Local governments’ indebtedness and its impact on real estate prices

Abstract

Chinese

This paper estimates the causal effect of public debt on real estate prices and rental prices. We identify shocks to debt of self-governed cities in Germany and control for potential benefits such as an increased supply of public goods, which may come together with increased indebtedness. Using spatial variation across self-governed cities, we can quantify the causal effect. We find that shocks to public debt capitalise into property prices, questioning the presence of debt illusion in Germany. Rental prices, on the other hand, do not seem affected by public debt but by the actual tax burden, indicating that renter illusion does not reflect an illusion but rational behaviour.

Keywords

debt illusion housing prices local government debt renter illusion

Introduction

The location, and by that local infrastructure, is probably the most important determinant of real estate prices. A comprehensive body of literature evaluates whether public goods such as education (Downes and Zabel, 2002; Gibbons and Machin, 2003) and transport facilities (Bajic, 1983; Bowes and Ihlanfeldt, 2001; Brandt and Maennig, 2012; Laakso 1992) capitalise into real estate prices. Hilber (2017) gives an excellent overview of the capitalisation hypothesis. The literature is suggestive of positive effects of infrastructure investment on property prices. However, there are different channels through which infrastructure improvements might affect real estate prices. Bowes and Ihlanfeldt (2001) disentangle the effect of rail transit stations into potentially positive effects of lower commuting costs and higher attractiveness for retail activity from potentially negative effects of higher emissions and higher crime rates. Brandt and Maennig (2012) find a positive effect of rail access for the German city Hamburg. The authors also take increased land transfer taxes due to higher property prices into account.

Another negative effect on housing prices may result from the costs of such projects. There is a substantial body of literature investigating effects of local fiscal variables. In a seminal contribution, Tiebout (1956) introduces the idea that local governments offer a basket of public goods and collect taxes to finance these public goods. Papers that quantify the effect of local fiscal variables on real estate prices therefore mostly concentrate on the effect of local taxes. The literature on the effects of local governments’ indebtedness is still rather limited. However, an understanding of the way public debt affects housing prices is crucial. The financial crisis and the following Great Recession originated in the housing market. Will debt-financed fiscal policies put further pressure on housing prices and by that, at least in part, counteract expansionary effects of fiscal policy?

This paper investigates the effect of local public debt on real estate prices and rental prices and is thereby related to literature on fiscal illusions. Banzhaf and Oates (2013) distinguish between debt illusion – the notion that households prefer debt finance over increasing taxes, and renter illusion – the notion that renters prefer financing local public expenditures via property taxes.

Empirical tests of the debt illusion mostly investigate whether residents prefer debt finance over increasing taxes and typically find a preference for debt finance (Banzhaf and Oates, 2013).¹ However, this may not necessarily reflect a debt illusion, e.g. if financing costs for local governments are lower than for households. Then, a preference for debt finance may be perfectly rational. Another way of testing for debt illusion is to test whether local public debt capitalises into real estate prices. Eichenberger and Stadelmann (2010) argue that rational households should anticipate that higher local debt levels will result in either higher local taxes or a lower supply of public goods at some point in the future. If Ricardian equivalence holds, local public debt shocks should therefore capitalise into house prices. However, the identification of debt shocks is not trivial. Studies that investigate the debt capitalisation hypothesis therefore typically analyse whether higher debt levels are associated with lower real estate prices and mostly find evidence for (potentially incomplete) capitalisation. Most related to this paper – trying to quantify the causal effect of local public debt on property prices – is MacKay (2014). MacKay identifies the growing awareness of unfunded pension liabilities in San Diego as a shock to the perception of local public debt. He finds that the increase in perceived public debt decreased house prices by a factor greater than one.

Empirical tests of the renter illusion hypothesis typically investigate whether the share of renters in a community is associated with a higher level of public expenditures. Even though empirical evidence in favour of a renter illusion is overwhelming, this is not necessarily irrational as tenants may also assess expenditures differently, e.g. because of differences in their bundle of consumption goods (Banzhaf and Oates, 2013).² Employing a different approach, Blom-Hansen (2005) uses survey data and finds that tenants do not have illusions with regard to property taxes but simply are ignorant.

