Table 5.

Complementary Log–Log Models on Exit from the Referee Pool

(1)(2)(3)(4)β1=β2
Referee avg. performance (season)−0.1047***−0.1532***−0.2032***−0.1674***
(0.0310)(0.0457)(0.0472)(0.0465)
DFB cup referee (d)−0.04360.09360.06870.1365***
(0.0724)(0.0903)(0.0615)(0.0465)
International referee (d)−0.0776−0.00230.02720.0041
(0.1124)(0.0535)(0.0372)(0.0298)
Age0.01090.02090.0311***0.0251
(0.0087)(0.0139)(0.0119)(0.0167)
PG: Type II errors−0.0789} p =0.1125
(0.1219)
PG: Type I errors−0.5050*
(0.2855)
PG: Type II errors (ATT ahead)0.0819} p =0.1299
(0.0682)
PG: Type II errors (ATT behind)−0.4281**
(0.2090)
PG: Type I errors (ATT ahead)−0.1714} p =0.0007
(0.2274)
PG: Type I errors (ATT behind)−1.4865***
(0.3692)
PG: Type II errors (members ahead)−0.0526} p =0.7423
(0.0726)
PG: Type II errors (members behind)0.0253
(0.2392)
PG: Type I errors (members ahead)−0.0341} p =0.1418
(0.0999)
PG: Type I errors (members behind)−3.4657**
(1.7507)
Tenure FEYesYesYesYes
N207207207207
(1)(2)(3)(4)β1=β2
Referee avg. performance (season)−0.1047***−0.1532***−0.2032***−0.1674***
(0.0310)(0.0457)(0.0472)(0.0465)
DFB cup referee (d)−0.04360.09360.06870.1365***
(0.0724)(0.0903)(0.0615)(0.0465)
International referee (d)−0.0776−0.00230.02720.0041
(0.1124)(0.0535)(0.0372)(0.0298)
Age0.01090.02090.0311***0.0251
(0.0087)(0.0139)(0.0119)(0.0167)
PG: Type II errors−0.0789} p =0.1125
(0.1219)
PG: Type I errors−0.5050*
(0.2855)
PG: Type II errors (ATT ahead)0.0819} p =0.1299
(0.0682)
PG: Type II errors (ATT behind)−0.4281**
(0.2090)
PG: Type I errors (ATT ahead)−0.1714} p =0.0007
(0.2274)
PG: Type I errors (ATT behind)−1.4865***
(0.3692)
PG: Type II errors (members ahead)−0.0526} p =0.7423
(0.0726)
PG: Type II errors (members behind)0.0253
(0.2392)
PG: Type I errors (members ahead)−0.0341} p =0.1418
(0.0999)
PG: Type I errors (members behind)−3.4657**
(1.7507)
Tenure FEYesYesYesYes
N207207207207

Notes: Complementary log–log estimates; average marginal effects; (d) for discrete change of dummy variable from 0 to 1. The dependent variable is a dummy that takes value 1 if a referee left the pool of active referees in a given season. Referees at the age limit are excluded. We furthermore exclude one referee as he was suspended due to connections to a betting scandal. The control variables measure average performance in the season—in standard deviations and with mean 0, whether a referee was also refereeing in the DFB Cup (German National Cup) or internationally (Champions League or Euro League), and the average age in the season. Tenure-fixed effects take the role of mixed proportional hazards. The last column presents the p-value for test of equality for the indicated coefficients. Standard errors in parentheses are clustered at the referee level.

***

p <0.01,

**

p <0.05,

*

p <0.1.

Table 5.

Complementary Log–Log Models on Exit from the Referee Pool

(1)(2)(3)(4)β1=β2
Referee avg. performance (season)−0.1047***−0.1532***−0.2032***−0.1674***
(0.0310)(0.0457)(0.0472)(0.0465)
DFB cup referee (d)−0.04360.09360.06870.1365***
(0.0724)(0.0903)(0.0615)(0.0465)
International referee (d)−0.0776−0.00230.02720.0041
(0.1124)(0.0535)(0.0372)(0.0298)
Age0.01090.02090.0311***0.0251
(0.0087)(0.0139)(0.0119)(0.0167)
PG: Type II errors−0.0789} p =0.1125
(0.1219)
PG: Type I errors−0.5050*
(0.2855)
PG: Type II errors (ATT ahead)0.0819} p =0.1299
(0.0682)
PG: Type II errors (ATT behind)−0.4281**
(0.2090)
PG: Type I errors (ATT ahead)−0.1714} p =0.0007
(0.2274)
PG: Type I errors (ATT behind)−1.4865***
(0.3692)
PG: Type II errors (members ahead)−0.0526} p =0.7423
(0.0726)
PG: Type II errors (members behind)0.0253
(0.2392)
PG: Type I errors (members ahead)−0.0341} p =0.1418
(0.0999)
PG: Type I errors (members behind)−3.4657**
(1.7507)
Tenure FEYesYesYesYes
N207207207207
(1)(2)(3)(4)β1=β2
Referee avg. performance (season)−0.1047***−0.1532***−0.2032***−0.1674***
(0.0310)(0.0457)(0.0472)(0.0465)
DFB cup referee (d)−0.04360.09360.06870.1365***
(0.0724)(0.0903)(0.0615)(0.0465)
International referee (d)−0.0776−0.00230.02720.0041
(0.1124)(0.0535)(0.0372)(0.0298)
Age0.01090.02090.0311***0.0251
(0.0087)(0.0139)(0.0119)(0.0167)
PG: Type II errors−0.0789} p =0.1125
(0.1219)
PG: Type I errors−0.5050*
(0.2855)
PG: Type II errors (ATT ahead)0.0819} p =0.1299
(0.0682)
PG: Type II errors (ATT behind)−0.4281**
(0.2090)
PG: Type I errors (ATT ahead)−0.1714} p =0.0007
(0.2274)
PG: Type I errors (ATT behind)−1.4865***
(0.3692)
PG: Type II errors (members ahead)−0.0526} p =0.7423
(0.0726)
PG: Type II errors (members behind)0.0253
(0.2392)
PG: Type I errors (members ahead)−0.0341} p =0.1418
(0.0999)
PG: Type I errors (members behind)−3.4657**
(1.7507)
Tenure FEYesYesYesYes
N207207207207

Notes: Complementary log–log estimates; average marginal effects; (d) for discrete change of dummy variable from 0 to 1. The dependent variable is a dummy that takes value 1 if a referee left the pool of active referees in a given season. Referees at the age limit are excluded. We furthermore exclude one referee as he was suspended due to connections to a betting scandal. The control variables measure average performance in the season—in standard deviations and with mean 0, whether a referee was also refereeing in the DFB Cup (German National Cup) or internationally (Champions League or Euro League), and the average age in the season. Tenure-fixed effects take the role of mixed proportional hazards. The last column presents the p-value for test of equality for the indicated coefficients. Standard errors in parentheses are clustered at the referee level.

***

p <0.01,

**

p <0.05,

*

p <0.1.

Close
This Feature Is Available To Subscribers Only

Sign In or Create an Account

Close

This PDF is available to Subscribers Only

View Article Abstract & Purchase Options

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Close