The state house prices make: the political elasticities of house prices and rents

ABSTARCT Fiscal policy allocation is not purely determined by the labour-capital conflict, but increasingly around cross-class housing coalitions. Although rising house prices are conventionally understood as drivers of fiscal austerity, this view has been challenged. Alternatively, governments may use fiscal policies to support house price growth to meet the primary economic interests of homeown-ers and compensate non-homeowners through the welfare system. Using an econometric analysis of 19 advanced economies between 1980 and 2018, we show house prices have positive effects on taxation revenue as well as fiscal spending on public investment, welfare and education. A second multi-level analysis provides a political explanation of this observed outcome by demonstrating parties respond to rising house prices by proposing more welfare and public investment spending in their manifestos. Conterminously rising house prices and rents also lead to greater welfare spending, suggesting governments use fiscal policy to protect those excluded from homeownership from labour market risks.


Introduction
The specific mechanisms influencing how governments and political actors implement fiscal tax and spend policies are a key area of interest for the fields of comparative and international political economy (Martin et al., 2009;Whiteside, 2023).The public finance literature largely focuses on how fiscal policy outcomes are influenced by political party appeals to the capital or labour interest groups aligned with their voter base (Prasad & Deng, 2009).More recent research in the "growth-model" literature highlights how capital and labour class divisions are becoming less clear.Here, influential cross-class sectoral coalitions determine the allocation of fiscal resources in ways that support specific areas of the economy in-line with their own economic objectives and interests (Baccaro et al., 2022;Haffert & Mertens, 2021).While this research has started to map the cross-class coalition in favour of homeownership and construction growth (e.g.Kohl & Spielau, 2022;Reisenbichler & Wiedemann, 2022), they do not sufficiently account for the importance of private homeownership politics in driving fiscal policy outcomes, which is where this paper makes an empirical contribution.
As has been long known in housing studies, the private owner-occupied home is the primary financial asset for most labour market participants in advanced economies, making house price growth a key economic concern for homeowners as a cross-class coalition interest group (Adkins et al., 2020;Stirling et al., 2023).Prevailing accounts as to how homeownership influences fiscal policy outcomes primarily focus on issues of welfare spending.Here, the high cost of private housing motivates homeowners to demand a low-taxation environment to maximize their disposable income, inimical to high levels of fiscal spending (Ansell, 2014;Kemeny, 2005;Schwartz & Seabrooke, 2008).As such, rising house prices reduce homeowner support for welfare policies they have little need to access, motivating conservative parties to adopt welfare state retrenchment policies (Ansell, 2014).However, rather than there being a clear trade-off between government spending and homeownership, recent research suggests there is a complementary relationship between the two (Anderson & Kurzer, 2019;Van Gunten & Kohl, 2019).The transition into homeownership has also not led to any significant increase in political support for right-leaning parties supporting fiscal retrenchment (Hadziabdic & Kohl, 2022).As such, there are a series of inherent tensions regarding any potential trade-off between house price appreciation and fiscal policy outcomes.
As a cross-class coalition sectoral interest group, homeowners may demand parties implement fiscal policies that support their primary economic interest of maintaining house price appreciation, such as improving public safety, education, healthcare, and infrastructure (Han & Shin, 2021;Hilber & Mayer, 2009;Larsen et al., 2019;Tiebout, 1956;Yoder, 2020).Furthermore, mortgaged homeowners may demand increased welfare and social investment spending to protect them from any mortgage default risk resulting from labour market shocks that could lead them to lose their homes (André et al., 2018;Lennartz & Ronald, 2017;Schwartz, 2018;Tranøy et al., 2020).As homeownership outsider groups (i.e.non-homeowning private and public tenants) are exposed to greater labour market risks than homeowners, political actors may look to generate electoral support from renters by allocating them fiscal resources through the welfare system.Furthermore, parties may well be willing to use fiscal policy to support the economic objectives of homeowners and those excluded from homeownership, as rising house prices are a significant source of both recurring and non-recurring tax revenue for local and national governments (Ansell, 2019;Lutz, 2008).
As the literature is divided on whether rising house prices are associated with the retrenchment or advancement of fiscal spending, this paper looks to formally evaluate whether, and to what extent, house prices effect fiscal policymaking in advanced economies.To examine this research question, this article proceeds in two parts.The first part performs an econometric analysis using error correction models to assess the long-run effects of house prices on fiscal policy outcomes.
This part of the study covers 19 advanced economies that introduced various reforms across the 1980-2018 time period associated with welfare state retrenchment.The results of our econometric analysis show house prices have positive and significant effects on total government revenues and expenditures, suggesting that house prices are an important fiscal capacity building tool for states.We also find that house prices have positive significant effects on both public investment and social spending.These results support suggestions that political actors appeal to the preferences of homeowning voters looking to maintain the value of the home as a financial asset.Finally, as we show conterminously rising house prices and rents are associated with increased welfare spending, this suggests governments are compensating those excluded from homeownership and protecting them from labour market risks via the welfare system.These results are supported by robustness checks using a different dataset for the 1980-2017 period, and a longer-run analysis from 1960-2017.However, these results may be driven by a low interest rate environment, which is coming under increasing pressure due to rising inflation.
The second part of our analysis investigates the political mechanisms underpinning our macro-level findings.Drawing on data from the Party Manifesto Project, we examine the "political elasticities" of house prices in terms of how parties react to house price dynamics in their election proposals.The results of this second analysis demonstrate rising house prices are associated with parties proposing increased welfare and public investment spending in their manifestos, independent of their party ideological position.Moreover, we demonstrate that when house price increases coincide with rising rents, parties tend to adopt more pro-welfare positions.These results provide significant evidence to suggest the political consequences of rising house prices incentivise parties to allocate fiscal spending on welfare and public investment programmes supporting house price appreciation.This, therefore, encourages support from homeownership insider groups, whilst also appealing to outsider groups by compensating them through the welfare system.
