Ivreg2 last estimates not found

 

the dataset you used to produce the margins results. Here is an Christopher F Baum & Vince Wiggins & Steven Stillman & Mark E Schaffer & Frank Windmeijer, 1999. As the default for > ivreg2 > is to produce large sample estimates, your test results and standard > errors > will not be affected by the transformation. Additional features include: Andrea, -ivendog- isn't equipped (yet?) to handle -xtivreg2-. I have a panel with the following structure: N=6 T=53 (monthly data). I think the "last estimates not found" error I'm getting traces to a reghdfe, requirements call. Sep 19, 2015 · ivreg2 with partial option. Here is a work-around using Clint's example: . 1. In any case, when you have such interactions, you' may want to use margins to examine the influence of dy/x2 (which is b2 + b3*x1) at different values of x1. reghdfe is a generalization of areg (and xtreg,fe, xtivreg,fe) for multiple levels of fixed effects (including heterogeneous slopes), alternative estimators (2sls, gmm2s, liml), and additional robust standard errors (multi-way clustering, HAC standard errors, etc). It provides additional features, such as testing for endogeneity, weak instruments, and overidentification. Re: st: overid after ivregress. 2. 非常感谢，我有点忘了，稍微再看一下该指令说明， 再对应楼主是问"面板"情况，您说的应该是对的 (我当初可能没注意楼主问的是面板资料情况)。. corr. 0. " OVERID: Stata module to conduct postestimation tests of overidentification ," Statistical Software Components S396902, Boston College Department of Economics, revised 29 Sep 2020. May 24, 2023 · You are not logged in. *! actest9 2. Stata 9 Andrea, -ivendog- isn't equipped (yet?) to handle -xtivreg2-. version (#) reghdfe has had so far two large rewrites, from version 3 to 4, and version 5 to version 6. Schaffer and Steven Stillman}, journal={Statistical Software Components}, year={2006 Jun 13, 2022 · Estimates efficient for homoskedasticity only Statistics robust to heteroskedasticity and clustering on ts and kernel-robust to common correlated disturbances (Driscoll-Kraay) kernel=Bartlett; bandwidth=1 time variable (t): ts group variable (i): id Number of clusters (ts) = 40 Number of obs = 578 F( 88, 39) = 1. residuals. More information below: New versions of and extensions to the Baum-Schaffer-Stillman packages. I am trying to replicate the ivreg output of a regression performing manually the first stage, predicting the instrument after the first stage and running the second stage regression with the instrument in place of the endogenous regressor in the structural model. The results should be very close to C Mar 12, 2020 · The p-value of the F-stat (Cragg-Donald or Kleibergen-Paap), is, I think, not available. Dear Mark, Thank you for your prompt answer I updated my versions of ivreg2 and ranktest as you suggested. Weak identification test (Cragg-Donald Wald F statistic): 31. Simplifying your code as suggested by @Nick Cox is best, however, to answer your question the solution is below: main do-file. what's the reason? > > Once again, many thanks. Sep 8, 2016 · Problem with ivreg2 (struct ms_vcvorthog undefined) 08 Sep 2016, 08:24. do. 8995 May 2, 2002 · But analysts’ optimism and use of less sophisticated valuation methods have been found to impede target price informativeness. Alternatively, you can explain the issue, and then discuss dy/dx2 at different values of x1. Whatever you put in -orthog ()- is what -xtivreg2- will treat as endogenous in the comparator regression. prais cinf unem we obtain different estimates and standard errors. (The standard errors are a little different across most but not all of these. We limit our exposition here to the linear case, which is what ivreg2 Sep 8, 2021 · I reran -ivregress- without the cluster option, and you were right, SEs are almost the same as that with -ivreghdfe-, which proves my suspicion that the command "vce (cluster id )" of -ivreghdfe- fail to implement what it is supposed to do. 0 Aug 12, 2016 · Or instead if someone know how to export the results of instrumental variable regression both first and second stages with ivgress 2sls or ivreg2 please help me. parameter estimates. Let’s fit a predicted curve to this plot to further investigate. Jan 19, 2015 · In this version: ivreg2 now supports factor variables, and can be used for ordinary least squares (OLS) estimation using the same command syntax as Stata's official regress and newey. Dec 16, 2014 · marginsplot looks for formats, labels, and value labels attached to the variables. ivreg2 may be used with time-series or panel data, in which case the data must be tsset before using ivreg2; see help tsset. For ftools the distribution date was Austin, Thank you for your reply! Yes, I was under that impresssion. 