This paper investigates whether shocks to public debt capitalise into property prices or materialise in rental prices. We thereby aim to quantify the causal effect of public debt and contribute to the literature on the debt and renter illusion. To do so, we identify shocks to a local government’s debt positions that are associated with government investment. Owing to our identification strategy, we focus on self-governed cities in North Rhine-Westphalia (NRW): first, because it is one of the most densely populated areas in Germany; second, because this allows for an attribution of governments debt which is not always clear because of implicit or explicit guaranties of higher order government entities. We analyse the effect of local public debt on apartment prices, as apartments are more common than houses.³ We control for potential benefits of public investment activity, as an increased supply of public goods, by analysing small border regions.⁴ Infrastructure improvements are typically not exclusive to properties within the boundary of a city but affect all the surrounding properties. By focusing on a narrow border region of two cities, all properties should be affected similarly by infrastructure improvements. The debt burden of the improved infrastructure, on the other hand, has to be taken on by residents of the respective city. Using spatial variation across self-governed cities, we quantify the causal effect of public debt on prices and rental prices.

An increase in local governments’ per capita indebtedness lowers apartment prices by a factor larger than one. Taking into account that typically more than one individual lives in an apartment, the per capita indebtedness of all individuals living in an apartment reduces the property value (and net worth of residing individuals) by a factor of about one, hinting at the absence of debt illusion. With regard to rental prices, these react to actual tax payments, not to shocks to indebtedness. We argue that this indicates the presence of renter illusion. However, renters may actually be able to avoid paying for public debt by relocating if high debt levels materialise in tax increases. In line with the literature, we therefore think that the renter illusion may actually not be an illusion but be perfectly rational.

Data

ImmobilienScout24 provides data on apartment prices and rental prices per square metre. ImmobilienScout24 is Germany’s largest online real estate marketplace with a self-reported market share of about 50% of all real estate transactions in Germany (Georgi and Barkow, 2010). At this online marketplace, potential sellers and property owners can place adverts to sell or rent out their properties. Therefore, real estate prices and rents in this study refer to asking prices, not transaction prices. Even though asking prices are less reliable than transaction prices, they seem to be a useful substitute given the lack of transaction data in Germany (Faller et al., 2009). Kholodilin et al. (2017) compare transaction and internet advertisement prices for Berlin and argue that differences are small. Information on apartment prices and rental prices are available at monthly frequency from January 2007 to June 2014. To approximate the transaction price and to eliminate potential biases resulting from differences in an object’s advertisement period we use only the last observation before an advertisement is set as inactive. Inactive objects do not enter any search queries and most probably indicate a transaction. We therefore use the last advertisement of objects in the period from January 2007 to December 2013.⁵ The data set includes information on an object’s asking price as well as various object-specific characteristics such as living space, the number of rooms and the year of construction.⁶ Summary statistics for objects advertised for sale or rent are reported in Table A1.⁷

Information on the fiscal positions of self-governed cities are available at IT.NRW, the statistical office in NRW. IT.NRW reports the credit positions of self-governed cities, subdivided into two categories. These are credit for investment and liquidity credit. Credit for investment, Fundierte Schulden, refers to debt accumulated to finance investment projects, mostly consisting of infrastructure spending such as new roads or investment related to cultural offerings such as museums or schools. Even though the debt break that was introduced in 2009 in Germany does not apply to local governments, pre-existing laws prohibited cities’ accumulation of debt. Investment credit may only be used to finance investment projects, not to finance running deficits. Liquidity credit, Kassenkredite, is used to buffer temporary liquidity shortages. Data on these two credit positions are available on annual frequency for the years 1995 to 2013. We divide local governments’ debt positions by population to yield per capita values. To take a city’s fiscal capacity into account, we use information on the city’s actual tax revenues per capita as well as the strength of revenues from taxes.⁸ All information is available at IT.NRW.

The neighbourhood may have an effect on apartment prices and rents. We therefore control for neighbourhood characteristics, average purchasing power per capita, the unemployment rate and the percentage of households where the head of the household has a migrant background, using a grid data with an edge length of 1 km. If neighbourhood characteristics exhibit discontinuities at the boundaries of cities, this would be a drawback for our analysis as discontinuities may indicate market segmentation. In our sample, however, neighbourhood characteristics seem continuous; we show the geographic distribution of purchasing power per capita in Figure A1. Information refers to the year 2010 and is obtained from microm ConsumerMarketing.⁹ We match this information to the real estate data set via the location.

Identification of shocks

To estimate the causal effect of local governments’ indebtedness on real estate prices we need to tackle two issues. First, we need to identify shocks to local governments’ debt positions. Second, we need to control for potential benefits of increased indebtedness, such as a potential increase in the supply of public goods.