Overall, our analysis makes a significant contribution to the literature on the effects of house prices on fiscal policymaking.Specifically, by demonstrating the political influence of homeowners as a cross-class economic interest group that motivates parties to allocate social and investment spending to support house price appreciation.Furthermore, by reversing the previously accepted major stylized fact of a 'trade-off ' between rising house prices and welfare expenditure, our results lend further support to the broader critiques of the asset-based welfare and housing 'trade-off ' literature (Beckmann, 2020;Van Gunten & Kohl, 2019).Instead, the results are much more in line with recent research, which links housing regimes to countries' growth models (Hofman & Aalbers, 2019;Hopkins & Voss, 2022;Wood & Stockhammer, 2024).In consumption, credit, or asset-led growth models homeowners are part of a broader growth coalition, which includes central banks and the financial sector, supporting government activity upholding housing asset prices, and hence, the dominant growth model.The politics of this growth regime are facilitated by the low-interest rate environment that is particular to this period, and may not be sustainable in the longer run.Housing is only one asset class whose price inflation may create new political cleavages.As such, the results of our analysis could potentially be extended to examine the political consequences of inequalities in other asset classes, most notably investments in stocks or bonds.
The rest of the paper is structured as follows.The first section examines the relationship between fiscal policymaking and private homeownership.The second section outlines the empirical design of the macro-econometric models, whilst the third section explains their results.The fourth section discusses the results of the analysis of how house prices influence party manifestos and is followed by a brief summarising conclusion.

Fiscal policymaking and the political economy of homeownership
Public finance theories have traditionally focused on the role of class-based coalitions in driving the tax and spend fiscal policies of national governments in advanced economies (Martin, 2015;Timmons, 2005).There is thought to be an antagonistic relationship between organised capital interests, as the main net contributors to the tax base, and labour interests who are considered the net beneficiaries of government redistribution policies (Meltzer & Richard, 1981).Subsequently, fiscal policy is determined by political parties looking to appeal to capital or labour interest groups aligned with specific party positions and the core of their voter base.However, rather than adopting progressive taxation policies against capital interests, countries with stronger welfare policies tend to have more regressive taxation on waged labour (Martin, 2015;Prasad & Deng, 2009).Furthermore, the notion of capital and labour as being antagonistic forces fails to account for how many modern voters have blended labour and capital interests.For example, the increase in private pension provision has seen many workers become shareholders (Van der Zwan, 2014).Therefore, class delineations are not the major driving factor behind fiscal policy positions (Haffert & Mertens, 2021).
The distributional conflict between specific class-based actors in the economy are not the main determinate of fiscal policymaking.Rather, the allocational conflict between cross-class political coalitions is oriented around the specific investments they have made into different economic sectors (Haffert & Mertens, 2021).Here, fiscal policies are determined by voter preferences around sectoral coalitions that look to allocate resources to support specific areas of the economy in-line with their own economic objectives, such as wages, employment, or returns on investment (Haffert & Mertens, 2021;Martin & Swank, 2012).This is often conceptualised in terms of the specific sector that voters, as employees or investors, have an economic interest in (Frieden, 1991).
Whilst Haffert & Mertens (2021) make an important contribution by demonstrating the importance of sectoral economic interests in shaping fiscal policy preferences and outcomes, their analysis focuses on the politics of specific growth models based around consumption and export-led growth.Although this is an important area for comparative political economy, it does not sufficiently acknowledge the important economic interest of the private owner-occupied home and its influence on voter preferences around fiscal policy allocation.This is a significant omission as the home is one of the most important economic interests for labour market participants in advanced economies, as for many it is their main private investment vehicle (Adkins et al., 2020).It also excludes the housing construction sector, which plays an important role in many national economies (Hayward, 2012).Homeowners are also an important political consistency in their own right and have economic interests that may influence the relationship between house prices and fiscal policymaking in different ways (Ansell, 2019;Schwartz & Seabrooke, 2008;Wiedemann, 2022).
Figure 1, adapted from Ansell et al. (2018), demonstrates the underlying causal mechanism linking house prices to fiscal spending.Here, house price changes are thought to influence individual policy preferences around fiscal policy, and parties adopt these policies to appeal to their voter base (Ansell et al., 2018).Therefore, at the macro level we should observe how house prices translate into fiscal policy allocation outcomes implemented by parties looking to meet the voting preferences of homeowners as a key constituency.
Generating stable taxation revenue is central to the provision of state welfare services (Beramendi & Rueda, 2007).House prices are an important source of tax revenue for local governments generated through recurring taxes, such as annual property or land taxes (Ansell, 2019;Lutz, 2008).However, house prices also contribute to the revenues of national governments, with direct non-recurring taxes levied on private homes at the point of transfer (e.g.purchase, inheritance or capital gains taxes) (Ansell, 2019).
Taxes are levied across the housing investment cycle, with VAT and transaction taxes during acquisition, property, income (minus mortgage relief ) and wealth taxes during the holding period and capital gains and inheritance taxes upon disposal (OECD, 2022).The OECD is only able to separate transaction taxes (incl.stamp duty, etc.) and recurrent real estate taxes as fiscal revenue.It thus only reports a lower-bound estimate of how important property taxation has been for government revenues: the global mean since 1960 amounts to almost 6% of all government revenue (even higher shares for local governments) and almost 2% of GDP (OECD, 2022), where Liberal Market Economies seem to rely much more on this form of taxation.These figures are also at the lower bound, as they ignore the indirect consumption taxes (and further feedback effects) generated from the home as a vehicle that facilitates debt-driven growth in many advanced economies, including those more traditionally associated with export-led growth (Wood & Stockhammer, 2024).
