Restricting the construction of second homes in tourist destinations: an effective intervention towards sustainability?

This study evaluates the impact of the second home restriction (Lex Weber) on construction investments in Switzerland. Different specifications and variations are defined, and the changes in investment behaviour are analysed. A first basic model considers the variation in investments through a discrete event study analysis, while a second model includes a continuous treatment variable to account for the different strong impacts of the law in various municipalities. In the second step, these two basic models are developed in a way that allows to include controls for the different cantonal legislation. The results show that with a delay of three years (due to the expiration of permits), the law had the desired impact on new construction investments. Additionally, the importance of the changes on different outcomes in the local economy and tourism industry are discussed.


Introduction
In Switzerland, the demand for second home ownership for leisure purposes grew very quickly in the late 1960s and early 1970s. The increasing construction activity resulted in the rapid consumption of construction grounds since hardly any restrictions and regulations for land use were applied. To address this increasingly important issue, several regional and local regulations were enacted, and spatial planning was undertaken. The last Federal Act on Spatial Planning (Spatial Planning Act, SPA) in 2011 requested cantons to identify possible ways to limit and guide the excessive construction of leisure homes in highly affected municipalities. Many of the measures enacted at the local level were financial disincentives, taxes and cross subsidies for primary homeowners and managed accommodations. Finally, on 11 March 2012, a federal popular initiative (Lex Weber) to limit second home construction in highly affected municipalities was approved by the majority of Swiss people and cantons. This was not expected to happen, since the most affected regions voted against the law, but people in the rather urban areas were favourable to this legal change. On this date, permits for new constructions were restricted to primary home purposes, and immediate transitory measures were applied. The federal council developed a regulation on new constructions, which was enacted by the end of 2012. Following its enactment, the parliament decreed the final second home law, which replaced the transitory measures starting in 2016, this as a formal change without any important change in the legal practice (Vinzens & Hefti, 2014).
A question that requires further analysis is what impact the number of second homes has on managed accommodations. One hypothesis is that among new second homeowners, overnight stays would be shifted away from hotels to their new flats and houses. The alternative to this hypothesis is that the second homeowners would be helpful in accessing additional markets for touristic areas; since these new homeowners would have high cohesion and identification Page 2 of 16 Stricker Swiss Journal of Economics and Statistics (2022) 158:9 with the region, they would bring new guests with them. This connected to a lower supply of second homes would incentive the construction of new managed accommodations and in this sense have an effect on the number of available hotel beds in the regions. Considering the effects of the new law on the housing prices, theory suggest an ambiguous effect. On the one hand, the prices of second homes would rise, due to the artificially capped supply on the market. On the other hand, a lower demand for new first homes would imply lower prices for first homeowners in that part of the segmented market. If the inter-and intra-cantonal migration is studied, then in consequence to the new law, more persons would choose the touristic areas as their main residence, since prices become more affordable for a larger group of potential inhabitants. Consequently this would cause a population growth in the studied municipalities. The supporting parties for this law additionally discussed that an intact landscape is the most important asset for the sustainable development of touristic activities; on the other hand, the main political debate argued that a strong reduction in construction activity would directly and indirectly impact the economic value chain, economically damaging the peripheral, touristic regions of the country. Hilber (2018), for instance, discusses the political opposition for this law in the most affected regions, where the initiative was in some cases perceived to be the result of the willingness to restrict second home construction among the residents of the Swiss Plateau being imposed on mountain dwellers. Figure 1 shows a map with the actual (2017 data) second home shares in the Swiss municipalities, with the red areas having restrictions on building new second homes. Furthermore, in their report, the Federal Office for Spatial Development ARE (2017) defines the spatial specialization and availability of construction grounds as well as land use in the various regions of Switzerland. This research aims to contribute to this political debate, suggesting that the reduction in new building construction activity has been focused on the construction of new dwellings and has not been as strong as initially claimed by the opponents of the law. In this sense, the main research questions is on how large the effects of the second home restriction law was on the investments in new housing construction business. Furthermore, based on the descriptive results, touristic overnight stays in Swiss municipalities did not change dramatically-only a slight shift in overnight stays from touristic central municipalities to the surroundings was identified in some important touristic regions.
The paper is structured as follows. Section 2 introduces the theoretical background, and Sect. 3 explains the institutional setting of the analysed areas and provides details  Economics and Statistics (2022) 158:9 about the legislation. Sections 4 and 5 explain the data and the empirical structure. In Sects. 6 and 7, the results on the impacts on new buildings are explained, and concluding remarks with a wider view of the law's impact, as well as the variations in the tourism industry, are presented.

