Evaluating museum free admission policy

Annals of Tourism Research 58 (2016) 156–170

Contents lists available at ScienceDirect

Annals of Tourism Research journal homepage: www.elsevier.com/locate/atoures

Research Notes and Reports

Evaluating museum free admission policy Chiang-Ming Chen, Yen-Chien Chen ⇑, Yi-Chun Tsai National Chi Nan University, Taiwan

This paper aims to evaluate the causal effect of free admission policy of public museums on the number of visits. Free charge of public museum is a topic that has been debated for a long time. The nature of museum has been recognized as an educational institution since the early nineteenth century (Hooper-Greenhill, 1991). Based on the view of educational function, public museums are suggested to be funded by central government, and politicians take museum free admission policy as one of the principal cultural policy achievements. National museums and galleries in England sponsored by Department for Culture, Media and Sport (DCMS) introduced free admission policy since December 2001. Compared to the period before 2001, visits increased by 158% in 2011 (DCMS, 2015). Being inspired in part by the free admission policy in England, visitors can enjoy 17 museums and galleries in Paris free of charge in the first half of 2008. According to the tourism report published by Office of Tourism and Congress of Paris, nine of the top visitor sites in 2009 are free museums. The issue of charging by museum is a source of political debate, however, the academic studies on the effect of charging or not charging are relative absent. Bailey and Falconer (1998) mentioned that many concerns should be taken into account when deciding whether public museums should charge. Maddison and Foster (2003) argued that museum free admission policy would lead to a certain level of cost of congestion, and the benefit of the increase in visits is hard to evaluate. Cowell (2007) used the data from DCMS and found that total number of museum visits is growing year by year from 2001 to 2004. Cowell (2007) also suggested that the future study on the impact of museum free admission policy needs to evaluate ‘‘the likely number of additional visits which is ascribed directly to free admission”. This paper, motivated by Cowell (2007)’s suggestion, uses the data of museums in Taiwan from 2004 to 2011, and employs the difference-in-difference (DID) methodology (Abadie, 2005) to evaluate the changes in the number of museum visits causally induced by the free admission policy. The finding in this study could be expected to add to the literature of museum admission policy. Our panel data of museum attendance is obtained from ‘‘Visitors to the Principal Scenic Spots in Taiwan 2004–2011” which is published by the Taiwan Tourism Bureau. Fig. 1 plots the monthly average number of museum visits in Taipei City and New Taipei City over the sample period. As Fig. 1 shows, the curve of museums in Taipei City and private museums lies mostly above the curve of public museums in New Taipei City before 2010. After the implementation of public museum free admission policy, the difference between both curves shrinks over time. The number of public museum attendances in New Taipei City even exceeds the corresponding value of Taipei City and New Taipei City during the first two months of 2010. ⇑ Corresponding author at: Department of Economics, National Chi Nan University, 1 University Road, Puli, Nantou 54561, Taiwan. Tel.: +886 49 2910960x4689; fax: +886 49 291443. E-mail addresses: [email protected] (C.-M. Chen), [email protected] (Y.-C. Chen), [email protected] (Y.-C. Tsai).

157

Research Notes and Reports / Annals of Tourism Research 58 (2016) 156–170

140,000 Taipei museum and New Taipei private museum

The number of visits

120,000 100,000

New Taipei Public musuem 80,000 60,000 40,000 20,000

Jul-11

Jan-11

Jul-10

Jan-10

Jul-09

Jan-09

Jul-08

Jan-08

Jul-07

Jan-07

Jul-06

Jan-06

Jul-05

Jan-05

Jul-04

Jan-04

0

year

Fig. 1. Monthly average number of museum visits in Taipei City and New Taipei City from January 2004 to December 2011.

In order to promote arts and culture, New Taipei City government inaugurated public museum free admission policy in January 2010. The common traditional methodology to measure the policy effect in previous studies is to compare the number of museum visits before and after policy implementation. However, such measurement may be spurious. For example, visits may be boosted by time trend or other factors during the observing period, not by free admission policy per se. The DID estimator comes from the comparison of the difference between treatment and control groups before and after policy implementation. The role of control group is to eliminate interventions caused by time trend or other factors which are common to treatment and control groups. The treatment group in our DID strategy is public museums in New Taipei City, and control group is private museums in New Taipei City and museums in Taipei City. The DID estimator for the change in the number of museum visits (denoted by Y) ascribed to free admission policy is:

    After Free Before Free After Free Free DFree ¼ Y treatment  Y control  Y Before treatment  Y control