The political economy of how house prices and homeownership impact fiscal spending policy outcomes is less clear.One key strand of literature emphasises the negative effects of private housing on fiscal policy, specifically welfare spending.Here, it is argued that the high cost of private housing leads homeowners to demand a low-taxation environment to maximize their disposable income, which makes high levels of fiscal spending unsustainable (Kemeny, 2005;Schwartz & Seabrooke, 2008).More specifically, as homeownership allows individuals to mitigate their own labour market risks, rising house prices reduce homeowner support for public social and redistribution policies they have little need to access (Ansell, 2014).This, in turn, motivates conservative parties to shift further towards the right and adopt welfare state retrenchment policies to generate political support from homeowners (Ansell, 2014).As such, there is thought to be a trade-off between rising house prices, homeownership rates and fiscal spending on welfare state provision, particularly pensions (Castles, 1998;Delfani et al., 2014;Kemeny, 2005).
Although these accounts provide a clear explanation as to how private homeownership has influenced welfare state restructuring in advanced economies, they have recently been challenged.There is a growing literature that demonstrates there is a complementary relationship between government spending and homeownership, rather than there being a clear trade-off between the two (Anderson & Kurzer, 2019).In contrast to Kemeny's thesis, higher homeownership rates have been associated with increased levels of government spending.In particular, homeownership has been considered a complement to many welfare state services, particularly pensions (Van Gunten & Kohl, 2019).The political consequences of homeownership in terms of driving support for welfare reform has also been challenged, as the transition into homeownership has not resulted in any significant increase in political support for right-leaning parties supporting welfare retrenchment (Schwartz, 2018).
The literature also suggests different ways in which homeowners may have a sectoral interest in increasing fiscal spending.As the private home is the main financial asset for many labour market participants, homeowners are highly sensitive to house price changes as their key economic concern (Adkins et al., 2020).Therefore, as a sectoral coalition oriented around maintaining upward pressure on house prices, homeowners may prefer higher levels of government spending to provide high quality public services that are capitalised in their house price gains, such as increased education spending to maintain the quality of nearby schools (Tiebout, 1956, Fischel, 2001, Hilber & Mayer, 2009).As such, national political parties may look to appeal to homeowners by promoting government spending policies maintaining house price appreciation, such as social investment in education or public investment on infrastructure (Lennartz & Ronald, 2017).Whilst individual local infrastructure projects may lower home values and cause case-specific NIMBY'ism, the expected overarching effect of massive infrastructure investment should be positive.
Mortgaged properties also require a strong welfare system to be considered viable financial assets (Schwartz, 2018).Since the 2008 Global Financial Crisis, the viability of the private home as a financial asset that can adequately mitigate labour market risks has been called into question (Mian & Sufi, 2016).In the aftermath of the crisis many homeowners faced sharp declines in house prices and were unable to either sell their homes or access mortgage credit to realise the equity available in their homes, in part due to experiencing unemployment themselves.As such, homeowners who relied on the home as a form of asset-based welfare were not always able to successfully use their gains in house price appreciation to mitigate income or employment risks.Furthermore, although homeowners have a strong interest in maintaining upward pressure on house prices, they are also fearful of defaulting on their mortgage, which could see the borrower lose any rights to access any potential financial gains from homeownership.The most common reasons for defaulting on a mortgage are unemployment or reduced earnings.Therefore, mortgaged homeowners may demand increased welfare spending to mitigate such labour market risks and protect their access to the home as a financial asset (André et al., 2018;Schwartz, 2018;Tranøy et al., 2020).
Although house price appreciation benefits voters in homeowner insider groups, it is detrimental to those in non-home owning outsider groups, such as aspiring homeowners and private tenants.Those excluded from homeownership may demand increased welfare spending as a form of social insurance to protect them from labour market risks, as they do not have a private home as a financial asset to access wealth in times of economic hardship.Those included in the outsider group are aspiring homeowners or those looking to trade-up their homes who may be priced out of the market.There are also private tenants who face increasing rents, as investors need to align rental yields with increasing house prices.In general, these voters tend to be younger, more mobile households on the lower-end of the income distribution.Even though these individuals might (counter-intuitively) not always have pro-distribution attitudes (e.g.Ansell & Cansunar, 2020), political parties may look to compensate voters in these homeownership outsider groups by increasing welfare spending.
When taken in combination, these mechanisms provide different logics through which higher house prices could be associated with higher levels of fiscal spending.More specifically, we suggest house price appreciation can provide governments with the capacity to generate higher revenues, which can be spent on either public investments and welfare, speaking to home-owning voters who wish to maintain upward pressure on their house prices and protect against mortgage defaults, as well as compensating those excluded from the financial gains from homeownership.However, the literature is divided on whether rising house prices are associated with the retrenchment or advancement of fiscal spending on welfare services.Therefore, to address this issue, this paper assesses whether and to what extent house prices effect fiscal policymaking in advanced economies?
The evaluation of this research question proceeds in two parts based on the argument from Ansell et al. (2018) that house prices produce fiscal policy outcomes based on political parties responding to homeowner preferences.The first part examines the macroeconomic effects of house prices on fiscal policy to assess the contradictory views as to how house prices may influence fiscal policy outcomes.The second part of this analysis traces the mechanisms behind the observed outcome of the macroeconomic analysis through an examination of what we call political house price elasticities.Here we examine how house price changes influence party manifesto positions on key elements of fiscal spending.Although the macro-level analysis draws on ECM models, thus avoiding the reverse causality interpretation issues of contemporaneous analyses, our argument assumes a positive feedback-loop where rising housing prices, mediated through party politics, lead to higher state revenues which, in turn, feed back into stable or rising house prices.While studies referenced above found sufficient empirical evidence for this latter feedback-loop, the following analysis will focus on putting evidence behind the first part of this political housing cycle.

