Skip to main content

Questions tagged [robust-standard-error]

Use this tag for questions related to any kind of robust standard error estimation, including but not limited to clusters-robust, heteroscedasticity/autocorrelation-robust, and related standard errors.

99 questions with no upvoted or accepted answers
Filter by
Sorted by
Tagged with
9 votes
0 answers
1k views

What GEE-exchangeable method can do that robust variance can't?

I asked a related question before here on the difference between GEE method with exchangeable varcov structure v. Robust standard errors known as Huber White method in group randomized trials. As ...
Sam's user avatar
  • 2,184
6 votes
0 answers
533 views

Robust standard errors under overlapping observations: confusing simulation results from alternative methods

I have a time series of $h$-step-ahead forecasts $\hat{y}_{t+h|t}$ for $t=1,\dots,T-h$ with $h>1$. I also have the corresponding realized values of their targets $y_{t+h}$ and the corresponding ...
Richard Hardy's user avatar
5 votes
0 answers
1k views

HAC standard errors: small-sample correction

The Python package statsmodels provides a use_correction option when computing HAC standard errors for an OLS model, which ...
Anthony's user avatar
  • 530
4 votes
0 answers
637 views

Correcting for robust/clustered standard errors within the lm function or replacing the results

Cross posted on Stackoverflow with a bounty of 200. EDIT: I think I have to clarify this question a little bit more. So what I am looking for, is a function in which I can provide both the vcov matrix ...
Tom's user avatar
  • 528
4 votes
0 answers
918 views

Robust regression for autocorrelation and heteroskedasticity - coefficients do not change, only standard errors change?

When using Newey-West robust standard errors to deal with heteroskedasticity and autocorrelation: http://support.sas.com/kb/40/098.html is it correct to state that the coefficients are not different ...
adrCoder's user avatar
  • 483
4 votes
0 answers
937 views

Correcting for heteroscedasticity in logistic regression

I am using a large health dataset as a part of a research project (N = ~18 000). My colleagues and I are investigating whether smoking predicts the presence or absence of a mental illness. We are ...
user80000's user avatar
3 votes
0 answers
93 views

Clustered vs. GMM-based standard errors: which ones to use in asset pricing?

This question was posted on Quantitative Finance Stack Exchange a while ago. While it was received positively there and generated a reasonable amount of views, no answers have been posted. Thus I am ...
Richard Hardy's user avatar
3 votes
0 answers
844 views

Why do heteroscedasticity-robust standard errors in logistic regression?

I am following a course on R. At the moment, we are working with logistic regression. The basic form we are taught is this one: ...
SnupSnurre's user avatar
3 votes
0 answers
302 views

Are HAC robust standard errors robust against autoregressive conditional heteroskedasticity?

Suppose I have a GARCH(p,q) model with constant conditional mean, \begin{aligned} y_t &= \mu + u_t, \\ u_t &= \sigma_t \varepsilon_t, \\ \sigma_t^2 &= \omega + \alpha_1 u_{t-1}^2 + \dotsc +...
Richard Hardy's user avatar
3 votes
0 answers
763 views

Should instruments always weakly increase standard errors?

In OLS, are standard error estimates using some instrument $z$, different from $x$ always weakly larger than standard errors when not using instruments? This has been discussed before: Why is the ...
Felix H's user avatar
  • 145
3 votes
0 answers
647 views

Robust standard errors for a time series regression

Suppose we have time series data $(p_t, (x_{i,t})_{i = 1...n})$, where $p_t$ is a price and $x_{i,t} \in \mathbb{R}^n$ regressors. From $p_t$ we construct $r_t = \frac{p_{t+30}}{p_t} -1$, the one-...
Michael's user avatar
  • 3,328
3 votes
0 answers
946 views

Double clustered standard errors and degrees of freedom in Wald-style F test for joint significance of regression coefficients with panel data

In panel regression, the Wald-style F test for joint significance of the regression coefficients is usually done with an adjustment for the degrees of freedom when robust/clustered standard errors are ...
Helix123's user avatar
  • 1,467
3 votes
0 answers
796 views

How to calculate robust standard error with offset?

