The Individual Visit Scheme and Its Development

Before the introduction of the IVS, mainland Chinese could visit Hong Kong on a restricted basis in the following three ways. The first was through family reunions, for those wishing to visit their relatives in Hong Kong. The second was as part of a group tour. If mainlanders wanted to go on a sightseeing trip to Hong Kong, they could join a group tour organized by one of the designated mainland tour companies. Visitors on a group tour had to arrive and depart together as a group. The last was through a business visit. This was for those who wanted to make a business trip to Hong Kong in a private capacity. All of these three forms of travel required applicants on the mainland to apply to the public security bureau office for an exit-entry permit for travelling to and from Hong Kong with an endorsement to visit relatives, join a group tour, or make a business trip.

In early 2003, an outbreak of SARS led to serious social panic and an economic recession in Hong Kong. The IVS was introduced in July 2003 under the framework of CEPA in order to restore social and economic confidence in the territory. The IVS was first applied to only four cities in Guangdong. It now covers 49 cities, including all provincial-level cities in Guangdong, the main cities in the Pan-Pearl River Delta region, and four municipalities. Table 1 presents the timeline of the implementation of the IVS.

Table 1: Timeline of the Implementation of the IVS

Source:

Commerce and Economic Development Bureau 2013, 3–4.

Under the IVS, mainland residents with a permanent household registration in the designated 49 cities are eligible to apply for the relevant exit endorsement, which is valid for three months or one year, and travel to Hong Kong as individuals. The holder of the endorsement is allowed to make one or two visits to Hong Kong and stay for not more than seven days per visit. Moreover, there is no quota on the number of endorsements that can be issued.Footnote ⁶

The most noticeable consequence of the IVS has been an enormous increase in the number of mainland visitors to Hong Kong (Figure 1). A year before the introduction of the IVS, this number was 6.8 million, or 41.2 per cent of the total number of visitor arrivals in Hong Kong. In 2008, the number had soared to 16.9 million, or 57.1 per cent of all visitor arrivals.Footnote ⁷ At the beginning of the IVS in 2003, the proportion of IVS visitors to total mainland visitors was only 7.9 per cent. In 2008, this had increased to 57.0 per cent. IVS visitors have become the main category of mainland visitors to Hong Kong.

Sources: Hong Kong Tourism Board 2003; 2014a.

Figure 1: Mainland Visitor Arrivals to Hong Kong, 1998–2013

The Hong Kong government has been engaged in active discussions with Beijing on how to make it easier for mainlanders to visit Hong Kong. In April 2009, a new measure was introduced that allowed eligible Shenzhen 深圳 permanent residents to apply for one-year multiple-entry IVS endorsements to come to Hong Kong. Under such an arrangement, one-year multiple-entry endorsement holders are allowed to travel to Hong Kong without any limitation on the number of visits in a year. Later, in December 2009, non-Guangdong permanent residents ordinarily residing in Shenzhen were allowed to apply for IVS endorsements in Shenzhen. In December 2010, most non-Guangdong permanent residents working in Shenzhen also became eligible to apply for these permits. All of these measures have made it more convenient for Shenzhen residents to visit Hong Kong, which has encouraged them to travel to Hong Kong more frequently, within a single day. In 2013, the number of mainland and IVS visitors increased to 40.7 million and 27.5 million, respectively. About half of all IVS visitors came from Shenzhen.Footnote ⁸

New IVS measures for Shenzhen residents have not only increased the number of visitors from Shenzhen but have also changed the duration of their stays. Compared with visitors from other mainland cities, Shenzhen visitors tend not to stay overnight, because Shenzhen is located immediately north of Hong Kong. In 2012, 90.9 per cent of one-year multiple-entry IVS visitors were same-day visitors, compared to 42.6 per cent of IVS visitors without one-year multiple-entry permits.Footnote ⁹ Therefore, the number of same-day mainland visitors (hereafter, same-day visitors) has ballooned from 2.8 million in 2003 to 23.7 million in 2013, contributing to a significant rise in the percentage of same-day visitors, from 32.8 per cent in 2003 to 58.1 per cent of total mainland visitors in 2013 (Figure 1). Same-day visits have replaced overnight visits as the most common form of travel for mainlanders.

