The elections in Cambodia are over, though doubts and recriminations remain. The
most technically complex and expensive election that Cambodia has ever had - and
is likely to have - is now history. It is not too early to ask what can be learned
from these historical elections.
Pre-election activity was marred and constrained by violence, but the political ferment
that occurred is unprecedented in Cambodia's (admittedly) limited experience with
In the 1993 elections only four parties won seats, two of them overwhelmingly. But
twenty parties were formed, party offices were established nation-wide, thousands
of political rallies and meetings were held, and hundreds of thousands of Cambodians
participated. Though the small parties did not win seats, they did win votes. Every
political party won votes in every Province (as did UNTAC, but that is another story).
Why did these small parties win the votes that they did? Was it the consequence of
successful campaign activity, or was it because in many provinces some voters did
not understand what they were doing?
Two conclusions are indicated by a first-cut statistical analysis of provincial and
nation-wide elections returns. First, the small parties received very little relative
benefit from their campaign activities. Second, in many provinces the distribution
of votes for the small parties is indistinguishable from a vote assignment that would
occur from a random process.
The first conclusion indicates that the small parties would have been almost as well
off had they not campaigned (in terms of the total number of votes that they got).
The second conclusion indicates that the small parties got votes (even in Provinces
where they had no offices and did no campaigning) because voters did not understand
what they were doing, rather than as a consequence of informed choice.
The first conclusion indicates the over-whelming advantages that the large parties
had in the minds and pencils of the voters, and the second indicates problems with
the education of the voters.
Undoubtedly the problem would have been much greater had there been not been voter
education (civic education) by UNTAC District Electoral Supervisors. The major parties
also helped, perhaps heeding the advice of the National Democratic Institute, by
posters and radio descriptions of how to vote - for their party. CPP posters blurred
all party symbols except their own in the upper left-hand corner. FUNCINPEC's instructions
were most imaginative. Voters were told to fold the ballot in half and half again,
to smooth the creases and re-open it. "X" marks the spot. The FUNCINPEC
Party symbol lay where the creases crossed!
This analysis depends on two kinds of statistical tests, regression analysis and
tests for randomness. The data used came from a variety of sources: interviews with
the small parties themselves, the Electoral Component and the Spokesman's Office
of the United Nations Transitional Authority in Cambodia.
Regression analysis was used to determine whether a statistically significant relationship
existed between the number of votes that any small party received and the number
of party offices that it opened, the number of political rallies that it held, and/or
the number of people who attended their rallies? A positive relationship would indicate
that the more a party campaigned, the more votes it could expect to get. A negative
relationship, the reverse.
There was no statistically significant relationship between the votes received by
any small party and the magnitude of their campaign activity in any Province.
If opening party offices or holding political rallies did not produce significantly
more votes, and further if parties were not penalized by not holding rallies or opening
offices, how can we explain the votes that they did get?
The data was perused to determine whether the votes among the small parties gave
any evidence of choice and preference, or were more likely to be random. This analysis
showed that in seven provinces, voting patterns for the small parties were more likely
to be a random product than to be the product of informed choice But in other provinces
clear choices were obvious. In other provinces the data was ambiguous, in these the
data might have been produced by a random process.
The regression model indicates that for the sixteen smallest parties campaign efforts
to get out the vote were of little benefit, those parties who bothered to campaign
were only marginally benefited. Those parties which did not campaign, did not seem
to be penalized. The failure of a party to campaign or to open offices did not result
in gaining no or fewer votes. In fact they got more or less the same number of votes,
on a nation-wide basis as did those parties which campaigned more vigorously. Statistical
analysis indicates that it is more likely (in many provinces) that they got votes
because the voters did not know what to do than as a consequence of informed choice.
With the nation-wide data some weak relationships were indicated. If we look only
at party offices established there were weak positive relationships between the total
number of votes received nation-wide by a small party and the number of political
offices opened (nation-wide) by a that small party.
A similar weak relationship held between the total number of votes and the total
number of supporters at rallies. Of the two, the stronger relationship was between
the total number of votes and the number of political offices. However, the marginal
utility of opening a party office was very small. The number of votes that a small
party could expect nation-wide, if it opened no party offices was between 11648 and
One potential explanation is that a fraction of voters in each province did not know
what they were doing. Were the small parties beneficiaries of by votes cast by people
who were unsure what to do, where to mark, how to mark?
One would like to think that all voters knew what to do when they went into the voting
booth. But informal interviews with District Electoral Supervisors, (DES) with UNTAC
Electoral Interpreters, International Observers, and Party Observers indicate that
lack of understanding was more prevalent than we might want to think. One BLDP Party
Observer said, "Some people just marked anywhere, they did not know what the
symbols meant." A DES from Rattanakiri said, "Some of the people just panicked
in the voting booth, they simply did not know what to do."