I think the default vcovHC in R's sandwich package does not handle offsets in Poisson models. We see this because robust (...
AdamO's user avatar
  • 63.2k
2 votes
1 answer
37 views

Maximum likelihood estimation with (robust) Huber-White standard errors appropriate for outlier management?

Is maximum likelihood estimation with robust Huber-White standard errors and a scaled test statistic — which is asymptotically equal to the Yuan-Bentler test statistic — appropriate for data with ...
Madamadam's user avatar
  • 247
2 votes
1 answer
491 views

How to correct for serial correlation in an ARDL model without increasing lags?

I'm currently looking to run an ARDL model - I'm able to compute results that show cointegration, however there is serial correlation when I run the Durbin-Watson and Breusch-Godfrey tests. To correct ...
woodvalestar's user avatar
2 votes
0 answers
201 views

How can I estimate robust standard errors for robust estimators? R

I am doing a robust regression and I want to estimate robust standard errors for my regression. I do know how to do a robust regression in R and how to estimate a robust standard errors for a ...
Cristina Mura's user avatar
2 votes
0 answers
63 views

Cluster robust standard error

$$\bf{y}=\bf{X\beta}+\bf{\epsilon}$$ In asymptotic setting, assume: (a) linearity (b) $\{y_i,\bf{x}_i\}$ ergodic stationarity (c) weak exogeneity $E(\bf{x}_i\epsilon_i)=0$ (d) Rank condition $E(\bf{x}...
jasmine's user avatar
  • 367
2 votes
0 answers
222 views

difference between standard error and robust standard error

What is the difference between standard error vs robust standard error? How is the robust standard error calculated in the Generalized Estimating Equation context?
Mia's user avatar
  • 21
2 votes
0 answers
822 views

Why do robust linear models give smaller standard errors?

I've always read and been told that for heteroskedastic errors a normal OLS fit will generate standard errors that are too small, leading to a false degree of confidence in coefficient estimates. ...
Teleopsis's user avatar
2 votes
0 answers
158 views

Relationship between expected absolute deviation and standard deviation

The standard deviation of a random variable $X$, $$\sigma = \sqrt{Var[{X}]} = \sqrt{E[X^2 - E[X]^2]}$$ is a very commonly used measure of "normal" distance from the mean of $X$. A much less frequently ...
Oddsee's user avatar
  • 71
2 votes
0 answers
126 views

Standard error clustering: Does it depend on the variable of interest?

Suppose I am interested in whether a students age affects their performance, and I run the following regression: $performance(i) = \beta_1 age(i) + \beta_2 Female(i)+\beta_3 classsize(i)+\varepsilon(...
safex's user avatar
  • 171
2 votes
0 answers
86 views

Is it a problem if cluster sizes vary wildly in a cluster-randomized experiment?

Consider a cluster-randomized experiment. There are 4,000 clusters and 2,000,000 observations. The dependent variable y is dichotomous, $Y \in \{0, 1\}$, and ...
Mark White's user avatar
  • 10.5k
2 votes
0 answers
36 views

OLS question about standard errors and Heteroscedasticity

In the case of OLS, what stops one from modelling heteroscedasticity and providing predictor dependent coefficients. So, different ranges of the predictor with their own standard errors? Is this ...
Forevertrip's user avatar
2 votes
0 answers
171 views

Calculating sandwich estimator

Considering design matrix $X \in \mathbb{R}^{n\times p}$ $(n>p)$ and response $y\in \mathbb{R}^{n}$. The sandwich estimator can be calculated directly using $$(X^TX)^{-1}X^T diag(r^2) X (X^TX)^{-...
cccfran's user avatar
  • 75
2 votes
0 answers
2k views

Is delta method better than bootstrap to generate standard error for marginal effects?

