Comparar texto

Encuentra la diferencia entre dos archivos de texto

Diff en tiempo real

Diff unificado

Contraer líneas

Resaltar cambios

Resaltado de sintaxis

Herramientas

Diffchecker Desktop The most secure way to run Diffchecker. Get the Diffchecker Desktop app: your diffs never leave your computer!Get Desktop

test file main vs 203

Created about a year agoDiff never expires

Líneas
Total
Eliminado

Palabras
Total
Eliminado

Para continuar usando esta función, actualice a Diffchecker Pro Ver precios

444 líneas

Líneas
Total
Añadido

Palabras
Total
Añadido

Para continuar usando esta función, actualice a Diffchecker Pro Ver precios

. clear all

. /*=========================================================================

> 1: Load data

> ===========================================================================*/

. use "example31.dta", clear

. /*=========================================================================

> 2: Run tests

> ===========================================================================*/

. graph drop _all

. *------------------------ 2.1: Replicate 2a and test basic funcionality ----------------------------------

. xtevent y eta , panelvar(i) timevar(t) policyvar(z) window(5)

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators Number of obs = 9,000

Absorbed variable: i No. of categories = 1,000

F(21, 7979) = 209.66

Prob > F = 0.0000

R-squared = 0.7545

Adj R-squared = 0.7231

Root MSE = 0.9954

------------------------------------------------------------------------------

y | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

_k_eq_m6 | .0692904 .1104722 0.63 0.531 -.147264 .2858448

_k_eq_m5 | .1651543 .1178905 1.40 0.161 -.0659419 .3962505

_k_eq_m4 | .1011972 .1131376 0.89 0.371 -.120582 .3229763

_k_eq_m3 | .1498392 .108524 1.38 0.167 -.0628963 .3625747

_k_eq_m2 | .0503755 .105976 0.48 0.635 -.1573651 .2581161

_k_eq_p0 | .9892317 .1045981 9.46 0.000 .7841921 1.194271

_k_eq_p1 | .9021193 .1058732 8.52 0.000 .6945801 1.109658

_k_eq_p2 | .997856 .1091899 9.14 0.000 .7838152 1.211897

_k_eq_p3 | .9620247 .114047 8.44 0.000 .7384629 1.185587

_k_eq_p4 | .9911646 .1189692 8.33 0.000 .7579537 1.224375

_k_eq_p5 | 1.183222 .1269706 9.32 0.000 .9343267 1.432118

_k_eq_p6 | 1.072797 .1182986 9.07 0.000 .8409004 1.304693

eta | .2616307 .0096806 27.03 0.000 .2426541 .2806073

t |

8 | .282378 .0446025 6.33 0.000 .1949454 .3698106

9 | .4604956 .0448172 10.27 0.000 .3726422 .5483491

10 | .6894495 .0451663 15.26 0.000 .6009118 .7779873

11 | .835558 .0456976 18.28 0.000 .7459787 .9251373

12 | 1.008256 .0464042 21.73 0.000 .9172915 1.09922

13 | 1.247446 .0472958 26.38 0.000 1.154734 1.340159

14 | 1.396196 .0481892 28.97 0.000 1.301732 1.490659

15 | 1.629461 .0491454 33.16 0.000 1.533123 1.725798

_cons | 1.314217 .1041623 12.62 0.000 1.110031 1.518402

------------------------------------------------------------------------------

F test of absorbed indicators: F(999, 7979) = 10.317 Prob > F = 0.000

. * Testing xtset options

. xtevent y eta, policyvar(z) window(5) plot

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators Number of obs = 9,000

Absorbed variable: i No. of categories = 1,000

F(21, 7979) = 209.66

Prob > F = 0.0000

R-squared = 0.7545

Adj R-squared = 0.7231

Root MSE = 0.9954

------------------------------------------------------------------------------

y | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

_k_eq_m6 | .0692904 .1104722 0.63 0.531 -.147264 .2858448

_k_eq_m5 | .1651543 .1178905 1.40 0.161 -.0659419 .3962505

_k_eq_m4 | .1011972 .1131376 0.89 0.371 -.120582 .3229763

_k_eq_m3 | .1498392 .108524 1.38 0.167 -.0628963 .3625747

_k_eq_m2 | .0503755 .105976 0.48 0.635 -.1573651 .2581161

_k_eq_p0 | .9892317 .1045981 9.46 0.000 .7841921 1.194271

_k_eq_p1 | .9021193 .1058732 8.52 0.000 .6945801 1.109658

_k_eq_p2 | .997856 .1091899 9.14 0.000 .7838152 1.211897

_k_eq_p3 | .9620247 .114047 8.44 0.000 .7384629 1.185587

_k_eq_p4 | .9911646 .1189692 8.33 0.000 .7579537 1.224375

_k_eq_p5 | 1.183222 .1269706 9.32 0.000 .9343267 1.432118

_k_eq_p6 | 1.072797 .1182986 9.07 0.000 .8409004 1.304693

eta | .2616307 .0096806 27.03 0.000 .2426541 .2806073

t |

8 | .282378 .0446025 6.33 0.000 .1949454 .3698106

9 | .4604956 .0448172 10.27 0.000 .3726422 .5483491

10 | .6894495 .0451663 15.26 0.000 .6009118 .7779873

11 | .835558 .0456976 18.28 0.000 .7459787 .9251373

12 | 1.008256 .0464042 21.73 0.000 .9172915 1.09922

13 | 1.247446 .0472958 26.38 0.000 1.154734 1.340159

14 | 1.396196 .0481892 28.97 0.000 1.301732 1.490659

15 | 1.629461 .0491454 33.16 0.000 1.533123 1.725798

_cons | 1.314217 .1041623 12.62 0.000 1.110031 1.518402

------------------------------------------------------------------------------

F test of absorbed indicators: F(999, 7979) = 10.317 Prob > F = 0.000

. xtevent y eta, policyvar(z) panelvar(i) window(5) plot

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators Number of obs = 9,000

Absorbed variable: i No. of categories = 1,000

F(21, 7979) = 209.66

Prob > F = 0.0000

R-squared = 0.7545

Adj R-squared = 0.7231

Root MSE = 0.9954

------------------------------------------------------------------------------

y | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

_k_eq_m6 | .0692904 .1104722 0.63 0.531 -.147264 .2858448

_k_eq_m5 | .1651543 .1178905 1.40 0.161 -.0659419 .3962505

_k_eq_m4 | .1011972 .1131376 0.89 0.371 -.120582 .3229763

_k_eq_m3 | .1498392 .108524 1.38 0.167 -.0628963 .3625747

_k_eq_m2 | .0503755 .105976 0.48 0.635 -.1573651 .2581161

_k_eq_p0 | .9892317 .1045981 9.46 0.000 .7841921 1.194271

_k_eq_p1 | .9021193 .1058732 8.52 0.000 .6945801 1.109658

_k_eq_p2 | .997856 .1091899 9.14 0.000 .7838152 1.211897

_k_eq_p3 | .9620247 .114047 8.44 0.000 .7384629 1.185587

_k_eq_p4 | .9911646 .1189692 8.33 0.000 .7579537 1.224375

_k_eq_p5 | 1.183222 .1269706 9.32 0.000 .9343267 1.432118

_k_eq_p6 | 1.072797 .1182986 9.07 0.000 .8409004 1.304693

eta | .2616307 .0096806 27.03 0.000 .2426541 .2806073

t |

8 | .282378 .0446025 6.33 0.000 .1949454 .3698106

9 | .4604956 .0448172 10.27 0.000 .3726422 .5483491

10 | .6894495 .0451663 15.26 0.000 .6009118 .7779873

11 | .835558 .0456976 18.28 0.000 .7459787 .9251373

12 | 1.008256 .0464042 21.73 0.000 .9172915 1.09922

13 | 1.247446 .0472958 26.38 0.000 1.154734 1.340159

14 | 1.396196 .0481892 28.97 0.000 1.301732 1.490659

15 | 1.629461 .0491454 33.16 0.000 1.533123 1.725798

_cons | 1.314217 .1041623 12.62 0.000 1.110031 1.518402

------------------------------------------------------------------------------

F test of absorbed indicators: F(999, 7979) = 10.317 Prob > F = 0.000

. xtevent y eta, policyvar(z) timevar(t) window(5) plot

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators Number of obs = 9,000

Absorbed variable: i No. of categories = 1,000

F(21, 7979) = 209.66

Prob > F = 0.0000

R-squared = 0.7545

Adj R-squared = 0.7231

Root MSE = 0.9954

------------------------------------------------------------------------------

y | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

_k_eq_m6 | .0692904 .1104722 0.63 0.531 -.147264 .2858448

_k_eq_m5 | .1651543 .1178905 1.40 0.161 -.0659419 .3962505

_k_eq_m4 | .1011972 .1131376 0.89 0.371 -.120582 .3229763

_k_eq_m3 | .1498392 .108524 1.38 0.167 -.0628963 .3625747

_k_eq_m2 | .0503755 .105976 0.48 0.635 -.1573651 .2581161

_k_eq_p0 | .9892317 .1045981 9.46 0.000 .7841921 1.194271

_k_eq_p1 | .9021193 .1058732 8.52 0.000 .6945801 1.109658

_k_eq_p2 | .997856 .1091899 9.14 0.000 .7838152 1.211897

_k_eq_p3 | .9620247 .114047 8.44 0.000 .7384629 1.185587

_k_eq_p4 | .9911646 .1189692 8.33 0.000 .7579537 1.224375

_k_eq_p5 | 1.183222 .1269706 9.32 0.000 .9343267 1.432118

_k_eq_p6 | 1.072797 .1182986 9.07 0.000 .8409004 1.304693

eta | .2616307 .0096806 27.03 0.000 .2426541 .2806073

t |

8 | .282378 .0446025 6.33 0.000 .1949454 .3698106

9 | .4604956 .0448172 10.27 0.000 .3726422 .5483491

10 | .6894495 .0451663 15.26 0.000 .6009118 .7779873

11 | .835558 .0456976 18.28 0.000 .7459787 .9251373

12 | 1.008256 .0464042 21.73 0.000 .9172915 1.09922

13 | 1.247446 .0472958 26.38 0.000 1.154734 1.340159

14 | 1.396196 .0481892 28.97 0.000 1.301732 1.490659

15 | 1.629461 .0491454 33.16 0.000 1.533123 1.725798

_cons | 1.314217 .1041623 12.62 0.000 1.110031 1.518402

------------------------------------------------------------------------------

F test of absorbed indicators: F(999, 7979) = 10.317 Prob > F = 0.000

. /* Must fail

> xtevent y eta, policyvar(z) panelvar(z) window(5)

> xtevent y eta, policyvar(z) timevar(z) window(5)

> */

. * Testing noci, nosupt, nozeroline, nonormlabel

. xtevent y eta, policyvar(z) timevar(t) window(5)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators Number of obs = 9,000

Absorbed variable: i No. of categories = 1,000

F(21, 7979) = 209.66

Prob > F = 0.0000

R-squared = 0.7545

Adj R-squared = 0.7231

Root MSE = 0.9954

------------------------------------------------------------------------------

y | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

_k_eq_m6 | .0692904 .1104722 0.63 0.531 -.147264 .2858448

_k_eq_m5 | .1651543 .1178905 1.40 0.161 -.0659419 .3962505

_k_eq_m4 | .1011972 .1131376 0.89 0.371 -.120582 .3229763

_k_eq_m3 | .1498392 .108524 1.38 0.167 -.0628963 .3625747

_k_eq_m2 | .0503755 .105976 0.48 0.635 -.1573651 .2581161

_k_eq_p0 | .9892317 .1045981 9.46 0.000 .7841921 1.194271

_k_eq_p1 | .9021193 .1058732 8.52 0.000 .6945801 1.109658

_k_eq_p2 | .997856 .1091899 9.14 0.000 .7838152 1.211897

_k_eq_p3 | .9620247 .114047 8.44 0.000 .7384629 1.185587

_k_eq_p4 | .9911646 .1189692 8.33 0.000 .7579537 1.224375

_k_eq_p5 | 1.183222 .1269706 9.32 0.000 .9343267 1.432118

_k_eq_p6 | 1.072797 .1182986 9.07 0.000 .8409004 1.304693

eta | .2616307 .0096806 27.03 0.000 .2426541 .2806073

t |

8 | .282378 .0446025 6.33 0.000 .1949454 .3698106

9 | .4604956 .0448172 10.27 0.000 .3726422 .5483491

10 | .6894495 .0451663 15.26 0.000 .6009118 .7779873

11 | .835558 .0456976 18.28 0.000 .7459787 .9251373

12 | 1.008256 .0464042 21.73 0.000 .9172915 1.09922

13 | 1.247446 .0472958 26.38 0.000 1.154734 1.340159

14 | 1.396196 .0481892 28.97 0.000 1.301732 1.490659

15 | 1.629461 .0491454 33.16 0.000 1.533123 1.725798

_cons | 1.314217 .1041623 12.62 0.000 1.110031 1.518402

------------------------------------------------------------------------------

F test of absorbed indicators: F(999, 7979) = 10.317 Prob > F = 0.000

. xteventplot, noci

option noci has been specified. Confidence intervals won't be displayed

. xteventplot, nosupt

option nosupt has been specified. Sup-t confidence intervals won't be displayed or calculated

. xteventplot, nozeroline

option nozeroline has been specified. The reference line at 0 won't be displayed

. xteventplot, nonormlabel

option nonormlabel has been specified. The label for the value of the dependent variable at event time correspon

> ding to the normalized coefficient won't be displayed

. * Test combinations

. xteventplot, noci nozeroline

option noci has been specified. Confidence intervals won't be displayed

option nozeroline has been specified. The reference line at 0 won't be displayed

. xteventplot, noci nonormlabel

option noci has been specified. Confidence intervals won't be displayed

option nonormlabel has been specified. The label for the value of the dependent variable at event time correspon

> ding to the normalized coefficient won't be displayed

. xteventplot, nosupt nozeroline

option nosupt has been specified. Sup-t confidence intervals won't be displayed or calculated

option nozeroline has been specified. The reference line at 0 won't be displayed

. xteventplot, nosupt nonormlabel

option nosupt has been specified. Sup-t confidence intervals won't be displayed or calculated

option nonormlabel has been specified. The label for the value of the dependent variable at event time correspon

> ding to the normalized coefficient won't be displayed

. * A common axis plot with labels

. gen y2 = y + 5

. xtevent y eta, policyvar(z) timevar(t) window(5)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators Number of obs = 9,000

Absorbed variable: i No. of categories = 1,000

F(21, 7979) = 209.66

Prob > F = 0.0000

R-squared = 0.7545

Adj R-squared = 0.7231

Root MSE = 0.9954

------------------------------------------------------------------------------

y | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

_k_eq_m6 | .0692904 .1104722 0.63 0.531 -.147264 .2858448

_k_eq_m5 | .1651543 .1178905 1.40 0.161 -.0659419 .3962505

_k_eq_m4 | .1011972 .1131376 0.89 0.371 -.120582 .3229763

_k_eq_m3 | .1498392 .108524 1.38 0.167 -.0628963 .3625747

_k_eq_m2 | .0503755 .105976 0.48 0.635 -.1573651 .2581161

_k_eq_p0 | .9892317 .1045981 9.46 0.000 .7841921 1.194271

_k_eq_p1 | .9021193 .1058732 8.52 0.000 .6945801 1.109658

_k_eq_p2 | .997856 .1091899 9.14 0.000 .7838152 1.211897

_k_eq_p3 | .9620247 .114047 8.44 0.000 .7384629 1.185587

_k_eq_p4 | .9911646 .1189692 8.33 0.000 .7579537 1.224375

_k_eq_p5 | 1.183222 .1269706 9.32 0.000 .9343267 1.432118

_k_eq_p6 | 1.072797 .1182986 9.07 0.000 .8409004 1.304693

eta | .2616307 .0096806 27.03 0.000 .2426541 .2806073

t |

8 | .282378 .0446025 6.33 0.000 .1949454 .3698106

9 | .4604956 .0448172 10.27 0.000 .3726422 .5483491

10 | .6894495 .0451663 15.26 0.000 .6009118 .7779873

11 | .835558 .0456976 18.28 0.000 .7459787 .9251373

12 | 1.008256 .0464042 21.73 0.000 .9172915 1.09922

13 | 1.247446 .0472958 26.38 0.000 1.154734 1.340159

14 | 1.396196 .0481892 28.97 0.000 1.301732 1.490659

15 | 1.629461 .0491454 33.16 0.000 1.533123 1.725798

_cons | 1.314217 .1041623 12.62 0.000 1.110031 1.518402

------------------------------------------------------------------------------

F test of absorbed indicators: F(999, 7979) = 10.317 Prob > F = 0.000

. loc lab : di %-9.2f `=e(y1)'