ð1Þ

The DID estimator (DFree ) can be captured by the estimates of c3 (the coefficient of interaction of Treatment and Free) in the Eq. (2):

Y ¼ c0 þ c1 Treatment þ c2 Free þ c3 Treatment  Free þ b0 X þ e

ð2Þ

where Treatment is a dummy variable, equal to 1 if the sample is belong to treatment group; Free is also a dummy variable, equal to one if observing time period is after the introduction of the free admission policy in January 2010; X is a vector of control variables including the relative humidity for each month (Wet), macroeconomic conditions and full set of monthly dummies. Macroeconomic conditions contain the monthly number of inbound tourists to Taiwan (ARR), GDP per capita (GDP) and the value of NTD per US dollar (EXG) are included in our empirical analysis. We first employ pooled OLS to estimate Eq. (2). Although we control for background factors as possible as we can, problems of serial correlation and omitted variables bias (OVB) could also exist. To address OVB, we further use random-effect and fixed-effect models to measure the DID estimator. Our analysis results based on 1,786 observations are reported in Table 1. Using pooled OLS to measure the DID estimator, one can see that the introduction of free admission policy increases the number of museum visits significantly by 15,111 (the coefficient of interaction of Free and Treatment). The DID estimator from pooled OLS by using panel data may have problems of serial correlation and OVB. The value of Hausman test is 13.13 with a P-value of 0.34, which implies that our data has limited problem of OVB and suggests that random-effect model is more appropriate to measure the DID esti-

158

Research Notes and Reports / Annals of Tourism Research 58 (2016) 156–170

Table 1 Estimates of free admission policy on museum visits. Variables

Random-effect model

Fixed-effect model

Coefficient

t-Statistics

Coefficient

t-Statistics

Coefficient

t-Statistics

Free Treatment Free*Treatment

23,238.28 12,394.95 15,111.46

3.82⁄⁄⁄ 3.33⁄⁄⁄ 2.50⁄⁄

5,640.58 5,583.88 6,835.239

1.52 0.38 1.90⁄

6,269.04 – 6,686.69

1.71⁄ – 1.87⁄

Weather factors Wet

3,218.658

12.47⁄⁄⁄

735.56

3.01

640.24

2.62⁄⁄⁄

Macroeconomic conditions ARR GDP EXG

0.23 0.05 1,707.88

4.66⁄⁄⁄ 0.05 1.11

0.07 0.30 3,241.38

2.25⁄⁄ 0.58 3.54

0.06 0.31 3,261.74

2.16⁄⁄ 0.59 3.59⁄⁄⁄

Season effects January February March April May June July August September October November Constant

25,059.16 31,170.02 7,319.09 10,643.00 7,850.23 7,850.03 24,981.32 22,352.30 6,626.77 15,377.34 8,967.87 306,866.40

3.05⁄⁄⁄ 3.93⁄⁄⁄ 1.00 1.42 1.05 1.02 3.28⁄⁄⁄ 3.08⁄⁄⁄ 0.90 2.11⁄⁄ 1.32 4.15⁄⁄⁄

513.23 9,178.25 163.04 5,086.85 1,951.20 5,876.66 20,939.58 16,852.38 2,858.43 4,761.254 4,164.701 175,706.00

0.10 1.90 ⁄ 0.04 1.14 0.44 1.27 4.60⁄⁄⁄ 3.90⁄⁄⁄ 0.65 1.10 1.03 3.75⁄⁄⁄

166.62 8,551.55 350.77 5,002.69 1,843.23 6,229.34 20,999.29 16,804.78 3,040.25 4,515.63 4,021.03 168,383.5

0.03 1.79⁄ 0.08 1.13 0.42 1.36 4.66⁄⁄⁄ 3.93⁄⁄⁄ 0.70 1.05 1.01 3.68⁄⁄⁄

Number of observation LM test (P-value) Hausman test (P-value)

Pooled OLS

⁄⁄

⁄⁄⁄

1,786 641.67(0.00) 13.13(0.34)