This hypothesis, and that is all that it is, is given additional weight by the fact
that there were so many bad ballots. There were thousands in every province. Provincial
percentages varied from 2.65 percent to 7.37 percent of votes cast. In absolute numbers
they ranged from 765 in Mondulkiri to 17,940 in Kampong Cham; the total was more
According to the statistical analysis, seven provinces had a pattern of voting for
the small parties which appears to be random. These are Koh Kong (3.57 %), Kampong
Speu (3.04%), Mondulkiri (7.37%), Preah Vihear (5.70%), Sihanoukville (3.68%), Stung
Treng (3.74%), Svay Rieng (2.90%). The numbers in the parentheses are the numbers
of ballots cast in those provinces for which no preference for a single party could
be determined -- these ballots had to be ignored in the UNTAC vote tally.
Those provinces in which the distribution might be random include Kampong Chnang
(4.07%), Kampot (4.13%), Prey Veng (3.24%), Siem Riep (3.60%), and Takeo (3.45%).
The remaining provinces are very unlikely to be random. These include Ratanakiri
(5.20%), Phnom Pehn (?) Kampong Thom (3.32%), Pursat (3.87%), Kandal (2.45%), Battambang
(2.68%), Kratie (2.40%), Kampong Cham (3 00%) and Beantey Meanchey (2.65%).
It seems in general that those provinces in which there were large percentages of
what UNTAC called "informal" ballots, are also the provinces in which the
voting patterns for the small parties indicate a lack of understanding on the part
of the voters.
The glaring exception is Rattanakiri in which 5.2% of the ballots were rejected,
and in which the small party voting does not appear to be random. Part of that is
the result of the voting for the Khmer National Congress Party. The KNCP had no offices
in the province, nor did it hold any rallies, but it was third in the tally. It also
was the party whose symbol appeared in the lower right hand corner of the ballot.
The interpretation in Rattanakiri was that KNCP received a lot of votes that were
intended for CPP, the voters just had the ballot upside-down.
These conclusions allow us to assess one of the major issues of the campaign period:
access to the voters by campaigning parties. In the brief six-week campaign period
a major issue was the physical safety of the participants. Second in importance,
and not unrelated, was access to the people of Cambodia: physical access for campaigning,
media access to propagandize.
The issue is a central one for democratic elections. In principle the more informed
the voters are about the plans, programs, and promises of a political party, the
more likely it is that voting will reflect informed choice.
How the smaller parties did was not the result of a dearth of ideas. Extensive interviews
conducted during the campaign period indicate that most of these small parties were
filled with serious and committed people, who had a variety of ideas for the future
of Cambodia. Many had spent time in America and France, and were filled with the
ambition to implement ideas of governance that they thought Cambodia could use. Some
had little idea what they were talking about, but many could articulate what democracy
was about better than most Americans. Men like Chak Sareun and Bit Seanglim and others,
had well articulated visions of Cambodia's future.
The two largest parties had huge advantages. Both had money. CPP had incumbency and
the "control of administrative structures", like the police, the army and
the judiciary. FUNCINPEC was seen as Prince Sihanouk's party. In addition, both parties
had, in the end, unfettered access to television broad-casting, as both had stations.
Though the Khmer Rouge boycotted the elections, they played an important role (besides
declining to disrupt the elections), but a contradictory one. CPP was aided by being
understood as willing to prevent the return of the Khmer Rouge, even if that meant
war. FUNCINPEC was aided by being perceived as able to solve the "Khmer Rouge
problem" without war.
The advantage of the two biggest parties can be summed up as "name recognition",
a simplistic phrase to cover a complex reality. But what is striking about the election
is that the efforts of the smallest parties to alter this reality by campaigning
was to almost no avail.
One feature of the lead-up to the elections in Cambodia was the proliferation of
party offices in Phnom Penh and in the Provinces. By the time of the elections UNTAC
provincial offices reported that over 2000 party offices had been established. Of
these offices, CPP had 644 and FUNCINPEC 762, for a combined total of 1406. BLDP
had 204 offices, and LDP, 146. The remaining 276 offices were shared among the 16
smallest parties. Thus 86 percent of the party offices were maintained by the largest
parties. Among the 16 small parties Molinaka had fifty-seven.
The data is over-whelmed by the two largest parties. In fact if all the parties,
large and small are considered together there are very strong relationships between
the number of party offices and the number of votes received. But if the small parties
are taken as a separate category, as seems sensible, such a linkage disappears.
Among the smallest parties, the over-all relationship between votes received and
number of party offices opened is implied by a simple examination of the data. Two
parties had only one office in all of Cambodia, but they did comparatively well in
spite of that. Cambodian Farmers' Independent Party got 37,474 total votes even though
it had only one office. This is almost as well as the Democratic Party's 41,799 with
their 56 offices, and much better than the National Khmer Party's 7827 votes with
its eight offices.
Fifteen political parties held mass political meetings during the campaign period.
The vast majority of the meetings were single party meetings; 2017 were held.
Again the two largest parties were the most visible. CPP held 1064 meetings, or 52.8%
of the total, FUNCINPEC held 346, about 17%. The next closest party was BLDP with
212 meetings. The remaining 267 meetings were distributed among the other 12 parties,
with the LDP having the largest total, 171. Five parties held no meetings reported
As these numbers indicate, CPP did less well than FUNCINPEC though it had far and
away the most meetings. Among the smaller parties, in provincial results, there is
not a significant relationship between the number of political meetings that a party
had and the number of votes that it received.