by the same author (Greene, 1990: Chapter 16, especially the section called "Longitudinal Data" pp. 480-505).  Another clear exposition is provided in Judge et al. (1980: Chapter 8, pp. 325-373; there is a newer edition of this text).  Hsiao (1986) is advanced but difficult.  The same may be said for Tuma and Hannan (1984: Chapter 13).  The new book by Baltagi (1995) is very useful too, and very advanced.  Early examples of applications in sociology can be found in Nielsen and Hannan (1977), Nielsen (1980, 1986), and Pampel and Williamson (1988).  See Nielsen and Alderson (1995) for an application to an unbalanced cross national data set with different numbers of observations over time for different countries.  Betz and Katz (1995) have recently criticized several studies, mainly in political science, that use a pooling model called the Parks method, which was provided as an option in the old SAS TSCSREG procedure, as reporting unrealistically small standard errors of estimates and exaggerating the significance of coefficient estimates.  Their criticism is specific to the Parks method and does not apply to methods discussed in this workshop, however.


Baltagi, Badi H.  1995.  Econometric Analysis of Panel Data.  New York: Wiley.
Beck, Nathaniel and Jonathan N. Katz.  1995.  "What To Do (and Not To Do) with Time-Series Cross-Section Data."  American Political Science Review 89:634-47.
Greene, William H.  1990.  Econometric Analysis.  New York: MacMillan.
Greene, William H.  1992.  LIMDEP User's Guide.  New York: Econometric Software.
Hannan, Michael T. and Alice A. Young.  1977.  "Estimation in Panel Models: Resulta on Pooling Cross-sections and Time-series."  Pp. 52-83 in David R. Heise (ed.), Sociological Methodology 1977.  San Francisco: Jossey-Bass.
Hsiao, Cheng.  1986.  Analysis of Panel Data.  New York: Cambridge University Press.
Janoski, Thomas and Alexander Hicks.  1994.  The Comparative Political Economy of the Welfare State.  New York: Cambridge University Press.
Judge, George G., William E. Griffiths, R. Carter Hill, and Tsoung-Chao Lee.  1980.  The Theory and Practice of Econometrics.  New York: Wiley.
Kessler, Ronald C. and David F. Greenberg.  1981.  Linear Panel Analysis: Models of Quantitative Change.  New York: Wiley.
Markus, Gregory B.  1979.  Analyzing Panel Data.  (Sage University Paper series on Quantitative Applications in the Social Sciences, 07-018).  Beverly Hills, CA: Sage.
Menard, Scott.  1991.  Longitudinal Research.  (Sage University Paper series on Quantitative Applications in the Social Sciences, 07-076).  Beverly Hills, CA: Sage.
Mundlak, Y.  1978.  "On the Pooling of Time Series and Cross Section Data."  Econometrica 46:69-85.
Nielsen, François.  1980.  "The Flemish Movement in Belgium after World War II: A Dynamic Analysis."  American Sociological Review 45:76-94.
Nielsen, François.  1986.  "Structural Conduciveness and Ethnic Mobilization: The Flemish Movement in Belgium."  Pp. 173-198 in Susan Olzak and Joane Nagel (eds.), Competitive Ethnic Relations.  New York: Academic Press.
Nielsen, François and Michael T. Hannan.  1977.  "The Expansion of National Educational Systems: Tests of a Population Ecology Model."  American Sociological Review 42:479-90.
Nielsen, François and Arthur S. Alderson.  1995.  "Income Inequality, Development, and Dualism: Results from an Unbalanced Cross-National Panel."  American Sociological Review 60:674-701.
Pampel, Fred and J. Williamson.  1988.  "Welfare Spending in Advanced Industrial Democracies, 1950-1980."  American Journal of Sociology 93:1424-56.
Rosenfeld, Rachel A. and François Nielsen.  1984.  "Inequality and Careers: A Dynamic Model of Socioeconomic Achievement."  Sociological Methods and Research 12:279-321.
Sayrs, Lois W.  1989.  Pooled Time Series Analysis.  (Sage University Paper series on Quantitative Applications in the Social Sciences, 07-070).  Beverly Hills, CA: Sage.
SAS Institute Inc.  1988.  SAS/ETS User's Guide.  (Version 6.  First Edition.)  Cary, NC: SAS Institute Inc.
Tuma, Nancy Brandon and Michael T. Hannan.  1984.  Social Dynamics: Models & Methods.  New York: Academic Press.

Last modified 24 March 1999