0 and it worked! Try to install an earlier version of reghdfe if you don't use Stata 16. I am so glad you pointed this out. 2010/4/6 Martin Weiss <[email protected]>: > > <> > > > " By the way, what's the meaning of FSRs The residuals do not appear randomly distributed. ivreg2 provides extensions to Stata's official ivregress and newey. I can save residuals from reghdfe without a problem. Jun 4, 2015 · First, the title should have read xtivreg2 instead of ivreg2. > > Phil > > > > 2010/4/6 Kit Baum <[email protected]>: > > <> > > Phil said > > > > > > I used the following ivreg2 to estimate an IV-regression, but could > > not produce the first-stage The issue was the most recent version of the reghdfe (6. ivreg2 cinf unem, bw(3) > . With -ivreg2- it runs ok! but the answer is not entirely satisfactory: > ivhettest supports ivreg, ivreg2 and regress but not ivregress (see the > help file Apr 22, 2024 · Stata Commands for Instrumental Variables. ivhettest not working after ivreg2 02 Jan 2017, 17:02. "estimation result not found" (Should have mentioned that) My instrument is a variable with 3 values (1, 2, 3). All Time Today Last Week Last Month. See log below. aweights, fweights, iweights, and pweights are allowed; see [U] 11. This seems to be related to line 2285 in the ivreghdfe. Nov 12, 2023 · ivreghdfe is crashing on me unless I revert to reghdfe version 5. 6 weight. This allows to link your profile to this item. Check out the help file examples of -ivreg2-, and you'll see an example in the list of how to mimic a Durbin-Wu-Hausman endogeneity test using this option. fitted. ) By contrast, if > we run > > . Given that T>N, I discarded FE model in favour of Zellner's seemingly unrelated regressions (sureg). net install reghdfe, from(https://raw. Specifically, I am using psacalc for this practice. 9. Both matrices by default include a units vector. It seems ivhettest does not handle factor variables. You can help adding them by using this form . a vector of predicted means. If you want to be sure, run two regressions, one with the variable as endogenous and one with it as exogenous, and do a difference-in-Sargan (or difference-in-J) by hand. ivreg cinf unem . ivreg returns an object of class "ivreg", with the following components: coefficients. reg cinf unem > > all yield exactly the same coefficient estimates. Prob > F = 0. 0000. The discrepancies vanish when I use the noconstant option. ivreg2: As a user-created command, ivreg2 extends the functionality of ivregress. perfect and coeflegend do not appear in the dialog box. Best. • Several advanced options, including options that will speed up estimation using ivreg2 by suppressing the calculation of various checks and statistics. You can browse but not post. 074 (ie. As you note, the saved first-stage estimation with ivreg2 creates an esample() variable: storage display value variable name type format label variable label ----- _est__ivreg2_ys byte %8. But you don't show us the exact command you used or the output provided by ivregress. We found no effect on special education If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. (I only can export result of first stage with ivregress2 by outreg2, I did search and read a lot on the internet but I still can't do with ivreg2 and ivregress by estout package ) I'm after Stata 13. 3. 6 4Feb2007 from SSC with Stata 13. Dec 3, 2023 · • 运行“estimates use mvmeta ” 出现 file mvmeta. ivreg2, xtivreg2, ranktest and xtoverid, and a new program Not sure what the problem is. I am trying to produce a table with outreg2 for estimates using xtivreg2 (on Stata 13. But the "savefirst“ option of ivreg2 could actually store the first stage result, so it's easy for ivreg2 to export it. *! ivreg2 4. this, just add a required argument to ajuste. Date. Writing an xtivreg2 program which does this all automatically is on my to-do list but I am not sure how soon I will have time to do it. Hope For simple status reports, set verbose to 1. It is now a mainstay of both econometric practice and econometrics textbooks. The main extensions 1. But they are very large in more complex examples. Dear fellow Stata users, I installed Stata in a new computer. I am trying to estimate the impact of opinion polls over the frequency of tweets about inequality by 6 parties. Naturally, the standard errors of my second stage regression do aweights are not allowed with the jackknife prefix; see[R] jackknife. We have no bibliographic references for this item. 