Although both same-day and overnight mainland visitors (hereafter, overnight visitors) visit mainly for leisure, their consumption behaviour is significantly different. First, overnight visitors spend several times more than same-day visitors. In 2013, the per capita spending of overnight visitors was HK$8,937, while that of same-day visitors was only HK$2,721. Second, the spending of overnight visitors is more diversified. In 2013, same-day visitors spent 92.4 per cent of their travel budgets on shopping. The spending of overnight visitors, meanwhile, was 71.6 per cent on shopping, 11.7 per cent on hotel bills, 9.4 per cent on meals outside hotels, and 2.7 per cent on entertainment.Footnote ¹⁰ Third, overnight visitors tend to buy luxury goods, while same-day visitors like to buy daily necessaries. In 2013, 41 per cent, 25 per cent and 21 per cent of overnight visitors bought garments, leather/synthetic goods, and jewellery and watches, respectively, while only 15 per cent, 9 per cent and 6 per cent of same-day visitors did so. Of the same-day visitors, 54 per cent and 26 per cent bought foodstuffs, alcohol and tobacco, as well as personal care goods, but 40 per cent and 12 per cent of overnight visitors bought these two types of daily necessaries.Footnote ¹¹ Because of their differences, these two types of mainland visitors entering under the IVS may have different impacts on Hong Kong's society and economy.

Two Impacts of the Individual Visit Scheme on Hong Kong

The most notable result of introducing the IVS has been a significant increase in mainland visitors to Hong Kong. The effect of the IVS on tourism was immediate. The number of mainland visitor arrivals doubled from 6.8 million in 2002 to 12.2 million in 2004 (Figure 1). The IVS boosted tourism and thus benefited the retail, hotel and other related industries in Hong Kong. Tourism is an important pillar of the Hong Kong economy. IVS visitors generated an additional HK$6.5 billion in tourist spending in 2004.Footnote ¹² The unemployment rate dropped from 7.9 per cent in 2003 to 6.8 per cent in 2004.Footnote ¹³ Obviously, the IVS policy was a factor in Hong Kong's economic recovery.

In 2012, spending by IVS visitors had directly generated HK$26.1 billion in value added and over 110,000 jobs. It captured 1.3 per cent of GDP and 3.1 per cent of total employment.Footnote ¹⁴ Although the contribution of IVS visitor spending to the total GDP and employment was not large, the IVS accounted for 11 per cent of the growth in employment from 2004 to 2007, and 25.8 per cent from 2007 to 2012.Footnote ¹⁵ Given that tourism is a labour-intensive industry, this rapid growth has created a number of job opportunities, especially for people with little in the way of skills or professional training.

However, same-day visitors have not contributed as significantly to Hong Kong's economy as have overnight visitors. The per capita spending of same-day visitors is lower. Although the total number of same-day visitors has exceeded that of overnight visitors since 2011, the total level of consumption of same-day visitors is still less than that of overnight visitors. In 2013, same-day visitors spent a total of HK$64.4 billion, less than half of the HK$152.7 billion spent by overnight visitors.Footnote ¹⁶ Hence, it can be assumed that same-day visitors have generated fewer jobs than overnight visitors. Moreover, same-day visitors spend most of their money on shopping, so industries other than retailing do not benefit from the increasing number of same-day visitors resulting from the one-year multiple-entry policy.

More importantly, the rise in the number of same-day visitors has created several social problems in Hong Kong. One such problem is the bulk buying of milk powder in Hong Kong, leading to a shortage of supply and to panic from local mothers unable to buy milk powder for their babies. In fact, the “milk powder issue” is seen as the tip of the iceberg for food safety in China. It reminds mainlanders about a scandal that erupted in September 2008 involving the Sanlu Group and concerning baby formula poisoned with melamine. Contaminated milk was later found in the products of 21 mainland milk product companies, including the biggest.Footnote ¹⁷ As a result, mainland parents have lost confidence in domestically produced milk and prefer to buy foreign brands of milk powder. The introduction of one-year multiple-entry permits has made it more convenient for Shenzhen residents to buy milk powder in Hong Kong during same-day trips.

