Dynamic Ordinary Least Squares (DOLS) Estimator
Dynamic Ordinary Least Squares Estimator · Also known as: DOLS, Stock-Watson dynamic OLS, dynamic least squares cointegration estimator, Dinamik OLS (DOLS)
Dynamic OLS is a cointegrating-regression estimator introduced by Stock and Watson (1993) that recovers the long-run relationship between I(1) variables. It augments the static regression with leads and lags of the differenced regressors, correcting endogeneity bias parametrically so that the long-run coefficient can be estimated by ordinary least squares.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use DOLS to estimate a long-run cointegrating relationship when your variables are integrated of order one, I(1), and cointegrated, in either a single time series or a panel. It needs a reasonable span of data (about 50 observations or more) and the number of leads and lags should be selected with BIC or AIC. For panels it accommodates heterogeneous panel structure (Kao & Chiang, 2001). It is preferred over fully modified OLS in smaller samples, where it typically shows lower bias.
Strengths & limitations
- Corrects endogeneity and serial-correlation bias parametrically, simply by adding leads and lags of the differenced regressors.
- Yields a consistent, asymptotically efficient estimate of the long-run cointegrating coefficient.
- Shows better finite-sample performance than fully modified OLS (FMOLS), with generally lower bias in smaller samples.
- Extends naturally to panels, allowing for heterogeneous panel structure (Kao & Chiang, 2001).
- Requires the variables to be I(1) and genuinely cointegrated; it is not valid otherwise.
- Results depend on the chosen number of leads and lags, which must be selected with an information criterion (BIC/AIC).
- Adding leads and lags consumes degrees of freedom, so it needs a reasonably long sample (about 50 observations or more).
- Mature software support is limited in Python; dedicated econometric packages are usually needed.
Frequently asked
How does DOLS differ from ordinary OLS?
Plain OLS on cointegrated I(1) levels gives biased long-run coefficients because the regressors are correlated with the error. DOLS augments the regression with leads and lags of the differenced regressors, which absorb the short-run dynamics and yield a consistent, efficient estimate of the long-run relationship.
How do I choose the number of leads and lags?
Select the lead and lag order with an information criterion such as BIC or AIC. The choice matters: too few terms fail to remove the endogeneity bias, while too many waste degrees of freedom in a finite sample.
When should I prefer DOLS over FMOLS?
DOLS and FMOLS both estimate cointegrating vectors, but DOLS generally has better finite-sample performance and lower bias in smaller samples, which is its main advantage over fully modified OLS.
Can DOLS be used with panel data?
Yes. Kao and Chiang (2001) extended DOLS to cointegrated panel regressions, where it accommodates heterogeneous panel structure across units.
Sources
- Stock, J. H. & Watson, M. W. (1993). A Simple Estimator of Cointegrating Vectors in Higher Order Integrated Systems. Econometrica, 61(4), 783–820. DOI: 10.2307/2951763 ↗
- Kao, C. & Chiang, M.-H. (2001). On the Estimation and Inference of a Cointegrated Regression in Panel Data. Advances in Econometrics, 15, 179–222. DOI: 10.1016/S0731-9053(00)15007-8 ↗
How to cite this page
ScholarGate. (2026, June 1). Dynamic Ordinary Least Squares Estimator. ScholarGate. https://scholargate.app/en/econometrics/dols-estimator
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Augmented Mean Group EstimatorEconometrics↔ compare
- CCEMG EstimatorEconometrics↔ compare
- OLS RegressionEconometrics↔ compare
- Panel Cointegration TestsEconometrics↔ compare
- Panel Fixed EffectsEconometrics↔ compare