CAPITULO IV: MARCO PROPOSITIVO
4.7 REGLAMENTO INTERNO DE TRABAJO PROPUESTO
The analysis presented in the previous section points to an endogeneity associated with the speed at which debates progress as the reason behind the differences in observed char- acteristics between the members whose questions are answered and those whose questions remain unanswered. To find out to what extent these differences can be attributed to the role of the Speaker, further difference-in-means tests will be conducted which compare only those questions which are marginally admitted or left out of the debate.
The reason for making this comparison is that while members submitting questions may have some knowledge of the debates in which a high proportion of the questions on the Order Paper will be reached, it can safely be assumed that they will not be able to predict exactly which question listed on the Order Paper for a debate will be the last question admitted. This implies that members will not be able to influence whether they fall into the treatment or control group in this comparison. On an Order Paper with a answered andu unanswered questions numbered 1, . . . , a+u, they will therefore have as much probability of their question being the question numbereda(the last one answered), as the question numbered a+ 1 (the first one unanswered). This makes allocation to
Table 3.4: Difference-in-means tests between the characteristics of the members whose ques- tions were the last ones to receive oral answers in House of Commons debates (Na = 2588)
and the corresponding characteristics of members whose questions were the first ones that were not reached during the debates and therefore received written answers (Nu = 2588).
The first column reports the mean characteristics among the members whose questions were the last to receive an oral answer in a particular debate (the ‘treatment group’). The second column gives the corresponding mean characteristics among the questions which were the first ones listed on the Order Paper not to receive oral answers (the ‘control group’). The final column reports the difference between these means.
Just-answered questions Just-unanswered questions Difference
Conservative Party 0.355 0.342 0.0128
(0.894)
Lib. Dem. Party 0.105 0.122 -0.0178
(-1.897) Changed party 0.0143 0.0116 0.00270 (0.695) Ministerial position 0.461 0.452 0.00850 (0.574) Age 53.39 52.81 0.583⇤ (2.188) Age squared 2929.3 2867.6 61.77⇤ (2.146) Years service 11.46 11.11 0.349 (1.385) Years svc. sqrd. 202.9 187.7 15.22 (1.722) Female 0.206 0.184 0.0220⇤ (1.992) Observations 5176
Standard errors clustered at member level
t-statistics reported in parentheses *p< 0.05, **p< 0.01, ***p< 0.001
treatment and control groups persist under this comparison, it will therefore be possible to conclude that it is an ex post manipulation of the speed of the debate after seeing the Order Paper (i.e. during the debate itself) which creates an environment where certain questions are more likely to be admitted than others.
Table 3.4 reports the results of difference-in-means tests where the ‘treatment group’ consists of the last questions answered across all debates and the ‘control group’ consists of the first questions unanswered across all debates.
Results in Table 3.4 are based on a much smaller sample size, given that attention is focused solely on questions marginally admitted and excluded from oral debates. As a
result, few of the results show statistical significance when compared with earlier findings. It is informative, however, to compare the magnitudes of the reported differences.
Once again, Conservative Party members are found to be more likely to have their questions just admitted to debates than have their questions just excluded. Among the ‘just-answered’ questions, 35.5% were asked by Conservative Party MPs, whereas among the ‘just-unanswered’ questions, 34.2% were asked by Conservative Party MPs. Given the small sample size, this 1.3% difference is not found to be statistically significant. In terms of magnitudes, however, it is not far short of the 2.4% and 1.5% differences reported in Tables 3.1 and 3.2 respectively. This offers some tentative evidence that endogeneity associated with the speed at which debates progress works in favor of Conservative Party MPs in offering them a higher probability of having their questions answered than would be expected under a true randomization.
Interestingly, Table 3.4 reports that the average age of members among the ‘just- answered’ questions is 0.58 years older than among the ‘just unanswered’ questions. This compares with a difference of 0.46 years in Table 3.1. This large effect is found to be statistically significant at the 5% level, even accounting for the much smaller sample size using this method. The endogeneity in the speed at which debates progress is therefore found to favor older members in allowing them a greater chance of having their questions answered once their questions are selected for the Order Paper. Among the ‘just-answered’ questions, members also have an average number of years’ service that is 0.35 years higher than among the ‘just-unanswered’ questions. This compares with a difference of 0.6 years in Table 3.1, but is not found to be statistically significant in this case, again owing to the small sample size.
Finally, Table 3.4 shows that female members are more likely to have their questions ‘just-answered’ than ‘just-unanswered’. 20.6% of the ‘just-answered’ questions were asked by female members, whereas only 18.4% of the ‘just-unanswered’ questions were asked by female members. The 2.2 percentage point difference between these two populations is found to be statistically significant at the 5% level. This suggests an endogeneity in the debating process which gives preferential treatment to female members when admitting marginal questions into debates.
Given that the members posing questions are limited in their involvement in debates, the two most likely causes of these differences are the ministers facing the questions in debates, and the Speaker. The ministers may have an incentive to give longer answers to early questions, for example, in the event that they see potentially hostile questions from opposition party members further down the Order Paper. The Speaker’s role is to correct such behavior and they may also have a further influence on which questions are answered by deciding how many follow-up questions to allow for each question listed on the Order Paper.
If ministers played the dominant role in influencing which questions receive an oral answer, it would surely be expected that a lower proportion of questions by opposition party members would receive oral answers. The fact that a higher proportion of questions asked by opposition party members are admitted into debates than are excluded suggests that the Speaker plays a more prominent role in this process than the ministers. While this could be interpreted as a bias against the Labour Party (which governed throughout the period covered by this dataset) it could also be interpreted as the Speaker counteracting the type of tactics which would otherwise be employed by ministers to manipulate the pattern of debates so as to face a lower proportion of hostile questions than would otherwise be the case.