Subsequently, IVS visitors prefer to buy not only milk powder but also other daily necessities in Hong Kong because of the assurance of quality. The huge demand on the mainland for Hong Kong goods and the introduction of one-year multiple-entry permits has given rise to parallel trading. Many Shenzhen residents take advantage of the multiple-entry policy to buy large quantities of goods in Hong Kong to sell on the mainland. According to Shenzhen customs, more than 20,000 parallel traders cross the border between Shenzhen and Hong Kong every day.Footnote ¹⁸ As a result, in 2013, Hong Kong suffered from a shortage of some brands of baby formula. The Hong Kong government then brought in an administrative measure that allows travellers to the mainland to carry only two tins of milk powder per day.

Parallel trading has also disturbed the daily life of local residents in the towns in the New Territories. Parallel traders (of whom around 40 per cent are mainlanders and the rest are Hong Kong residents)Footnote ¹⁹ are most active in Sheung Shui 上水 because it is only one railway station away from Lo Wu 罗湖, which lies on the border between Hong Kong and Shenzhen. They buy goods in Sheung Shui and then return to Shenzhen by train. Their activities are a nuisance to local residents and are believed to have pushed up the prices of various daily necessities in the district. The average price of everyday goods in Sheung Shui is reportedly 10 to 20 per cent higher than in nearby districts.Footnote ²⁰ The “Reclaim Sheung Shui” movement, a protest against parallel traders, erupted in September 2012. People engaged in parallel trading activities were forced to spread out to other areas. Such “reclaim” movements then followed them to other districts such as Shatin 沙田 and Mongkok 旺角.

Many large luxury, electronic, and pharmaceutical companies have established stores in districts frequently visited by IVS visitors. Commercial premises in these districts tend to command higher rents, forcing small businesses serving locals to relocate. As the commercial premises have become more crowded with mainland visitors, shops have become more oriented towards mainlanders. For example, in a promotion, Mannings pharmacy gave coupons only to mainland passport-holders. Dolce & Gabbana banned Hong Kong people from taking photographs of its shop front, but allowed mainlanders to do so. In response, thousands of local protesters took photographs outside the store and demanded an apology from Dolce & Gabbana.

Given that anchor babies are given the right of abode in Hong Kong, the IVS also offers opportunities for mainland women to give birth in Hong Kong. The number of births by mainland women not married to a Hong Kong permanent resident increased from 1,250 in 2002 to 35,736 in 2011.Footnote ²¹ This enormous increase has made it more difficult for local pregnant women to find a hospital bed. Although the parents of anchor babies do not live in Hong Kong, more than half of them plan to send their children to Hong Kong.Footnote ²² The number of children crossing the border to attend schools in Hong Kong has doubled in the past few years, contributing to a shortage of places in local kindergartens and primary schools, particularly in areas near Shenzhen.Footnote ²³ In 2013, hundreds of local parents took to the streets urging the government to promise that the children of Hong Kong parents would all be able to get into kindergartens and primary schools in their local district.

All of these developments since the introduction of one-year multiple-entry permits in 2009 have exacerbated tensions between local Hong Kong people and mainlanders.Footnote ²⁴ Some local people have begun to label mainlanders “locusts” because they feel that mainlanders “rob” them of resources such as local hospital beds, school places and daily necessities. In 2014, an “anti-locust” protest broke out in Tsim Sha Tsui 尖沙咀, a key shopping district in the city, demanding that the government curb the number of mainland visitors. The participants shouted anti-mainlander slogans such as “Go back to China” and “Reclaim Hong Kong.” The target of protests has shifted from parallel traders to IVS visitors and even all mainland visitors. There has been increasing public dissatisfaction with the way that the Hong Kong government has dealt with the social problems created by the rapid growth of same-day visitors since 2009.

In short, as IVS visitors continuously poured into Hong Kong, which seemed to have reached optimal capacity, further increasing their number in 2009 clearly created more problems than benefits. One can see that, on the positive side, the increase in mainland visitors to Hong Kong, particularly overnight visitors, has stimulated an economic recovery and reduced the unemployment rate. However, on the negative side, since the introduction of the one-year multiple-entry policy in 2009, the rapid growth in the number of same-day visitors has created several social problems, such as the rise of parallel trading and the increase of anchor babies, augmenting tensions between local Hong Kong people and mainlanders. The local people expect, and have repeatedly asked, the Hong Kong government to solve all of these problems, but the government seems to have turned a deaf ear to these demands or has simply failed to find solutions. As a result, public dissatisfaction with the government's performance has been increasing. In the next section, we discuss the relationship between the failure of the government to address these problems and the decline in the political trust of Hong Kong people.