. loc lab=strtrim("`lab'")

. xteventplot, nonormlabel ylab(-3 0 `"0 (`lab')"' 3) name(g1) /* " */

option nonormlabel has been specified. The label for the value of the dependent variable at event time correspon

> ding to the normalized coefficient won't be displayed

. xtevent y2 eta, policyvar(z) timevar(t) window(5)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators Number of obs = 9,000

Absorbed variable: i No. of categories = 1,000

F(21, 7979) = 209.66

Prob > F = 0.0000

R-squared = 0.7545

Adj R-squared = 0.7231

Root MSE = 0.9954

------------------------------------------------------------------------------

y2 | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

_k_eq_m6 | .0692904 .1104722 0.63 0.531 -.147264 .2858448

_k_eq_m5 | .1651543 .1178905 1.40 0.161 -.0659419 .3962505

_k_eq_m4 | .1011971 .1131376 0.89 0.371 -.120582 .3229763

_k_eq_m3 | .1498392 .108524 1.38 0.167 -.0628963 .3625747

_k_eq_m2 | .0503755 .105976 0.48 0.635 -.1573651 .2581162

_k_eq_p0 | .9892317 .1045981 9.46 0.000 .7841921 1.194271

_k_eq_p1 | .9021193 .1058732 8.52 0.000 .6945801 1.109658

_k_eq_p2 | .997856 .1091899 9.14 0.000 .7838152 1.211897

_k_eq_p3 | .9620247 .114047 8.44 0.000 .7384629 1.185587

_k_eq_p4 | .9911646 .1189692 8.33 0.000 .7579538 1.224375

_k_eq_p5 | 1.183222 .1269706 9.32 0.000 .9343268 1.432118

_k_eq_p6 | 1.072797 .1182986 9.07 0.000 .8409004 1.304693

eta | .2616307 .0096806 27.03 0.000 .2426541 .2806073

t |

8 | .2823779 .0446025 6.33 0.000 .1949453 .3698105

9 | .4604956 .0448172 10.27 0.000 .3726422 .5483491

10 | .6894495 .0451663 15.26 0.000 .6009118 .7779873

11 | .835558 .0456976 18.28 0.000 .7459787 .9251372

12 | 1.008256 .0464042 21.73 0.000 .9172915 1.09922

13 | 1.247446 .0472958 26.38 0.000 1.154734 1.340159

14 | 1.396196 .0481892 28.97 0.000 1.301732 1.490659

15 | 1.629461 .0491454 33.16 0.000 1.533123 1.725798

_cons | 6.314217 .1041623 60.62 0.000 6.110031 6.518402

------------------------------------------------------------------------------

F test of absorbed indicators: F(999, 7979) = 10.317 Prob > F = 0.000

. loc lab : di %-9.2f `=e(y1)'

. loc lab=strtrim("`lab'")

. xteventplot, nonormlabel ylab(-3 0 `"0 (`lab')"' 3) name(g2) /* " */

option nonormlabel has been specified. The label for the value of the dependent variable at event time correspon

> ding to the normalized coefficient won't be displayed

. graph combine g1 g2

. drop y2

. * Test if/in

. xtevent y eta if i<100, panelvar(i) timevar(t) policyvar(z) window(5) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators Number of obs = 891

Absorbed variable: i No. of categories = 99

F(21, 771) = 19.21

Prob > F = 0.0000

R-squared = 0.7712

Adj R-squared = 0.7359

Root MSE = 1.0356

------------------------------------------------------------------------------

y | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

_k_eq_m6 | -.2882097 .351599 -0.82 0.413 -.9784146 .4019952

_k_eq_m5 | -.6923773 .3927855 -1.76 0.078 -1.463433 .0786785

_k_eq_m4 | -.0932655 .3525298 -0.26 0.791 -.7852976 .5987666

_k_eq_m3 | -.068185 .3384769 -0.20 0.840 -.7326305 .5962605

_k_eq_m2 | -.3234206 .3406932 -0.95 0.343 -.992217 .3453757

_k_eq_p0 | .5722541 .3342813 1.71 0.087 -.0839554 1.228464

_k_eq_p1 | .9961526 .3417043 2.92 0.004 .3253715 1.666934

_k_eq_p2 | .7954004 .3524217 2.26 0.024 .1035806 1.48722

_k_eq_p3 | .8880688 .3628256 2.45 0.015 .1758256 1.600312

_k_eq_p4 | 1.039313 .3719825 2.79 0.005 .3090946 1.769532

Diferencias guardadas

Texto original

Abrir archivo

. 
. clear all

. 
. /*=========================================================================
>                         1: Load data 
> ===========================================================================*/   
. use "example31.dta", clear

. 
. 
. /*=========================================================================
>                         2: Run tests
> ===========================================================================*/
.         
. graph drop _all

. 
. *------------------------ 2.1: Replicate 2a and test basic funcionality ----------------------------------
. 
. xtevent y eta , panelvar(i) timevar(t) policyvar(z) window(5)

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(21, 7979)       = 209.66
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7545
                                                    Adj R-squared     = 0.7231
                                                    Root MSE          = 0.9954

. 
. * Testing xtset options
. 
. xtevent y eta, policyvar(z) window(5) plot

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. xtevent y eta, policyvar(z) panelvar(i) window(5) plot

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. xtevent y eta, policyvar(z) timevar(t) window(5) plot

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. 
. 
. /* Must fail
> xtevent y eta, policyvar(z) panelvar(z) window(5) 
> xtevent y eta, policyvar(z) timevar(z) window(5) 
> */
. 
. * Testing noci, nosupt, nozeroline, nonormlabel
. xtevent y eta, policyvar(z) timevar(t) window(5)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. xteventplot, noci

option noci has been specified. Confidence intervals won't be displayed

. xteventplot, nosupt

option nosupt has been specified. Sup-t confidence intervals won't be displayed or calculated

. xteventplot, nozeroline

option nozeroline has been specified. The reference line at 0 won't be displayed

. xteventplot, nonormlabel

option nonormlabel has been specified. The label for the value of the dependent variable at event time correspon
> ding to the normalized coefficient won't be displayed

. 
. * Test combinations
. xteventplot, noci nozeroline

option noci has been specified. Confidence intervals won't be displayed

option nozeroline has been specified. The reference line at 0 won't be displayed

. xteventplot, noci nonormlabel

option noci has been specified. Confidence intervals won't be displayed

option nonormlabel has been specified. The label for the value of the dependent variable at event time correspon
> ding to the normalized coefficient won't be displayed

. xteventplot, nosupt nozeroline

option nosupt has been specified. Sup-t confidence intervals won't be displayed or calculated

option nozeroline has been specified. The reference line at 0 won't be displayed

. xteventplot, nosupt nonormlabel

option nosupt has been specified. Sup-t confidence intervals won't be displayed or calculated

option nonormlabel has been specified. The label for the value of the dependent variable at event time correspon
> ding to the normalized coefficient won't be displayed

. 
. * A common axis plot with labels
. gen y2 = y + 5

. xtevent y eta, policyvar(z) timevar(t) window(5)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. loc lab : di %-9.2f `=e(y1)'

. loc lab=strtrim("`lab'")

. xteventplot, nonormlabel ylab(-3 0 `"0 (`lab')"' 3) name(g1) /* " */

option nonormlabel has been specified. The label for the value of the dependent variable at event time correspon
> ding to the normalized coefficient won't be displayed

. xtevent y2 eta, policyvar(z) timevar(t) window(5)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

------------------------------------------------------------------------------
          y2 | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    _k_eq_m6 |   .0692904   .1104722     0.63   0.531     -.147264    .2858448
    _k_eq_m5 |   .1651543   .1178905     1.40   0.161    -.0659419    .3962505
    _k_eq_m4 |   .1011971   .1131376     0.89   0.371     -.120582    .3229763
    _k_eq_m3 |   .1498392    .108524     1.38   0.167    -.0628963    .3625747
    _k_eq_m2 |   .0503755    .105976     0.48   0.635    -.1573651    .2581162
    _k_eq_p0 |   .9892317   .1045981     9.46   0.000     .7841921    1.194271
    _k_eq_p1 |   .9021193   .1058732     8.52   0.000     .6945801    1.109658
    _k_eq_p2 |    .997856   .1091899     9.14   0.000     .7838152    1.211897
    _k_eq_p3 |   .9620247    .114047     8.44   0.000     .7384629    1.185587
    _k_eq_p4 |   .9911646   .1189692     8.33   0.000     .7579538    1.224375
    _k_eq_p5 |   1.183222   .1269706     9.32   0.000     .9343268    1.432118
    _k_eq_p6 |   1.072797   .1182986     9.07   0.000     .8409004    1.304693
         eta |   .2616307   .0096806    27.03   0.000     .2426541    .2806073
             |
           t |
          8  |   .2823779   .0446025     6.33   0.000     .1949453    .3698105
          9  |   .4604956   .0448172    10.27   0.000     .3726422    .5483491
         10  |   .6894495   .0451663    15.26   0.000     .6009118    .7779873
         11  |    .835558   .0456976    18.28   0.000     .7459787    .9251372
         12  |   1.008256   .0464042    21.73   0.000     .9172915     1.09922
         13  |   1.247446   .0472958    26.38   0.000     1.154734    1.340159
         14  |   1.396196   .0481892    28.97   0.000     1.301732    1.490659
         15  |   1.629461   .0491454    33.16   0.000     1.533123    1.725798
             |
       _cons |   6.314217   .1041623    60.62   0.000     6.110031    6.518402
------------------------------------------------------------------------------
F test of absorbed indicators: F(999, 7979) = 10.317          Prob > F = 0.000

. loc lab : di %-9.2f `=e(y1)'

. loc lab=strtrim("`lab'")

. xteventplot, nonormlabel ylab(-3 0 `"0 (`lab')"' 3) name(g2) /* " */

option nonormlabel has been specified. The label for the value of the dependent variable at event time correspon
> ding to the normalized coefficient won't be displayed

. graph combine g1 g2

. drop y2

. 
. * Test if/in
. xtevent y eta if i<100, panelvar(i) timevar(t) policyvar(z) window(5) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =    891
Absorbed variable: i                                No. of categories =     99
                                                    F(21, 771)        =  19.21
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7712
                                                    Adj R-squared     = 0.7359
                                                    Root MSE          = 1.0356

. xtevent y eta in 1/600 , panelvar(i) timevar(t) policyvar(z) window(5) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =    270
Absorbed variable: i                                No. of categories =     30
                                                    F(21, 219)        =   6.20
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7716
                                                    Adj R-squared     = 0.7195
                                                    Root MSE          = 0.9710

. xtevent y eta in 1/600 if i<30 , panelvar(i) timevar(t) policyvar(z) window(5) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =    261
Absorbed variable: i                                No. of categories =     29
                                                    F(21, 211)        =   5.76
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7595
                                                    Adj R-squared     = 0.7037
                                                    Root MSE          = 0.9726

. 
. * Test nofe, note
. xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) nofe plot

No proxy or instruments provided. Implementing OLS estimator

Source |       SS           df       MS      Number of obs   =     9,000
-------------+----------------------------------   F(21, 8978)     =    332.33
       Model |  14084.5946        21  670.694983   Prob > F        =    0.0000
    Residual |  18119.0165     8,978  2.01815733   R-squared       =    0.4374
-------------+----------------------------------   Adj R-squared   =    0.4360
       Total |  32203.6112     8,999  3.57857664   Root MSE        =    1.4206

. xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) nofe note plot

No proxy or instruments provided. Implementing OLS estimator

Source |       SS           df       MS      Number of obs   =     9,000
-------------+----------------------------------   F(13, 8986)     =    405.50
       Model |  11906.8789        13  915.913763   Prob > F        =    0.0000
    Residual |  20296.7323     8,986  2.25870601   R-squared       =    0.3697
-------------+----------------------------------   Adj R-squared   =    0.3688
       Total |  32203.6112     8,999  3.57857664   Root MSE        =    1.5029

. 
. * Test smoothest line
. xtevent eta, panelvar(i) timevar(t) policyvar(z) window(4)

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     = 11,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(20, 9980)       = 145.18
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.8646
                                                    Adj R-squared     = 0.8507
                                                    Root MSE          = 1.2464

------------------------------------------------------------------------------
         eta | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    _k_eq_m5 |  -2.828136   .1064315   -26.57   0.000    -3.036763   -2.619509
    _k_eq_m4 |  -1.677028   .1240764   -13.52   0.000    -1.920243   -1.433813
    _k_eq_m3 |  -1.336378   .1222929   -10.93   0.000    -1.576097   -1.096659
    _k_eq_m2 |  -.8921123   .1185683    -7.52   0.000     -1.12453   -.6596945
    _k_eq_p0 |   1.458456   .1185385    12.30   0.000     1.226096    1.690815
    _k_eq_p1 |   1.640089   .1193893    13.74   0.000     1.406062    1.874117
    _k_eq_p2 |   1.818459   .1215983    14.95   0.000     1.580102    2.056816
    _k_eq_p3 |   2.036383   .1257121    16.20   0.000     1.789962    2.282804
    _k_eq_p4 |    2.11163   .1315385    16.05   0.000     1.853788    2.369472
    _k_eq_p5 |   2.373969   .1132438    20.96   0.000     2.151988     2.59595
             |
           t |
          7  |   -.069634   .0557916    -1.25   0.212    -.1789968    .0397288
          8  |  -.1508391   .0558944    -2.70   0.007    -.2604034   -.0412748
          9  |  -.2464859   .0560989    -4.39   0.000    -.3564511   -.1365207
         10  |  -.3086252   .0563966    -5.47   0.000    -.4191738   -.1980765
         11  |  -.4407218   .0567634    -7.76   0.000    -.5519896   -.3294541
         12  |  -.5387938   .0572022    -9.42   0.000    -.6509217    -.426666
         13  |  -.6787502   .0577788   -11.75   0.000    -.7920082   -.5654922
         14  |  -.7380675    .058395   -12.64   0.000    -.8525335   -.6236015
         15  |  -.7471374   .0590178   -12.66   0.000    -.8628243   -.6314506
         16  |  -.8863071   .0597648   -14.83   0.000    -1.003458    -.769156
             |
       _cons |   2.388192   .1077882    22.16   0.000     2.176906    2.599479
------------------------------------------------------------------------------
F test of absorbed indicators: F(999, 9980) = 29.433          Prob > F = 0.000

. xteventplot, smpath(scatter)

Note: Smoothest line drawn for system confidence level = 95%
Wald Critical Value 18.3070381
Order 0
Wald value 2888.84103
Order 1
Wald value 519.967356
Order 2
Wald value 133.928595
Order 3
Wald value 112.44878
Order 4
Wald value 102.223602
Order 5
Wald value 56.6864304
Order 6
Wald value 22.8128592
Order 7
Wald value 22.7419107
Order 8
Wald value 9.27938467
(setting technique to nr)
Iteration 0:  f(p) =   46164996  (not concave)
Iteration 1:  f(p) =   46109859  (not concave)
Iteration 2:  f(p) =   46102811  (not concave)
Iteration 3:  f(p) =   46102316  (not concave)
Iteration 4:  f(p) =   46092836  (not concave)
(switching technique to bfgs)
Iteration 5:  f(p) =   46092567  
Iteration 6:  f(p) =   46092566  
Iteration 7:  f(p) =   46092565  
Iteration 8:  f(p) =   46092565  
Iteration 9:  f(p) =   46092562  
(switching technique to nr)
Iteration 10: f(p) =   46092545  
Iteration 11: f(p) =   46091975  (not concave)
Iteration 12: f(p) =   46023841  (not concave)
Iteration 13: f(p) =   46021814  (not concave)
Iteration 14: f(p) =   46021637  (not concave)
(switching technique to bfgs)
Iteration 15: f(p) =   46021616  
Iteration 16: f(p) =   46021616  
Iteration 17: f(p) =   46021614  
Iteration 18: f(p) =   46021613  
Iteration 19: f(p) =   46021536  
(switching technique to nr)
Iteration 20: f(p) =   46021421  
Hessian is not positive semidefinite
(setting technique to nr)
Iteration 0:  f(p) =   46175610  (not concave)
Iteration 1:  f(p) =   46102071  (not concave)
Iteration 2:  f(p) =   46097347  (not concave)
Iteration 3:  f(p) =   46097077  (not concave)
Iteration 4:  f(p) =   46097013  (not concave)
(switching technique to bfgs)
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 5:  f(p) =   46097002  
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 6:  f(p) =   46097002  (backed up)
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 7:  f(p) =   46097002  (backed up)
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 8:  f(p) =   46097002  (backed up)
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 9:  f(p) =   46097002  (backed up)
(switching technique to nr)
Hessian is not positive semidefinite

The optimization to calculate the smoothest path returned an error code.
Smoothest path won't be displayed.
Try changing the optimization options. For example, try -smpath(scatter, tech(dfp)-.

