PDF
Peter Aucoin; Lori Turnbull
Five of ten provincial governments – New Brunswick, Prince Edward Island, Quebec, Ontario, and British Columbia – are
considering changes to the way in which votes cast during provincial elections
are translated into seats in their provincial legislatures. Federally, the Law
Commission of Canada has recommended specific changes to the electoral system.
This article considers one possible consequence of major electoral reform
– the more frequent occurrence of minority or coalition government
– and suggests a need to rethink certain traditional Canadian conventions
of responsible government, namely the virtual right of Canadian prime ministers
and premiers to dissolve the legislature and call elections when they see fit,
even after their governments have lost the confidence of the legislature. It
recommends new protocols to govern the respective powers of first ministers and
governors general or lieutenant governors in the operation of responsible
government.
The prospect of greater frequency of
single-party minority governments (or coalition-party majority governments
which are more susceptible to collapse into a single-party minority government
than are single-party majority governments) demands a reconsideration of the
virtual right of a first minister to call an election when her or his
government is defeated on a vote of confidence in the legislature. Under the
constitution, a prime minister or premier merely advises the Crown (the Governor
General or Lieutenant Governor) to dissolve the House and thereby call an
election. No constitutional expert denies that the Crown in Canada possesses a residual power to deny a first minister’s request for dissolution.
Yet, as the tradition has evolved in Canada, the first minister has assumed a
virtual right to dissolution, even when their government is defeated in the
legislature on a matter of confidence. There is one major exception to this
right, but, even in this circumstance, unfortunately, there is an entirely
unsatisfactory degree of uncertainty about what the Crown should do in granting
or refusing dissolution.
At a time when most Canadian governments are
addressing the so-called “democratic deficit”,1 a
deficit that is partly due to a perceived excessive concentration of power in
first ministers, the issue of a first minister’s prerogative to call an
election following a defeat in the legislature has not yet been addressed in
the agenda of democratic reform that is now in fashion in several jurisdictions.
Indeed, the one remedy to check the perceived excess of power vested in first
ministers to call elections, namely the proposal to legislate fixed election
dates, is amazingly silent on this most critical issue.
In British Columbia, a first jurisdiction to
legislate fixed election dates, the statutory provision to accommodate the
requirement of responsible government, that is, that an election may be called
as a consequence of the defeat of a government in the legislature on a matter
of confidence, provides for no protocol on the matter. Nothing is said about
the fact that there is an alternative to an election in this circumstance. The
alternative, of course, is a new government summoned by the governor and formed
from the existing legislature without a new election. In this instance, the
first minister of the defeated government will have already submitted her or
his resignation to the governor, thus bringing that government to an end, or
will do so when a new government is formed.
The Canadian Tradition
In Canada, the responsibilities of a
Governor General or Lieutenant Governor (hereafter a governor) are now usually
thought of as ceremonial. The principles of representative democracy, to
say nothing of precepts of direct democracy, are deemed to have overtaken a
governor’s discretionary powers in the actual conduct of government.
Given this understanding, it is generally assumed that a first minister’s
request for dissolution should almost always be granted. In fact, there is now
considered to be only one situation where it is not certain whether a governor
might not, even should not, grant dissolution.
The situation occurs when a government has
lost a vote of confidence in the House in the period immediately following an
election that it had called. In this instance, it is accepted that the governor
should seek to determine whether a new government can be formed by the leader
of the Official Opposition without the legislature being dissolved and a new
election being held. Unfortunately, how long this “period immediately
following an election” needs to be before the right of a first minister
to dissolution is restored has never been definitely established.
Canada’s
foremost authority on the subject, the late Senator Eugene Forsey, defended the
right and power of a governor to refuse a request for dissolution.2
In certain circumstances, he argued, a governor’s exercise of this
discretion might be the only constitutional check on the first minister. The
power to refuse dissolution, in other words should not be rejected as
illegitimate or improper from a democratic perspective. Indeed, the essential
role of the Crown under responsible government is to protect and preserve the
constitution of responsible government itself.
