Local average treatment effect
In econometrics and related empirical fields, the local average treatment effect (LATE), also known as the complier average causal effect (CACE), is the effect of a treatment for subjects who comply with the experimental treatment assigned…
≈ 14 min read
Names in other languages
No names in other languages are held for this entry yet.
Article bodies are English and are not machine-translated. A name in another language is not a translation of the article.
In econometrics and related empirical fields, the local average treatment effect (LATE), also known as the complier average causal effect (CACE), is the effect of a treatment for subjects who comply with the experimental treatment assigned to their sample group. It is not to be confused with the average treatment effect (ATE), which includes compliers and non-compliers together.
Compliance refers to the human-subject response to a proposed experimental treatment condition. Similar to the ATE, the LATE is calculated but does not include non-compliant parties. If the goal is to evaluate the effect of a treatment in ideal, compliant subjects, the LATE value will give a more precise estimate.
However, it may lack external validity by ignoring the effect of non-compliance that is likely to occur in the real-world deployment of a treatment method. The LATE can be estimated by a ratio of the estimated intent-to-treat effect and the estimated proportion of compliers, or alternatively through an instrumental variable estimator.
The LATE was first introduced in the econometrics literature by Guido W. Imbens and Joshua D. Angrist in 1994, who shared one half of the 2021 Nobel Memorial Prize in Economic Sciences.…
In econometrics and related empirical fields, the local average treatment effect (LATE), also known as the complier average causal effect (CACE), is the effect of a treatment for subjects who comply with the experimental treatment assigned to their sample group. It is not to be confused with the average treatment effect (ATE), which includes compliers and non-compliers together.
Compliance refers to the human-subject response to a proposed experimental treatment condition. Similar to the ATE, the LATE is calculated but does not include non-compliant parties. If the goal is to evaluate the effect of a treatment in ideal, compliant subjects, the LATE value will give a more precise estimate.
However, it may lack external validity by ignoring the effect of non-compliance that is likely to occur in the real-world deployment of a treatment method. The LATE can be estimated by a ratio of the estimated intent-to-treat effect and the estimated proportion of compliers, or alternatively through an instrumental variable estimator.
The LATE was first introduced in the econometrics literature by Guido W. Imbens and Joshua D. Angrist in 1994, who shared one half of the 2021 Nobel Memorial Prize in Economic Sciences. As summarized by the Nobel Committee, the LATE framework "significantly altered how researchers approach empirical questions using data generated from either natural experiments or randomized experiments with incomplete compliance to the assigned treatment.
At the core, the LATE interpretation clarifies what can and cannot be learned from such experiments." The phenomenon of non-compliant subjects (patients) is also known in medical research. In the biostatistics literature, Baker and Lindeman (1994) independently developed the LATE method for a binary outcome with the paired availability design and the key monotonicity assumption.
Baker, Kramer, Lindeman (20161) summarized the history of its development. Various papers called both Imbens and Angrist (1994) and Baker and Lindeman (1994) seminal. An early version of LATE involved one-sided noncompliance (and hence no monotonicity assumption).
In 1983 Baker wrote a technical report describing LATE for one-sided noncompliance that was published in 2016 in a supplement. In 1984, Bloom published a paper on LATE with one-sided compliance. For a history of multiple discoveries involving LATE see Baker and Lindeman (2024).
The typical terminology of the Rubin causal model is used to measure the LATE, with units indexed i = 1, \ldots, N and a binary treatment indicator, z_i for unit i . The term Y_i(z_i) is used to denote the potential outcome of unit i under treatment z_i . In an ideal experiment, all subjects assigned to the treatment will comply with the treatment, while those that are assigned to control will remain untreated. In reality, however, the compliance rate is often imperfect, which prevents researchers from identifying the ATE.
In such cases, estimating the LATE becomes the more feasible option. The LATE is the average treatment effect among a specific subset of the subjects, who in this case would be the compliers.
ATE= E[Y_i(1)-Y_i(0)]=E[Y_i(1)]-E[Y_i(0)]=E[Y_i(1)|Z_i=1]-E[Y_i(0)|Z_i=0] Compliers are subjects who will take the treatment if and only if they were assigned to the treatment group, i.e., the subpopulation with d_i(1)=1 and d_i(0)=0 . Non-compliers are composed of the three remaining subgroups: Always-takers are subjects who will always take the treatment even if they were assigned to the control group, i.e., the subpopulation with d_i(z)=1 Never-takers are subjects who will never take the treatment even if they were assigned to the treatment group, i.e., the subpopulation with d_i(z)=0 Defiers are subjects who will do the opposite of their treatment assignment status, i.e., the subpopulation with d_i(1)=0 and d_i(0)=1 Non-compliance can take two forms: one-sided (always-takers and never-takers) and two-sided (defiers).
In the case of one-sided non-compliance, a number of the subjects who were assigned to the treatment group remain untreated. Subjects are thus divided into compliers and never-takers, such that d_i(0) = 0 for all i , while d_i(1)= 0 or 1 .