Error code = 4. See mf_optimize##r_error to see what that means.
Error code = 4. See mf_optimize##r_error to see what that means.

. xteventplot, smpath(line)

The optimization to calculate the smoothest path returned an error code.
Smoothest path won't be displayed.
Try changing the optimization options. For example, try -smpath(scatter, tech(dfp)-.

Error code = 4. See mf_optimize##r_error to see what that means.
Error code = 4. See mf_optimize##r_error to see what that means.

. xteventplot, smpath(line, technique("nr 10 bfgs 10"))

Note: Smoothest line drawn for system confidence level = 95%
Wald Critical Value 18.3070381
Order 0
Wald value 2888.84103
Order 1
Wald value 519.967356
Order 2
Wald value 133.928595
Order 3
Wald value 112.44878
Order 4
Wald value 102.223602
Order 5
Wald value 56.6864304
Order 6
Wald value 22.8128592
Order 7
Wald value 22.7419107
Order 8
Wald value 9.27938467
(setting technique to nr)
Iteration 0:  f(p) =   46164996  (not concave)
Iteration 1:  f(p) =   46109859  (not concave)
Iteration 2:  f(p) =   46102811  (not concave)
Iteration 3:  f(p) =   46102316  (not concave)
Iteration 4:  f(p) =   46092836  (not concave)
Iteration 5:  f(p) =   46092567  (not concave)
Iteration 6:  f(p) =   45994866  (not concave)
Iteration 7:  f(p) =   45898041  (not concave)
Iteration 8:  f(p) =   45896408  (not concave)
Iteration 9:  f(p) =   45881715  (not concave)
(switching technique to bfgs)
Iteration 10: f(p) =   45880998  
Iteration 11: f(p) =   45880929  
Iteration 12: f(p) =   45868963  
Iteration 13: f(p) =   45863305  
Iteration 14: f(p) =   45862989  (backed up)
Iteration 15: f(p) =   45862989  (backed up)
Iteration 16: f(p) =   45862989  (backed up)
Iteration 17: f(p) =   45862960  
Iteration 18: f(p) =   45862959  (backed up)
Iteration 19: f(p) =   45858646  
(switching technique to nr)
Iteration 20: f(p) =   45858577  (not concave)
Iteration 21: f(p) =   45857638  (not concave)
Iteration 22: f(p) =   45857570  (not concave)
Iteration 23: f(p) =   45855841  (not concave)
Iteration 24: f(p) =   45855821  (not concave)
Iteration 25: f(p) =   45855444  (not concave)
Iteration 26: f(p) =   45855149  (not concave)
Iteration 27: f(p) =   45663340  (not concave)
Iteration 28: f(p) =   45654588  (not concave)
Iteration 29: f(p) =   45653253  (not concave)
(switching technique to bfgs)
Iteration 30: f(p) =   45653209  
Iteration 31: f(p) =   45653131  
BFGS stepping has contracted, resetting BFGS Hessian
Iteration 32: f(p) =   45649348  
Iteration 33: f(p) =   45649291  (backed up)
Iteration 34: f(p) =   45645823  (backed up)
Iteration 35: f(p) =   45645218  (backed up)
Iteration 36: f(p) =   45606313  (backed up)
Iteration 37: f(p) =   45601956  (backed up)
Iteration 38: f(p) =   45600241  (backed up)
Iteration 39: f(p) =   45600185  (backed up)
(switching technique to nr)
Iteration 40: f(p) =   45600169  (backed up)
Hessian is not positive semidefinite
(setting technique to nr)
Iteration 0:  f(p) =   46175610  (not concave)
Iteration 1:  f(p) =   46102071  (not concave)
Iteration 2:  f(p) =   46097347  (not concave)
Iteration 3:  f(p) =   46097077  (not concave)
Iteration 4:  f(p) =   46097013  (not concave)
Hessian is not positive semidefinite

The optimization to calculate the smoothest path returned an error code.
Smoothest path won't be displayed.
Try changing the optimization options. For example, try -smpath(scatter, tech(dfp)-.

Error code = 4. See mf_optimize##r_error to see what that means.
Error code = 4. See mf_optimize##r_error to see what that means.

. 
. * Test more suptreps
. 
. cap graph drop g1

. cap graph drop g2

. 
. xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(3)

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     = 13,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(21, 11979)      = 572.74
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7653
                                                    Adj R-squared     = 0.7453
                                                    Root MSE          = 1.0058

. xteventplot, suptreps(20) name(g1)

. xteventplot, suptreps(1e6) name(g2)

. 
. graph combine g1 g2, rows(1)

. 
. graph drop g1

. graph drop g2

. 
. * Test savek
. xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) savek(a)

No proxy or instruments provided. Implementing OLS estimator

. des a_eq*, s

Variable      Storage   Display    Value
    name         type    format    label      Variable label
----------------------------------------------------------------------------------------------------------------
a_eq_m6         double  %10.0g                Event time <= - 6
a_eq_m5         double  %10.0g                Event-time = - 5
a_eq_m4         double  %10.0g                Event-time = - 4
a_eq_m3         double  %10.0g                Event-time = - 3
a_eq_m2         double  %10.0g                Event-time = - 2
a_eq_m1         double  %10.0g                Event-time = - 1
a_eq_p0         double  %10.0g                Event-time = + 0
a_eq_p1         double  %10.0g                Event-time = + 1
a_eq_p2         double  %10.0g                Event-time = + 2
a_eq_p3         double  %10.0g                Event-time = + 3
a_eq_p4         double  %10.0g                Event-time = + 4
a_eq_p5         double  %10.0g                Event-time = + 5
a_eq_p6         double  %10.0g                Event time >= + 6

. des a_evtime, s

. drop a*

. 
. * Test savek with suboption noestimate
. xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) savek(a, noe)

No proxy or instruments provided. Implementing OLS estimator

. des a_eq*, s

. des a_evtime, s

. drop a*

. 
. * Test different prefix
. xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) savek(b)

No proxy or instruments provided. Implementing OLS estimator

. 
. *savek + suboption replace 
. cap noi xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) savek(b)

No proxy or instruments provided. Implementing OLS estimator

You specified to save the event-time dummy variables using the prefix b, but you already have event-time dummy v
> ariables saved with that prefix.

Use the replace suboption to replace the existing variables.

. xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) savek(b, replace)

No proxy or instruments provided. Implementing OLS estimator

. 
. * Test factor variables in varlist
. cap gen pois=rpoisson(5)

. xtevent y eta i.pois, panelvar(i) timevar(t) policyvar(z) window(5) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(36, 7964)       = 122.65
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7549
                                                    Adj R-squared     = 0.7230
                                                    Root MSE          = 0.9955

. cap drop pois

. 
. * Test time series variables in varlist
. 
. xtevent y l.eta , panelvar(i) timevar(t) policyvar(z) window(5) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(21, 7979)       = 178.89
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7410
                                                    Adj R-squared     = 0.7079
                                                    Root MSE          = 1.0225

. xtevent y f.eta , panelvar(i) timevar(t) policyvar(z) window(5) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(21, 7979)       = 178.44
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7408
                                                    Adj R-squared     = 0.7076
                                                    Root MSE          = 1.0229

. 
. * Test asymmetric window
. xtevent y eta , panelvar(i) timevar(t) policyvar(z) window(-4 2) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     = 13,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(21, 11979)      = 572.89
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7583
                                                    Adj R-squared     = 0.7377
                                                    Root MSE          = 1.0034

. 
. * Test finding and estimating with the widest window 
. xtevent y eta , panelvar(i) timevar(t) policyvar(z) impute(stag) window(max) plot

No proxy or instruments provided. Implementing OLS estimator
The calculated window by window(max) is (-18,16), plus the endpoints -19 and 17.

Linear regression, absorbing indicators             Number of obs     = 20,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(56, 18944)      = 702.23
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.8023
                                                    Adj R-squared     = 0.7913
                                                    Root MSE          = 1.0022

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
   _k_eq_m19 |   .6423926   .2608484     2.46   0.014     .1311064    1.153679
   _k_eq_m18 |  -.1775268   .2104908    -0.84   0.399    -.5901075    .2350538
   _k_eq_m17 |   .1263326   .1698441     0.74   0.457     -.206577    .4592422
   _k_eq_m16 |   .0380571   .1450188     0.26   0.793    -.2461928    .3223069
   _k_eq_m15 |    .005294   .1337594     0.04   0.968    -.2568864    .2674743
   _k_eq_m14 |  -.1028837   .1236077    -0.83   0.405    -.3451658    .1393984
   _k_eq_m13 |  -.1811439   .1143673    -1.58   0.113     -.405314    .0430262
   _k_eq_m12 |   .1848555   .1070519     1.73   0.084    -.0249757    .3946868
   _k_eq_m11 |   .0700347   .1022051     0.69   0.493    -.1302964    .2703658
   _k_eq_m10 |   .1192835   .0973245     1.23   0.220    -.0714811    .3100481
    _k_eq_m9 |  -.1123919   .0934684    -1.20   0.229    -.2955984    .0708146
    _k_eq_m8 |  -.0008947   .0905761    -0.01   0.992    -.1784318    .1766425
    _k_eq_m7 |   .0825609   .0871064     0.95   0.343    -.0881755    .2532973
    _k_eq_m6 |   .1123127   .0850227     1.32   0.187    -.0543394    .2789648
    _k_eq_m5 |   .1134179    .082994     1.37   0.172    -.0492577    .2760935
    _k_eq_m4 |   .1225368   .0815805     1.50   0.133    -.0373683    .2824418
    _k_eq_m3 |   .1437796   .0806316     1.78   0.075    -.0142655    .3018246
    _k_eq_m2 |   .0626728   .0801585     0.78   0.434     -.094445    .2197905
    _k_eq_p0 |   1.030078   .0803795    12.82   0.000      .872527    1.187629
    _k_eq_p1 |   .9553605    .081658    11.70   0.000     .7953035    1.115418
    _k_eq_p2 |   1.046724   .0825916    12.67   0.000     .8848376    1.208611
    _k_eq_p3 |   1.036179    .084252    12.30   0.000     .8710377     1.20132
    _k_eq_p4 |    1.00154   .0863377    11.60   0.000     .8323109     1.17077
    _k_eq_p5 |   1.153019   .0880671    13.09   0.000     .9803992    1.325638
    _k_eq_p6 |   1.000366   .0902809    11.08   0.000     .8234072    1.177324
    _k_eq_p7 |   1.030018   .0935097    11.02   0.000     .8467308    1.213306
    _k_eq_p8 |     .97777   .0975634    10.02   0.000     .7865371    1.169003
    _k_eq_p9 |   1.035819   .1016011    10.19   0.000     .8366717    1.234966
   _k_eq_p10 |   1.111073   .1076606    10.32   0.000     .9000485    1.322097
   _k_eq_p11 |   .9799806      .1153     8.50   0.000     .7539823    1.205979
   _k_eq_p12 |    1.16532   .1240796     9.39   0.000     .9221132    1.408527
   _k_eq_p13 |   .8056877    .143361     5.62   0.000     .5246875    1.086688
   _k_eq_p14 |   1.129944   .1649623     6.85   0.000     .8066028    1.453284
   _k_eq_p15 |   .7956263   .2124408     3.75   0.000     .3792235    1.212029
   _k_eq_p16 |   .7021863   .3051937     2.30   0.021     .1039793    1.300393
   _k_eq_p17 |   1.457152   .5985379     2.43   0.015     .2839648     2.63034
         eta |   .2519384   .0047391    53.16   0.000     .2426494    .2612275
             |
           t |
          2  |   .1434719   .0450851     3.18   0.001     .0551011    .2318427
          3  |   .3312757   .0451509     7.34   0.000     .2427759    .4197754
          4  |   .5528668   .0452825    12.21   0.000      .464109    .6416247
          5  |   .7741836   .0454452    17.04   0.000     .6851069    .8632602
          6  |   .9513304   .0455863    20.87   0.000     .8619772    1.040684
          7  |   1.116552   .0457312    24.42   0.000     1.026915    1.206189
          8  |   1.394627   .0458923    30.39   0.000     1.304674     1.48458
          9  |   1.574857   .0460787    34.18   0.000     1.484539    1.665176
         10  |   1.803088   .0462321    39.00   0.000     1.712469    1.893707
         11  |   1.950842    .046446    42.00   0.000     1.859804    2.041881
         12  |   2.123341   .0466535    45.51   0.000     2.031896    2.214786
         13  |   2.361859    .046878    50.38   0.000     2.269974    2.453744
         14  |   2.511809   .0470803    53.35   0.000     2.419527    2.604091
         15  |   2.747289   .0472441    58.15   0.000     2.654686    2.839891
         16  |   2.968836   .0475149    62.48   0.000     2.875702     3.06197
         17  |   3.174446   .0477045    66.54   0.000     3.080941    3.267952
         18  |   3.336821   .0479173    69.64   0.000     3.242898    3.430743
         19  |   3.544168   .0481713    73.57   0.000     3.449748    3.638588
         20  |   3.743031   .0486024    77.01   0.000     3.647766    3.838296
             |
       _cons |  -.1941443   .1921508    -1.01   0.312     -.570777    .1824883
------------------------------------------------------------------------------
F test of absorbed indicators: F(999, 18944) = 21.600         Prob > F = 0.000

. 
. * Test finding and estimating with the widest window with balanced time periods for all units 
. * expect an error message because balanced window is too narrow 
. cap noi xtevent y eta , panelvar(i) timevar(t) policyvar(z) impute(stag) window(balanced) plot

No proxy or instruments provided. Implementing OLS estimator
The calculated window by window(balanced) is (-1,-1), plus the endpoints -2 and 0.

Left window can not be positive and right window can not be negative.

Check for first-treated units and last-treated units. Both types of units might have few common periods around t
> reatment time which causes a narrow calculated window.

. 
. * Test normalizations
. 
. xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) norm(-1) plot

No proxy or instruments provided. Implementing OLS estimator

. xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) norm(-2) plot

No proxy or instruments provided. Implementing OLS estimator

. xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) norm(-6) plot

No proxy or instruments provided. Implementing OLS estimator

. * Should fail
. * xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) norm(-7) plot
. xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) norm(1) plot

No proxy or instruments provided. Implementing OLS estimator

. xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) norm(5) plot

No proxy or instruments provided. Implementing OLS estimator

. * Should fail
. * xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) norm(7) plot
. graph drop _all

. 
. * Test exclusion of unbalanced units with ambiguous eventtime
. 
. gen z2 = z

. replace z2 = . if i==1 & t==7 
(1 real change made, 1 to missing)

. xtevent y eta, policyvar(z2) window(5)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

Unit 1 not used because of ambiguous event-time due to missing values in policyvar.