Note: ⁄, ⁄⁄ and ⁄⁄⁄ indicate statistical significance at 10%, 5% and 1% levels, respectively. Dependent variable: number of museum visits. Mean of number of museum visits = 46,393.7

mator. Therefore, the causal estimates in our analysis would be driven by estimates of random-effect model. The estimates of random-effect model show that the free admission policy leads to an increase of 6835 museum visits significantly. It accounts for 14.73 percent of monthly average museum attendances (=46,394). We also report the fixed-effect estimations in Table 1. One can see that the gap in the coefficients of interaction between random-effect and fixed-effect model is small. In summary, our robust estimates suggest that the free admission policy can directly boost the number of museum visits. For other background factors, our random-effect estimations show humidity level (Wet) have a negative influence on museum visits; ARR increases museum attendances, while EXG is associated negatively with museum attendances. This paper evaluates the impact of the free admission policy of New Taipei City’s museum, introduced in January 2010, on the number of museum visits. Using panel data from January 2004 to December 2011, we adopt a difference-in- difference (DID) methodology to estimate the causal effect of free museum policy. We find that an increase of 14.73% of museum visits can be ascribed to the free museum admission policy. This finding can be a benchmark for future research on cost and benefit of free museum admission policy. References Abadie, A. (2005). Semiparametric difference-in-difference estimators. Review of Economic Studies, 72, 1–19. Bailey, S. J., & Falconer, P. (1998). Charging for admission to museums and galleries: A framework for analyzing the Impact on Access. Journal of Culture Economics, 22, 167–177. Cowell, B. (2007). Measuring the impact of free admission. Cultural Trends, 16(3), 203–224.

Research Notes and Reports / Annals of Tourism Research 58 (2016) 156–170

159

Department for Culture, Media and Sport (2015). Policy Paper: 2010 to 2015 government policy: Museums and galleries. Hooper-Greenhill, E. (1991). Museum and gallery education.Leicester: Leicester University Press. Maddison, D., & Foster, T. (2003). Valuing congestion costs in the British museum. Oxford Economic Papers, 55(1), 173–190.

Received in 7 November 2015. Revised 20 January 2016. Accepted 14 March 2016. Ó 2016 Elsevier Ltd. All rights reserved. Available online 31 March 2016 http://dx.doi.org/10.1016/j.annals.2016.03.002 _____________________________________________________________________________________________________________________

Someone’s been sleeping in my bed Logi Karlsson ⇑, Sara Dolnicar 1 The University of Queensland, Australia

Shortly after Goldilocks fell asleep in a stranger’s bed she was woken up by a very surprised bear family. Goldilocks had to make a run for it. Did she have to? Not in times of the ‘‘sharing economy” where people choose to share their unused properties with strangers (Botsman & Rogers, 2011). In the sharing economy, peer-to-peer accommodation networks like Airbnb and Roomorama would have enabled Goldilocks to have a good night’s sleep, a free bowl of porridge, a good social experience and many happy follow-up visits to the bear family. Peer-to-peer networks have developed as a consequence of Web 2.0 (O’Reilly, 2007). Similar to ebay.com, the world’s largest online marketplace (Nair, 2014), peer-to-peer networks facilitate ‘‘business” transactions. What separates peer-to-peer networks from electronic markets is that the main aim is sharing and borrowing, not buying (Gansky, 2010). Unused capacity, such as a spare bedroom, can be listed online for other to borrow, either for a fee (e.g. www.airbnb.com), or for free (e.g. www.couchsurfing.com). Time magazine refers to the sharing economy as one of ten ideas which will change the world (Walsh, 2011). The sharing economy displays the key characteristic of a disruptive innovation (Bower & Christensen, 1995): initially it performs worse than mainstream providers because the offer is not speaking to present customer demand. Disruptive innovation caters for future customer needs. Once the market has transformed, mainstream providers have typically missed the opportunity to catch up with satisfying the needs under the new circumstances. The accommodation sector is being radically transformed by peer-to-peer networks, with Airbnb.com a pioneer (Zervas, Proserpio, & Byers, 2015). Tourists can live with locals or in the houses of locals instead of staying in hotels. Airbnb’s success points to high demand due to attractive prices (Tussyadiah & Pesonen, 2015), connecting with local people (Guttentag, 2013) and exploring off the beaten track experiences (Guttentag, 2013). Less is known, however, about the supply side. Why do people allow strangers to sleep in their houses, their spare bedrooms or even their own beds in their absence? In the traditional accommodation sector the reason for providing accommodation to tourists is obvious: profit. This is not the case, however, in peer-to-peer networks. There is a very practical possibility that strangers could steal furnishings, damage the property or—in the worst case—burn it down. Why then are more and more people willing to serve as hosts and let strangers into their houses and their beds? The general, not tourism-specific, literature suggests that the supply side is motivated by the financial incentive (Stephany, 2015) and the motivation to share unused capacity so less is ⇑ Corresponding author. E-mail addresses:[email protected] (L. Karlsson), [email protected] (S. Dolnicar). 1

Tel.: +61 733656702.