Back

Note: This article represents the view of the author and not the University of Hong Kong.

On 2 August, the Public Opinion Programme at the University of Hong Kong released the latest popularity figures of CE Donald Tsang and the HKSAR Government. The author's comment then was:

"Both the popularity rating and support rate of Donald Tsang have dropped, giving an important signal to the government. In between the two latest surveys, RTHK stopped its horseracing radio programme, Tsang was accused of intervening an application by New World Development for turning a car park into a showroom for selling vehicles two years ago, and the government was criticized for reacting too slowly in banning the import of pork from Sichuan. All these could have contributed to Tsang's popularity drop. The overall popularity of the government, however, is still on the rise. Satisfaction rate has increased by 4 percentage points over the last month, 13 percentage points over the last quarter, and 24 percentage points counting from the resignation of Tung Chee-Hwa. In other words, the government's popularity has basically recovered to that shortly after the handover, but the fact that CE's popularity has dropped is a sign that honeymoon is over. What happens next will depend very much on the government's actual policies."

Shortly afterwards, quite a number of commentators and programme hosts began to discuss whether Donald Tsang's honeymoon period was indeed over. Although it is an open question of what constitutes "honeymoon period", the author nevertheless likes to elaborate his commentary a bit more, in order to enrich the discussion.

The concept of 'honeymoon' originates from Western marriage traditions. The new couple has to drink honey or honey wine within 30 days after their wedding. Therefore the first month after their marriage is called 'honeymoon'. In other words, one month after the couple's wedding would not be considered as 'honeymoon period' anymore no matter how much the couple loves each other. Of course, the meaning of every concept may be changed after some kind of conventional setting.

Political scientists and communications researchers have extended the concept of honeymoon to become "political honeymoon period" during which the public, the press and the legislature are likely to "go easy" on the newly elected leader. Some scholars also believed that the other side of the "honeymoon period" is the "lame-duck period", during which nobody seems to bother the outgoing leaders during their final days in office.

Leaders in the Western countries regard policymaking within hundred days as political honeymoon period, there is some modern saying about this, that political honeymoon period depends on individual practices. Looking back to our CE Donald Tsang, he became the Acting CE after Tung Chee-Wah's resignation on 12 March, and later on 21 June he officially became the CE. To calculate the number of days between the two surveys were conducted in late July and mid March, there are actually more than 130 days. Even counting from late June, it is still more than one month. This is not a short period of time at all.

Simply by looking at the trends of popularity ratings, according to our own records, his rating has increased to 70.4 marks after becoming the acting CE in mid March, and this is the beginning of the honeymoon period. Until his official report to the office in June, his rating has even climbed to 72.3 marks which is his record high, and gradually decrease to 69.0 marks in late July. Is this decrease caused by an occasional standard error or truly reflecting the reality?

First of all, let us look at the random sampling error. When Public Opinion Programme (POP) released the figures, the range of statistical sampling error has already been stated as appears in this paper as well, and it is also explained repetitively, '"95% confidence level" means that if we were to repeat a certain survey 100 times, using the same questions each time but with different random samples, we would expect 95 times getting a figure within the error margins specified'. From the results, all sampling error of percentages should be less than plus/minus 3 percentage points, but in the sampling error of ratings, that should be less than plus/minus 1.2 percentage points, but not plus/minus 3 percentage points Many people have misunderstanding about this. The rating of CE Donald Tsang in late July has dropped by 2.4 marks when comparing with the figures registered in early July, and decreased by 3.3 marks when comparing with late June's figures. The drop is obvious and clear, it cannot be explained by the sampling errors. Also, the support rating of Donald Tsang in late July has dropped by 5 percentage points when comparing with the figures in early July, this also exceeds the range of the sampling errors.

The more important is that when the author judges Donald Tsang's honeymoon period is over, it is in fact a cross reference of similar survey results in early August. They both show that two popularity ratings continue to drop, hence it is not just an occasional incident.

Of course, the continual drop and explanation of the trends may be two totally different stories. The author has stated some incidents which may affect Donald Tsang's popularity between the period of the most recent surveys. These incidents include the RTHK incident, the New World Development incident, and the threat of Streptococcus Suis infection. These are only personal judgment, and may have missed out something important. Nevertheless, these are almost the major explanations.

Summer holidays should originally be a relaxing time for public opinions, but things can happen unexpectedly, and the CE's popularity is not very stable. Therefore, the CE and government should not neglect these little signals observed from public sentiments. If the CE is not aware of this because of someone's personal belief, this is extremely irrational. Some commentaries say that Donald Tsang 'has done nothing wrong, where does his mistake come from?' is in fact negligent.

Fortunately, CE Donald Tsang is never careless, he seems to realize under the condition that his popularity is not stabilized and the epidemic situation is not yet under control, he can never enjoy his 'long vacation'. CE immediately met media executives, and also issued the 'Letter to Hong Kong', apart from elaborating his policy directions, he also drew people's attention to his forthcoming policy address in two months time. This is a much more brilliant attempt than just saying 'public opinion has ups and downs'.

No doubt, everyone is expecting Donald Tsang's first Policy Address. From Donald Tsang's previous experience and working style dealing with the Budget, the author is also waiting for 'the second spring of public opinion'. However, from the author's observation, the situation that Donald Tsang is facing is a two-sided danger. If the CE is not dealing with each alarm with serious attitude, the final victim would still be the people. Of course, all smart leaders will not be bounded by public opinion figures, and Donald Tsang has done a good job in promoting people-based governance and populism, and dismissed the discussion on 'honeymoon period' in a simple way.

The author has no intention and no need to exaggerate in order to attract mass attention. The concept of 'honeymoon period' actually depends on your own interpretation. By using the new marriage honeymoon period as a metaphor, honeymoon period can be seen as a sensational and a mutual affectionate stage under a fresh relationship and environment. After the honeymoon period, whether the relationship of the couple can last for a long time really depends on the actual comprise in their daily lives.

There was a famous saying by American President Ford shortly after he took office: "I do not want a honeymoon with you. I want a good marriage." Perhaps over-indulgence in one's honeymoon period would ultimately ruin one's marital life.

Table: Key Schedule of Donald Tsang's popularity trends

* "95% confidence level" means that if we were to repeat a certain survey 100 times, using the same questions each time but with different random samples, we would expect 95 times getting a figure within the error margins specified.