Linear regression, absorbing indicators             Number of obs     =  8,991
Absorbed variable: i                                No. of categories =    999
                                                    F(21, 7971)       = 209.76
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7546
                                                    Adj R-squared     = 0.7233
                                                    Root MSE          = 0.9952

. drop z2

. 
. * Test reghdfe
. xtevent y eta , panelvar(i) timevar(t) policyvar(z) window(3) reghdfe plot

No proxy or instruments provided. Implementing OLS estimator
(MWFE estimator converged in 2 iterations)

HDFE Linear regression                            Number of obs   =     13,000
Absorbing 2 HDFE groups                           F(   9,  11979) =     341.82
                                                  Prob > F        =     0.0000
                                                  R-squared       =     0.7653
                                                  Adj R-squared   =     0.7453
                                                  Within R-sq.    =     0.2043
                                                  Root MSE        =     1.0058

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
           i |      1000           0        1000     |
           t |        13           1          12     |
-----------------------------------------------------+

. 
. * Test reghdfe, proxy and absorbing a variable 
. gen k=round(x) //generate a categorical variable. Use it as a control

. xtevent y eta, policyvar(z) window(3) proxy(x) nofe note addabsorb(k) reghdfe

Using options panelvar and timevar from xtset

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.
(dropped 2 singleton observations)
(MWFE estimator converged in 1 iterations)

IV (2SLS) estimation
--------------------

Estimates efficient for homoskedasticity only
Statistics consistent for homoskedasticity only

Number of obs =    12998
                                                      F(  9, 12960) =     0.53
                                                      Prob > F      =   0.8521
Total (centered) SS     =  39316.59193                Centered R2   = -4.5e+02
Total (uncentered) SS   =  39316.59193                Uncentered R2 = -4.5e+02
Residual SS             =  17592667.23                Root MSE      =    36.84

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
           x |  -128.9603   4714.768    -0.03   0.978    -9370.598    9112.677
         eta |   3.057847   103.4982     0.03   0.976    -199.8138    205.9294
    _k_eq_m4 |  -3.015101   105.0347    -0.03   0.977    -208.8986    202.8684
    _k_eq_m3 |  -3.201484   119.8647    -0.03   0.979     -238.154     231.751
    _k_eq_p0 |  -1.328407   88.66188    -0.01   0.988    -175.1187    172.4619
    _k_eq_p1 |   .9751639   6.763573     0.14   0.885    -12.28243    14.23276
    _k_eq_p2 |   -4.00355   192.9401    -0.02   0.983    -382.1945    374.1874
    _k_eq_p3 |   .9918023   12.95882     0.08   0.939    -24.40938    26.39299
    _k_eq_p4 |   -2.13739   139.3771    -0.02   0.988    -275.3369    271.0621
------------------------------------------------------------------------------
Underidentification test (Anderson canon. corr. LM statistic):           0.001
                                                   Chi-sq(1) P-val =    0.9781
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic):                0.001
Stock-Yogo weak ID test critical values: 10% maximal IV size             16.38
                                         15% maximal IV size              8.96
                                         20% maximal IV size              6.66
                                         25% maximal IV size              5.53
Source: Stock-Yogo (2005).  Reproduced by permission.
------------------------------------------------------------------------------
Sargan statistic (overidentification test of all instruments):           0.000
                                                 (equation exactly identified)
------------------------------------------------------------------------------
Instrumented:         x
Included instruments: eta _k_eq_m4 _k_eq_m3 _k_eq_p0 _k_eq_p1 _k_eq_p2 _k_eq_p3
                      _k_eq_p4
Excluded instruments: _fd1z
Partialled-out:       _cons
                      nb: total SS, model F and R2s are after partialling-out;
                          any small-sample adjustments include partialled-out
                          variables in regressor count K
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
           k |        29           0          29     |
-----------------------------------------------------+

. 
. *Test additional cluster and robust specifications
. xtevent y eta, policyvar(z) window(3) proxy(x) nofe note addabsorb(k) reghdfe robust cluster(i)

Using options panelvar and timevar from xtset

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.
(dropped 2 singleton observations)
(MWFE estimator converged in 1 iterations)

IV (2SLS) estimation
--------------------

Estimates efficient for homoskedasticity only
Statistics robust to heteroskedasticity and clustering on i

Number of clusters (i) =          1000                Number of obs =    12998
                                                      F(  9,   999) =     0.58
                                                      Prob > F      =   0.8152
Total (centered) SS     =  39316.59193                Centered R2   = -4.5e+02
Total (uncentered) SS   =  39316.59193                Uncentered R2 = -4.5e+02
Residual SS             =  17592667.23                Root MSE      =    36.84

------------------------------------------------------------------------------
             |               Robust
           y | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
           x |  -128.9603   4423.488    -0.03   0.977    -8809.354    8551.433
         eta |   3.057847   97.09499     0.03   0.975    -187.4757    193.5914
    _k_eq_m4 |  -3.015101   98.55369    -0.03   0.976    -196.4111    190.3809
    _k_eq_m3 |  -3.201484   112.5611    -0.03   0.977    -224.0847    217.6817
    _k_eq_p0 |  -1.328407    83.1373    -0.02   0.987    -164.4722    161.8154
    _k_eq_p1 |   .9751639   6.380948     0.15   0.879    -11.54643    13.49676
    _k_eq_p2 |   -4.00355   180.9694    -0.02   0.982    -359.1274    351.1203
    _k_eq_p3 |   .9918023   12.13761     0.08   0.935    -22.82634    24.80994
    _k_eq_p4 |   -2.13739   130.7144    -0.02   0.987    -258.6437    254.3689
------------------------------------------------------------------------------
Underidentification test (Kleibergen-Paap rk LM statistic):              0.001
                                                   Chi-sq(1) P-val =    0.9767
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic):                0.001
                         (Kleibergen-Paap rk Wald F statistic):          0.001
Stock-Yogo weak ID test critical values: 10% maximal IV size             16.38
                                         15% maximal IV size              8.96
                                         20% maximal IV size              6.66
                                         25% maximal IV size              5.53
Source: Stock-Yogo (2005).  Reproduced by permission.
NB: Critical values are for Cragg-Donald F statistic and i.i.d. errors.
------------------------------------------------------------------------------
Hansen J statistic (overidentification test of all instruments):         0.000
                                                 (equation exactly identified)
------------------------------------------------------------------------------
Instrumented:         x
Included instruments: eta _k_eq_m4 _k_eq_m3 _k_eq_p0 _k_eq_p1 _k_eq_p2 _k_eq_p3
                      _k_eq_p4
Excluded instruments: _fd1z
Partialled-out:       _cons
                      nb: total SS, model F and R2s are after partialling-out;
                          any small-sample adjustments include partialled-out
                          variables in regressor count K
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
           k |        29           0          29     |
-----------------------------------------------------+

. *equivalent specification (it admits use of vce)
. xtevent y eta, policyvar(z) window(3) proxy(x) nofe note addabsorb(k) reghdfe vce(robust cluster i)

Using options panelvar and timevar from xtset

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.
(dropped 2 singleton observations)
(MWFE estimator converged in 1 iterations)

IV (2SLS) estimation
--------------------

Estimates efficient for homoskedasticity only
Statistics robust to heteroskedasticity and clustering on i

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
           k |        29           0          29     |
-----------------------------------------------------+

. 
. /*
> *Test other standard-error specifications (not allowed)
> xtevent y eta, policyvar(z) window(3) proxy(x) nofe note addabsorb(k) reghdfe vce(bootstrap) //will show an er
> ror message
> */
. 
. *Imputation of policyvar without verifying staggered adoption conditions
. xtevent y eta, policyvar(z) timevar(t) window(5) impute(nuchange)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators            Number of obs     =  20,000
Absorbed variable: i                               No. of categories =   1,000
                                                   F(32, 18968)      = 1227.00
                                                   Prob > F          =  0.0000
                                                   R-squared         =  0.8019
                                                   Adj R-squared     =  0.7911
                                                   Root MSE          =  1.0025

. 
. *outer imputation of policyvar verifying staggered adoption conditions
. xtevent y eta, policyvar(z) timevar(t) window(5) impute(stag)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. 
. *outer and inner imputation of policyvar verifying staggered adoption conditions
. xtevent y eta, policyvar(z) timevar(t) window(5) impute(instag)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. 
. *outer and inner mputation of policyvar. Adds the imputed policyvar to the database
. xtevent y eta, policyvar(z) timevar(t) window(5) impute(instag, saveimp)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. drop z_imputed

. 
. *imputation fails if staggered conditions are not satisfied. It reverts to no imputation
. replace z=0.5 in 7
(1 real change made)

. xtevent y eta, policyvar(z) timevar(t) window(5) impute(instag)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator
The policy variable is not binary. Assuming non-staggered adoption (no imputation).
If event dummies and variables are saved, event-time will be missing.

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(21, 7979)       = 209.77
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7545
                                                    Adj R-squared     = 0.7232
                                                    Root MSE          = 0.9954

. replace z=1 in 7
(1 real change made)

. 
. *Difference in averages between the post and pre-period
. xtevent y eta , panelvar(i) timevar(t) policyvar(z) window(5) diffavg

No proxy or instruments provided. Implementing OLS estimator

Difference in pre and post-period averages from lincom:

( 1)  - .1666667*_k_eq_m6 - .1666667*_k_eq_m5 - .1666667*_k_eq_m4 - .1666667*_k_eq_m3 - .1666667*_k_eq_m2 +
       .1428571*_k_eq_p0 + .1428571*_k_eq_p1 + .1428571*_k_eq_p2 + .1428571*_k_eq_p3 + .1428571*_k_eq_p4 +
       .1428571*_k_eq_p5 + .1428571*_k_eq_p6 = 0

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
         (1) |      .9247     .07987    11.58   0.000        .7682       1.081
------------------------------------------------------------------------------

. xtevent y eta , panelvar(i) timevar(t) policyvar(z) window(4) diff

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     = 11,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(21, 9979)       = 364.11
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7597
                                                    Adj R-squared     = 0.7351
                                                    Root MSE          = 0.9988

Difference in pre and post-period averages from lincom:

( 1)  - .2*_k_eq_m5 - .2*_k_eq_m4 - .2*_k_eq_m3 - .2*_k_eq_m2 + .1666667*_k_eq_p0 + .1666667*_k_eq_p1 +
       .1666667*_k_eq_p2 + .1666667*_k_eq_p3 + .1666667*_k_eq_p4 + .1666667*_k_eq_p5 = 0

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
         (1) |      .9229     .05989    15.41   0.000        .8055        1.04
------------------------------------------------------------------------------

. xtevent y eta , panelvar(i) timevar(t) policyvar(z) window(5) norm(1) diff

No proxy or instruments provided. Implementing OLS estimator

Difference in pre and post-period averages from lincom:

( 1)  - .1666667*_k_eq_m6 - .1666667*_k_eq_m5 - .1666667*_k_eq_m4 - .1666667*_k_eq_m3 - .1666667*_k_eq_m2 -
       .1666667*_k_eq_m1 + .1428571*_k_eq_p0 + .1428571*_k_eq_p2 + .1428571*_k_eq_p3 + .1428571*_k_eq_p4 +
       .1428571*_k_eq_p5 + .1428571*_k_eq_p6 = 0

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
         (1) |      .9247     .07987    11.58   0.000        .7682       1.081
------------------------------------------------------------------------------

. xtevent y eta , panelvar(i) timevar(t) policyvar(z) window(5) norm(2) diff

No proxy or instruments provided. Implementing OLS estimator

Difference in pre and post-period averages from lincom:

( 1)  - .1666667*_k_eq_m6 - .1666667*_k_eq_m5 - .1666667*_k_eq_m4 - .1666667*_k_eq_m3 - .1666667*_k_eq_m2 -
       .1666667*_k_eq_m1 + .1428571*_k_eq_p0 + .1428571*_k_eq_p1 + .1428571*_k_eq_p3 + .1428571*_k_eq_p4 +
       .1428571*_k_eq_p5 + .1428571*_k_eq_p6 = 0

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
         (1) |      .9247     .07987    11.58   0.000        .7682       1.081
------------------------------------------------------------------------------

. xtevent y eta , panelvar(i) timevar(t) policyvar(z) window(5) norm(-2) diff

No proxy or instruments provided. Implementing OLS estimator

Difference in pre and post-period averages from lincom:

( 1)  - .1666667*_k_eq_m6 - .1666667*_k_eq_m5 - .1666667*_k_eq_m4 - .1666667*_k_eq_m3 - .1666667*_k_eq_m1 +
       .1428571*_k_eq_p0 + .1428571*_k_eq_p1 + .1428571*_k_eq_p2 + .1428571*_k_eq_p3 + .1428571*_k_eq_p4 +
       .1428571*_k_eq_p5 + .1428571*_k_eq_p6 = 0

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
         (1) |      .9247     .07987    11.58   0.000        .7682       1.081
------------------------------------------------------------------------------

. 
. *Trend adjustment. Default method is GMM
. xtevent y eta, policyvar(z) timevar(t) window(5) trend(-3)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. 
. *Trend adjustment. Use OLS instead
. xtevent y eta, policyvar(z) timevar(t) window(5) trend(-3, method(ols))

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(20, 7980)       = 220.17
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7545
                                                    Adj R-squared     = 0.7231
                                                    Root MSE          = 0.9954

. 
. *Compare: 1) no adjustment; 2) adjustment by GMM and; 3) adjustment by OLS
. xtevent y eta, policyvar(z) timevar(t) window(5)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. xteventplot, name(g1)

. xtevent y eta, policyvar(z) timevar(t) window(5) trend(-3, method(gmm))

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. xteventplot, name(g2)

. xtevent y eta, policyvar(z) timevar(t) window(5) trend(-3, method(ols))

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. xteventplot, name(g3)

. 
. graph combine g1 g2 g3, rows(2)

. 
. graph drop _all

. 
. *Sun and Abraham Estimator (2021) 
. *Generate cohort indicator
. * This works because of staggered adoption
. gen timet=t if z==1
(17,111 missing values generated)

. by i: egen time_of_treat=min(timet)
(13,720 missing values generated)

. *Generate control cohort indicator. We use the never treated units as the control cohort. 
. gen never_treat=time_of_treat==.

. *Could also use last trated as the control group
. egen last_treat_time = max(time_of_treat)

. gen last_treat = (time_of_treat == last_treat_time)

. replace last_treat = . if time_of_treat == .
(13,720 real changes made, 13,720 to missing)

. gen cohort_for_last_treat = time_of_treat
(13,720 missing values generated)

. replace cohort_for_last_treat = . if last_treat
(340 real changes made, 340 to missing)

. 
. *Estimate the event-time coefficients with the Sun-and-Abraham Estimator.
. xtevent y eta , policyvar(z) window(5) vce(cluster i) impute(nuchange) cohort(variable time_of_treat) control_
> cohort(variable never_treat)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

You have specified the cohort or the sunabraham option
Event-time coefficients will be estimated with the Interaction Weighted Estimator of Sun and Abraham (2021)

Linear regression, absorbing indicators            Number of obs     =  20,000
Absorbed variable: i                               No. of categories =   1,000
                                                   F(32, 999)        = 1319.19
                                                   Prob > F          =  0.0000
                                                   R-squared         =  0.8019
                                                   Adj R-squared     =  0.7911
                                                   Root MSE          =  1.0025
                                  (Std. err. adjusted for 1,000 clusters in i)
------------------------------------------------------------------------------
             |               Robust
           y | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    _k_eq_m6 |   .0457452   .0747931     0.61   0.541    -.1010245    .1925149
    _k_eq_m5 |   .1189648   .0836275     1.42   0.155    -.0451409    .2830704
    _k_eq_m4 |   .1165655   .0842832     1.38   0.167    -.0488269    .2819579
    _k_eq_m3 |   .1449931   .0775793     1.87   0.062    -.0072439    .2972301
    _k_eq_m2 |   .0638224   .0835375     0.76   0.445    -.1001067    .2277515
    _k_eq_p0 |   1.027701   .0821568    12.51   0.000      .866481     1.18892
    _k_eq_p1 |   .9636293   .0898445    10.73   0.000     .7873238    1.139935
    _k_eq_p2 |    1.08301   .0852336    12.71   0.000     .9157528    1.250268
    _k_eq_p3 |   1.046513   .0892559    11.72   0.000      .871362    1.221663
    _k_eq_p4 |   .9859918   .0880341    11.20   0.000     .8132388    1.158745
    _k_eq_p5 |    1.13976   .0920912    12.38   0.000     .9590461    1.320475
    _k_eq_p6 |   .9998216   .0864235    11.57   0.000     .8302292    1.169414
         eta |   .2525003    .004768    52.96   0.000     .2431439    .2618567
             |
           t |
          2  |    .131826   .0457042     2.88   0.004     .0421387    .2215133
          3  |   .3236485   .0456104     7.10   0.000     .2341453    .4131517
          4  |   .5422926   .0461881    11.74   0.000     .4516557    .6329296
          5  |   .7656458   .0462201    16.57   0.000     .6749462    .8563454
          6  |   .9441123   .0466213    20.25   0.000     .8526255    1.035599
          7  |   1.110485   .0462219    24.03   0.000     1.019782    1.201188
          8  |   1.391892   .0456709    30.48   0.000      1.30227    1.481514
          9  |   1.569177   .0458476    34.23   0.000     1.479208    1.659145
         10  |   1.798732   .0457135    39.35   0.000     1.709027    1.888438
         11  |   1.944693    .045547    42.70   0.000     1.855314    2.034071
         12  |   2.118919   .0472219    44.87   0.000     2.026253    2.211584
         13  |   2.358402   .0486343    48.49   0.000     2.262965    2.453839
         14  |   2.508353   .0471617    53.19   0.000     2.415806    2.600901
         15  |   2.743864   .0474198    57.86   0.000      2.65081    2.836918
         16  |    2.96558   .0471802    62.86   0.000     2.872996    3.058164
         17  |   3.171225    .048172    65.83   0.000     3.076695    3.265755
         18  |   3.334319   .0476294    70.01   0.000     3.240854    3.427784
         19  |   3.537133   .0477691    74.05   0.000     3.443394    3.630872
         20  |   3.737882   .0478568    78.11   0.000     3.643971    3.831794
             |
       _cons |   .2224166   .0715959     3.11   0.002      .081921    .3629121
------------------------------------------------------------------------------