Shocks to public indebtedness

Before we can analyse the effect of public indebtedness, we need to ensure that changes in our debt variable represent actual shocks. We focus on self-governed cities in NRW. The focus on one state, NRW, ensures that higher-order government debt is the same for all municipalities in our sample. The focus on self-governed cities ensures that there are no smaller government entities that are allowed to accumulate public debt.¹⁰ Self-governed cities report indebtedness subdivided into two categories: investment credit and liquidity credit (section ‘Data’). In this study, we focus on investment credit, as our strategy to control for potential positive effects of indebtedness (section ‘Supply of public goods’) is only valid for this debt position.

Measuring shocks to public debt is not a trivial task. There are no surveys or financial market indicators that indicate expectations of future public debt. This is also true for public debt on the local level. To identify shocks to public debt nevertheless, we employ assumptions with regard to the process that generates the expected path of public debt. We employ two different assumptions: adaptive expectations and rational expectations.

Under adaptive expectations, the best guess for a variable’s value in the next period – in our case public indebtedness – is that of the current period (Ezekiel, 1938). Generating shocks to local government’s indebtedness is then straightforward.

Under rational expectations, an expected value should be an unbiased estimator of its ex-post realisations (Muth, 1961). This raises the question of what is a good forecast for public debt. In the forecasting literature, simple AR processes typically represent the benchmark forecasts, which are hard to outperform in terms of forecast accuracy (Rapach and Strauss, 2007, 2009; Stock and Watson, 2003, 2004). We therefore hold to this approach and assume an AR process.¹¹

We estimate an AR model for the logarithm of investment credit. We use all 17 self-governed cities in NRW that are directly adjacent to another self-governed city.¹² Information on the credit positions is available for the years from 1995 to 2013; we estimate the AR model for the logarithm of investment credit for these years. We include one lag (according to the Schwarz Information Criterion). We have to reject the hypothesis of homogeneity in the AR parameter and the fixed effect. We cannot reject the hypothesis of cross-sectional independence and therefore estimate a panel AR with fixed effects and heterogeneity in the AR coefficient.¹³ Estimation results and summary statistics for the shocks are reported in Tables A2 and A3 in the Appendix.¹⁴

Supply of public goods

To control for potential positive effects of public debt because of an increased supply of public goods, we concentrate on border regions, as proposed by Black (1999), of adjacent self-governed cities and shocks to investment credit. As discussed in section ‘Data’, investment credit is driven by actual government investment such as expenditures for new infrastructure. However, infrastructure investment typically affects real estate prices in a surrounding area, not only properties within the city boundaries. If we assume that there are two adjacent properties, both on different sides of a border separating two cities, it is easy to see that the relevant infrastructure for those two properties is almost identical.

Starting from the left, Figure 1 shows the self-governed cities of Duisburg, Mülheim an der Ruhr and Essen. The bands surrounding the two borders are the 3-km buffers, each representing one border group consisting of two adjacent cities. The two pairs of dots represent theoretical real estate objects; circles surrounding these objects have a radius of 1 km and should be indicative of the relevant infrastructure. Assume that one city increases its infrastructure spending at the expense of increased indebtedness, e.g. to build a new tramline within its city’s borders. It is easy to see that two adjacent properties should benefit equally from the improved infrastructure. However, the financial burden of such a project is typically not split between the two cities. Therefore, comparing the evolution of property prices in a border group and linking this to shocks to the investment credit position allows us to estimate the causal effect of public debt that is not associated with any benefits on real estate prices.

Figure 1.

Border groups.

Estimation

In our estimation, we follow a flexible approach. We assume that only the price trend within a border region is similar and that prices within a border group depend on the city in which the property is located. However, the government may be able to supply public goods that are not financed by investment credit but are bound to the place of residence. Examples for this may be differences in cost recovery for rubbish collection or differences in how generous cities are with respect to social benefits. We implement this by allowing for a discontinuity in apartment prices at the border. We essentially estimate whether differences in shocks to investment credit explain time variation in the size of the discontinuity at the boundary in the different border groups.