不过，就如同我 May 29, 2019 · I'm running ivreghdfe with 3 fixed effects and trying to save the residuals. github install sergiocorreia/reghdfe, version ("5. Mar 6, 2024 · 发表回复. As you point out, it takes some effort or experimentation to extract this information from the - ivreg2- help file. -ivregress- is "too new" for -overid-. Nov 18, 2023 · Run IV/2SLS with many levels of fixed effects (i. Here is an example: In this example, the differences between the two sets of estimates are minimal. Here is to Cheers, Mark > Also, one mystery is that I used the same syntax in another > project last night, it did work. LM statistic): 58. it to work properly with svy bootstrap. Most time is usually spent on three steps: map_precompute (), map_solve () and the regression step. ivreg28 supports the same command syntax as official ivreg and supports (almost) all of its options. When installing ivreg2 the first time, I used find it and then installed one of the available options. r(199); Edit: the same model is estimated with no problems in Stata 14. Although user could add "savefirst" option in xtivreg2 command, and it runs without an error, it still no result is stored like "ivreg2 xxx, savefirst" does. Show. So it seems the be driven by R always using small sample corrections while Stata only using those when explicitly asked for. ivreg2+reghdfe) - Issues · sergiocorreia/ivreghdfe Nov 22, 2019 · Problem with sureg: last estimates not found. 11 22Nov2019 *! authors cfb & mes *! see end of file for version comments * Variable naming: * lhs = LHS endogenous * endo = X1, RHS endogenous (instrumented) = #K1 * inexog = X2 = Z2 = included exogenous (instruments) = #K2 = #L2 * exexog = Z1 = excluded exogenous (instruments) = #L1 * iv = {inexog exexog} = all instruments * rhs = {endo inexog} = RHS regressors * no 0 or 1 at Aug 5, 2019 · Note this doesn't change the estimates substantively. May 15, 2019 · // set up some example data sysuse auto, clear drop if missing(rep78) // run the regressions on the sample data levelsof rep78, local(rep) levelsof foreign, local(fgn) generate res = . I The newest version of ivreg2 incorporates additional code to compute eigenvalues of G T before reporting other estimates. I've tried uninstalling everything (ftools, reghdfe, ivreghdfe) and reinstalling from github to no avail. bvid1 to create a dummy for each groups, it is not very feasible. The standard regression functionality (parameter estimation, inference, robust covariances, predictions, etc. 1 and 2 are treatments, 3 is control. 0000 Total (centered) SS = 2033. foreach r of local rep { foreach f of local fgn { capture regress weight length if rep78==`r' & foreign==`f' if _rc != 0 display "FAILURE rep78 `r' foreign `f' return code " _rc else { quietly predict temp May 31, 2021 · The ivreg package (by John Fox , Christian Kleiber, and Achim Zeileis) provides a comprehensive implementation of instrumental variables regression using two-stage least-squares (2SLS) estimation. Something for my (unfeasibly long) to-do list. 08. These are reported by the Stata program ivreg2 when the ffirst option is specified, and include the partial R2 and first-stage F-statistics on excluded instruments. Oct 14, 2019 · Thank you. If m is the number of lags you would specify in -newey- or - newey2-, then use m for the -ivregress, vce (hac nw m)- command, but n=m+1 for the -ivreg2, bw (n)- command. Feb 20, 2010 · Luckily, a routine for first-stage F-stats in models with multiple endogenous variables is now programmed in ivreg2 . Unlike ivreg2 or ivregress, ivreg2h allows the syntax ivreg2h depvar exogvar (endogvar=), as after augmentation with the generated regressors, the order condition for identification will be satisfied. FYI, official ivreg is still available in Stata 10 and Stata 11: -help ivreg Aug 23, 2018 · I also encountered a problem when trying to estimate IV with regdhfe. I'm not quite sure how to tell the version of each package, but here are the distribution dates from the ado describe command: For ivreghdfe the distribution date was 20211214. Schaffer, v1. 