Political Trust in Hong Kong and the Individual Visit Scheme

Political trust is defined by many scholars as a basic evaluative orientation towards the government.Footnote ²⁵ Arthur Miller and Ola Listhaug state that political trust reflects evaluations of whether or not political authorities and institutions are performing in accordance with the normative expectations held by the public.Footnote ²⁶ Thomas Rudolph and Jillian Evans share a similar view that political trust is an affective orientation towards the government which reflects the satisfaction of citizens with government policies. Political trust occurs when policies are relatively effective and consistent with public expectations.Footnote ²⁷ Marc Hetherington and Jason Husser further point out that political trust is the people's evaluation of the government's performance relative to their normative expectations of how the government ought to perform. This conceptual definition of political trust matches the findings of existing studies showing that trust is a function of perceived government performance and some long-term factors.Footnote ²⁸

Institutional and cultural approaches are two major theoretical traditions that have developed to explain political trust.Footnote ²⁹ Adherents of the institutional approach believe that political trust is endogenous. Based upon the perspective of rational choice, they argue that political trust arises from the rational responses of individuals to the performance of political institutions.Footnote ³⁰ Hence, institutionalists typically emphasize the importance of policy outcomes, especially economic outcomes.Footnote ³¹ In contrast, adherents of the cultural approach assert that trust in political institutions is exogenous, originating from interpersonal trust which has been learned through socialization early in life and later projected onto political institutions.Footnote ³² In the view of culturalists, people's predispositions towards political institutions are conditioned by cultural values and beliefs, which are transmitted from one generation to the next and vary across cultures and societies.

In recent years, both approaches have been used to explain political trust in Hong Kong. Institutional factors have been found to be more powerful than cultural factors in explaining the case of Hong Kong. Empirical studies by Timothy Wong, Po-san Wan and Michael Hsiao have demonstrated that cultural factors, including interpersonal trust, traditionalism and post-materialism, have weak and inconsistent effects on trust in different political institutions in Hong Kong, but institutional factors such as government performance and life satisfaction are significant determinants.Footnote ³³ Soo Jiuan Tan and Siok Kuan Tambyah have similarly found that generalized trust is not correlated with trust in political and public institutions in Hong Kong.Footnote ³⁴ The superiority of the institutional approach over the cultural approach is also valid in other Asian societies. Soonhee Kim found that in Japan and South Korea satisfaction with the government's performance was significantly correlated with public trust in the government but the value of self-expression was not.Footnote ³⁵ Therefore, it is more suitable to use the institutional approach to explain Hong Kong people's trust in political institutions.

Applying the institutional approach to the context of tourism, Robin Nunkoo and Stephen Smith suggest that the outcome of policies on tourism has determined public trust in government actors.Footnote ³⁶ Political trust varies according to the benefits and costs of developing tourism. Individuals would have a higher degree of trust in their government if they perceived that existing tourism policies brought about higher benefits and lower costs. Since the IVS is one of the most important tourism policies pursued by the Hong Kong government in recent decades, we believe that it should have an important effect in explaining political trust in Hong Kong.

The institutional approach can be further divided into macro-oriented and micro-oriented theories.Footnote ³⁷ The former group of theories stresses that political trust is a collective property broadly shared by all members of a society, and that popular trust in political institutions is thus the consequence of policy and political performance, and especially of economic outcomes. By contrast, the latter group recognizes that political trust varies among individuals within a society based on differences in individual circumstances and values. We formulated hypotheses on the relationship between the IVS and political trust based on these two groups of institutional theories.