I read here, here, here, here, and elsewhere that " Parametric bootstrap closely related to objective Bayes. (That’s why it’s a good importance sampling choice.) When it applies, parboot approach ...
Krantz's user avatar
  • 585
2 votes
0 answers
818 views

Explanation and formula for Modified Z-score?

I'm struggling understanding and finding a reference for Modified (robust) z-score. I found a reference that says I have to adjust by 1.25 or 1.48 depending on the value of the median absolute ...
Jose G's user avatar
  • 468
2 votes
0 answers
5k views

What is Logistic regression with clustered standard errors?

I am learning mixed effects logistic regression from this link. In section "Analysis methods you might consider", the author listed several options: Mixed effects logistic regression, the focus ...
Haitao Du's user avatar
  • 37k
2 votes
0 answers
827 views

Multinomial tests of heterogeneity using robust standard errors?

I am looking at differences in frequencies of categorical variables collected at different sites, adjusting for stratification factors of age and sex. I would like to use robust standard error ...
AdamO's user avatar
  • 63.2k
2 votes
0 answers
155 views

Robust quadratic form loss

I am using a loss function in quadratic form: $$ l(w) = w^TAw \\ s.t. w^TBw=1 $$ where both $A$ and $B$ are symmetric positive definite. $A$ and $B$ are between-class scatter matrix and within-class ...
MarsPlus's user avatar
  • 151
2 votes
0 answers
171 views

Measuring robustness of network constructed with python mapper

I am trying to visualize a large multidimensional data set with the help of the Python Mapper (open source software package using the Mapper-Algorithm, a method of Topological Data Analysis). http:/...
manz's user avatar
  • 21
2 votes
0 answers
839 views

Comparing Classical and Robust (Huber-White/sandwich/heteroscedasticity consistent) Standard Errors in Linear Multiple Regression

I'm running a linear multiple regression model of the type $y_i = \beta_0 + \beta_1 X_{i1} + \beta_2 X_{i2} + \beta_3 X_{i3} + u_i$. I came across King and Roberts' 2015 paper called "How Robust ...
Sannita's user avatar
  • 271
1 vote
0 answers
17 views

Large $N$, small $T$ in SUR: workaround using system GMM

Consider a system of linear equations as in seemingly unrelated regression (SUR). If the number of equations $N$ is large relative to the sample size $T$, the weighting matrix in SUR (i.e. the error ...
Richard Hardy's user avatar
1 vote
0 answers
184 views

How to compute robust standard errors for a linear mixed effects model with two non-nested random effects?

I asked this question earlier this week but I hadn't really done my research yet or inquired into my error. So my data is heteroscedastic which has necessitated computation of robust standard errors ...
Shannon Cahalan's user avatar
1 vote
0 answers
56 views

Robust Standard Errors as Remedy for Violation of Assumptions in Multi-Level Model

So I ran a multi-level model using the nlme/lme4 R packages. Testing the assumptions, I found that level 1 as well as level 2 residuals are not normally distributed, also there's heteroscedasticity on ...
Ben's user avatar
  • 23
1 vote
0 answers
34 views

Why are FGLS standard errors so low when estimating linear regression on panel data?

I am currently writing my master thesis an the effect of academic publication on anomaly returns. The idea is that in finance literature anomaly - factors/portfolios are proposed. After the factor is ...
CHRISS's user avatar
  • 11
1 vote
0 answers
117 views

How to conduct a power analysis for a linear regression with clustered standard errors?

I would like to conduct a posthoc power analysis for a linear regression model to see the power of my analysis given the effect size I am finding. However, due to the clustering of standard errors, I ...
Chonasi's user avatar
  • 11
1 vote
0 answers
25 views

estimating variance when parameters are estimated

I am trying to understand the sandwich variance estimator. Mainly my confusion is why it is OK to use estimated parameters. I think it boils down to this question. If $f(X_i; \theta),i=1,\ldots,n$ are ...
kimla.2's user avatar
  • 21
1 vote
0 answers
87 views

When using both Sattertwaite degrees of freedom adjustment AND sandwich estimator of the covariance, how do they interact?