. *Use reghdfe as the underlying estimation command
. xtevent y eta , policyvar(z) window(5) vce(cluster i) impute(nuchange) cohort(variable time_of_treat) control_
> cohort(variable never_treat) reghdfe

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

You have specified the cohort or the sunabraham option
Event-time coefficients will be estimated with the Interaction Weighted Estimator of Sun and Abraham (2021)

HDFE Linear regression                                  Number of obs = 20,000
Absorbing 2 HDFE groups                                 F(13, 999)    = 619.55
                                                        Prob > F      = 0.0000
                                                        R-squared     = 0.8019
                                                        Adj R-squared = 0.7911
                                                        Root MSE      = 1.0026
                                  (Std. err. adjusted for 1,000 clusters in i)
------------------------------------------------------------------------------
             |               Robust
           y | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    _k_eq_m6 |   .0457452   .0729014     0.63   0.530    -.0973123    .1888027
    _k_eq_m5 |   .1189648   .0815853     1.46   0.145    -.0411334    .2790629
    _k_eq_m4 |   .1165655    .082259     1.42   0.157    -.0448547    .2779857
    _k_eq_m3 |   .1449931   .0757206     1.91   0.056    -.0035966    .2935829
    _k_eq_m2 |   .0638224   .0815323     0.78   0.434    -.0961718    .2238166
    _k_eq_p0 |   1.027701   .0802026    12.81   0.000     .8703158    1.185086
    _k_eq_p1 |   .9636293   .0876691    10.99   0.000     .7915927    1.135666
    _k_eq_p2 |    1.08301   .0831531    13.02   0.000     .9198353    1.246185
    _k_eq_p3 |   1.046513   .0871074    12.01   0.000     .8755782    1.217447
    _k_eq_p4 |   .9859918   .0858694    11.48   0.000     .8174868    1.154497
    _k_eq_p5 |    1.13976   .0897592    12.70   0.000     .9636221    1.315899
    _k_eq_p6 |   .9998216   .0842215    11.87   0.000     .8345503    1.165093
         eta |   .2525003   .0046472    54.33   0.000     .2433809    .2616196
       _cons |   2.072326   .0593099    34.94   0.000     1.955939    2.188712
------------------------------------------------------------------------------

. * Automatic generation of cohort and control_cohort variables
. xtevent y eta , policyvar(z) window(5) vce(cluster i) impute(nuchange) cohort(create) reghdfe

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

You have specified the cohort or the sunabraham option
Event-time coefficients will be estimated with the Interaction Weighted Estimator of Sun and Abraham (2021)

Control cohort not specified. Using values with cohort variable == . as the control cohort

. 
. *Overlay trend plot
. xtevent y eta, policyvar(z) timevar(t) window(5) trend(-3, method(gmm) saveov)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. xteventplot, overlay(trend)

. 
. /*
> *Overlay trend plot fails because suboption "saveov" was not specified
> xtevent y eta, policyvar(z) timevar(t) window(5) trend(-3, method(gmm))
> xteventplot, overlay(trend)
> */
. 
. * Overlay static plot
. xtevent y eta, policyvar(z) timevar(t) window(5)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. xteventplot, overlay(static)
Estimating static model...

Using options panelvar and timevar from xtset

option static specified. Estimating static model

Plotting options ignored

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(10, 7990)       = 439.29
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7542
                                                    Adj R-squared     = 0.7231
                                                    Root MSE          = 0.9954

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
         eta |   .2611098   .0093119    28.04   0.000     .2428561    .2793635
           z |   .9033452   .0666819    13.55   0.000     .7726312    1.034059
             |
           t |
          8  |   .2844437   .0445517     6.38   0.000     .1971108    .3717765
          9  |    .462642   .0446181    10.37   0.000     .3751789    .5501051
         10  |   .6945258   .0447295    15.53   0.000     .6068444    .7822073
         11  |   .8423474    .044911    18.76   0.000     .7543101    .9303846
         12  |   1.016765   .0450953    22.55   0.000     .9283665    1.105164
         13  |     1.2599   .0453833    27.76   0.000     1.170937    1.348863
         14  |   1.409257   .0456502    30.87   0.000     1.319771    1.498744
         15  |   1.645633   .0458656    35.88   0.000     1.555724    1.735541
             |
       _cons |   1.384714   .0317884    43.56   0.000       1.3224    1.447027
------------------------------------------------------------------------------
F test of absorbed indicators: F(999, 7990) = 10.334          Prob > F = 0.000

. 
. * Test graphic options
. gen y2 = y - z

. xtevent y2 eta, policyvar(z) timevar(t) window(5)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(21, 7979)       = 156.66
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7090
                                                    Adj R-squared     = 0.6718
                                                    Root MSE          = 0.9954

------------------------------------------------------------------------------
          y2 | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    _k_eq_m6 |   .0692904   .1104722     0.63   0.531     -.147264    .2858448
    _k_eq_m5 |   .1651543   .1178905     1.40   0.161    -.0659419    .3962505
    _k_eq_m4 |   .1011972   .1131376     0.89   0.371     -.120582    .3229763
    _k_eq_m3 |   .1498392    .108524     1.38   0.167    -.0628963    .3625747
    _k_eq_m2 |   .0503755    .105976     0.48   0.635    -.1573651    .2581161
    _k_eq_p0 |  -.0107683   .1045981    -0.10   0.918    -.2158079    .1942714
    _k_eq_p1 |  -.0978807   .1058732    -0.92   0.355    -.3054199    .1096585
    _k_eq_p2 |   -.002144   .1091899    -0.02   0.984    -.2161848    .2118968
    _k_eq_p3 |  -.0379753    .114047    -0.33   0.739    -.2615371    .1855866
    _k_eq_p4 |  -.0088354   .1189692    -0.07   0.941    -.2420463    .2243754
    _k_eq_p5 |   .1832222   .1269706     1.44   0.149    -.0656733    .4321176
    _k_eq_p6 |   .0727966   .1182986     0.62   0.538    -.1590996    .3046928
         eta |   .2616307   .0096806    27.03   0.000     .2426541    .2806073
             |
           t |
          8  |    .282378   .0446025     6.33   0.000     .1949454    .3698106
          9  |   .4604956   .0448172    10.27   0.000     .3726422    .5483491
         10  |   .6894495   .0451663    15.26   0.000     .6009118    .7779873
         11  |    .835558   .0456976    18.28   0.000     .7459787    .9251373
         12  |   1.008256   .0464042    21.73   0.000     .9172915     1.09922
         13  |   1.247446   .0472958    26.38   0.000     1.154734    1.340159
         14  |   1.396196   .0481892    28.97   0.000     1.301732    1.490659
         15  |   1.629461   .0491454    33.16   0.000     1.533123    1.725798
             |
       _cons |   1.314217   .1041623    12.62   0.000     1.110031    1.518402
------------------------------------------------------------------------------
F test of absorbed indicators: F(999, 7979) = 10.317          Prob > F = 0.000

. xteventplot, smplotopts(lcolor(green)) smpath(line)

Note: Smoothest line drawn for system confidence level = 95%
Wald Critical Value 21.0260698
Order 0
Wald value 12.4887323

. xteventplot, ciplotopts(lcolor(green))

. xteventplot, suptciplotopts(lcolor(green))

. xteventplot, scatterplotopts(mcolor(green))

. xteventplot, overlay(static) staticovplotopts(lcolor(red))
Estimating static model...

Using options panelvar and timevar from xtset

option static specified. Estimating static model

Plotting options ignored

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(10, 7990)       = 327.96
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7086
                                                    Adj R-squared     = 0.6719
                                                    Root MSE          = 0.9954

------------------------------------------------------------------------------
          y2 | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
         eta |   .2611098   .0093119    28.04   0.000     .2428561    .2793635
           z |  -.0966548   .0666819    -1.45   0.147    -.2273688    .0340591
             |
           t |
          8  |   .2844437   .0445517     6.38   0.000     .1971108    .3717765
          9  |    .462642   .0446181    10.37   0.000     .3751789    .5501051
         10  |   .6945258   .0447295    15.53   0.000     .6068444    .7822073
         11  |   .8423474    .044911    18.76   0.000     .7543101    .9303846
         12  |   1.016765   .0450953    22.55   0.000     .9283665    1.105164
         13  |     1.2599   .0453833    27.76   0.000     1.170937    1.348863
         14  |   1.409257   .0456502    30.87   0.000     1.319771    1.498744
         15  |   1.645633   .0458656    35.88   0.000     1.555724    1.735541
             |
       _cons |   1.384714   .0317884    43.56   0.000       1.3224    1.447027
------------------------------------------------------------------------------
F test of absorbed indicators: F(999, 7990) = 10.334          Prob > F = 0.000

. 
. xtevent y2 eta, policyvar(z) timevar(t) window(5) trend(-3, saveov)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

------------------------------------------------------------------------------
          y2 | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    _k_eq_m6 |  -.3045795    .235951    -1.29   0.197    -.7671052    .1579461
    _k_eq_m5 |  -.1339416   .1179329    -1.14   0.256    -.3651209    .0972377
    _k_eq_m4 |  -.1231248   .1503292    -0.82   0.413    -.4178093    .1715598
    _k_eq_m3 |   .0002913   .1117368     0.00   0.998    -.2187421    .2193246
    _k_eq_m2 |  -.0243985   .0930362    -0.26   0.793    -.2067737    .1579767
    _k_eq_p0 |   .0640057   .1386137     0.46   0.644    -.2077133    .3357247
    _k_eq_p1 |   .0516673   .1846966     0.28   0.780    -.3103863    .4137208
    _k_eq_p2 |   .2221779   .2361656     0.94   0.347    -.2407685    .6851243
    _k_eq_p3 |   .2611206   .2901196     0.90   0.368    -.3075895    .8298308
    _k_eq_p4 |   .3650344    .345605     1.06   0.291    -.3124416    1.042511
    _k_eq_p5 |    .631866    .402778     1.57   0.117    -.1576841    1.421416
    _k_eq_p6 |   .5962144   .4556476     1.31   0.191    -.2969739    1.489403
         eta |   .2616307   .0096806    27.03   0.000     .2426541    .2806073
             |
           t |
          8  |    .282378   .0446025     6.33   0.000     .1949454    .3698106
          9  |   .4604956   .0448172    10.27   0.000     .3726422    .5483491
         10  |   .6894495   .0451663    15.26   0.000     .6009118    .7779873
         11  |    .835558   .0456976    18.28   0.000     .7459787    .9251373
         12  |   1.008256   .0464042    21.73   0.000     .9172915     1.09922
         13  |   1.247446   .0472958    26.38   0.000     1.154734    1.340159
         14  |   1.396196   .0481892    28.97   0.000     1.301732    1.490659
         15  |   1.629461   .0491454    33.16   0.000     1.533123    1.725798
             |
       _cons |   1.314217   .1041623    12.62   0.000     1.110031    1.518402
------------------------------------------------------------------------------
F test of absorbed indicators: F(999, 7979) = 10.317          Prob > F = 0.000

. xteventplot, overlay(trend) trendplotopts(lcolor(red))

. xtevent y eta, policyvar(z) proxy(x) window(5)

Using options panelvar and timevar from xtset

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =      9,000
Group variable: i                               Number of groups  =      1,000

R-squared:                                      Obs per group:
     Within  = 0.2538                                         min =          9
     Between = 0.4789                                         avg =        9.0
     Overall = 0.3968                                         max =          9

Wald chi2(21)     =   47985.01
corr(u_i, Xb) = -0.0138                         Prob > chi2       =     0.0000

. xteventplot, overlay(iv) scatterplotopts(mcolor(green red))

. xtevent y eta , panelvar(i) timevar(t) policyvar(z) window(5)

No proxy or instruments provided. Implementing OLS estimator

. xteventplot, textboxoption(color(blue) size(large))

. drop y2

. 
. ******** Repeated cross-sectional data 
. use "small_repeated_cross_sectional_example31.dta", clear

. xtset, clear

. *OLS, impute and trend adjustment
. xtevent y, panelvar(state) t(t) policyvar(z) window(5) trend(-3, method(ols)) impute(instag) repeatedcs

Option repeatedcs was specified. Using state as the panel variable and t as the time variable.

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     = 20,000
Absorbed variable: state                            No. of categories =     46
                                                    F(30, 19924)      = 501.73
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.4891
                                                    Adj R-squared     = 0.4872
                                                    Root MSE          = 1.5321

. xteventplot

. *IV
. xtevent y, panelvar(state) t(t) policyvar(z) window(5) impute(stag) proxy(x) repeatedcs

Option repeatedcs was specified. Using state as the panel variable and t as the time variable.