In the case of two-sided non-compliance, a number of the subjects assigned to the treatment group fail to receive the treatment, while a number of the subjects assigned to the control group receive the treatment. In this case, subjects are divided into the four subgroups, such that both d_i(0) and d_i(1) can be 0 or 1.
Given non-compliance, certain assumptions are required to estimate the LATE. Under one-sided non-compliance, non-interference and excludability is assumed. Under two-sided non-compliance, non-interference, excludability, and monotonicity is assumed.
The first part of this assumption stipulates that the actual treatment status, d_i , of subject i depends only on the subject's own treatment assignment status, z_i . The treatment assignment status of other subjects will not affect the treatment status of subject i . Formally, if z_i=z_i' , then D_i(\mathbf{z})=D_i(\mathbf{z}') , where \mathbf{z} denotes the vector of treatment assignment status for all individuals. The second part of this assumption stipulates that subject i 's potential outcomes are affected by its own treatment assignment, and the treatment it receives as a consequence of that assignment.
The treatment assignment and treatment status of other subjects will not affect subject i 's outcomes. Formally, if z_i=z_i' and d_i=d_i' , then Y_i(z,d)=Y_i(z',d) . The plausibility of the non-interference assumption must be assessed on a case-by-case basis.
The excludability assumption requires that potential outcomes respond to treatment itself, d_i , not treatment assignment, z_i . Formally Y_i(z,d)=Y_i(d) . So under this assumption, only d matters. The plausibility of the excludability assumption must also be assessed on a case-by-case basis.
The LATE = \frac{ITT}{ITT_D} , whereby ITT=E[Y_i(z=1)]-E[Y_i(z=0)] ITT_D = E[d_i(z=1)]-E[d_i(z=0)] The ITT measures the average effect of experimental assignment on outcomes without accounting for the proportion of the group that was actually treated (i.e., an average of those assigned to treatment minus the average of those assigned to control). In experiments with full compliance, the ITT = ATE .
The ITT_D measures the proportion of subjects who are treated when they are assigned to the treatment group, minus the proportion who would have been treated even if they had been assigned to the control group, i.e., ITT_D = the share of compliers. Under one-sided noncompliance, all subjects assigned to control group will not take the treatment, therefore: E[d_i(z=0)]=0 , so that ITT_D = E[d_i(z=1)]=P[d_i(1)=1] If all subjects were assigned to treatment, the expected potential outcomes would be a weighted average of the treated potential outcomes among compliers, and the untreated potential outcomes among never-takers, such that \begin{align} {\displaystyle E[Y_{i}(z=1)]=E[Y_{i}(d(1),z=1)]} = E[Y_i(z=1,d=1)|d_i(1)=1]*P[d_i(1)=1] & \\ +E[Y_i(z=1,d=0)|d_i(1)=0]* (1-P[d_i(1)=1]) \end{align} If all subjects were assigned to control, however, the expected potential outcomes would be a weighted average of the untreated potential outcomes among compliers and never-takers, such that \begin{align} {\displaystyle E[Y_{i}(z=0)]=E[Y_{i}(d=0,z=0)]} = E[Y_i(z=0,d=0)|d_i(1)=1]*P[d_i(1)=1] & \\ +E[Y_i(z=0,d=0)|d_i(1)=0]* (1-P[d_i(1)=1]) \end{align} Through substitution, the ITT is expressed as a weighted average of the ITT among the two subpopulations (compliers and never-takers), such that \begin{alignat}{2} ITT= E[Y_i(z=1)]-E[Y_i(z=0)] = E[Y_i(z=1,d=1)-Y_i(z=0,d=0)|d_i(1)=1]*P[d_i(1)=1]+&\\ E[Y_i(z=1,d=0)-Y_i(z=0,d=0)|d_i(1)=0]*P[d_i(1)=0] \end{alignat} Given the exclusion and monotonicity assumption, the second half of this equation should be zero.
As such, \begin{align} \frac{ITT}{ITT_D}= & \frac {E[Y_i(z=1,d=1)-Y_i(z=0,d=0)|d_i(1)=1]*P[d_i(1)=1]}{P[d_i(1)=1]} \\ = & E[Y_i(d=1)-Y_i(d=0)|d_i(1)=1] \\ = & LATE \end{align} The ATE is calculated by the average of Y_i(d=1)- Y_i(d=0) ATE= \frac{3+2+4+3+6+6+4+4+3}{9}=\frac{35}{9}=3.9 LATE is calculated by ATE among compliers, so LATE = \frac{3+4+6+4+4}{5}=4.2 ITT is calculated by the average of Y_i(z=1)-Y_i(z=0) , so ITT = \frac{3+0+4+0+6+0+4+4+0}{9}=\frac{21}{9}=2.3 ITT_D is the share of compliers ITT_D = \frac{5}{9} \frac{ITT}{ITT_D}= \frac{21/9}{5/9}=\frac{21}{5}=4.2=LATE
Others: LATE in instrumental variable framework
LATE can be thought of through an IV framework. Treatment assignment z_i is the instrument that drives the causal effect on outcome Y_i through the variable of interest d_i , such that z_i only influences Y_i through the endogenous variable d_i , and through no other path. This would produce the treatment effect for compliers.