At the macro level, we hypothesized that the number of mainland visitors to Hong Kong is correlated with the trust that the Hong Kong people have in their government, because the most notable and direct result of introducing the IVS has been a drastic increase in the number of mainland visitors to Hong Kong. As discussed in the previous section, the increase in overnight visitors has contributed to an economic recovery and a reduction in the unemployment rate in Hong Kong. However, the growth in the number of same-day visitors has created several social problems and increased tensions between local people and mainlanders. The fact that the Hong Kong government has failed to solve these problems effectively has inevitably undermined its standing in the eyes of citizens. Hence, the number of overnight visitors is likely to have had a different impact on political trust than the number of same-day visitors. The former should have had a positive effect on the perceived government performance and the latter a negative one. We therefore formulated hypotheses for these two groups of visitors and their numbers separately. Two different hypotheses are offered on the number of mainland visitors:

The opportunities brought about by regional integration have not translated equally into benefits for all classes of Hong Kong people.Footnote ³⁸ Spending by mainland visitors has benefited selected industries, namely, the retail trade, accommodation services (hotels), and food services (restaurants).Footnote ³⁹ Compared with more highly educated people, those with less education are more likely to work in these industries. They should receive more benefits from the IVS policy and in turn be more tolerant of the social problems created by the policy. According to the micro-oriented approach, the perception that individuals have of the benefits and costs of the increasing number of mainland visitors should be dependent on their own life experiences. Therefore, the impact that the mainland visitors have had on political trust should vary across different educational groups. Two other hypotheses are offered:

Data, Variables and Methods

The study aims to examine the effect of the macro environment on individual attitudes, so both macro- and individual-level data are used. The individual-level data used in this study came from a longitudinal telephone survey research project that has been conducted semi-annually since April 1998 by the Hong Kong Institute of Asia-Pacific Studies of the Chinese University of Hong Kong. It is an ongoing longitudinal study. The goal is to chart changes in public attitudes towards the practice of “one country, two systems” in Hong Kong after the 1997 handover.

The sampling for every semi-annual survey involves two steps. First, a fixed set of telephone numbers is randomly selected from the latest Hong Kong residential telephone directory. To cover unlisted and new numbers, the last two digits of the selected telephone numbers are replaced with two randomly selected digits within the range of 00–99. Second, after successfully reaching the selected resident, a family member aged 18 or above is selected for an interview in accordance with the last birthday rule. For every valid telephone number, a maximum of three contacts at different times is made before the number is classified as “unanswered.” The response rate for each survey has been around 50 per cent and the samples have ranged in size from 800 to 1,000.

By October 2013, 32 surveys had been completed. The analysis in this article is based on a pooled dataset of all 32 surveys. Since a random sampling method in selecting eligible respondents was employed in all of the surveys, we believe that the views of the respondents basically reflect the views of Hong Kong citizens as a whole.

As mentioned before, political trust can be broadly viewed as evaluations of political authorities and institutions. However, political authorities and institutions cover many different branches, and there is no consensus on the number and objects of branches included in the measurement of political trust.Footnote ⁴⁰ If we create a generalized political trust index consisting of all relevant political institutions, this measurement would bring together so many institutions that it could not represent trust in political institutions.Footnote ⁴¹ Recognizing this, we limited the objects of trust to the executive branch of government. We measured the Hong Kong people's trust in their government by directly asking whether the respondents trusted the Hong Kong government. The response categories were coded on a three-point scale ranging from 1 (distrust) to 3 (trust). The higher the score, the higher the level of political trust.

According to the institutional approach, political trust arises from the rational responses of individuals to the performance of the government. Government administration covers different domains, so we selected four indicators of government performance as control variables. On the performance of the Hong Kong government, the respondents were asked to evaluate the following domains: dealing with democratic reform; dealing with freedom of speech and assembly; dealing with housing; and dealing with economic development. The response categories for all of the indicators were coded on a three-point scale, ranging from 1 (dissatisfied) to 3 (satisfied). The higher the score, the more favourable the evaluation. In addition, gender, age, and level of education were selected as control variables, because some socio-demographic variables were also found to play a role in shaping political trust.Footnote ⁴²

At the macro level, the respective numbers of overnight and same-day visitors to Hong Kong were the independent variables in our hypotheses. The number of IVS visitors was also included to examine the effect of the IVS policy. These three variables of mainland visitors were measured annually in terms of millions of persons. The information was drawn from the Hong Kong Tourism Board, which is a government-subvented body tasked with marketing and promoting Hong Kong as a travel destination worldwide.