Macroeconomic empirical design
This section provides an overview of the empirical design used to evaluate the different mechanisms by which house prices influence fiscal policy outcomes.Each equation and model specification uses an autoregressive distributed lag (ARDL) model with an OLS estimator to examine the variables independent of their order of integration (Charemza & Deadman, 1997).We reparametrize the ARDL as an error correction model (ECM) that produces short-run and long-run coefficients, but we only report the long-run effects in main analysis (for the full results see the supplementary material), as they crowd-out the short-run effects (Engle & Granger, 1987).Focusing on the long-run effects also addresses potential endogeneity issues, as the construction of the long-run coefficient does not require any interpretation of a contemporaneous variable whose causality is potentially bidirectional and relies on specific causal assumptions in its interpretation (Lin & Tomaskovic-Devey, 2013).As such, the ECM does not rely on any specific or potentially implausible causal assumptions about the contemporaneous relationship between house prices and fiscal policy outcomes (Lin & Tomaskovic-Devey, 2013).Furthermore, as we argue the effects of house prices on social spending operate via a non-contemporaneous revenue and expenditure mechanism (e.g.Blanchard & Perotti, 2002), we only consider lagged effects, which, at the minimum, are considered 'weakly exogenous' and are not subject to endogeneity in dynamic specifications (Bellemare et al., 2017).Two lags were selected for each model specification as they had the smallest MBIC, MAIC, and MQIC values, as suggested by Andrews & Lu (2001).
The results tables report robust panel-corrected standard errors (PCSE) based on the Prais-Winsten transformation (Beck & Katz, 1995;Plümper et al., 2005).Each of the PCSE specifications for each equation uses a first order autocorrelation structure to address potential problems of serial correlation and assume panel-level heteroskedastic errors to mitigate for heteroskedasticity.Each model specification is also run with fixed effects using country-specific intercepts to account for time-invariant characteristics at the country level.The sample consists of an unbalanced panel dataset of 19 advanced economies (Australia, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Netherlands, New Zealand, Norway, Portugal Spain, Sweden, Switzerland, the United Kingdom and the United States of America) from 1980 to 2018.These cases and the time period of the analysis were selected as each country is an advanced economy that has transitioned away from Keynesian-style welfare systems since the 1980s (Clayton & Pontusson, 1998).
This macroeconomic analysis consists of seven model specifications, each with distinct dependent variables to evaluate the different mechanisms by which house prices may influence government fiscal policies.The first specification evaluates the fiscal capacity building effects of house prices by using total government revenue as the dependent variable.The second specification examines total government expenditure before breaking it down into investment and welfare categories in the remaining specifications.Here, gross fixed capital formation (GFCF) and transport and communications investment are the dependent variables in the third and fourth specifications respectively.Whilst the remaining dependent variables are: total social spending (specification 5); social spending per capita (specification 6) and education spending (specification 7).
The main independent variable of interest for each model specification is real house prices, measured as an index level of country-specific house prices relative to 2015.Mortgage debt could have also been used, as it would be closer to the growth-model literature, but mortgage debt is a function of house price growth, not the other way around, and house prices are central to macroeconomic outcomes in both export and debt-driven economies (Wood & Stockhammer, 2024).Homeownership rates could have been used as an alternative measure, but previous research has shown that it is 'increases in asset prices rather than homeownership per se appear to drive partisan policymaking' (Ansell, 2014, pp. 399-400).Furthermore, the availability of annual homeownership data is inconsistent, meaning that some data needs to be interpolated, which may produce misleading results.Therefore, this analysis focuses on the political consequences of house prices, rather than homeownership rates.Although Ansell (2014, p. 396) suggests "house price levels cannot be usefully compared cross-sectionally", requiring first differences to be used, level property price indices are commonly used in cross-sectional macroeconomic time series analyses of house prices that account for country-specific levels using fixed effects, which are used in this analysis (Geng, 2018;Hofmann, 2004;Stockhammer & Wildauer, 2018).Furthermore, using first differences only accounts for short-run effects that cannot be used to estimate the long-run effects of the ECM.If house prices build the fiscal capacity of governments, we would expect to find positive effects of house prices on government revenue.If governments allocate fiscal spending towards the interests of homeowners as an economic interest group, then we would expect to observe a positive relationship between house prices and total government spending, the public investment variables (GFCF and transport and communications), as well as the welfare spending variables (total social spending; social spending per capita; social investment).However, as previous analyses emphasise a trade-off between house prices and welfare provision, we may find a negative relationship between the variables.
The first series of model specifications also includes two control variables.The first control variable is a proxy for globalization, measured as the sum of total real exports and imports over real GDP, as suggested by Roine et al. (2009).This accounts for the total share of international trade in relation to the size of a nation's economy.This variable is included as governments seeking trade expansions may be motivated to retrench welfare programs and reduce taxation rates to encourage capital investment and foster international competitiveness.Therefore, if globalization has contributed to welfare state retrenchment, then we would expect to observe a negative relationship between the variables.
The second control variable accounts for the political effects of domestic politics on welfare state retrenchment (Kus, 2006).Domestic politics is measured using the ideological center of gravity for each government in each country on a continuous ten-point scale, with 1 being the most left-wing government and 10 being the most right-wing.Traditionally, more right-leaning conservative parties tend to be associated with increases in public investment to foster support from the insider homeowner group, whereas left-leaning parties look to increase welfare spending to capture support from the outsider group.However, there has been a distinct decline in partisanship effects on welfare provision in recent decades (Bandau & Ahrens, 2020).