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Using reghdfe for fixed effects estimation with repeated cross-sectional data.
(MWFE estimator converged in 4 iterations)

IV (2SLS) estimation
--------------------

Estimates efficient for homoskedasticity only
Statistics consistent for homoskedasticity only

Number of obs =    19005
                                                      F( 12, 18929) =   123.84
                                                      Prob > F      =   0.0000
Total (centered) SS     =   47851.7849                Centered R2   =   0.0360
Total (uncentered) SS   =   47851.7849                Uncentered R2 =   0.0360
Residual SS             =  46127.15415                Root MSE      =    1.561

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
       state |        46           0          46     |
           t |        19           1          18     |
-----------------------------------------------------+

. xteventplot

. xteventplot, overlay(iv)

. 
. *static ols
. xtevent y, panelvar(state) t(t) policyvar(z) impute(stag) static repeatedcs

Option repeatedcs was specified. Using state as the panel variable and t as the time variable.

option static specified. Estimating static model

Plotting options ignored

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     = 20,000
Absorbed variable: state                            No. of categories =     46
                                                    F(20, 19934)      = 744.21
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.4866
                                                    Adj R-squared     = 0.4849
                                                    Root MSE          = 1.5356

. *static IV
. xtevent y, panelvar(state) t(t) policyvar(z) impute(stag) proxy(x) static repeatedcs

Option repeatedcs was specified. Using state as the panel variable and t as the time variable.

option static specified. Estimating static model

Plotting options ignored

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

Fixed-effects (within) IV regression            Number of obs     =     19,005
Group variable: state                           Number of groups  =         46

R-squared:                                      Obs per group:
     Within  = 0.4256                                         min =        249
     Between = 0.9838                                         avg =      413.2
     Overall = 0.4862                                         max =        486

Wald chi2(20)     =   51510.98
corr(u_i, Xb) = -0.0229                         Prob > chi2       =     0.0000

. 
. *get unit time effects
. get_unit_time_effects y u eta, panelvar(state) timevar(t) saving("effect_file.dta", replace)

Linear regression, absorbing indicators            Number of obs     =  20,000
Absorbed variable: unittimeinteraction             No. of categories =     920
                                                   F(2, 19078)       = 1504.21
                                                   Prob > F          =  0.0000
                                                   R-squared         =  0.5781
                                                   Adj R-squared     =  0.5577
                                                   Root MSE          =  1.4229

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
           u |   -.001644   .0051352    -0.32   0.749    -.0117095    .0084214
         eta |   .2550361   .0046498    54.85   0.000     .2459221    .2641501
       _cons |   2.251179   .0100644   223.68   0.000     2.231452    2.270906
------------------------------------------------------------------------------
F test of absorbed indicators: F(919, 19078) = 18.418         Prob > F = 0.000
file effect_file.dta saved

. 
. *get_unit_time_effects + xtevent 
. get_unit_time_effects y u eta, panelvar(state) timevar(t) saving("effect_file.dta", replace)

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
           u |   -.001644   .0051352    -0.32   0.749    -.0117095    .0084214
         eta |   .2550361   .0046498    54.85   0.000     .2459221    .2641501
       _cons |   2.251179   .0100644   223.68   0.000     2.231452    2.270906
------------------------------------------------------------------------------
F test of absorbed indicators: F(919, 19078) = 18.418         Prob > F = 0.000
file effect_file.dta saved

. bysort state t (z): keep if _n==1
(19,080 observations deleted)

. keep state t z

. merge m:1 state t using "effect_file.dta"

Result                      Number of obs
    -----------------------------------------
    Not matched                             0
    Matched                               920  (_merge==3)
    -----------------------------------------

. drop _merge

. xtevent _unittimeeffects, panelvar(state) t(t) policyvar(z) window(5)

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =    414
Absorbed variable: state                            No. of categories =     46
                                                    F(20, 348)        =  79.91
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.8504
                                                    Adj R-squared     = 0.8225
                                                    Root MSE          = 0.3233

------------------------------------------------------------------------------
_unittimee~s | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
    _k_eq_m6 |    .040263   .1525886     0.26   0.792     -.259849    .3403749
    _k_eq_m5 |     .19858   .1646535     1.21   0.229    -.1252611    .5224211
    _k_eq_m4 |    .224199   .1603164     1.40   0.163     -.091112    .5395099
    _k_eq_m3 |   .2520825   .1566749     1.61   0.109    -.0560663    .5602312
    _k_eq_m2 |   .1545788   .1539657     1.00   0.316    -.1482416    .4573991
    _k_eq_p0 |   1.144426   .1539818     7.43   0.000     .8415734    1.447278
    _k_eq_p1 |   1.220535   .1567996     7.78   0.000     .9121414     1.52893
    _k_eq_p2 |   1.071692   .1607148     6.67   0.000     .7555971    1.387786
    _k_eq_p3 |   1.102224   .1685971     6.54   0.000     .7706266    1.433822
    _k_eq_p4 |   1.269389   .1780274     7.13   0.000     .9192437    1.619534
    _k_eq_p5 |   1.113971   .1894849     5.88   0.000     .7412915    1.486651
    _k_eq_p6 |    1.33035   .1752273     7.59   0.000     .9857118    1.674987
             |
           t |
          8  |   .1635674   .0675319     2.42   0.016     .0307455    .2963894
          9  |   .4627374   .0678297     6.82   0.000     .3293296    .5961452
         10  |   .6030953   .0683276     8.83   0.000     .4687083    .7374823
         11  |   .7896641   .0690418    11.44   0.000     .6538725    .9254558
         12  |   1.071347   .0700234    15.30   0.000     .9336251     1.20907
         13  |   1.211792   .0712616    17.00   0.000     1.071634    1.351949
         14  |    1.42655   .0727431    19.61   0.000     1.283479    1.569622
         15  |   1.692904   .0744535    22.74   0.000     1.546468    1.839339
             |
       _cons |  -.9467688   .1473088    -6.43   0.000    -1.236496   -.6570413
------------------------------------------------------------------------------
F test of absorbed indicators: F(45, 348) = 1.154             Prob > F = 0.240

. xteventplot

. 
. *------------------------ 2.2: Replicate 2b and test basic funcionality without controls ---------------------
> -------------
. 
. * load panel dataset
. use "example31.dta", clear

. 
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(20, 7980)       = 168.25
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7320
                                                    Adj R-squared     = 0.6978
                                                    Root MSE          = 1.0399

. 
. * Testing xtset options
. 
. xtevent y , policyvar(z) window(5) plot

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. xtevent y , policyvar(z) panelvar(i) window(5) plot

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. xtevent y , policyvar(z) timevar(t) window(5) plot

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. xteventplot, noci

option noci has been specified. Confidence intervals won't be displayed

. xteventplot, nosupt

option nosupt has been specified. Sup-t confidence intervals won't be displayed or calculated

. xteventplot, nozeroline

option nozeroline has been specified. The reference line at 0 won't be displayed

. xteventplot, nonormlabel

option nonormlabel has been specified. The label for the value of the dependent variable at event time correspon
> ding to the normalized coefficient won't be displayed

. 
. * Test if/in
. xtevent y if i<100, panelvar(i) timevar(t) policyvar(z) window(5) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =    891
Absorbed variable: i                                No. of categories =     99
                                                    F(20, 772)        =  14.33
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7459
                                                    Adj R-squared     = 0.7071
                                                    Root MSE          = 1.0907

. xtevent y in 1/600 , panelvar(i) timevar(t) policyvar(z) window(5) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =    270
Absorbed variable: i                                No. of categories =     30
                                                    F(20, 220)        =   5.07
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7506
                                                    Adj R-squared     = 0.6950
                                                    Root MSE          = 1.0124

. xtevent y in 1/600 if i<30 , panelvar(i) timevar(t) policyvar(z) window(5) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =    261
Absorbed variable: i                                No. of categories =     29
                                                    F(20, 212)        =   4.66
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7373
                                                    Adj R-squared     = 0.6778
                                                    Root MSE          = 1.0141

. 
. * Test nofe, note
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) nofe plot

No proxy or instruments provided. Implementing OLS estimator

Source |       SS           df       MS      Number of obs   =     9,000
-------------+----------------------------------   F(20, 8979)     =    223.83
       Model |  10714.0449        20  535.702247   Prob > F        =    0.0000
    Residual |  21489.5662     8,979  2.39331398   R-squared       =    0.3327
-------------+----------------------------------   Adj R-squared   =    0.3312
       Total |  32203.6112     8,999  3.57857664   Root MSE        =     1.547

. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) nofe note plot

No proxy or instruments provided. Implementing OLS estimator

Source |       SS           df       MS      Number of obs   =     9,000
-------------+----------------------------------   F(12, 8987)     =    294.16
       Model |  9081.69274        12  756.807728   Prob > F        =    0.0000
    Residual |  23121.9184     8,987  2.57281834   R-squared       =    0.2820
-------------+----------------------------------   Adj R-squared   =    0.2810
       Total |  32203.6112     8,999  3.57857664   Root MSE        =     1.604

. 
. * Test smoothest line
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(3)

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     = 13,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(20, 11980)      = 472.86
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7371
                                                    Adj R-squared     = 0.7148
                                                    Root MSE          = 1.0644

. xteventplot, smpath(scatter)

The optimization to calculate the smoothest path returned an error code.
Smoothest path won't be displayed.
Try changing the optimization options. For example, try -smpath(scatter, tech(dfp)-.

Error code = 9. See mf_optimize##r_error to see what that means.
Error code = 9. See mf_optimize##r_error to see what that means.

. xteventplot, smpath(line)

The optimization to calculate the smoothest path returned an error code.
Smoothest path won't be displayed.
Try changing the optimization options. For example, try -smpath(scatter, tech(dfp)-.

Error code = 9. See mf_optimize##r_error to see what that means.
Error code = 9. See mf_optimize##r_error to see what that means.

. /* maximum order allowed is 10
> xteventplot, smpath(line, maxorder(25))
> xteventplot, smpath(line, maxorder(30))
> */
. xteventplot, smpath(line, technique("nr 10 bfgs 10"))

Note: Smoothest line drawn for system confidence level = 95%
Wald Critical Value 15.5073131
Order 0
Wald value 1464.06384
Order 1
Wald value 269.975294
Order 2
Wald value 193.714313
Order 3
Wald value 137.593424
Order 4
Wald value 128.489737
Order 5
Wald value 83.1331862
Order 6
Wald value 58.1316458
Order 7
Wald value 48.5284897
Order 8
Wald value 8.0986e-11
(setting technique to nr)
Iteration 0:  f(p) =  4.840e+08  (not concave)
Iteration 1:  f(p) =  4.838e+08  (not concave)
Iteration 2:  f(p) =  4.838e+08  (not concave)
Iteration 3:  f(p) =  4.838e+08  (not concave)
Iteration 4:  f(p) =  4.838e+08  
Iteration 5:  f(p) =  4.838e+08  (not concave)
Iteration 6:  f(p) =  4.838e+08  (not concave)
Iteration 7:  f(p) =  4.838e+08  (not concave)
Iteration 8:  f(p) =  4.837e+08  (not concave)
Iteration 9:  f(p) =  4.835e+08  (not concave)
(switching technique to bfgs)
Iteration 10: f(p) =  4.831e+08  
Iteration 11: f(p) =  4.831e+08  (backed up)
Iteration 12: f(p) =  4.831e+08  
Iteration 13: f(p) =  4.831e+08  
Iteration 14: f(p) =  4.831e+08  
Iteration 15: f(p) =  4.831e+08  
Iteration 16: f(p) =  4.831e+08  
Iteration 17: f(p) =  4.831e+08  (backed up)
Iteration 18: f(p) =  4.831e+08  (backed up)
Iteration 19: f(p) =  4.831e+08  (backed up)
(switching technique to nr)
Iteration 20: f(p) =  4.831e+08  (not concave)
Iteration 21: f(p) =  4.830e+08  (not concave)
Iteration 22: f(p) =  4.830e+08  (not concave)
Iteration 23: f(p) =  4.830e+08  (not concave)
Iteration 24: f(p) =  4.830e+08  (not concave)
Iteration 25: f(p) =  4.830e+08  
Hessian is not positive semidefinite
(setting technique to nr)
Iteration 0:  f(p) =  4.840e+08  (not concave)
Iteration 1:  f(p) =  4.838e+08  (not concave)
Iteration 2:  f(p) =  4.838e+08  (not concave)
Iteration 3:  f(p) =  4.838e+08  (not concave)
Iteration 4:  f(p) =  4.838e+08  
Hessian is not positive semidefinite

The optimization to calculate the smoothest path returned an error code.
Smoothest path won't be displayed.
Try changing the optimization options. For example, try -smpath(scatter, tech(dfp)-.

Error code = 4. See mf_optimize##r_error to see what that means.
Error code = 4. See mf_optimize##r_error to see what that means.

. 
. 
. * Test more suptreps
. 
. cap graph drop g1

. cap graph drop g2

. 
. xtevent y, panelvar(i) timevar(t) policyvar(z) window(3)

No proxy or instruments provided. Implementing OLS estimator

. xteventplot, suptreps(20) name(g1)

. xteventplot, suptreps(1e6) name(g2)

. 
. graph combine g1 g2, rows(1)

. 
. graph drop g1

. graph drop g2

. 
. * Test savek
. xtevent y, panelvar(i) timevar(t) policyvar(z) window(5) savek(a)

No proxy or instruments provided. Implementing OLS estimator

. des a_eq*, s

. des a_evtime, s

. drop a*

. 
. * Test factor variables in varlist
. cap gen pois=rpoisson(5)

. xtevent y i.pois, panelvar(i) timevar(t) policyvar(z) window(5) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(35, 7965)       =  96.32
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7323
                                                    Adj R-squared     = 0.6976
                                                    Root MSE          = 1.0403

. cap drop pois

. 
. * Test asymmetric window
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(-4 2) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     = 13,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(20, 11980)      = 476.98
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7303
                                                    Adj R-squared     = 0.7074
                                                    Root MSE          = 1.0599

. 
. * Test normalizations
. 
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) norm(-1) plot

No proxy or instruments provided. Implementing OLS estimator

. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) norm(-2) plot

No proxy or instruments provided. Implementing OLS estimator

. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) norm(-6) plot

No proxy or instruments provided. Implementing OLS estimator

. * xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(5) norm(-7) plot
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) norm(1) plot

No proxy or instruments provided. Implementing OLS estimator

. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) norm(5) plot

No proxy or instruments provided. Implementing OLS estimator

. 
. graph drop _all

. 
. * Test exclusion of unbalanced units with ambiguous eventtime
. gen z2 = z

. replace z2 = . if i==1 & t==7 
(1 real change made, 1 to missing)

. xtevent y , policyvar(z2) window(5)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

Unit 1 not used because of ambiguous event-time due to missing values in policyvar.

Linear regression, absorbing indicators             Number of obs     =  8,991
Absorbed variable: i                                No. of categories =    999
                                                    F(20, 7972)       = 168.20
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7321
                                                    Adj R-squared     = 0.6979
                                                    Root MSE          = 1.0398

. drop z2

. 
. * Overlay static plot
. xtevent y , policyvar(z) timevar(t) window(5)

Using options panelvar and timevar from xtset

No proxy or instruments provided. Implementing OLS estimator

. xteventplot, overlay(static) 
Estimating static model...

Using options panelvar and timevar from xtset

option static specified. Estimating static model

Plotting options ignored

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(9, 7991)        = 364.88
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7300
                                                    Adj R-squared     = 0.6959
                                                    Root MSE          = 1.0432

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
           z |   1.593592   .0649456    24.54   0.000     1.466282    1.720903
             |
           t |
          8  |   .2722747   .0466872     5.83   0.000     .1807557    .3637937
          9  |   .4402111   .0467515     9.42   0.000      .348566    .5318563
         10  |    .669563   .0468665    14.29   0.000     .5776925    .7614335
         11  |   .7950925   .0470328    16.91   0.000     .7028959    .8872891
         12  |   .9565168   .0472055    20.26   0.000     .8639818    1.049052
         13  |   1.175626   .0474565    24.77   0.000     1.082599    1.268653
         14  |   1.320488   .0477255    27.67   0.000     1.226934    1.414043
         15  |   1.564441   .0479705    32.61   0.000     1.470406    1.658476
             |
       _cons |   1.329013   .0332486    39.97   0.000     1.263837    1.394189
------------------------------------------------------------------------------
F test of absorbed indicators: F(999, 7991) = 12.551          Prob > F = 0.000

. 
. *------------------------ 2.3: Replicate 2c  ----------------------------------
. 
. * Replicate 2c
. xtevent y x , panelvar(i) timevar(t) policyvar(z) window(5) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(21, 7979)       = 174.69
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7390
                                                    Adj R-squared     = 0.7056
                                                    Root MSE          = 1.0263

. 
. * areg y x _k_eq* i.t , absorb(i) cluster(i)
. 
. *------------------------ 2.4: Replicate 2d and test basic funcionality  ----------------------------------
. 
. * Replicate 2d
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) proxy(x) plot

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =      9,000
Group variable: i                               Number of groups  =      1,000

R-squared:                                      Obs per group:
     Within  = 0.2966                                         min =          9
     Between = 0.4626                                         avg =        9.0
     Overall = 0.3932                                         max =          9

Wald chi2(20)     =   50246.32
corr(u_i, Xb) = 0.1187                          Prob > chi2       =     0.0000

. 
. * Test alternative ways of specifying iv
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) proxy(x) proxyiv(1) plot

Proxy for the confound specified. Implementing FHS estimator

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =      9,000
Group variable: i                               Number of groups  =      1,000

Wald chi2(20)     =   50246.32
corr(u_i, Xb) = 0.1187                          Prob > chi2       =     0.0000

. xteventplot, smpath(scatter)

Note: Smoothest line drawn for system confidence level = 95%
Wald Critical Value 19.6751376
Order 0
Wald value 93.4338443
Order 1
Wald value 93.4338443
Order 2
Wald value 83.1263635
Order 3
Wald value 61.472432
Order 4
Wald value 61.4721624
Order 5
Wald value 59.4494704
Order 6
Wald value 57.4880328
Order 7
Wald value 42.447924
Order 8
Wald value 42.4280227
Order 9
Wald value 33.4822543
Order 10
Wald value 35.9566899

Could not find a polynomial with order<=maxorder through the Wald confidence region.
Smoothest path won't be displayed.