The Canadian convention, however, also
allows a governor almost complete discretion in the exercise of the
Crown’s power to grant dissolution. Ironically, if not perversely, the
discretion to grant dissolution in almost any circumstance has served to
undermine the potency of the governor’s check on a first minister. The
reason is simple. Since governors are expected to grant dissolution in
virtually all circumstances, a governor who refused dissolution in any other
circumstance other than the one exception noted above would risk politicizing
the matter of the Crown’s powers. To make matters worse, refusal to grant
dissolution even in this one exception would be equally fraught with the risk
of politicization, since the governor would need to define the period of time
required for the first minister to regain the right to dissolution and that
definition, naturally, would inevitably be subject to political debate. The
legacy of the 1926 King-Byng affair would merely add fuel to the fire.
The New Zealand Solution
As in Canada, the essential principle of
responsible government in New Zealand is that the Governor General acts in
accordance with the advice of the prime minister or the prime minister and
ministers so long as they have the confidence of the House of Representatives.3
If the government loses this confidence in New Zealand, however, the prime
minister advises the Governor General that the government will resign. In
this situation, the Governor General must ascertain whether a new government
can be formed with the confidence of the House. If one can be formed, the
defeated government resigns and the new government assumes office. If a new
government is not possible, an election is called, and the incumbent government
continues in office, but only as a “caretaker government”. The
responsibility of the Governor General in this circumstance is “to
ascertain where the support of the House lies”; it is “not the role
of the Governor General to form the government or to participate in any negotiations”
leading to the possible formation of a new government.
The crucial difference between the Canadian
and New Zealand traditions is that the New Zealand convention requires the
prime minister of a defeated government to offer its resignation, and then to
wait to see whether a new government can be formed from the legislature. In New Zealand, in short, the prime minister does not have a right to dissolve the House
following a loss of confidence, even if the government has been in office for
some time. In Canada, in contrast, it is assumed that a first minister whose
government has been defeated in the House has a virtual right to dissolution,
unless it had been granted dissolution in the recent past (which, as noted, is
an undefined period of time). In other words, once a minority government has
survived for some undefined period of time it has every right to dissolve the
legislature if it is defeated. It need not worry about the possibility of
another government being formed from the same legislature.
The Canadian Convention: One More
Democratic Deficit?
Given the likelihood of electoral reform in
the near future, and the resulting increase in the occurrence of minority
governments, Canadians should debate whether the first minister of a defeated
government should have the virtual right to dissolve the legislature. Surely if
it is appropriate to debate fixed election dates that would remove from the
first minister the right to call an election at her or his discretion when the
majority or minority government which he or she leads has the confidence of the
House then it is even more pressing to pose the question of a first minister
whose government no longer enjoys the confidence of the legislature. In posing
this question, moreover, it is necessary to consider the wisdom of the Crown
possessing an undefined discretion to grant or refuse dissolution in a
representative democracy.
In our view, the Canadian tradition is
defective on three fronts.
First, the Canadian convention is not
effective in addressing the possibility of “a diet of
dissolutions”, as would occur if the first minister of a continuously
defeated government sought continuously to go to the polls in the hope of a
better electoral outcome. (In somewhat the same vein, the current convention is
obviously not effective in checking a first minister who times an election
merely for partisan-political purposes.) The Canadian convention forces a
governor to refuse dissolution in order to check a first minister as a matter
of the governor’s discretion. There is no established rule or protocol
for a governor to follow or to insist that a first minister follow. As a
consequence, the Canadian convention inevitably also forces a governor to state
her or his reasons for refusing dissolution, thereby drawing her or him into the
vortex of partisan politics in what will often, if not invariably, be a
confusing or complicated political situation.
Second, the Canadian convention does not
constrain the discretion of a governor in granting dissolution. A governor can
grant dissolution in virtually any circumstance and, in this case, can do so
without the need to give reasons. The convention of no constraints on a
governor’s discretion to grant dissolution can thereby result in a
legislature being denied the opportunity to support the formation of a new
government after it has explicitly withdrawn its confidence in the incumbent
government.