In addition to the potential outcomes framework mentioned above, LATE can also be estimated through the Structural Equation Modeling (SEM) framework, originally developed for econometric applications. SEM is derived through the following equations: D_i = \alpha_0 + \alpha_1 Z_i + \xi_{1i} Y_i = \beta_0 + \beta_1 Z_i + \xi_{2i} The first equation captures the first stage effect of z_i on d_i , adjusting for variance, where \alpha_1=Cov(D,Z)/var(Z) The second equation \beta_1 captures the reduced form effect of z_i on Y_i , \beta_1=Cov(Y,Z)/var(Z) The covariate adjusted IV estimator is the ratio \tau_{LATE}=\frac{\beta_1}{\alpha_1}=\frac{Cov(Y,Z)/Var(Z)}{Cov(D,Z)/Var(Z) } = \frac{Cov(Y,Z)}{Cov(D,Z)} Similar to the nonzero compliance assumption, the coefficient \alpha_1 in first stage regression needs to be significant to make z a valid instrument.
However, because of SEM’s strict assumption of constant effect on every individual, the potential outcomes framework is in more prevalent use today. Although one's compliance score cannot be directly observed, the probability of compliance can be estimated by observing the compliance condition from the same strata, in other words those that share the same covariate profile.
The compliance score is treated as a latent pretreatment covariate, which is independent of treatment assignment Z . For each unit i , compliance score is denoted as P_{Ci}=Pr(D_1>D_0|X=x_i) , where x_i is the covariate vector for unit i . In one-sided noncompliance case, the population consists of only compliers and never-takers. All units assigned to the treatment group that take the treatment will be compliers.
Thus, a simple bivariate regression of D on X can predict the probability of compliance. In two-sided noncompliance case, compliance score is estimated using maximum likelihood estimation.
By assuming probit distribution for compliance and of Bernoulli distribution of D, where \hat{\Pr{c_i}}=\hat{\Pr}(D_1>D_0|X=x_i)=F(\hat{\theta}_{A,C,x_i})(1-F(\hat{\theta}_{A|A,C,x_i}))^3 . and \theta is a vector of covariates to be estimated, F(.) is the cumulative distribution function for a probit model ICSW estimator By the LATE theorem, average treatment effect for compliers can be estimated with equation: \tau_{LATE}=\frac{\sum_{i=1}^n {Z_i}{Y_i}/\sum_{i=1}^n {Z_i}-\sum_{i=1}^n {(1-Z_i)}{Y_i}/\sum_{i=1}^n {(1-Z_i)}}{ \sum_{i=1}^n {Z_i}{D_i}/\sum_{i=1}^n {Z_i}-\sum_{i=1}^n {(1-Z_i)}{D_i}/\sum_{i=1}^n {(1-Z_i)}} Define \hat{w_{Ci}}=1/\hat{Pr_{Ci}} the ICSW estimator is simply weighted by: \tau_{ATE}=\frac{\sum_{i=1}^n \hat{W_i}{Z_i}{Y_i}/\sum_{i=1}^n \hat{W_i}{Z_i}-\sum_{i=1}^n \hat{W_i}{(1-Z_i)}{Y_i}/\sum_{i=1}^n {\hat{W_i}(1-Z_i)}}{ \sum_{i=1}^n \hat{W_i}{Z_i}{D_i}/\sum_{i=1}^n \hat{W_i}{Z_i}-\sum_{i=1}^n \hat{W_i}{(1-Z_i)}{D_i}/\sum_{i=1}^n \hat{W_i}{(1-Z_i)}} This estimator is equivalent to using 2SLS estimator with weight . Core assumptions under reweighting An essential assumption of ICSW relying on treatment homogeneity within strata, which means the treatment effect should on average be the same for everyone in the strata, not just for the compliers.
If this assumption holds, LATE is equal to ATE within some covariate profile. Denote as: \text{for all }x \in Supp(X), E[Y_1-Y_0|D_1>D_0] Notice this is a less restrictive assumption than the traditional Ignorability assumption, as this only concerns the covariate sets that are relevant to compliance score, which further leads to heterogeneity, without considering all sets of covariates.
The second assumption is consistency of \hat{Pr_{Ci}} for Pr_{Ci} and the third assumption is the nonzero compliance for each strata, which is an extension of IV assumption of nonzero compliance over population. This is a reasonable assumption as if compliance score is zero for certain strata, the inverse of it would be infinite.
ICSW estimator is more sensible than that of IV estimator, as it incorporate more covariate information, such that the estimator might have higher variances. This is a general problem for IPW-style estimation. The problem is exaggerated when there is only a small population in certain strata and compliance rate is low.
One way to compromise it to winsorize the estimates, in this paper they set the threshold as =0.275. If compliance score for lower than 0.275, it is replaced by this value. Bootstrap is also recommended in the entire process to reduce uncertainty(Abadie 2002).
Text derived from https://en.wikipedia.org/w/index.php?oldid=1368691707 under CC BY-SA 4.0; modifications have been made.
References
- ↩Wikipedia “Local average treatment effect” publication date (×1) · confidence 0.95 revision URL