07 *! see end of file for version comments program define actest9, rclass sortpreserve version 9. Note that, unlike the command ivqreg written by Do Won Kwak, ivqreg2 does not implement the estimator Martin and Mark, many thanks for the kind explanation. May 26, 2023 · 26 May 2023, 05:22. Jul 2, 2019 · Unable to display all first-stage regressions. ivregress: ivregress is a built-in command provided by Stata for instrumental variables regression. The overidentifying restrictions test reported after LIML estimation is the Anderson-Rubin (1950) overidentification statistic in a homoskedastic context. ) is derived from and supersedes the Aug 3, 2022 · 1. ) estat weakrobust performs tests of the form 1 = r1, where r1 is a constant vector of the same dimension as 1. 1 or later. 2. For this reason, marginsplot needs. (The standard errors > are a little different across most but not all of these. estimation results instead of perform the estimation. Andrew Musau. github. My apologies. When calculating the delta for beta=0, the command I use is: psacalc delta variable_of_interest. values. It also allows you to accept potential citations to this item that we are uncertain about. ajuste should not be using the svy prefix if you want. stata用ivreghdfe命令的时候报错last estimates not found,求助！. The F-stat is usually compared to critical value thresholds given in Stock and Yogo (2005), which are helpfully provided in the output of ivreghdfe/ivreg2, rather than reported as a p-value. (In practice, the exogenous covariates X2 are partialled out for simplicity. 2 IV and GMM estimation GMM was introduced in Hansen (1982). We consider the model. F( 6, 62) = 1045. aweights are not allowed with the jackknife prefix; see[R] jackknife. Description. 12 and livreg2 Mata library 1. Oct 6, 2023 · I got the following error: Code: command ivreg2 is unrecognized. One catch: as usual, when doing fixed effects using the demeaning transform, the reported standard errors and statistics using the demeaned data will have the wrong degrees of freedom because the program reporting the results doesn the command gmm estimates parameters by GMM you can specify the moment conditions as substitutable expressions a substitutable expression in Stata is like any mathematical expression, except that the parameters of the model are enclosed in braces {} alternatively, you may use command program to create a The ` variables in varlist1 and instlist comprise Z, the matrix of instruments. It works for the above (#20) two sets of 2SLS regressions (attachment Result#22). 回帖后跳转到最后一页. This code just modifies ivreg2 adding an absorb() option that uses reghdfes Mata functions (see this link for the line-by-line differences). reg cinf unem all yield exactly the same coefficient estimates. regression. Jun 29, 2017 · Hi, Is there any way of saving the first stage coefficients from ivreg2 to a dta. Before discussing further the motivation for various weak instrument diagnostics, we define the setting for IV estimation as a Generalized Method of Moments (GMM) optimization problem. So I will show you my example, including the command and its output, which doesn't have this problem. All varlists may contain time-series operators, but factor variables are not currently supported; see help varlist. Jul 23, 2014 · When I estimate a simple 2SLS model with ivregress (or ivreg) and ivreg2, I get discrepant estimates. I then installed ivregdhfe and everything worked perfect. Handle: RePEc:boc:bocode:s396802. 3). This module estimates the structural quantile functions defined by Chernozhukov and Hansen (2008) using the method of Machado and Santos Silva (forthcoming in the Journal of Econometrics ). This will not work for estimating a random effects iv model. ivreg2 cinf unem > . Jun 5, 2017 · Dear All, I am recently learning Oster (2016)'s approach for robustness checks with respect to unobservable selection bias. hascons, vce(), noheader, depname(), and weights are not allowed with the svy prefix; see[SVY] svy. There may be insufficient room to store results using -estimates store-, or names of endogenous regressors may be too long to store the results. In the command pane: Aug 22, 2023 · For ivreg2 I installed using ssc install ivreg2. When MYSELF estimate a simple 2SLS model with ivregress (or ivreg) and ivreg2, I gain discrepant estimates. thank you in advance. However, when I try to run the DMEXOGXT command after running the XTIVREG2, which I use insteda of XTIVREG, since I want to use robust and cluster, DMEXOGXT does not work and I get the following message; dmexogxt only works after xtivreg, fe last estimates not found r(301); 2) Is there an equivalent command for IVHETTEST for panel data after Jan 1, 2010 · Estimates indicate that a child's exposure to higher NC Pre-K funding was positively associated with that child's academic