One of the few party manifesto-based studies suggests that house-price increases systematically move right-wing parties leftwards, as they look to capture voters in constituencies not already satisfied by higher house prices (Beckmann, 2020).Parties of the left, by contrast, are associated with housing welfare spending (Beckmann, 2020).Finally, incumbent political parties, regardless of their ideological position, may be incentivised to maintain house price appreciation via increases in government spending, as both have been shown to generate political support for incumbent governments (Afonso & Sousa, 2012;Larsen et al., 2019;Shi & Svensson, 2006).In short, the expected partisan effects and asymmetries that transmit higher house prices into government spending are not clear.Our second analysis uses the same model specifications as the first, but includes an additional variable controlling for the interaction of government partisanship and house prices.Existing accounts (e.g.Ansell, 2014) examining data from 1975 to 2001 find that rising house prices only influence social spending when governments become more right-leaning.Therefore, we expect to observe a negative relationship between the variables.
We could have included economic growth as a potential additional control variable.However, the data on government expenditure items is a subset of GDP data.Therefore, including a measure of economic growth as an additional variable created multicollinearity issues for the models.House prices also move pro-cyclically with macroeconomic growth dynamics and the business cycle has even been described as moving concurrently with, and driven by house prices (Davis & Heathcote, 2005;Leamer, 2007).Therefore, economic growth is, at least approximately, accounted for in the models through house price dynamics.
Our third analysis accounts for 'the other' side of the housing market, as it includes a real rent cost variable alongside the house price variable, but does not include any controls.These specifications also account for synchronous changes in private ownership and rental costs by examining the effects of an interaction of the two variables.If governments are using fiscal policy to meet the economic preferences of renters as a sectoral coalition to mitigate the negative economic effects of being excluded from homeownership, then we would expect to observe a positive relationship between the rising house prices and rent costs interaction term and the social spending variables.As the model specifications only use three control variables at a maximum, they may be considered to be underspecified.However, the right-hand side of the equation includes time lags of the dependent variables as part of the ARDL, which accounts for potential omitted variable bias issues and mitigates the need for additional control variables to be included in the analysis (Wooldridge, 2016).
We perform eight specific robustness checks to account for specific issues pertaining to this analysis.The first robustness check analyses the same sample as the main equations, but uses share of GDP variables to assess the effects of house prices on government revenue, government expenditure, social spending and public investment.This is performed to account for whether the initial results may be due to GDP growth alone.Although previous studies of social or government spending often use shares of GDP as variables, we did not use them in our analysis as it is unclear whether the results are due to effects or changes in the numerator (i.e. total social or government spending) or the denominator (GDP).Given the slow-moving nature of the numerators in this analysis, much of the volatility and year-to-year changes may actually be driven by the GDP denominator.Thus, potentially providing misleading results.This is a particularly important issue for this analysis, as house prices have been demonstrated to have significant multidirectional effects on GDP (Goodhart & Hofmann, 2008).Despite these issues, we include the share of GDP variables as a robustness check to align our results with those of previous studies.The second robustness check performs a longer-run analysis of the share variables from 1960 to 2017, using a different dataset from the Macrohistory Database (Jordà et al., 2017, Knoll et al., 2017), to corroborate or challenge the results of our main analysis.
The third robust check uses the main data, but also includes additional control variables to account for immigration, construction, and labour market reforms, the latter of which is a common-cause concern.The fourth robustness check focuses on the effects of house prices on the biggest contributors to social expenditure, healthcare and pensions.The fifth reports separate results for the effects of house prices on government spending broken down into six other different spending categories, including defence, public order, housing, and the environment.The next two robustness checks focus on issues pertaining to different levels of government.The sixth robustness check examines the effects of house prices on local and central government spending to check whether countries' constitutional set-up could be confounding the main mechanism.Similarly, the seventh robustness check includes Lijphart's five scale variable to proxy for degree of federalism as a control in the main analysis to account for potential variations in political and taxation systems, which have important consequences for housing and real estate taxes.The final robustness check replaces the national house price index with a city-level price index (Amaral et al., 2021) to assess whether there are potentially heterogeneous within country effects.
All nominal values were adjusted for inflation using CPI data obtained from the OECD.Each variable, except for the gravity variable, was transformed using the natural-logarithmic scale, as the coefficient generated from such log transformations may be interpreted as a close approximation for the percentage change to provide elasticities for each variable, which is more relevant to this analysis than a single unit change (Gelman & Hill, 2007).The use of such a transformation technique is commonly used in econometric analyses.As the gravity variable is not transformed as a natural log, this is a log-level relationship where a one unit shift to the right of a government's ideological centre of gravity may lead to a percentage change in fiscal outcomes.See section A of the supplementary material for a complete account of data sources, variable definitions and transformations.Whilst section B of the supplementary material provides the summary statistics for each variable.