. /*
> * This should not work now, for leads write 1
> cap gen f1z=f1.z
> cap noi xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) proxy(x) proxyiv(f1z)
> cap drop f1z
> */
. 
. *Generate an instrument for the proxy. This instrument is collinear with the event-time dummies.
. gen lead1=f1.d.z 
(1,000 missing values generated)

. *expect an error message: instrument is collinear  
. *xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) proxy(x) proxyiv(lead1)
. drop lead1

. 
. * Other leads
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(4) proxy(x) proxyiv(2) plot

Proxy for the confound specified. Implementing FHS estimator

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =     11,000
Group variable: i                               Number of groups  =      1,000

R-squared:                                      Obs per group:
     Within  = 0.3931                                         min =         11
     Between = 0.4362                                         avg =       11.0
     Overall = 0.4113                                         max =         11

Wald chi2(20)     =   64046.60
corr(u_i, Xb) = 0.1277                          Prob > chi2       =     0.0000

. xteventplot, smpath(scatter)

Note: Smoothest line drawn for system confidence level = 95%
Wald Critical Value 16.9189776
Order 0
Wald value 121.723003
Order 1
Wald value 121.723003
Order 2
Wald value 108.241241
Order 3
Wald value 78.3429876
Order 4
Wald value 74.8193942
Order 5
Wald value 62.9956812
Order 6
Wald value 61.998723
Order 7
Wald value 48.467384
Order 8
Wald value 37.3701674
Order 9
Wald value 35.7685454
Order 10
Wald value 36.4441307

Could not find a polynomial with order<=maxorder through the Wald confidence region.
Smoothest path won't be displayed.

. xtevent y , panelvar(i) timevar(t) policyvar(z) window(4) proxy(x) proxyiv(3) plot

Proxy for the confound specified. Implementing FHS estimator

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -3 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =     11,000
Group variable: i                               Number of groups  =      1,000

R-squared:                                      Obs per group:
     Within  = 0.3882                                         min =         11
     Between = 0.4504                                         avg =       11.0
     Overall = 0.4171                                         max =         11

Wald chi2(20)     =   63535.91
corr(u_i, Xb) = 0.1239                          Prob > chi2       =     0.0000

. xteventplot, smpath(scatter)

Note: Smoothest line drawn for system confidence level = 95%
Wald Critical Value 16.9189776
Order 0
Wald value 120.75243
Order 1
Wald value 120.75243
Order 2
Wald value 108.683546
Order 3
Wald value 80.6549725
Order 4
Wald value 75.90084
Order 5
Wald value 62.821969
Order 6
Wald value 62.8217534
Order 7
Wald value 51.7739477
Order 8
Wald value 42.4734094
Order 9
Wald value 38.8996058
Order 10
Wald value 40.4261902

Could not find a polynomial with order<=maxorder through the Wald confidence region.
Smoothest path won't be displayed.

. 
. * Test additional instruments
. 
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) proxy(x) proxyiv(1 2) plot

Proxy for the confound specified. Implementing FHS estimator

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =      9,000
Group variable: i                               Number of groups  =      1,000

Wald chi2(20)     =   50246.32
corr(u_i, Xb) = 0.1187                          Prob > chi2       =     0.0000

. xteventplot, smpath(scatter, technique("nr 10 bfgs 10"))

Could not find a polynomial with order<=maxorder through the Wald confidence region.
Smoothest path won't be displayed.

. xtevent y , panelvar(i) timevar(t) policyvar(z) window(4) proxy(x) proxyiv(3 4) plot

Proxy for the confound specified. Implementing FHS estimator

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -3 was selected to be normalized to zero.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -4 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =     11,000
Group variable: i                               Number of groups  =      1,000

R-squared:                                      Obs per group:
     Within  = 0.3865                                         min =         11
     Between = 0.4531                                         avg =       11.0
     Overall = 0.4178                                         max =         11

Wald chi2(19)     =   63360.21
corr(u_i, Xb) = 0.1224                          Prob > chi2       =     0.0000

. xteventplot, smpath(scatter, technique("nr 10 bfgs 10"))

Note: Smoothest line drawn for system confidence level = 95%
Wald Critical Value 15.5073131
Order 0
Wald value 120.362278
Order 1
Wald value 120.362278
Order 2
Wald value 120.362278
Order 3
Wald value 115.968694
Order 4
Wald value 112.540054
Order 5
Wald value 111.690995
Order 6
Wald value 73.14401
Order 7
Wald value 51.7301481
Order 8
Wald value 43.6471586
Order 9
Wald value 39.1677076
Order 10
Wald value 41.0125623

Could not find a polynomial with order<=maxorder through the Wald confidence region.
Smoothest path won't be displayed.

. 
. * Test both normalizations
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) norm(-3) proxy(x) proxyiv(1) plot

Proxy for the confound specified. Implementing FHS estimator

The coefficient at -3 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -1 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =      9,000
Group variable: i                               Number of groups  =      1,000

R-squared:                                      Obs per group:
     Within  = 0.3111                                         min =          9
     Between = 0.4443                                         avg =        9.0
     Overall = 0.3883                                         max =          9

Wald chi2(20)     =   51302.84
corr(u_i, Xb) = 0.1277                          Prob > chi2       =     0.0000

. 
. * Test additional proxys
. cap gen x2=rnormal()

. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) proxy(x x2) proxyiv(1 2) plot

Proxy for the confound specified. Implementing FHS estimator

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =      9,000
Group variable: i                               Number of groups  =      1,000

R-squared:                                      Obs per group:
     Within  = 0.2096                                         min =          9
     Between = 0.4143                                         avg =        9.0
     Overall = 0.3286                                         max =          9

Wald chi2(20)     =   44715.46
corr(u_i, Xb) = 0.1041                          Prob > chi2       =     0.0000

. * Must fail
. * xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) proxy(x x2) proxyiv(1)
. cap drop x2

. 
. * Testing xtset options
. 
. xtevent y , policyvar(z) window(5) proxy(x) plot

Using options panelvar and timevar from xtset

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =      9,000
Group variable: i                               Number of groups  =      1,000

Wald chi2(20)     =   50246.32
corr(u_i, Xb) = 0.1187                          Prob > chi2       =     0.0000

. xteventplot, levels(90 95 99)

. xtevent y , policyvar(z) panelvar(i) window(5)  proxy(x)

Using options panelvar and timevar from xtset

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =      9,000
Group variable: i                               Number of groups  =      1,000

Wald chi2(20)     =   50246.32
corr(u_i, Xb) = 0.1187                          Prob > chi2       =     0.0000

. xtevent y , policyvar(z) timevar(t) window(5)  proxy(x)

Using options panelvar and timevar from xtset

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =      9,000
Group variable: i                               Number of groups  =      1,000

Wald chi2(20)     =   50246.32
corr(u_i, Xb) = 0.1187                          Prob > chi2       =     0.0000

. 
. /* Must fail
> xtevent y , policyvar(z) panelvar(z) window(5) proxy(x)
> xtevent y , policyvar(z) timevar(z) window(5) proxy(x)
> */
. 
. * Testing noci, nosupt, nozeroline, nonormlabel
. xtevent y , policyvar(z) timevar(t) window(5) proxy(x)

Using options panelvar and timevar from xtset

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =      9,000
Group variable: i                               Number of groups  =      1,000

Wald chi2(20)     =   50246.32
corr(u_i, Xb) = 0.1187                          Prob > chi2       =     0.0000

. xteventplot, nosupt

option nosupt has been specified. Sup-t confidence intervals won't be displayed or calculated

. xteventplot, noci

option noci has been specified. Confidence intervals won't be displayed

. xteventplot, nozeroline

option nozeroline has been specified. The reference line at 0 won't be displayed

. xteventplot, nonormlabel

option nonormlabel has been specified. The label for the value of the dependent variable at event time correspon
> ding to the normalized coefficient won't be displayed

. 
. * Test if/in
. xtevent y if i<100, panelvar(i) timevar(t) policyvar(z) window(5) proxy(x) plot

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =        891
Group variable: i                               Number of groups  =         99

R-squared:                                      Obs per group:
     Within  =      .                                         min =          9
     Between = 0.2013                                         avg =        9.0
     Overall = 0.1137                                         max =          9

Wald chi2(20)     =      58.74
corr(u_i, Xb) = -0.6954                         Prob > chi2       =     0.0000

. xtevent y in 1/600 , panelvar(i) timevar(t) policyvar(z) window(5) proxy(x) plot

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 3 selected.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -3 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =        270
Group variable: i                               Number of groups  =         30

R-squared:                                      Obs per group:
     Within  =      .                                         min =          9
     Between = 0.2068                                         avg =        9.0
     Overall = 0.0840                                         max =          9

Wald chi2(20)     =      38.08
corr(u_i, Xb) = -0.7204                         Prob > chi2       =     0.0126

. xtevent y in 1/600 if i<30 , panelvar(i) timevar(t) policyvar(z) window(5) proxy(x) plot

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 3 selected.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -3 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =        261
Group variable: i                               Number of groups  =         29

R-squared:                                      Obs per group:
     Within  =      .                                         min =          9
     Between = 0.0966                                         avg =        9.0
     Overall = 0.0243                                         max =          9

Wald chi2(20)     =     154.53
corr(u_i, Xb) = -0.6305                         Prob > chi2       =     0.0000

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
           x |  -.8596922   1.361934    -0.63   0.528    -3.529034     1.80965
    _k_eq_m6 |  -3.717503   3.894232    -0.95   0.340    -11.35006    3.915051
    _k_eq_m5 |  -3.034587   2.748779    -1.10   0.270    -8.422094    2.352921
    _k_eq_m4 |  -2.107002   2.087867    -1.01   0.313    -6.199146    1.985142
    _k_eq_m2 |  -1.069151    1.55633    -0.69   0.492    -4.119502    1.981199
    _k_eq_p0 |   2.402322   4.108095     0.58   0.559    -5.649395    10.45404
    _k_eq_p1 |   2.829266   3.046377     0.93   0.353    -3.141522    8.800055
    _k_eq_p2 |   1.849055   2.180036     0.85   0.396    -2.423738    6.121847
    _k_eq_p3 |    2.81236   2.661867     1.06   0.291    -2.404804    8.029523
    _k_eq_p4 |   .5229343   1.894417     0.28   0.783    -3.190056    4.235924
    _k_eq_p5 |  -.9970253   2.799874    -0.36   0.722    -6.484678    4.490628
    _k_eq_p6 |  -.3820892   1.754115    -0.22   0.828    -3.820092    3.055914
             |
           t |
          8  |  -.2382977   .6826405    -0.35   0.727    -1.576249    1.099653
          9  |   .0576936   .6688667     0.09   0.931    -1.253261    1.368648
         10  |   .4155803   .7284304     0.57   0.568    -1.012117    1.843278
         11  |   .7196852   .7691893     0.94   0.349    -.7878981    2.227269
         12  |   -.029409   1.102559    -0.03   0.979    -2.190385    2.131566
         13  |   1.018625   .7369307     1.38   0.167     -.425733    2.462982
         14  |   .0843227   1.393674     0.06   0.952    -2.647228    2.815873
         15  |    .642588   1.079589     0.60   0.552    -1.473368    2.758544
             |
       _cons |   3.486757   2.103578     1.66   0.097    -.6361801    7.609694
-------------+----------------------------------------------------------------
     sigma_u |  2.8508066
     sigma_e |  2.4485654
         rho |  .57546876   (fraction of variance due to u_i)
------------------------------------------------------------------------------
 F test that all u_i=0: F(28,212) =     1.40              Prob > F    = 0.0973
------------------------------------------------------------------------------
Endogenous: x
Exogenous:  _k_eq_m6 _k_eq_m5 _k_eq_m4 _k_eq_m2 _k_eq_p0 _k_eq_p1 _k_eq_p2
            _k_eq_p3 _k_eq_p4 _k_eq_p5 _k_eq_p6 8.t 9.t 10.t 11.t 12.t 13.t
            14.t 15.t _fd3z

. 
. * Test nofe, note
. xtevent y, panelvar(i) timevar(t) policyvar(z) window(5) nofe proxy(x)

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.
note: 1.t identifies no observations in the sample.
note: 2.t identifies no observations in the sample.
note: 3.t identifies no observations in the sample.
note: 4.t identifies no observations in the sample.
note: 5.t identifies no observations in the sample.
note: 6.t identifies no observations in the sample.
note: 15.t omitted because of collinearity.
note: 16.t identifies no observations in the sample.
note: 17.t identifies no observations in the sample.
note: 18.t identifies no observations in the sample.
note: 19.t identifies no observations in the sample.
note: 20.t identifies no observations in the sample.

Instrumental variables 2SLS regression

Source |       SS       df       MS         Number of obs   =      9,000
-------------+------------------------------      F( 20,  8979)   =     246.36
       Model |  12679.2548    20  633.962738      Prob > F        =     0.0000
    Residual |  19524.3564  8979  2.17444664      R-squared       =     0.3937
-------------+------------------------------      Adj R-squared   =     0.3924
       Total |  32203.6112  8999  3.57857664      Root MSE        =     1.4746

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
           x |   .1866905   .1539382     1.21   0.225    -.1150635    .4884445
    _k_eq_m6 |  -.4400942      .6054    -0.73   0.467    -1.626816    .7466279
    _k_eq_m5 |  -.0131096   .2651722    -0.05   0.961    -.5329077    .5066886
    _k_eq_m4 |  -.1266771   .1993939    -0.64   0.525    -.5175346    .2641804
    _k_eq_m3 |   .1100165   .2208765     0.50   0.618    -.3229518    .5429848
    _k_eq_p0 |   1.075928   .3009703     3.57   0.000     .4859574    1.665898
    _k_eq_p1 |   .9444141   .3278846     2.88   0.004     .3016854    1.587143
    _k_eq_p2 |   1.027645   .3469694     2.96   0.003     .3475055    1.707784
    _k_eq_p3 |   .9607365   .3593791     2.67   0.008     .2562715    1.665201
    _k_eq_p4 |   .9472728     .35444     2.67   0.008     .2524895    1.642056
    _k_eq_p5 |     1.2183    .366638     3.32   0.001     .4996055    1.936994
    _k_eq_p6 |   1.025804   .3653197     2.81   0.005     .3096943    1.741914
             |
           t |
          1  |          0  (empty)
          2  |          0  (empty)
          3  |          0  (empty)
          4  |          0  (empty)
          5  |          0  (empty)
          6  |          0  (empty)
          7  |   -1.58012   .1467227   -10.77   0.000     -1.86773    -1.29251
          8  |  -1.309948   .1323895    -9.89   0.000    -1.569461   -1.050434
          9  |  -1.146683   .1211621    -9.46   0.000    -1.384188   -.9091773
         10  |  -.8969929   .0955315    -9.39   0.000    -1.084256   -.7097294
         11  |  -.7714088   .0866672    -8.90   0.000    -.9412962   -.6015214
         12  |   -.599894   .0732455    -8.19   0.000    -.7434719   -.4563161
         13  |  -.4000918   .0733866    -5.45   0.000    -.5439462   -.2562374
         14  |  -.2312987   .0661804    -3.49   0.000    -.3610274     -.10157
         15  |          0  (omitted)
         16  |          0  (empty)
         17  |          0  (empty)
         18  |          0  (empty)
         19  |          0  (empty)
         20  |          0  (empty)
             |
       _cons |   3.311779   .3668735     9.03   0.000     2.592624    4.030935
------------------------------------------------------------------------------
Endogenous: x
Exogenous:  _k_eq_m6 _k_eq_m5 _k_eq_m4 _k_eq_m3 _k_eq_p0 _k_eq_p1 _k_eq_p2
            _k_eq_p3 _k_eq_p4 _k_eq_p5 _k_eq_p6 7.t 8.t 9.t 10.t 11.t 12.t
            13.t 14.t _fd1z

. 
. * Note: Warning for the omitted time dummy comes from ivregress 2sls
. 
. xtevent y, panelvar(i) timevar(t) policyvar(z) window(5) nofe note proxy(x)

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Instrumental variables 2SLS regression

Source |       SS       df       MS         Number of obs   =      9,000
-------------+------------------------------      F( 12,  8987)   =     317.63
       Model |  10790.8053    12  899.233776      Prob > F        =     0.0000
    Residual |  21412.8059  8987  2.38264225      R-squared       =     0.3351
-------------+------------------------------      Adj R-squared   =     0.3342
       Total |  32203.6112  8999  3.57857664      Root MSE        =     1.5436