We also take into account that the similarity of the infrastructure relevant for a property may decrease with increasing distance to the boundary. To take this into account, we include distance to the border as a control variable. Distance is allowed to have a different effect for each city in each border group. To capture the causal effect of public debt, we estimate equation (1).

\begin{matrix} P_{igct} = α_{Z} Z_{igct} + α_{gt} + α_{gc | c = 2} + α_{F} F_{gct} \\ + α_{D} D_{igct} + ε_{igct} \end{matrix}

(1)

The variable $P$ represents information on square metre prices or square metre rental prices. $Z$ contains object-specific¹⁵ and neighbourhood¹⁶ characteristics, $F$ represents information on fiscal variables of the city in which the apartment is located. This includes the two shocks to the cities’ investment and liquidity credit positions as introduced in section ‘Shocks to public indebtedness’ as well as the cities’ actual tax revenues and the strength of revenues from taxes, all in per capita terms. As we expect fiscal variables to affect apartment prices in absolute terms, e.g. an increase in debt of 1 euro lowers apartment prices by X euro, we do not employ a logarithmic transformation for the fiscal variables.¹⁷ $i, g, c$ and $t$ identify, respectively, the object, the border group¹⁸, the city within a border group and the year of observation. $D_{igct}$ represents the apartment’s distance to the border. The coefficient $α$ represents the price of the respective characteristic: $α_{gt}$ represents time fixed effects on the border group level; $α_{gc | c = 2}$ controls for the city within the the border group, allowing for a time invariant discontinuity in apartment prices with respect to distance to the border. As we are interested in the effect of city-wide shocks to investment credit and these are homogenous for all properties within a city, potentially resulting in serial correlation on the city level, we cluster the error term $ε$ on the city level.

Before we estimate equation (1), we have to set two thresholds. First, we need to set a threshold distance to the boundary below which apartments have to fall to be included in a border group and thus in the analysis. Second, we need to ensure that the border region is inhabited as we do not want to extrapolate a price effect if there are no observations very close to the boundary. We therefore set a threshold for the minimum number of observations in both cities close to the boundary. This threshold refers to the number of observations within 250 m of the border in both cities. In our baseline estimate, we include objects within 3 km of the boundary and set a threshold of 50 for the minimum number of observations.

As we are looking at a densely populated region with a cluster of self-governed cities, some apartments are included in more than one border group, given each border group includes all apartments within a certain threshold distance to the border with another self-governed city. To avoid complications arising from inclusion of observations multiple times, we include each apartment only once – in the border group where the distance to the border is the shortest.

Results

Ricardian equivalence implies that a shock to per capita public debt of 1 euro lowers inhabitants’ net worth by 1 euro. However, hedonic price functions for property prices typically are estimated given a logarithmic transformation, as estimating elasticities better fits the data.¹⁹ We therefore present the results for both left-hand side variables, square metre prices and square metre rental prices, as well as their logarithmic transformations and compare the results. The estimation results are reported in Tables 1 to 4.

Table 1.

Estimation results for apartment prices.

Expectations	Rational expectations			Rational expectations			Rational expectations			Adaptive expectations
Minimum observations within 250 m	50^b			25^a			50^b			50^b
Maximum distance to boundary	3 km			3 km			1.5 km			3 km
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
Dummy indicating border group and year	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Dummy for second city in border group	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Living space_i	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Distance to boundary_ij	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Distance to boundary_ij²		Yes	Yes		Yes	Yes		Yes	Yes		Yes	Yes
Distance to boundary_ij³			Yes			Yes			Yes			Yes
Shock to investment credit	−0.0163***	−0.0190***	−0.0203***	−0.0236***	−0.0271***	−0.0281***	−0.0228**	−0.0213**	−0.0214***	−0.0085	−0.0118*	−0.0132**
	(0.0044)	(0.0054)	(0.0051)	(0.0081)	(0.0085)	(0.0083)	(0.0102)	(0.0083)	(0.0082)	(0.0066)	(0.0065)	(0.0063)
Shock to liquidity credit	0.0057	0.0041	0.0038	−0.0175	−0.018	−0.0184	−0.0584**	−0.0648**	−0.0641**	0.0089	0.0042	0.0029
	(0.0287)	(0.0280)	(0.0283)	(0.0261)	(0.0260)	(0.0263)	(0.0251)	(0.0253)	(0.0258)	(0.0120)	(0.0101)	(0.0102)
Strength of tax revenues	0.0905	0.0571	0.0512	0.4306	0.4287	0.4071	1.1570**	1.0081*	1.0753*	−0.1247	−0.1594	−0.1691
	(0.5446)	(0.5525)	(0.5617)	(0.4145)	(0.4215)	(0.4298)	(0.5086)	(0.6046)	(0.5983)	(0.6395)	(0.6373)	(0.6439)
Actual tax revenues	−0.1106	−0.0906	−0.0948	−0.3606	−0.3664	−0.3574	−0.9942***	−0.8796**	−0.9244**	0.074	0.0914	0.0894
	(0.4179)	(0.4246)	(0.4303)	(0.3296)	(0.3367)	(0.3414)	(0.3724)	(0.4398)	(0.4359)	(0.5016)	(0.4989)	(0.5031)
Observations	12,521	12,521	12,521	16,301	16,301	16,301	6021	6021	6021	12,521	12,521	12,521