Theoretical background
From a theoretical point of view, this research contributes to the previous literature in three main ways. First, it adds to the existing research on the geographic distribution of second homes and on the characteristics of second homeowners. For example, Gutierrez (1999) evaluates the impact of wealth effects on the demand for second homes. Di et al. (2001) define what second homes are and how they are distributed in the US. Additionally, they focus on the characteristics of second homeowners and the factors that drive the increasing demand for second homes. Planidea (2012) report some of the possible implications of the accepted second home restriction initiative in Switzerland, with a focus on the canton of Graubünden. BAK Basel (2013) analyses the topic in a more general way for all of Switzerland, hypothetically assuming different strong interpretations of the law. The author forecasts the effects of the second home act on tourism, construction activity, real estate and the other components of economy of the alpine areas. Four different scenarios of law implementation are outlined to conclude the study. Second, the study contributes to the literature examining the implications of second home construction for housing demand and more general implications. For example, Mayo (1981) focuses on the income elasticity of demand and the difference between renter and owners. Goodman (1988) identifies housing price and income as the main decision variables of housing demand. Akintoye and Skitmore (1994) divide demand into commercial and industrial construction and private construction and find a larger price elasticity for the private sector. Belsky et al. (2006) study the case of second homeownership in the US. The innovative aspect of their model is that it accounts for the possibility of multiple home ownership, which other models of income elasticity and the demand for housing have not done. It is important to distinguish between the two different parts of demand, with second home ownership categorized as an investment. The concept that second homes do not produce a flow of housing services that they consume supports this idea. There is both a flow of consumption and investment in the case of housing demand, and which of the two prevails might be different for a person's first or second home. Additionally, in the Canadian context, Demers (2005) defines a model to forecast housing expenditures.
Last, from a methodological point of view, the present work contributes to the recent broad field of difference-indifference (DiD) and event study design, such as for that by Abadie (2005). An important application is the work of Card (1992), which was used as a referring in developing the empirical strategy for this research. Regarding in particular the event study design, Abraham and Sun (2020) discuss treatment with heterogeneous effects. Further, the work of Ly and Paty (2020) provides an example of how to address spatial heterogeneity and different local time trends in the application of DiD and event studies to nationwide legal change. Bertrand et al. (2004) describe important implications for dealing with standard errors in the DiD technique.

Institutional setting
Before the second home law implemented in 2012, no concrete restrictions and rules at the federal level had been applied for municipalities when construction permits were issued. The first wave of second home construction starting in the 1960s showed that the high level of construction activity was not only a positive development. The lack of any regulation led to the development of a spatial planning act that was applied in 1980. With the end of the oil crisis, construction activity entered a second wave, and guidelines and recommendations were formulated. More recent debate concerned property for foreign homeowners as well as construction restrictions. The popular vote on 11 March 2012 introduced the new article 75b to the federal constitution, which explicitly limits the share of second homes among the total homes to a maximum of 20% in every municipality. The new law took effect the same day through an immediate stop of emitting construction permissions without a restriction to use note. In this sense the law considers all from this date on newly requested construction permissions, not the ones placed before the date. The legal change needed to be mandatorily enacted in a similar way across all municipalities and cantons with a sharp immediate stop of new second home constructions. Additionally, the article stipulates that municipalities are in charge of annually publishing their first home plans and reporting their applied measures to enact the law (Vinzens & Hefti, 2014).
To guide the calculation of the share of second homes in a municipality, article 2 in the regulation defines second homes as habitations that are not permanently used by persons with legal residence in the municipality and that are not used by persons for working or education purposes. Criteria in civil law define where a person's place of legal residence is. A further important point is the rule on existing flats, houses and hotels. Existing first and second homes are not restricted to any particular use and can be freely converted for different purposes, requalified and restructured as long as their original sizes are maintained, with reasonably sized enlargements being permitted. Hotels can be transformed into second homes if they have been operative for at least 25 years and if it can be proven that their operation cannot be profitably maintained. In municipalities where second Page 4 of 16 Stricker Swiss Journal of Economics and Statistics (2022) 158:9 homes are restricted, only managed touristic second homes (resorts and vacation villages) can be newly constructed. They need to be professionally managed for example as a part of a vacation resort or hotel village and not just be rented through private owners for a limited number of days in a year (Vinzens & Hefti, 2014). The report of BAK Basel (2013) focuses on the possible impact of the second home restriction in the alpine regions. First, based on the economic trend in these areas, stagnating employment is predicted. The new law will have the greatest impact in peripheral areas that are specialized in tourism. According to the previous study, the main impact on employment will occur in the gastronomy and construction sectors. An earlier report by Credit Suisse (2005) analyses the importance and distribution of second homes and the associated construction activity on the construction industry and examines the impact of the abolishment of measures contained in an earlier restriction of house and land ownership for foreigners in Switzerland (Lex Koller).
Additionally, to this end, several institutions were asked by the Federal Office for Spatial Development (ARE) to determine possible consequences of the second home law for the local economy. An example is the analysis of the law with a focus on businesses presented by BHP (2019) 1 that proposed managerial adaptation processes for some example touristic, construction and real estate firms and touristic areas. Rütter Soceco (2019) 2 presents a more economic perspective, defining hotspot areas with high second home market margins and describing the potential for a 20% drop in construction investments in these hotspots. Further data analysis on the industry mix and labour productivity of the affected areas was performed. Such analysis is important from the perspective of new potentials and dynamic business changes in the affected industries, and the present work complements previous analyses. Last, the presentation by Infraconsult (2019) 3 proposes an evaluation of the implementation of the law at the municipal level, focusing in particular on different strongly affected typologies of the spatial dimension.