Third, the Canadian convention is defective
as it relates to political practice because it creates a disincentive for
members of the opposition (or even the members of a minor party in a coalition
government) to defeat a government on a vote of confidence even when a majority
of the legislature favours a change in government but when a majority do not
want an election to be held simply to determine the fate of the government. On
the assumption that there can be democratically valid reasons for a majority in
the legislature not wanting to have an election at a particular point in time
when it also wants a change in government, the Canadian convention works
against good government.
(Conversely, there is no public purpose
served in a first minister being able to use the threat of dissolution against
opposition parties, or even against one or more minor parties in the coalition
government headed by the first minister, to keep these members of the
legislature from voting non-confidence in the government. The claim that this
power is checked in the sense that the electorate can hold a first minister
accountable is technically accurate. But the check in question, namely the
electoral defeat of a government, is not an appropriate check. It requires that
the electorate focus first and foremost on the first minister’s
responsibility for the calling of an election rather than on the record of the
government. Responsible democratic government is not well served by forcing the
issue in this way.)
Changing the Convention: A Proposed Set
of Protocols
Fixed election dates deal with one
democratic deficit respecting dissolution, namely, the power of a prime
minister whose government commands the confidence of the legislature to play
partisan-politics with the timing of elections, with all of the perverse
consequences for good government that can ensue. Fixed election dates bring
their own problems, of course, but they are more easily dealt with than the
open-ended discretion of a first minister’s discretion to call elections.
A new set of protocols to govern what happens when a government loses the
confidence of the legislature is required to do what fixed election dates
cannot do, namely, stipulate the responsibilities and powers of the first
minister and the governor respectively.
The New Zealand model offers an attractive
alternative to the Canadian one: it requires that the possibility of a new
government being formed from the existing legislature be explored before the
legislature is dissolved and an election is called. In doing so, it not only
avoids the possibility that a first minister and a governor become embroiled in
a political struggle and perhaps public debate over the proper understanding of
the requirements of the constitution of responsible government in the situation
at issue, it also avoids the possibility that a governor be seen to have
exercised her or his discretion to grant or deny dissolution for
partisan-political reasons (including the matter of who appointed her or him
governor).
To repeat: the New Zealand protocol is as
follows:
1. When a government loses the
confidence of the legislature, the prime minister advises the Governor General
that the government will resign.
2. The Governor General ascertains
from opposition members whether a new government can be formed with the
confidence of the House.
3. If a new government can be formed,
the defeated government resigns and the new government assumes office.
4. If a new government is not
possible, an election is called, and the incumbent government continues in
office through the election period, but only as a “caretaker
government”.
We propose that the Canadian jurisdictions
adopt the protocols of the New Zealand model. Coupled with fixed election
dates, they offer a democratic reform that respects the formal constitutional
structure, the democratic principles of responsible government, and the demand
that the excessive concentration of power in first ministers be checked.
If fixed election dates are not accepted as
a way to check the power of a first minister whose government has the
confidence of the legislature, we propose that dissolution only be granted a
first minister whose government has the confidence of the legislature following
a resolution in the legislature. This would at least constitute a check on a
first minister who heads a minority government. As a consequence, an election
called by a first minister in this context would require, at a minimum, the
majority opposition approve the resolution and thus also accept responsibility
for calling the election. A defeated resolution would require that a minority
government either carry on or resign so that a new government could be formed.
Notes
1. See Peter Aucoin and Lori Turnbull,
“The democratic deficit: Paul Martin and parliamentary reform,” Canadian
Public Administration, Winter 2003, Vol. 46, number 4, pp. 427-449.
2. Eugene Forsey, The Royal Power of Dissolution
of Parliament in the British Commonwealth (Toronto: Oxford University
Press, 1943).
3. See New Zealand, Department of the Prime
Minister and Cabinet, Cabinet Office, Cabinet Manual 2001 (Wellington: Cabinet Office, 2001), Part 4, Elections, Transitions and Government Formation,
pp. 53-62.
|