Notes: *(p < 0.10), **(p < 0.05), ***(p < 0.01). i indicates the border group. j indicates the city in the border group. Standard errors clustered on the city level in parentheses. Local governments’ fiscal variables are in euros. Additional control variables are: object’s age and age², the apartment type (ten categories: ‘no information’, ‘top floor’, ‘loft’, ‘maisonette’, ‘penthouse’, ‘terrace flat’, ‘floor apartment’, ‘mezzanine’ and ‘basement’, reference category is ‘apartment’); the apartment condition (nine categories: ‘no information’, ‘first occupation’, ‘as new’, ‘renovated’, ‘in need of renovation’, ‘modernised’, ‘first occupancy after modernisation’, ‘redeveloped’, reference category is ‘well maintained’); whether there is an elevator, a garden, a balcony or a built-in kitchen (each variable with the characteristics yes and no information, reference category is not present; and dummy variables indicating the apartment’s year of construction (ten categories: ‘before 1920’, ‘1921–1945’, ‘1946–1950’, ‘1951–1960’, ‘1961–1970’, ‘1971–1980’, ‘1981–1990’, ‘1991–1995’, ‘1996–2005’, reference category is ‘2006–not built yet’); as well as grid-level information on the unemployment rate; a dummy variable if the unemployment rate is not available; the purchasing power; the percentage of households where the head of household has a migrant background; and a dummy variable if the information is not available. ^aBorder groups are: Duisburg|Muelheim, Duisburg|Oberhausen, Essen|Muelheim, Essen|Oberhausen, Muelheim|Oberhausen, Gelsenkirchen|Bochum, Bochum|Dortmund and Bochum|Herne. ^bBorder groups are: Duisburg|Muelheim, Essen|Muelheim, Muelheim|Oberhausen, Bochum|Dortmund and Bochum|Herne.

Table 2.

Estimation results for log apartment prices.

Expectations	Rational expectations			Rational expectations			Rational expectations			Adaptive expectations
Minimum observations within 250 m	50^b			25^a			50^b			50^b
Maximum distance to boundary	3 km			3 km			1.5 km			3 km
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
Dummy indicating border group and year	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Dummy for second city in border group	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Living space_i	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Distance to boundary_ij	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Distance to boundary_ij²		Yes	Yes		Yes	Yes		Yes	Yes		Yes	Yes
Distance to boundary_ij³			Yes			Yes			Yes			Yes
Shock to investment credit	−0.0113*	−0.0137**	−0.0141**	−0.0191**	−0.0220**	−0.0224**	−0.0159**	−0.0139*	−0.0143*	−0.0057	−0.0086	−0.0093
	(0.0059)	(0.0068)	(0.0064)	(0.0088)	(0.0093)	(0.0090)	(0.0079)	(0.0075)	(0.0076)	(0.0068)	(0.0069)	(0.0067)
Shock to liquidity credit	−0.0201	−0.0209	−0.0213	−0.0356*	−0.0361*	−0.0362*	−0.0710***	−0.0744***	−0.0738***	0.0133	0.0100	0.0086
	(0.0205)	(0.0200)	(0.0201)	(0.0212)	(0.0209)	(0.0209)	(0.0264)	(0.0258)	(0.0266)	(0.0084)	(0.0099)	(0.0105)
Strength of tax revenues	−0.2766	−0.2882	−0.3197	0.0564	0.0744	0.0329	0.2463	0.1273	0.2372	−0.5428	−0.5542	−0.5824
	(0.4686)	(0.4810)	(0.4778)	(0.3498)	(0.3582)	(0.3608)	(0.6585)	(0.7340)	(0.7327)	(0.4640)	(0.4639)	(0.4693)
Actual tax revenues	0.1541	0.1597	0.1768	−0.0981	−0.1164	−0.0917	−0.3299	−0.2371	−0.3131	0.3804	0.3832	0.3966
	(0.3528)	(0.3630)	(0.3594)	(0.2789)	(0.2867)	(0.2876)	(0.4791)	(0.5316)	(0.5321)	(0.3547)	(0.3531)	(0.3560)
Observations	12,521	12,521	12,521	16,301	16,301	16,301	6021	6021	6021	12,521	12,521	12,521

Table 3.