achievement 6 years later. However, I get a different error code invsym(): matrix has Sep 14, 2021 · I would like to reconfirm my interpretation of the post estimationn results: 1. May 20, 2020 · Dear Andrew: Thank you! Even if I use outreg2 and ivreg2 I get into trouble. Hi all Feb 2, 2022 · When I run ivreg2 y (x=z), small in Stata the results seem alike. a vector of residuals. com > wrote: > Dear all, I am using -ivhettest- by Mark E. lclogit2 and nlcom - "last estimates not found" 24 May 2023, 07:19. I have included several packages using the ssc install and find it commands. In the command pane: rvfplot, yline(0) As the scatter of this plot is not uniform, it indicates heteroskedasticity in the model. . LIML is a k-class estimator with k=the LIML eigenvalue lambda; 2SLS is a k-class estimator with k=1; OLS is a k-class esimator with k=0. However, I like to implement three-way clustering, but ivregdhfe only supports two-way clustering because it is based on ivreg2. >10 and greater than all critical values) 3. ivreg2 cinf unem . ivreg2 is compatible with Stata version 10. y = Xβ + u, u ∼ (0, Ω) with X (N × k) and define a matrix Z (N × `) where ` ≥ k. LIML, OLS and IV/2SLS are examples of k-class estimators. 使用ivregdhfe,不加absorb命令可以正常回归，但是加上之后一直报错，请各位老师帮忙看看这是怎么回事呢！. 12. Corpus ID: 117279769; XTOVERID: Stata module to calculate tests of overidentifying restrictions after xtreg, xtivreg, xtivreg2, xthtaylor @article{Schaffer2006XTOVERIDSM, title={XTOVERID: Stata module to calculate tests of overidentifying restrictions after xtreg, xtivreg, xtivreg2, xthtaylor}, author={Mark E. weights. 13 22Jan2015 *! based on actest 2. file? I am using regsave for the second stage coefficients, but even after reading similar post here on Statalist I can't find out how to do it for the first stage? I had the same sets of code for 4 different sets of residuals I was trying to create and all ran until the last one and it is now stuck on 2008 and 2009 like the above code was. And by "post" it means that it'd only work if there was an estimation performed in the first place. ) By contrast, if we run . eststo model_1: regress short_phys ever_attend_tk, vce (robust) May 12, 2020 · As the description of xtivreg2 says, it is the wrapper of ivreg2. When used, absorb() will also activate the small, noconstant and nopartialsmall options of ivreg2 (basically to force small sample adjustments, which are required as we might have a substantial number of fixed effects). In other words, can I still include fixed effect with cross-section group without using dummy variable approach with xi:ivreg2 Sep 17, 2014 · Hmm, not sure why this didn't work for me the first time, but thanks. is the last version of ivreg2 which runs on Stata 8. Jun 11, 2014 · The call to ajuste looks like replay syntax to svy. ivreg28 provides extensions to Stata's official ivreg and newey. com/serg The estimates of 1, 2, 1, and 2 can be tested for consistency with a null hypothesis about 1, our parameter of interest. The resulting estimates are saved in the ereturn list and as a set of estimates named GenInst. I use Stata. The k variables in varlist1 and varlist2 comprise X. Anyway, this works. The spirit is right, but you need to pass the contents of depvar as well. Tags: None. last estimates not found r(301); I then installed -dmexogxt- and tried: dmexogxt lreptextg but got the following error: dmexogxt only works after xtivreg, fe last estimates not found r(301); Something similar seems to occur when I try to use the -overid, all- command, although for me this is not essential, I only use it as a check. Then estmate using -ivreg2-, and test using -ivhettest-. postestimation- informs that -estat overid- can be used to obtain the. 1 and ivreg2 3. I am an Economist at the Federal Reserve Board. The same update includes other useful routines, like two-way clustering. 0 after uninstalling every package. Hello everyone, I use this code using the data panel, but the results don't come out and it says that Unable to display first-stage estimates; macro e (first) is missing. 1 - Windows machine). 