In longitudinal experimental trials, where typically (relatively) small sample number of participants are examined, two solutions are proposed to address two kinds of issues: robust estimator of the ...
Kaloobin's user avatar
1 vote
0 answers
90 views

HAC estimator in GARCH models

In my time-series class we learned that the HAC estimator is only applicable to correct the standard error (SE) if the underlying series is stationary. Now, GARCH models are unconditionally stationary ...
DLTS's user avatar
  • 63
1 vote
0 answers
68 views

For repeated data, why we don't use just OLS with sandwich SEs but rather GLS or mixed models?

If I have repeated observations and want to summarize the means at each time point, the OLS will give me the true, raw means, while the GLS will give me means dependent on the selected covariance ...
Blitzkordk's user avatar
1 vote
0 answers
667 views

Robust standard errors for mixed effects models in R

I understand that you cannot fit a glmer(via "lme4") then test using robust SE estimator. Ie ...
jean10's user avatar
  • 11
1 vote
0 answers
196 views

Correcting for Heteroscedasticity in multiple imputed datasets

I have a question regarding the homogeneity of variance in three regression models of diffrent datasets belonging to the same multiple imputed data. As I used multiple imputation I have to check ...
Ecidem's user avatar
  • 11
1 vote
0 answers
74 views

Test statistic for regression analysis with robust standard errors

I am looking for the test statistic for a regression analysis (using lm in R) with robust standard errors. As far as I understood the F-statistic from the original model is invalid but I was wondering ...
mafrikone's user avatar
1 vote
0 answers
27 views

Heteroscedasticity-consistent standard error is not the same as OLS standard error

Problem I perform the following regression $$ y_i = a + b x_i +\epsilon, \quad \quad i = 1, ..., 21$$ I managed to estimate $\hat a = −0.59$ and $\hat b = 0.069$ with standard error $s.e.(\hat b) = 0....
Bruh's user avatar
  • 27
1 vote
0 answers
26 views

Standard error for panel data models and heteroscedastic robust SE

There is something I do not understand on a deeper level on standard errors for panel data models. I know the reason for not using usual heteroscedastic-robust standard-errors is because of auto ...
Daniel Ortiz's user avatar
1 vote
0 answers
158 views

How to manually inflate standard errors to approximate clustered SEs

I'm reading a handout on clustering here It's not clear to me how to compute $\rho_x$ or $\rho_\epsilon$. What is meant by within-cluster correlation of the regression, or within-cluster error ...
Parseltongue's user avatar
  • 1,010
1 vote
0 answers
390 views

How to decide between different robust standard errors?

Specifying my model I ran into some very mild heteroscedasticity problems. Given its superior small-sample properties (my dataset contains 79 observations) I used the HC3 specification of the White ...
philipp.kn_98's user avatar
1 vote
0 answers
17 views

In regression modeling, are there any caveats to always using robust standard errors?

Aside from efficiency issues, is there anything else to this?
user321627's user avatar
  • 4,386
1 vote
0 answers
86 views

Why different regression result with real and excess return as the dependent variable?

I want to test how granted patent, R&D & their ratio predicts real & excess market return. I am doing 2 robust linear regressions: \begin{align} \text{Real Return} \left( t + 1 \right) =...
Lamya's user avatar
  • 11
1 vote
0 answers
52 views

Pooled panel regression with group-wise clustering by time

How can I compute t-statistics for the coefficients of the pooled panel regression model below such that I account for group-wise clustering by time? $$ Y_{it} = \alpha + \beta x_{it} + \epsilon_{it} ...
Jack's user avatar
  • 305