Estimation results for apartment rental prices.

Expectations	Rational expectations			Rational expectations			Rational expectations			Adaptive expectations
Minimum observations within 250 m	50^b			25^a			50^b			50^b
Maximum distance to boundary	3 km			3 km			1.5 km			3 km
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
Dummy indicating border group and year	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Dummy for second city in border group	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Living space_i	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Distance to boundary_ij	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Distance to boundary_ij²		Yes	Yes		Yes	Yes		Yes	Yes		Yes	Yes
Distance to boundary_ij³			Yes			Yes			Yes			Yes
Shock to investment credit	0.0088	0.0090	0.0074	0.0438	0.0436	0.0407	0.0171	0.0147	0.0125	0.0304***	0.0310***	0.0297**
	(0.0221)	(0.0217)	(0.0222)	(0.0392)	(0.0393)	(0.0387)	(0.0517)	(0.0526)	(0.0528)	(0.0110)	(0.0113)	(0.0117)
Shock to liquidity credit	−0.0601*	−0.0652*	−0.0680**	−0.1293***	−0.1275***	−0.1261***	0.0189	0.0133	0.0108	0.0592**	0.0553**	0.0541**
	(0.0334)	(0.0334)	(0.0328)	(0.0462)	(0.0469)	(0.0473)	(0.0281)	(0.0274)	(0.0279)	(0.0255)	(0.0252)	(0.0246)
Strength of tax revenues	1.9262***	1.8434***	1.8001***	1.5467	1.4263	1.4303	0.8899	0.8712	0.9881	1.2896***	1.1998**	1.1378**
	(0.5536)	(0.5319)	(0.5539)	(0.9542)	(0.9678)	(0.9906)	(0.9347)	(1.0002)	(0.9814)	(0.4522)	(0.4673)	(0.4517)
Actual tax revenues	−1.8319***	−1.7709***	−1.7410***	−1.3451*	−1.2651	−1.2738	−1.0761	−1.0610	−1.141	−1.2679***	−1.2043***	−1.1600***
	(0.4861)	(0.4691)	(0.4866)	(0.7722)	(0.7828)	(0.8015)	(0.8385)	(0.8822)	(0.8684)	(0.2415)	(0.2523)	(0.2429)
Observations	36,797	36,797	36,797	50,470	50,470	50,470	16,255	16,255	16,255	36,797	36,797	36,797

Table 4.

Estimation results for log apartment rental prices.

Expectations	Rational expectations			Rational expectations			Rational expectations			Adaptive expectations
Minimum observations within 250 m	50^b			25^a			50^b			50^b
Maximum distance to boundary	3 km			3 km			1.5 km			3 km
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
Dummy indicating border group and year	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Dummy for second city in border group	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Living space_i	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Distance to boundary_ij	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Distance to boundary_ij²		Yes	Yes		Yes	Yes		Yes	Yes		Yes	Yes
Distance to boundary_ij³			Yes			Yes			Yes			Yes
Shock to investment credit	0.0032	0.0033	0.0029	0.0086	0.0086	0.008	0.0048	0.0042	0.0038	0.0063***	0.0064***	0.0062***
	(0.0041)	(0.0041)	(0.0042)	(0.0068)	(0.0068)	(0.0067)	(0.0099)	(0.0100)	(0.0101)	(0.0022)	(0.0022)	(0.0024)
Shock to liquidity credit	−0.0103*	−0.0113*	−0.0119*	−0.0254***	−0.0252***	−0.0250***	0.0049	0.0040	0.0036	0.0129***	0.0123***	0.0119***
	(0.0061)	(0.0063)	(0.0062)	(0.0080)	(0.0081)	(0.0081)	(0.0045)	(0.0046)	(0.0047)	(0.0039)	(0.0039)	(0.0038)
Strength of tax revenues	0.2567**	0.2454**	0.2348**	0.2322	0.2122	0.2122	0.0708	0.0708	0.0878	0.1558*	0.1430*	0.1285
	(0.1107)	(0.1077)	(0.1127)	(0.1617)	(0.1647)	(0.1705)	(0.2110)	(0.2228)	(0.2185)	(0.0836)	(0.0856)	(0.0808)
Actual tax revenues	−0.2590***	−0.2502***	−0.2429**	−0.2049	−0.1914	−0.1927	−0.1203	−0.1206	−0.1323	−0.1673***	−0.1578***	−0.1476***
	(0.0942)	(0.0919)	(0.0959)	(0.1286)	(0.1311)	(0.1357)	(0.1794)	(0.1878)	(0.1849)	(0.0464)	(0.0481)	(0.0447)
Observations	36,797	36,797	36,797	50,470	50,470	50,470	16,255	16,255	16,255	36,797	36,797	36,797