One key issue pertaining to this long-run macroeconomic time series analysis is that there may be a non-stationary (i.e. unit root) upward trend for many of the variables (such as house prices or government spending), which has the potential to produce spurious results.However, ARDL and ECM models can be used to account for variables with such trends (Charemza & Deadman, 1997).The pre-testing results (see section C of the supplementary material) show the equations in this analysis consist of a combination of stationary and non-stationary variables.Many political scientists believe that combined models using variables with mixed orders of integration are invalid, as this constitutes an unbalanced equation (Keele et al., 2016).However, statisticians argue this is not the case as different orders of integration do not constitute an unbalanced equation, which 'is mis-specified and prone to type I error' (Enns et al., 2021 p.1).Rather, equations with combinations of different orders of integration can be analysed using either an ARDL or ECM model, depending on whether the variables are cointegrated or not (Enns et al., 2021).Cointegration occurs when variables move in similar patterns (i.e. both with an upward trend) and demonstrate a long-run equilibrium relationship between the variables (Charemza & Deadman, 1997).Whilst it has been argued that cointegration requires that each variable has the same order of integration (e.g.Charemza & Deadman, 1997), equations with combined integration processes can be cointegrated in the long-run (Wlezien, 2000).Therefore, such combined cointegration processes can be analysed using an ECM to capture the long-run relationship of the variables (Enns et al., 2021).The pre-testing results in the supplementary material demonstrate that despite having a combination of stationary and non-stationary variables, they are cointegrated, thus, supporting the use of the ECM in this analysis.As we reparametrize the ARDL equations as an ECM, the generic form of our models is:

The macro-political economy of house prices and fiscal policy
The long-run effects are calculated by taking the ratio of the sum of the coefficients of the two lags of the independent variable to the adjustment speed (error correction).We calculate the adjustment speed by subtracting the sum of the  coefficients of the two lags of the dependent variable from 1.As such, our adjustment speed will produce positive values, rather than the negative values normally associated with ECMs.We recalculate the standard errors of the long-run coefficients to assess their statistical significance using the Stata command nlcom, as suggested by Vlandas (2017).
Figure 2 provides a graphical representation of the results of the first set of model specifications examining the long-run effects of house prices, globalisation and the ideological centre of gravity on each of the seven dependent variables.The coefficient value of each variable is denoted by the marker symbol and lines either side of each marker symbol represent the 90 and 95% confidence intervals.
The results of the ECM model specifications in the first analysis show house prices have positive and statistically significant long-run effects at the 1% level on each of the dependent variables.As the coefficients may be interpreted as elasticities, a 1% increase in house prices is associated with a 0.4806% increase in government revenue, which suggests house prices are a significant tax raising tool for governments.A 1% change in house prices also translates into increased total government expenditure (0.6530).When broken down further, we observe that a 1% increase in house prices leads to a 0.7700% increase in public investment and a 0.6692% increase in transport and communications expenditure.Whilst a 1% increase in house prices is associated with a 0.5802% increase in social spending; a 0.3617% increase in social spending per capita; and a 0.5960% increase in education spending.These results challenge previous studies suggesting there are negative effects of house prices on fiscal policy to meet voter preferences on welfare prevision 1 .Rather, our results suggest that homeowners are a key economic constituency in their own right and rising house prices encourage governments to allocate significant levels of fiscal spending support homeowner's economic interests.
The results also show that globalisation has mixed effects on the dependent variables.A 1% increase in the value of the globalisation variable is associated with a 0.8399% increase in government revenues as well as a 0.7590 increase in government expenditure, and both are significant at the 1% level.The effects on the total social spending and social spending per capita variables are significant at the 5% level and have coefficients of 0.3656 and 0.3116, respectively.The effects of globalisation on public investment spending are significant at the 1% level and have a coefficient of 1.2712.However, there were no significant effects of globalisation on either education or transport and communication spending.Overall, these results support accounts suggesting that claims of globalisation eroding the welfare state have been overstated (e.g.Onaran & Boesch, 2014).They also suggest that governments are funnelling the revenues generated from the gains of increased international trade into the welfare system, potentially to compensate those excluded from the benefits of globalisation.
Although the coefficients representing the long-run effects of the ideological centre of gravity of government on each of the dependent variables are negative, none of the relationships are statistically significant, except for the transport and communication variable (at the 10% level).Here, a 1 unit shift towards the right on the ideological centre of gravity of an elected government suggests a 0.0682% decrease in spending on transport and communications infrastructure.Overall, these results suggest changes in tax and spend policies across the sample are largely a-political, thus challenging claims of welfare state retrenchment being a specific conservative or neoliberal political project (Deacon, 2007;Kus, 2006).
Figure 3 shows the results of the second set of model specifications examining the long-run effects of the interaction of house prices and the ideological centre of gravity of the government on the dependent variables.House prices, globalisation and the ideological centre of gravity of the government are included in the model, but the results are not reported here (see appendix E for a full breakdown of the results).The results show that the house prices and the ideological centre of gravity interaction term has a positive and statistically significant long-run effect on total social spending at the 10% level.The coefficient suggests that a 1% increase in house prices combined with a one unit ideological shift to the right of the government is associated with a 0.0558% increase in total social spending.We observe similar results when social spending per capita is used as the dependent variable, where the coefficient has a similar order of magnitude (0.0515) and is significant at the 10% level.The interaction term also has positive and significant effects on education spending at the 10% level, with a coefficient of 0.1386.These results are contrary to expectations based on previous analyses (e.g.Ansell, 2014) and suggest that right leaning governments increase fiscal spending towards homeowners as an economic interest group.Although the results are significant only at the 10% level, there is often an overemphasis on the importance of the statistical significance of variables.It is more relevant for this analysis that the direction of the coefficient is aligned with the broader results of this analysis, providing additional support for the findings (Ziliak & McCloskey, 2008).The interaction term did not report any statistically significant effects on the government revenue, government expenditure, public investment or transport and communication variables.