Empirical strategy
The formal model is used to analyse the effect of the second home restriction as a two-way fixed effect regression which compares the changing investments in construction in treated and control municipalities. The baseline event study model compares the investment in new home constructions in restricted municipalities with that ones in other municipalities. Additionally, the effects of the second home restriction on renovation investments, housing prices, migration, residential population, wealth of homeowners and the supply of hotel beds are analysed in separate models.
A first baseline model is formulated as an event study design, taking year-specific time dummies and a continuous treatment variable to identify the different treatment intensities in the period after the law's introduction. This procedure is discussed by Schmidheiny and Siegloch (2020) as well as Card (1992). Formally, this procedure defines the following model: where (ln I it ) is the natural logarithm of new construction investments in municipality (i) in the years 2005, …, 2017. (SHShare i ) is used to represent the actual municipal second home share and (treat i ) the 2017 ARE second home statistics. 4 Furthermore, ( X it ) describes a set of control variables and ( θ t ) represents the time effects. Additionally, a variable (γ ct ) that controls for the different construction regulations at the regional (canton) level is included with the aim of capturing the time-varying changes in these regional laws. Meanwhile, the second specification with a discrete treatment includes uniquely the dummy variable ( treat i ), with a value of 1 for municipalities to which a second home restriction applies and 0 otherwise, to identify the effect of the legal change: The results of the two models should be increasingly visible at least after three years, for the fact that existing construction permits from the period preceding the new law continued to be valid for another three years in the most affected areas, depending on the cantonal legislation. Therefore, the full effect of the new law would be noticed only with the complete expiration of the last "old" construction permits in the years after the actual treatment. A summary of the duration of the construction permits defined by the single cantonal laws is provided in "Appendix 1".