Let us first examine the presence of debt illusion by investigating the effect of shocks to indebtedness on apartment prices. The most reliable estimates, at least in our opinion, include the square metre price as left-hand side variable. Additionally to that, we assume rational expectations for the two credit positions. The estimation results are reported in columns (1)–(3) of Table 1. We find that shocks to investment credit, which we, because of our identification strategy, interpret as shocks to public indebtedness not associated with any benefit, are highly significant and lower apartment prices in all cases. An increase in this debt position by 1 euro lowers square metre apartment prices by about 2 cents. For an average apartment of about 80 m², which is the average in our sample, this translates into a price reduction of about 1.6 euro. Shocks to liquidity credit are insignificant.

We test for the robustness of this result by varying the two aforementioned thresholds. Using the sample including all border groups with at least 25 observations within 250 m of the boundary (columns (4)–(6)), the quantitative effect slightly increases and remains significant. We also lower the maximum distance to the boundary. The effect remains significant when halving the maximum distance to the boundary to 1.5 km (columns (7)–(9)). We assume the less reliable, at least in our opinion, adaptive expectations (columns (10)–(12)), lower the estimated effect as well as its significance. In the case where we allow for a cubic effect of distance to the boundary, however, the effect still is significant at the 5% level.

Turning to the estimations with the logarithmic transformation of square metre prices as dependent variable (Table 2), the results seem consistent with the previous estimations. In the baseline estimation, the effect of shocks on investment credits are negative and significant. A debt shock of 1000 euro lowers apartment prices by about 1.4%, which corresponds to a price effect for the average apartment of about 1400 euro.²⁰ The effect is robust with respect to varying the two thresholds values (25 observations within 250 m and 1.5 km maximum distance to the boundary). For adaptive expectations, the prefix may be seen as an indication of a negative effect. However, coefficients are not significant at conventional levels.

Let us now turn to renter illusion by discussing the price effect of shocks on investment credit to rental prices of apartments. The estimation results can be found in Tables 3 and 4. In the estimations with square metre rental prices as left-hand side variable (Table 3) the coefficients for investment shocks are positive in all cases, however they are mostly insignificant. Only in the case of assuming adaptive expectations we find significant positive effects (columns (10)–(12)). This pattern is also found when we use log square metre rental prices as the left-hand side variable (Table 4).

We therefore argue that debt does not affect rental prices. The, in some cases, positive effect of shocks to the investment credit position on rental prices could be interpreted as evidence that we may not have been able to control for all of the possible benefits of an increase in this debt position. Therefore, our estimates suggesting a negative effect on apartment prices could be seen as a lower bound. With regard to fiscal variables, rental prices seem to be negatively correlated with the tax burden (Table 3), of which property taxes are an important share. We interpret this as evidence that tenants are not necessarily subject to an illusion, as rental prices are not affected by shock to investment credit. Renters do care about their tax burden, as property taxes can be shifted from owners to tenant if specified in the rental contract, which typically is the case in Germany.

Conclusion

In this paper, we investigate the causal effect of government debt on apartment prices and rental prices. We do this with the example of self-governed cities in North Rhine-Westphalia. These cities act as both districts and municipalities, which allows us to identify shocks to residents’ per capita local public debt position. We thereby contribute to the literature on debt illusion and renter illusion. By identifying shocks to the credit positions of local governments and by controlling for potential benefits that may be associated with shocks to the investment credit positions of local governments, we isolate the causal effect of public indebtedness on apartment prices.

Apartment prices strongly react to shocks to public debt. An increase in public debt of 1 euro lowers the square metre apartment price by about 2 cents, which translates into a decrease in the apartment price of about 1.6 euro. This is indicative of public debt capitalising into property prices. The size of the effect speaks rather against the presence of debt illusion.

In the rental market, shocks to public debt do not seem to affect prices. If the debt burden materialises in an increased tax burden, this lowers rental prices. We therefore tend to think that renter illusion may actually not be an illusion. Tax increases seem to be offset by lower rental prices. Owing to higher mobility, tenants may be able to avoid paying for public debt.

Footnotes

Appendix

Table A3.

Summary statistics for shocks to the cities’ credit positions.