Apart from the numbers of mainland visitors, five macro-level factors were selected as control variables. The first two variables had economic relevance. They were GDP per capita and the unemployment rate, both of which are collected and measured annually by the Hong Kong Census and Statistics Department. The third variable was the Centa-City Index (CCI), which is a monthly index based on all transaction records as registered with the Land Registry to reflect secondary private residential property price movements.Footnote ⁴³ High property prices have been one of the most important social issues that the Hong Kong government has had to address. It is believed that fluctuations in property prices reflect the performance of the government and thus affect the public's evaluation of the government.Footnote ⁴⁴ The last two variables were time period dummies of two incidents, namely, the recovery from SARS in 2004 and the resignation of Tung Chee-hwa 董建华 as chief executive in 2005.

To simultaneously model the effect of macro-level variables and their interaction with individual characteristics on individual-level political trust, we employed a multilevel modelling framework to analyse our combined macro-micro dataset. Specifically, we exercised random-effects generalized least squares (GLS) regressions with STATA. This statistical model was suitable for our research because it allowed for estimates to be made of variations in the political trust of individuals in response to changes in the number of mainland visitors, with the inclusion of macro- and individual-level control variables.

Findings

Figure 2 indicates the trend in political trust in Hong Kong from 1998 to 2013. Generally, public trust in the Hong Kong government decreased continuously from 2.44 points in April 1998 to 1.83 points in April 2003, owing to the economic recession following the 1997 Asian financial crisis and the outbreak of SARS in 2003. The IVS was introduced in July 2003, so more overnight visitors visited Hong Kong after that date. The level of political trust then began to rise after October 2003. Political trust rose rapidly and reached its highest level at around 2.5 points in 2006 to early 2008. However, political trust declined after October 2008. In April 2009, one-year multiple-entry permits were introduced for Shenzhen permanent residents and the number of same-day visitors increased sharply, but the fall in the level of political trust never stopped. By October 2013, the level of political trust had decreased to 2.07 points. The trend in political trust seems to support our hypotheses. Random-effects GLS regressions were exercised for further examination.

Figure 2: Political Trust in Hong Kong, 1998–2013

The results of the random-effects GLS regressions are reported in Table 2. Political trust was our dependent variable. Socio-demographic factors including gender, age, and level of education were the control variables. In order to assess the extent to which the government's performance in dealing with the IVS policy affects individual trust in the government, we needed to control the factors of government performance in other domains. Three individual-level evaluation variables, including individual evaluations of democratic reform, freedom, and housing policy, and macro-level CCI were included as control variables.

Table 2: Random-effects GLS Regression on Political Trust

Notes:

Standard errors and variances are not presented. * p < 0.05, ** p < 0.01, *** p < 0.001, two tailed tests.

In model 1, time period dummies were the independent variables. Time periods were divided into the following three groups: (1) before the introduction of the IVS (April 1998–April 2003; (2) the introduction of the IVS (October 2003–April 2009); and (3) after the introduction of the one-year multiple-entry policy (October 2009–October 2013). The introduction of the IVS dummy was chosen as the reference group. The coefficients of two period dummies were negative at either the 0.01 or 0.001 significance level, implying that the level of political trust in the period of the introduction of IVS was higher than before the introduction of IVS and after the introduction of the one-year multiple-entry policy, all other things being constant. This difference in period was independent of the government's performance in the areas of democratic reform, freedom, and housing policies.

To examine whether the IVS policy contributed to this difference in period, in model 2 we added the number of IVS visitors and its interaction with the “after the introduction of the one-year multiple-entry policy” period dummy as the independent variable. Both of these variables were statistically significant. The coefficient of the number of IVS visitors was positive at the 0.05 level and its interaction was negative at the 0.001 level. More importantly, the coefficient of the interaction term was larger than that of the number of IVS visitors. This implies that the number of IVS visitors had two different kinds of impact on political trust. The number of IVS visitors had a positive impact on political trust before the introduction of the one-year multiple-entry policy, but the impact became negative after the introduction of that policy. Moreover, it is interesting to note that the two period dummies did not continue to be negative when the number of IVS visitors was controlled.Footnote ⁴⁵ These findings support the view that the trend in the number of IVS visitors to Hong Kong helps to explain the rise in political trust at the beginning of the IVS policy and its fall since the introduction of the one-year multiple-entry policy.