Data and summary statistics
To analyse the impact of the second home restriction law in Switzerland, an annual panel dataset for Swiss municipalities out of several statistics and data sources was constructed. First, the housing construction statistics data from the FSO 5 provide data on investments in new constructions and renovations for private real estate at the municipal level. The considered data include the years 2005-2017 (which has been the most recent available data when the analysis has been conducted) and take into account the numerous mergers of municipalities taking place during this period. The data are standardized with respect to the official register of municipalities of 2017. Furthermore, the 2017 ARE second home statistics, which include the share of secondary homes in each municipality for the first time since the law's introduction for each single entity, are included in the dataset. The data is revealed annually in this statistic, starting from 2017. More in details, only after 2017, it is possible that some municipalities close to a 20% share as a consequence of a varying number of second homes change from the treated to the control group or vice versa. This dataset identifies the treated and the control groups, with a threshold at the 20% share level of second homes. In order to improve the comparability of the control group, all municipalities below the 10% second homes share were dropped from the sample.
In addition to the previously mentioned data, a variation of the model includes several research variables of interest in alternative to the investments at the municipal level. First, it includes the filtered average square metre prices for housing at the district level. These data come from Comparis 6 and are based on the prices demanded on online announcements placed on the main platforms. This variable is included as an index to study the different appreciation patterns of housing in the districts. Second, data from the FSO Population and Households Statistics 7 on the arrival of new residents from the internal and from the remainder of the cantons moving to the municipality are used as further dependent variables. This variable helps to understand the changing settlement behaviour of persons. In addition, the dataset includes the number of available beds in hotels and structured accommodations according to the FSO Tourist accommodation statistics 8 to understand the importance of tourism activity and how it changed after the application of the new law. Additionally, the number of empty, unused dwellings from the FSO Buildings and dwellings statistics 9 is included in this research paper to describe a potential excessive supply of apartments on the local market. Finally, to include the wealth of first and second homeowners, microdata from the FSO and ARE Mobility and Transport Microcensus (MTMC) 10 are included. A variable is constructed based on the average classes of the household incomes of first and second homeowners in each municipality and their relative importance based on the second home share compared to the primary home share as a control for income in the regressions. Since the MTMC is conducted only every five years, a five-year fixed income variable for the years 2005-2009, 2010-2014 and 2015-2017 is assumed. This variable is constructed as midpoints of the five defined intervals. There are slight variations within the grouped years because not all of the household income variables are available for each single municipality in ever year. For this reason, they might not be considered in the dataset in a determined year. The averages and standard deviations of the considered variables for the municipalities subject to the second home restriction and the control municipalities are reported in Table 1. Figure 2 compares the average investments in new constructions (in natural logarithms of thousands of Swiss Francs) for the two groups of municipalities. This comparison of investments among the groups of municipalities to which a second home restriction applies as well as to the remainder, undermines, that the previous similar investment trends are diverging starting from 2012. Immediately after  the 2008 crisis (which led to a temporary reduction in new second home construction investments) and later, especially in the post-treatment period, restricted municipalities registered stagnating investments for new constructions after the introduction of the second home law. A larger difference was registered after 2015, when even the latest pre-reform construction permits in the touristic cantons had expired. Additionally to the new construction investments, an examination of the effects on renovation investments has been conducted (see Fig. 5 in "Appendix 2").
In order to validate the application of the event study design, the year-specific effects have been singularly analysed and later been included in the graphic results illustration. More detailed applications and variations of these techniques can be found in Fuest et al. (2018) as well as in Autor (2003).