Figure 4 shows the results of the third set of model specifications examining the long-run effects of the interaction of house prices and rents on the 7 dependent variables.The results show that the interaction term has positive and significant effects on some of the social spending variables, but there are no significant effects on government revenue, public investment, transport and communications or education.The results suggest a combined 1% increase in house prices and rents is associated with a 0.9652% increase in government expenditure and is significant at the 10% level.The effects of the interaction term on total social spending are of a similar magnitude (0.9429) and are significant at the 5% level.The interaction term also shows positive and significant effects on the social spending per capita variable at the 10% level with a coefficient of 0.6748.These results suggest that when house prices and rents rise governments use fiscal policies to allocate resources to those excluded from homeownership to support their economic interests via the welfare system.The results of the various specifications across all models are robust under the seven robust checks reported in the Appendix F.
In sum, the results of our macro-analysis produce four key findings.First, we demonstrate rising house prices contribute to raising taxation revenues.Second, as house prices show positive effects on different forms of public investment, social and education spending, we argue that governments allocate fiscal resources to meet the economic demands from home-owning voters to maintain upward pressure on the asset-prices of their homes.Third, as conterminously rising house prices and rents are associated with higher levels of social spending, this suggests governments look to compensate and protect the economic interests of those excluded from homeownership through the welfare system.Fourth, and finally, the political ideology of the specific government formation has no direct effect on social spending across the sample, and in the cases where there is a significant interaction between political formation and house prices, we show that more right-leaning governments are inclined to increase social welfare spending when house prices increase.

Political house price elasticities
What are the political mechanisms behind these macroeconomic relationships?This section seeks to explain the observed outcomes of the macroeconomic analysis through an examination of how house price changes influence party manifesto positions on key elements of fiscal spending, which we describe as political house price elasticities.We assess whether house price appreciation incentivises parties to allocate investment and social spending to meet the economic interests of homeowners.Existing theories on partisan constraint would suggest these effects operate asymmetrically (Beckmann, 2020).Here, conservative right-leaning parties may be considered more likely to propose public investment expenses in-line with their general support for homeowners as a voter base.In contrast, left-leaning parties are thought to either propose welfare expansions or reduce cuts on social transfers to support homeownership outsider groups.We explore the "political elasticities" of house price changes by combining party manifesto data with the macroeconomics of house-price cycles in a multi-level time series regression for 229 different parties nested in 20 countries between 1980 and their most recent election.
This analysis is conducted in two parts.The first only considers the effects of house prices on party manifestos to evaluate the influence of homeowners as an economic interest group.Whilst the second also includes rent prices to account for whether the economic interests of those excluded from homeownership are considered by political parties.The main explanatory variable in the first analysis examines changes in real house prices for the years between elections.We use this variable to account for whether developments in house prices since the last legislative term influence the party manifesto for the upcoming election.Our analysis uses six specifications, each with their own dependent variable, to examine the different ways in which house prices may influence party manifestos.The first two model specifications measure party preferences for welfare spending using both 'anti-welfare' and 'pro-welfare' variables from the Manifesto Project (CMP) (Lehmann et al., 2024).These variables are predefined CMP content codings and measure the percentage share of quasi-sentences devoted to issues either favouring or opposing welfare spending.The third specification uses a left-right index from the same source that combines different variables to approximate a position of the manifesto on a left-right spectrum.The fourth specification examines whether party positions may be considered pro-homeownership based on supplementary housing codings of manifestos (Kohl, 2020).The fifth and sixth specifications operationalize public and education investment using the "investment in infrastructure and technology" and the "education spending" items, as suggested by Kraft (2018) and are analogous to the variables in the macro-level analysis in the previous section.
We include control variables at the level of manifestos, the political system and macro-economic conditions.In our analysis manifestos are generally nested in countries, election years and parties.We use parties, rather than the more abstract party families, to control for programmatic changes within parties.We control for the length of manifestos, the number of parties in the political system at each election, and the party vote share at the previous election.At the macro-level, we control for (interpolated) homeownership levels, total population and GDP per capita (both transformed using natural logs).See appendix A for a full list of all data sources used for this part of the analysis.We also interact house price changes with homeownership rates to account for whether the political effects of rising house prices vary in countries with high levels of homeownership.A second interaction of house prices with the political orientation of manifestos at the last election is also included to test for asymmetric partisan effects, but remains insignificant throughout.
Due to the hierarchical nature of the data, we use a multi-level OLS regression, except for the pro-homeownership specification, which uses a logistic model due to the binary nature of the dependent variable.As robustness checks we estimate a fixed effects model accounting for time, country and party effects, which also includes robust standard errors (see appendix F).We also estimate our sample for two time periods: the first includes pre-1980, when there was a welfare expansion coinciding with house price stagnation (not shown), and the second one shown here looks at the period after 1980 when the converse occurred.The overall results of our robustness checks are similar to those in the full-sample multilevel model in Table 1, thus providing additional support for our analysis.
The first three columns in Table 1 report the elasticity of party positions on welfare to house price increases.The results of the first two specifications show house price increases in the years prior to an election are significantly associated with party proposals to expand welfare programs and reduce cutbacks on social transfers.Whilst specification 3 shows rising house prices tend to shift the general political orientation of party manifestos to the left.Columns 4-6, in turn, report the house-price effects on public investment and homeownership programs.The results show no significant direct effects of house prices on party positions towards homeownership or education spending, but there are positive and significant effects on infrastructure and technology investment spending.
The interactions of house prices with both homeownership rates and previous party positions are not significant (not shown).This suggests there is no asymmetry resulting from levels of homeownership or party ideology, which challenges existing findings (Beckmann, 2020) that have fewer subsamples and less complete coverage than ours.