As discussed before, we believe that the numbers of overnight and same-day visitors have two different kinds of impact on political trust, and thus a change in the ratio of the number of overnight visitors to same-day visitors may lead to a shift in the impact of the IVS policy. Hence, in model 3, we added the respective numbers of these two mainland visitor groups instead of period dummies, and the number of IVS visitors and their interaction term. Both surpassed the 0.001 significance level. The level of political trust in Hong Kong would be higher if more overnight visitors and fewer same-day visitors were to visit Hong Kong. These findings support H1a and H1b.

One may question whether the rise in political trust after 2003 is the result of the recovery from SARS and the resignation of Tung as chief executive rather than of the inflow of overnight visitors to Hong Kong. Hence, in model 4, we added the recovery from SARS and Tung's resignation as control variables. If this counter-argument is correct, then these two dummies would be positive and significant; however, the results reveal that they are not. More importantly, the coefficient of the number of overnight visitors remains positive and significant, implying that the relationship between overnight visitors and political trust is independent of these two incidents.

In model 5, we added GDP per capita and the unemployment rate in order to determine whether the positive impact of overnight visitors was mediated by improvement in the economy. Although GDP per capita failed to pass the significance level, the unemployment rate was found to be negative at the 0.001 significance level. Individuals will have stronger political trust when the unemployment rate is lower. At the same time, the coefficient of the number of overnight visitors turned out to be insignificant. Previous studies have found that the inflow of IVS visitors, especially overnight visitors, helped to reduce the unemployment rate.Footnote ⁴⁶ Accordingly, the inflow of overnight visitors has contributed to the rise in political trust by reducing the unemployment rate.

The effect of macro-level factors on political trust is always indirect and mediated at the micro level by the perception of individuals.Footnote ⁴⁷ In model 6, therefore, we replaced GDP per capita and the unemployment rate with the individual evaluations of the government's performance in dealing with economic development. The coefficient of the evaluation of economic development was positive and surpassed the 0.001 significance level. The number of overnight visitors became insignificant again in model 6. Since the number of overnight visitors was found to be positively correlated with individual evaluations of economic development,Footnote ⁴⁸ the inflow of overnight visitors helped to build political trust by improving evaluations of the government's performance in dealing with economic development. In short, the findings in models 5 and 6 support the argument that the inflow of overnight visitors improved political trust by increasing the employment rate and, thus, the public's satisfaction with the government's performance in economic development.

We hypothesized in H2a and H2b that the impact of mainland visitors on political trust depends on the level of education of individuals, and then added the interaction terms for the number of mainland visitors using education dummies. Control variables in model 3 were also included. Table 3 reports the results of random-effects GLS regressions with these interaction terms. We added the interaction with education dummies for the number of overnight and same-day visitors in models 7 and 8, respectively. Two interaction terms with a senior secondary level dummy were found to be statistically significant and positive. In addition to a positive coefficient for overnight visitors, the results in model 6 imply that the growth rate in the political trust of people educated to a senior secondary level from the increase in overnight visitors was higher than that of those who had received a tertiary education. Unlike that for overnight visitors, the coefficient of the number of same-day visitors was negative. The results in model 8 show that the reduction in the rate of political trust of people with a senior secondary level of education from the increase in the number of same-day visitors was lower than that of those with a tertiary education.

Table 3: Random-effects GLS Regression on Political Trust with Interaction Term

Notes:

Standard errors and variances are not presented. * p < 0.05, ** p < 0.01, *** p < 0.001, two tailed tests.

Interaction terms with the dummy for junior secondary education or below failed to pass the significance level in models 7 and 8. This may be because there was no significant difference between people with a tertiary or junior secondary level of education or below with regard to the impact of mainland visitors. We therefore removed the interaction terms with the dummy for a junior secondary education or below in models 9 and 10. Interaction terms with the dummy for a senior secondary education continued to be positive and statistically significant. This reveals that the positive impact of the increasing number of overnight visitors on political trust has been stronger for people with a senior secondary education than for those with a tertiary or junior secondary education or below, but that the negative impact of the increase in same-day visitors was weaker for people with a senior secondary education. Therefore, H2a and H2b are partly supported.