这是加了absorb的↓这是不加的↓,经管之家 (原人大经济论坛) As the default for ivreg2 is to produce large sample estimates, your test results and standard errors will not be affected by the transformation. overidentification test. prais cinf unem > > we obtain different estimates and standard errors. This will not work for > estimating a random effects iv model. Feb 6, 2019 · However, I have more than 3000 groups, if I simply use i. 187 (not significant) 2. The minimum eigenvalue should be compared to Dec 9, 2018 · Thanks to Kit Baum, ivqreg2 is now available in SSC. The call to clear prior to calling estimates use drops the dataset from memory. In the meantime, you can use ivreg or ivreg2 to do the estimation. And I have two questions when implementing this: (1) Negative Delta. To match the HAC SE estimates of -newey- or -newey2 That is, > > . This means that svy thinks that you want it to replay. However, the overidentification statistic and the Durbin-Wu-Hausman statistic can’t be reported because. I switched back to reghdfe 5. ivreg2 cinf unem, bw(3) . ado file where it references that command. That is, the do-file needs to see two arguments. Underidentification test (Anderson canon. time it shows the elapsed time at different steps of the estimation. My main research interests are in Empirical Banking and Corporate Finance. 13. 2 syntax [varlist(default=none ts)] [if] [in] [aw fw pw iw] /// [, LAGs(numlist integer>0 sort) STRICTexog q0 bp small /// Robust CLuster(varlist) BW(string) kernel Jul 4, 2019 · I worked it out by two ways: mata and chi2 test: Here's my codes: *. particularly useful when the user has a rank-deficient estimate of the covariance matrix of orthogonality conditions (common with the cluster option and single-ton dummy variables). (1) 2SLS, IVs: iv_prefect0522 iv_05prob2 eststo clear foreach var of varlist y* I'm using Stata 13. On Thu, Jul 15, 2010 at 10:20 AM, xueliansharon < xuelianstata@gmail. Thu, 15 Jul 2010 16:36:51 -0500. To get around. I did both of these today on Stata version 18. DfBeta is a postestimation command. ivreg cinf unem > . Its main capabilities: two-step feasible GMM estimation; continuously updated GMM estimation (CUE); LIML and k-class estimation; automatic 面板数据DWH的命令不是这个吧，stata提示：ivendog works only after ivreg, ivreg2; use dmexogxt after 1. | Robust. 0g esample() from estimates store And savefirst with xtivreg2 does not create a corresponding variable. I am very new to Stata and have worked through the book "Data Analysis Using Stata" and read tons of info online but I am still not clear on a lot of things. This one's easy to answer, but the answer is not entirely satisfactory: ivhettest supports ivreg, ivreg2 and regress but not ivregress (see the help file). 5e+08 Prob > F = 0. e. Last edited by Marry Lee; 06 Oct 2023, 03:21 . Although I estimate a simple 2SLS exemplar with ivregress (or ivreg) and ivreg2, I get dissimilar estimates. See my revised answer. Let’s add a line to confirm. 0 and I am getting two errors 1) return code 111, that is one of the independent variables in the previously issued estimation command is not found during the execution of -ivhettest-. 988801 Centered R2 = 0. However, you have an easy alternative, which is to use the -orthog()- option of -xtivreg2- to conduct your endogeneity tests. ) an indicator that is 1 for 1 and 1 for 2. Therefore I create (ib3. 19 Sep 2015, 03:30. Try dropping one or more estimation results using -estimates drop-, using the -savefprefix- option, or using shorter variable names. Saved Results reported. ster not found; • last estimates not found; • 求助 使用循环命令显示last estimates not found怎么办; • 新手求教，请问输入rvfplot，yline（0）之后报错显示last estimates not found; • 用stata估计残差的时候，出现last estimates not found是为 The last two sections provide a summary of ivreg2 estimation options and syntax diagrams for all programs in the extended ivreg2 suite. Try this demo: That is, . Mar 7, 2007 · Abstract. Dear all, I’m doing a IV regression with the endog (varlist_en) option and cluster at country level. Warning: estimated covariance matrix of moment conditions not of full rank. ssc install github. Besides, -help ivregress. If I want to make a MS Word Table of four sets of 2SLS regressions, It would not work. To get the first stage just need to tweak the above code as: estadd scalar APF=first[7,1]: first_iq ; est restore first_iq ; estout, c(b) stats(APF) ; – user1189728. The estimators we have discussed are available from Baum, Schaffer and Stillman’s ivreg2 package (ssc describe ivreg2). I have literally the exact same lines of code above in my question. 93. IV estimation as a GMM problem. Ben Jann's -center- command is very handy for demeaning. used to produce the margins results. Just in case you need it, here is what I did: *download the zip file from the repository. Oct 17, 2021 · That suggests that something went wrong in the first-stage regression, perhaps. The plot also appears to show a curve. bo wn yd yj lf rl ek od wg ai