. 
. * Test savek
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) savek(a) proxy(x)

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =      9,000
Group variable: i                               Number of groups  =      1,000

Wald chi2(20)     =   50246.32
corr(u_i, Xb) = 0.1187                          Prob > chi2       =     0.0000

. des a_eq*, s

. des a_evtime, s

. drop a_*

. 
. 
. * Test factor variables in varlist
. cap gen pois=rpoisson(5)

. xtevent y i.pois, panelvar(i) timevar(t) policyvar(z) window(5) proxy(x)

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =      9,000
Group variable: i                               Number of groups  =      1,000

R-squared:                                      Obs per group:
     Within  = 0.2971                                         min =          9
     Between = 0.4633                                         avg =        9.0
     Overall = 0.3938                                         max =          9

Wald chi2(35)     =   50202.66
corr(u_i, Xb) = 0.1183                          Prob > chi2       =     0.0000

------------------------------------------------------------------------------
           y | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
           x |   .1466324   .1040605     1.41   0.159    -.0573225    .3505872
             |
        pois |
          1  |    .191326   .1659783     1.15   0.249    -.1339854    .5166374
          2  |   .1336157   .1605032     0.83   0.405    -.1809649    .4481962
          3  |   .2575876   .1580215     1.63   0.103    -.0521289    .5673041
          4  |   .1961903   .1541175     1.27   0.203    -.1058746    .4982551
          5  |   .1813632   .1568042     1.16   0.247    -.1259674    .4886937
          6  |   .1954914   .1554352     1.26   0.208    -.1091559    .5001387
          7  |   .2128354   .1577168     1.35   0.177    -.0962839    .5219547
          8  |   .2858568   .1586829     1.80   0.072     -.025156    .5968696
          9  |   .1841903   .1660965     1.11   0.267    -.1413527    .5097334
         10  |   .1963874    .170818     1.15   0.250    -.1384098    .5311845
         11  |   .1167189   .1937772     0.60   0.547    -.2630774    .4965151
         12  |   .3815787   .2766065     1.38   0.168      -.16056    .9237174
         13  |   -.426867    .358012    -1.19   0.233    -1.128558    .2748236
         14  |  -.0947799   .4775869    -0.20   0.843    -1.030833    .8412731
         15  |   .3674068   .9509024     0.39   0.699    -1.496328    2.231141
             |
    _k_eq_m6 |  -.2100656   .3003827    -0.70   0.484    -.7988049    .3786738
    _k_eq_m5 |   -.030045   .2151736    -0.14   0.889    -.4517775    .3916875
    _k_eq_m4 |   -.087816   .1631552    -0.54   0.590    -.4075944    .2319623
    _k_eq_m3 |   .0795204   .1679079     0.47   0.636     -.249573    .4086139
    _k_eq_p0 |    1.19254   .2192058     5.44   0.000     .7629042    1.622175
    _k_eq_p1 |   1.089885    .248956     4.38   0.000       .60194     1.57783
    _k_eq_p2 |   1.200801   .2758655     4.35   0.000     .6601149    1.741488
    _k_eq_p3 |   1.181829   .3004698     3.93   0.000     .5929189    1.770739
    _k_eq_p4 |    1.20383   .3014786     3.99   0.000     .6129428    1.794717
    _k_eq_p5 |   1.450539   .3133095     4.63   0.000     .8364635    2.064614
    _k_eq_p6 |    1.33898   .3380326     3.96   0.000     .6764484    2.001512
             |
           t |
          8  |   .2739489   .0474901     5.77   0.000       .18087    .3670277
          9  |    .437539   .0496756     8.81   0.000     .3401765    .5349014
         10  |   .6778654   .0608031    11.15   0.000     .5586936    .7970373
         11  |   .8033137   .0672734    11.94   0.000     .6714602    .9351672
         12  |   .9715076    .079529    12.22   0.000     .8156336    1.127382
         13  |   1.173289   .0795799    14.74   0.000     1.017315    1.329263
         14  |   1.334305   .0961122    13.88   0.000     1.145929    1.522682
         15  |    1.56285   .0989666    15.79   0.000     1.368879    1.756821
             |
       _cons |    1.33087   .3164631     4.21   0.000     .7106139    1.951126
-------------+----------------------------------------------------------------
     sigma_u |   1.104182
     sigma_e |   1.040541
         rho |   .5296472   (fraction of variance due to u_i)
------------------------------------------------------------------------------
 F test that all u_i=0: F(999,7965) =     9.64            Prob > F    = 0.0000
------------------------------------------------------------------------------
Endogenous: x
Exogenous:  1.pois 2.pois 3.pois 4.pois 5.pois 6.pois 7.pois 8.pois 9.pois
            10.pois 11.pois 12.pois 13.pois 14.pois 15.pois _k_eq_m6 _k_eq_m5
            _k_eq_m4 _k_eq_m3 _k_eq_p0 _k_eq_p1 _k_eq_p2 _k_eq_p3 _k_eq_p4
            _k_eq_p5 _k_eq_p6 8.t 9.t 10.t 11.t 12.t 13.t 14.t 15.t _fd1z

. cap drop pois

. 
. * Test asymmetric window
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(-4 2) proxy(x) plot

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =     13,000
Group variable: i                               Number of groups  =      1,000

R-squared:                                      Obs per group:
     Within  = 0.4622                                         min =         13
     Between = 0.3968                                         avg =       13.0
     Overall = 0.4219                                         max =         13

Wald chi2(20)     =   65240.77
corr(u_i, Xb) = 0.1036                          Prob > chi2       =     0.0000

. xtevent y , panelvar(i) timevar(t) policyvar(z) window(-2 2) proxy(x) plot

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

The corresponding coefficient of lead 1 and the normalized coefficient were the same. Lead 1 has been changed to
>  2.

The coefficient at -1 is normalized to zero.

For estimation with proxy variables, an additional coefficient needs to be normalized to zero.

The coefficient at -2 was selected to be normalized to zero.

Fixed-effects (within) IV regression            Number of obs     =     15,000
Group variable: i                               Number of groups  =      1,000

R-squared:                                      Obs per group:
     Within  = 0.5170                                         min =         15
     Between = 0.4448                                         avg =       15.0
     Overall = 0.4710                                         max =         15

Wald chi2(20)     =   90489.01
corr(u_i, Xb) = 0.1196                          Prob > chi2       =     0.0000

. 
. * Test overlay plots
. graph drop _all

.         
. xteventplot, y

. xteventplot, proxy

. xteventplot, overlay(iv)

. xteventplot

. xteventplot, overlay(static)
Estimating static model...

option static specified. Estimating static model

Plotting options ignored

Proxy for the confound specified. Implementing FHS estimator

proxyiv=select. Selecting lead order of differenced policy variable to use as instrument.

Lead 1 selected.

Fixed-effects (within) IV regression            Number of obs     =     15,000
Group variable: i                               Number of groups  =      1,000

R-squared:                                      Obs per group:
     Within  = 0.4926                                         min =         15
     Between = 0.4570                                         avg =       15.0
     Overall = 0.4685                                         max =         15

Wald chi2(16)     =   86121.88
corr(u_i, Xb) = 0.0905                          Prob > chi2       =     0.0000

. 
. 
. 
. 
. 
. *------------------------ 2.5: Replicate 2e and test basic funcionality  ----------------------------------
. 
. * Replicate 2e
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) trend(-3, saveov) plot

No proxy or instruments provided. Implementing OLS estimator

. 
. * Test overlay plot
. xteventplot, overlay(trend)

. 
. 
. /* Must fail
> xtevent y eta, policyvar(z) panelvar(z) window(5) trend(-3 5)
> xtevent y eta, policyvar(z) timevar(z) window(5) trend(-3 5)
> */
. 
. * Testing noci, nosupt, nozeroline, nonormlabel
. xteventplot, noci

option noci has been specified. Confidence intervals won't be displayed

. xteventplot, nosupt

option nosupt has been specified. Sup-t confidence intervals won't be displayed or calculated

. xteventplot, nozeroline

option nozeroline has been specified. The reference line at 0 won't be displayed

. xteventplot, nonormlabel

option nonormlabel has been specified. The label for the value of the dependent variable at event time correspon
> ding to the normalized coefficient won't be displayed

. 
. * Must fail
. * xtevent y eta if i<100, panelvar(i) timevar(t) policyvar(z) window(5) trend(-8 5)
. 
. 
. * Test nofe, note
. xtevent y, panelvar(i) timevar(t) policyvar(z) window(5) nofe trend(-3) plot

No proxy or instruments provided. Implementing OLS estimator

. xtevent y, panelvar(i) timevar(t) policyvar(z) window(5) nofe note trend(-3) plot

No proxy or instruments provided. Implementing OLS estimator

. 
. * Test savek
. xtevent y, panelvar(i) timevar(t) policyvar(z) window(5) savek(a) trend(-3)

No proxy or instruments provided. Implementing OLS estimator

. des a_eq*, s

. des a_evtime, s

. drop a_*

. 
. * Test factor variables in varlist
. cap gen pois=rpoisson(5)

. xtevent y i.pois, panelvar(i) timevar(t) policyvar(z) window(5) trend(-3)

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(35, 7965)       =  96.39
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7324
                                                    Adj R-squared     = 0.6976
                                                    Root MSE          = 1.0402

. cap drop pois

. * Test time series variables in varlist
. 
. xtevent y l.eta , panelvar(i) timevar(t) policyvar(z) window(5) trend(-3)

No proxy or instruments provided. Implementing OLS estimator

. 
. * Test asymmetric window
. * Must fail
. * xtevent y eta , panelvar(i) timevar(t) policyvar(z) window(-4 2) trend(-3 5)
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(-4 6) trend(-3) plot

No proxy or instruments provided. Implementing OLS estimator

Linear regression, absorbing indicators             Number of obs     =  9,000
Absorbed variable: i                                No. of categories =  1,000
                                                    F(20, 7980)       = 162.37
                                                    Prob > F          = 0.0000
                                                    R-squared         = 0.7425
                                                    Adj R-squared     = 0.7097
                                                    Root MSE          = 1.0393

. 
. * Test overlay plot
. xtevent y , panelvar(i) timevar(t) policyvar(z) window(5) trend(-3, saveov) plot

No proxy or instruments provided. Implementing OLS estimator

. xteventplot, overlay(trend)

. 
. *------------------------ 2.6: Hypotheses tests  ----------------------------------
. 
. xtevent y eta, panelvar(i) timevar(t) policyvar(z) window(3)

No proxy or instruments provided. Implementing OLS estimator

. 
. xteventtest, coefs(1 2)

( 1)  _k_eq_p1 = 0
 ( 2)  _k_eq_p2 = 0

F(  2, 11979) =   85.53
            Prob > F =    0.0000

. xteventtest, coefs(1 2) cumul

Test sums of coefficients

( 1)  _k_eq_p1 + _k_eq_p2 = 0

F(  1, 11979) =  170.94
            Prob > F =    0.0000

. xteventtest, coefs(-2 -3)

( 1)  _k_eq_m2 = 0
 ( 2)  _k_eq_m3 = 0

F(  2, 11979) =    2.58
            Prob > F =    0.0761

. xteventtest, allpre

Test for all pre-event coefficients = 0

( 1)  _k_eq_m4 = 0
 ( 2)  _k_eq_m3 = 0
 ( 3)  _k_eq_m2 = 0

F(  3, 11979) =    1.74
            Prob > F =    0.1566

. xteventtest, allpre cumul

Test for all pre-event coefficients = 0

Test sums of coefficients

( 1)  _k_eq_m4 + _k_eq_m3 + _k_eq_m2 = 0

F(  1, 11979) =    4.98
            Prob > F =    0.0256

. xteventtest, allpost

Test for all post-event coefficients = 0

( 1)  _k_eq_p0 = 0
 ( 2)  _k_eq_p1 = 0
 ( 3)  _k_eq_p2 = 0
 ( 4)  _k_eq_p3 = 0
 ( 5)  _k_eq_p4 = 0

F(  5, 11979) =   44.66
            Prob > F =    0.0000

. xteventtest, allpost cumul

Test for all post-event coefficients = 0

Test sums of coefficients

( 1)  _k_eq_p0 + _k_eq_p1 + _k_eq_p2 + _k_eq_p3 + _k_eq_p4 = 0

F(  1, 11979) =  219.10
            Prob > F =    0.0000

. xteventtest, coefs(1 2) testopts(coef)

( 1)  _k_eq_p1 = 0
 ( 2)  _k_eq_p2 = 0

F(  2, 11979) =   85.53
            Prob > F =    0.0000

Constrained coefficients

------------------------------------------------------------------------------
             | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
    _k_eq_m4 |  -.4309127   .0620137    -6.95   0.000    -.5524574    -.309368
    _k_eq_m3 |  -.4184633   .0771514    -5.42   0.000    -.5696773   -.2672494
    _k_eq_m2 |  -.4695899    .075589    -6.21   0.000    -.6177416   -.3214382
    _k_eq_p0 |   .3974088   .0729196     5.45   0.000      .254489    .5403287
    _k_eq_p1 |          0  (omitted)
    _k_eq_p2 |  -2.22e-16          .        .       .            .           .
    _k_eq_p3 |     .33181   .0782141     4.24   0.000     .1785132    .4851068
    _k_eq_p4 |    .329905    .055141     5.98   0.000     .2218306    .4379795
         eta |   .2725277   .0067675    40.27   0.000     .2592635    .2857918
             |
           t |
          6  |   .1872903   .0450053     4.16   0.000     .0990816    .2754989
          7  |   .3551906    .045056     7.88   0.000     .2668825    .4434987
          8  |   .6479289   .0451421    14.35   0.000     .5594521    .7364057
          9  |   .8307415   .0452795    18.35   0.000     .7419953    .9194878
         10  |   1.064703   .0454506    23.43   0.000     .9756219    1.153785
         11  |   1.215921   .0457167    26.60   0.000     1.126318    1.305524
         12  |   1.392897   .0460076    30.28   0.000     1.302724     1.48307
         13  |   1.639145   .0463299    35.38   0.000      1.54834     1.72995
         14  |   1.792725   .0466725    38.41   0.000     1.701248    1.884201
         15  |   2.033178   .0469996    43.26   0.000      1.94106    2.125295
         16  |    2.25316   .0474322    47.50   0.000     2.160195    2.346126
         17  |   2.458771   .0478287    51.41   0.000     2.365028    2.552513
             |
       _cons |   1.468441   .0662724    22.16   0.000      1.33855    1.598333
------------------------------------------------------------------------------

. xteventtest, linpretrend

Specification test for linear pre-trend
           chi2(  1) =.43
         Prob > chi2 =.5101

. xteventtest, overidpre(2)

Overidentification test for pretrends: 2 pre-event coefficients are 0

( 1)  _k_eq_m4 = 0
 ( 2)  _k_eq_m3 = 0

F(  2, 11979) =    2.44
            Prob > F =    0.0870

. xteventtest, overidpost(3)

Overidentification test for effects leveling off: 3 last post-event coefficients are equal

( 1)  - _k_eq_p3 + _k_eq_p4 = 0
 ( 2)  - _k_eq_p2 + _k_eq_p4 = 0

F(  2, 11979) =    0.07
            Prob > F =    0.9368

. xteventtest, overid

Overidentification test for pretrends: 3 pre-event coefficients are 0

Overidentification test for effects leveling off: 2 last post-event coefficients are equal

Overidentification test for pretrends: 3 pre-event coefficients are 0

( 1)  _k_eq_m4 = 0
 ( 2)  _k_eq_m3 = 0
 ( 3)  _k_eq_m2 = 0

F(  3, 11979) =    1.74
            Prob > F =    0.1566

Overidentification test for effects leveling off: 2 last post-event coefficients are equal

( 1)  - _k_eq_p3 + _k_eq_p4 = 0

F(  1, 11979) =    0.10
            Prob > F =    0.7478

Joint overidentification test

( 1)  _k_eq_m4 = 0
 ( 2)  _k_eq_m3 = 0
 ( 3)  _k_eq_m2 = 0
 ( 4)  - _k_eq_p3 + _k_eq_p4 = 0

F(  4, 11979) =    1.34
            Prob > F =    0.2519

. 
. cap log close

Texto modificado