Results
Following the empirical strategy introduced in Sect. 4, the results in Table 2 in this section report the estimates for the models. In detail it reports the results for two different specifications. The first baseline models (1) and (3) are extended to include canton by year effects in models (2) and (4) to control for the different cantonal legislations. All scenarios are further analysed through a continuous treatment (accounting for the share of second homes in the municipalities) and a discrete treatment (separating only affected from non-affected municipalities without considering the importance of the second home market for the town). The results suggest a negative impact on investments due to second home restriction. This negative effect is greater in the long run. As shown in the data section, this is mainly due to some investors bringing forward their investments, and therefore, they invest in finishing and starting the last constructions initiated before the law's application.
Considering the cumulative treatment effects of the second home law in each of the following years, the marginal effect increases during the entire observed posttreatment period. The results show that starting with the years following 2013 (one year after the treatment), the negative marginal effect of the second home restriction increases year by year, because of the cumulative effect for every year increasing faster in the post-treatment period. This means that the effect of the second home restriction continues to grow and that the new posttreatment equilibrium has not yet been reached.
The results in Table 2 suggest that the observed effect of the restriction to construct new second homes in the touristic areas of Switzerland had the desired effect in reducing the construction activity in the affected municipalities. Nevertheless, it took several years to notice the full magnitude of the obtained results. As much as emerges from the results of the discrete treatment effect model (3) the effect is continuously growing over the observed period and coefficients are significant at a p value of 0.99 starting with the fourth year after the law was enacted. Figure 3 illustrates the obtained results from this model, including the coefficients of the years in the period precedent to the law's enactment.
Focusing on the results of the specifications (1) and (2) in the results in Table 2, then a growing negative trend in construction investments in new buildings in municipalities with a higher existing second home share is observed. Results show that the higher a second home share is, the more important this submarket of the construction industry is in the considered municipality. A more than proportional reduction for highly affected municipalities can be observed in the results of the continuous treatment effect specifications. Figure 4 evidences the cumulative effects of the specification (1) in the results in Table 2, including the coefficients for the pre-treatment period.
To check for a possible shift of construction activity away from new constructions towards investments in renovations, the models are rerun to analyse the effect of the law on renovation investments and on the remainder of the potential outcomes of the new second home law discussed in the introduction section (detailed results for all the specified models and a complete list for the above illustrated results can be found in "Appendix 3"-Tables 4, 5). These detailed results suggest no evidence for a possible shift (and consequently increase) of construction activity towards requalification yet. Considering the effect of the legal change on the housing prices, then two effects can be noticed: First, housing prices grew in the affected municipalities after the law's application; this might be an anticipation of the demand for the in future potentially scarce second homes. In a second moment, the housing prices were lowered, this probably as a result of the increasing number of first homes placed on the market at a lower price in the meantime, influencing the overall housing prices in second home restricted municipalities. Additionally, the interpretation of the changes in population before and after the law's application shows that the population in the peripheral touristic areas grew less than in the areas closer to the urban centres and this might even has an effect on the housing market. Moreover, the initiative was not able to cancel these trends for both the population growth and the inter-cantonal migration towards the centres. If migration within the canton is considered separately, then the increasing segmentation of the housing market makes residents more often move to these touristic hotspot municipalities. In this sense, an inversion of the previous trend that workers in touristic hotspots commute in from always more distant regions can be noticed. A further point of discussion were the effects of the new law on the available infrastructure in hotels and managed accommodations. In a first moment, the results show that the number of available hotel beds decreased in the affected municipalities   Page 9 of 16 Stricker Swiss Journal of Economics and Statistics (2022) 158:9 just after the second home restriction's application. This could be partially because non-profitable hotels could still be transformed into second homes as an exception of the enacted law. In a second moment, starting with 2017, the number of available hotel beds increased again. Such a delay in growing supply is plausible for the fact that new hotel projects require a longer time period to be developed than private housing. No significant effects on the homeowners' household income was noticed in the municipalities to which a second home restriction applies. In the end, it can be expected that in the case of a future lack of new second homes, it could become more attractive for interested persons to buy existing second homes and renovate the structures according to their needs.

Conclusions
The results of this research confirm the negative impact of the second home law on new construction investments. Proportionally, municipalities where there is a higher second home share and, consequently, where the business of second home construction used to be an important economic branch have been more affected by the law. The higher the proportion of already existing second homes, the larger the effect is. On the other hand, no significant effects of higher renovation investments were observed in the analysed period. Considering  the ageing of the investors in the second vacation home construction wave in the 1980s, it can be expected that many old second homes will soon be inherited or alternatively placed on the market and will be subject to larger renovation investments by new homeowners in the next years.
A great worry of touristic leaders and policymakers in the hotspot municipalities during the political debate was the expected negative impact of this law on several branches of the local economy in this mostly peripheral regions. This is because they expected a negative impact on tourism of the second home restriction in general. As such, in "Appendix 4", data from the FSO tourist accommodation statistics on overnight stays in touristic accommodations in the Swiss municipalities were studied. The observed effects are low-level variations, and hardly any of the considered municipalities reported large decreases or increases in the number of overnight stays, which would have caused a large negative or positive impact on the industry of managed accommodations.   Economics and Statistics (2022) 158:9 In conclusion, the second home restriction has helped fight excessive construction and building in extensively occupied touristic towns in the mountains, significantly reducing investments in new constructions without leading to greater negative impacts on tourism in the short-run period. Such a segmentation of the market helps to lower the housing prices in touristic areas, making it affordable for the residents and local working population to settle in these towns. The questions of whether the second home act has had the desired impact on landscape protection in the Swiss mountain regions, and more generally, what have been the middle-and long-term effects on the local economy in the alpine regions can be investigated in future studies. Last, it could be interesting to study economic impacts on borderline municipalities that have alternated between being subject to and not subject to the second home restriction when a larger time series of annual second home registry data are available.

Appendix 1: Construction permit's validity according to cantonal laws
See Table 3.