The second part of this analysis uses the same model specifications as the first part, but includes real rent increases and their interaction with real house prices as independent variables to account for the 'other' side of the housing market and homeownership outsider groups.We include these variables to assess whether parties look to allocate resources to economically compensate aspiring homeowners and private tenants via the welfare system.Table 2 summarises the results of the second part of the analysis and show rent increases by themselves have no significant effects on any of the dependent variables.However, the interaction of house prices and rent costs show significant and positive effects on pro-welfare positions, yet negative and significant effects on the right-left index.These results suggest that when rising house prices are accompanied by rising rent costs, parties shift towards the left and increase the allocation of fiscal resources towards welfare spending to compensate homeownership outsider groups.
Overall, the results of the "political elasticities" of house price changes provide an important means of understanding our macroeconomic findings showing that rising house prices are associated with an increase in the allocation of fiscal resources towards supporting homeowner's economic interests.Rather than turning to fiscal retrenchment and austerity in times of house price inflation, parties try to garner political support from homeowners by allocating fiscal spending to support house price appreciation.Whilst simultaneously appealing to voters in homeownership outsider groups negatively affected by house price increases through the welfare system.We demonstrate the political symmetry of these effects, as parties across the political spectrum tend to shift leftwards with house price increases and propose more of both welfare and public investment spending to appeal to both homeownership insider and outsider voting groups.Thus, lending support to previous studies showing that how price appreciation shifts political parties leftwards (Beckmann, 2020).

Conclusion
Cross-class coalitions have a significant influence over how fiscal policies are allocated in ways aligned with their economic interests.Rather than encouraging parties to adopt fiscal austerity and welfare retrenchment (e.g.Ansell, 2014), we demonstrate rising house prices motivate governments to allocate fiscal resources towards increased welfare and (social) investment spending policies to maintain upward pressure on house prices, which is the key economic concern for homeowners as a cross-class coalition.Whilst parties may have their own ideological views on fiscal policy allocations, we argue that house prices may override their political ideology for more pragmatic reasons.In particular, we argue sustaining upward pressure on house prices by increasing investment spending, and absorbing housing market risks via welfare spending, is necessary for maintaining political legitimacy with homeowners as a key cross-class voting coalition.Several economic analyses support the positive reaction of house prices to various kinds of government expenditure (Gupta et al., 2019;Khan & Reza, 2017), but often lack an explanation for the macroeconomic outcome they observe.
Our findings add to recent scepticism in housing studies about easy trade-off relationships between mortgage debt, homeownership, and house price growth on the one hand, and public welfare on the other.They are rather in line with recent growth-model approaches which put homeownership at the centre of consumptionand debt-driven growth, backed by a political coalition in favour of stable and rising house prices.Additional government revenues, mechanically supported by rising house prices through various channels, are then used for maintaining house price growth through government investment expenditure, but also through social investment spending.Even unproductive welfare expenditure may rise, as the growth coalition may want to broaden its political base by compensating losers in the housing-oriented growth model.In this view, the positive political house price elasticities constitute a political feedback mechanism supporting the homeownership insider group.However, these results may be driven by a low interest rate environment, which has come under increasing pressure in recent years due to rising inflation.It is also an empirically open question how the homeownership insider group relates to the construction-led growth coalition (Kohl & Spielau, 2022), which could be in favour of additional supply and hence potential asset price shocks.
Another key question emerging from this analysis is the question of scale, as local politics is often related to concerns around housing markets and public amenities.Therefore, the results of our national-level analysis provide a starting point for a new research agenda to be developed exploring the sub-national level relationship between local house prices and public services, especially from a comparative perspective.
The results demonstrate the important political consequences from being able to access the financial asset-price gains from homeownership.However, the political consequences of the distribution of housing asset price gains remains an under-explored research area, particularly the politics of inheritance around the private home, and the distributional problems of those excluded from accessing the home as a financial asset.Furthermore, private homeownership is but one form of asset class that may establish cross-class coalitions with significant electoral consequences.As such, the results of our analysis suggest future research could examine the fiscal policy effects of other asset classes, particularly the inequalities of savings and private pension investments in stocks or bonds.
Our research also highlights the importance of non-homeowning outsider groups as a cross-class economic interest group, which parties appeal to through the allocation of fiscal resources through welfare spending.However, their political preferences and the politics of rental markets are notable absences in existing research, which is largely oriented around homeowner insider groups.Homeowners are an important political group in many countries, but there are substantial numbers of non-homeowning voters who go beyond the stereotype of poor tenants.These range from the nursing home population to aspiring homeowners living with their parents, or even high-income earners who are not yet settled, or who have insufficient savings for a deposit.In most North-Western European cities, including London, approximately half of voters are in the non-homeowning outsider group, whilst many major Swiss and German cities have homeownership rates below 20%.This neglected "other side" of the housing market also coincides with the urban-rural cleavage line that has regained importance in recent elections.Given the rising housing costs for non-home owning households, we suggest that future research should account for the politics of rental markets, both public and private, and their subsequent effects on voting preferences.

Note
1.Although our results challenge those of Ansell (2014), we can replicate Ansell's (2014) macroeconomic results if we examine the same countries in the same time period  using a short-run analysis only.However, as we use levels, our analysis takes longer-run effects into account that override the short-run effects captured by Ansell's (2014) analysis.Our analysis also covers a longer time horizon.

Figure 1 .
Figure 1.Mechanisms linking house prices and fiscal policy outcomes.

Figures 2 -Figure 2 .
Figures 2-4 provide graphical representations of the estimates of the long-run coefficients from the econometric models performed in this analysis.See section D of the supplementary material for a full breakdown of the econometric results in table form.The generic form of each ARDL equation with two lags of each variable is:

Figure 3 .
Figure 3. House price and government ideology results.

Figure 4 .
Figure 4. House prices and rents results.

Table 1 .
the effects of house price growth on party positions in manifestos.

Table 2 .
Rent-price effects party positions in manifestos.