	Rational expectation shock		Adaptive expectation shock
	Investment credit	Liquidity credit	Investment credit	Liquidity credit
Mean	55.05	202.08	8.35	221.09
Median	−6.78	83.05	−22.9	207.55
Min	−528.47	−654.79	−622.92	−512.77
Max	1607.58	2040.16	1676.45	771.16
Standard deviation	262.03	408.64	236.25	228.42
Skewness	3.09	1.96	3.8	0.06
Kurtosis	15.37	8	28.63	3.47

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

an de Meulen

Micheli

Schaffner

(2014) Documentation of German real estate market data – Sample of real estate advertisements on the internet platform ImmobilienScout24. RWI Materialien 80. Essen: RWI.

Bajic

(1983) The effects of a new subway line on housing prices in metropolitan Toronto. Urban Studies 20(2): 147–158.

Banzhaf

Oates

(2013) On fiscal illusion in local public finance: Re-examining Ricardian equivalence and the renter effect. National Tax Journal 66(3): 511–540.

Bauer

Feuerschütte

Kiefer

, et al. (2013) Ein hedonischer Immobilienpreisindex auf Basis von Internetdaten: 2007–2011. AStA Wirtschafts- und Sozialstatistisches Archiv 7(1–2): 1–26.

Black

(1999) Do better schools matter? Parental valuation of elementary education. The Quarterly Journal of Economics 114(2): 577–599.

Blom-Hansen

(2005) Renter illusion: Fact or fiction? Urban Studies 42(1): 127–140.

Bowes

Ihlanfeldt

(2001) Identifying the impacts of rail transit stations on residential property values. Journal of Urban Economics 50(1): 1–25.

Brandt

Maennig

(2012) The impact of rail access on condominium prices in Hamburg. Transportation 39(5): 997–1017.

Budde

Eilers

(2014) Sozioökonomische Daten auf Rasterebene – Datenbeschreibung der microm-Rasterdaten (Socioeconomic grid data – Description of microm grid data). RWI Materialien 77. Essen: RWI.

10.

Dollery

Worthington

(1996) The empirical analysis of fiscal illusion. Journal of Economic Surveys 10(3): 261–297.

11.

Downes

Zabel

(2002) The impact of school characteristics on house prices: Chicago 1987–1991. Journal of Urban Economics 52(1): 1–25.

12.

Eichenberger

Stadelmann

(2010) How federalism protects future generations from today’s public debts. Review of Law & Economics 6(3): 395–420.

13.

Ezekiel

(1938) The cobweb theorem. The Quarterly Journal of Economics 52(2): 255–280.

14.

Faller

Helbach

Vater

, et al. (2009) Möglichkeiten zur Bildung eines Regionalindex Wohnkosten unter Verwendung von Angebotsdaten.Technical Report. Berlin: German Council for Social and Economic Data (RatSWD).

15.

Georgi

Barkow

(2010) Wohnimmobilien-Indizes: Vergleich Deutschland – Großbritannien (Residential real estate indices: A comparison Germany – Great Britain). Project Report. Available at: http://www.zia-deutschland.de/fileadmin/Redaktion/Meta_Service/PDF/ZIA_Broschuere_WohnimmobilienIndices.pdf (accessed 18 April 2017).

16.

Gibbons

Machin

(2003) Valuing English primary schools. Journal of Urban Economics 53(2): 197–219.

17.

Hilber

CAL

(2017) The economic implications of house price capitalization: A synthesis. Real Estate Economics 45(2): 301–339.

18.

Kholodilin

Mense

Michelsen

(2017) The market value of energy efficiency in buildings and the mode of tenure. Urban Studies 54: 3218–3238.

19.

Laakso

(1992) Public transport investment and residential property values in Helsinki. Scandinavian Housing and Planning Research 9(4): 217–229.

20.

MacKay

(2014) Implicit debt capitalization in local housing prices: An example of unfunded pension liabilities. National Tax Journal 67(1): 77–112.

21.

Muth

(1961) Rational expectations and the theory of price movements. Econometrica 29(3): 315–335.

22.

Rapach

Strauss

(2007) Forecasting real housing price growth in the eighth district states. Federal Reserve Bank of St. Louis. Regional Economic Development 3(2): 33–42.

23.

Rapach

Strauss

(2009) Differences in housing price forecastability across US states. International Journal of Forecasting 25(2): 351–372.

24.

Stock

Watson

(2003) How did leading indicator forecasts perform during the 2001 recession? Economic Quarterly – Federal Reserve Bank of Richmond 89(3): 71–90.

25.

Stock

Watson

(2004) Combination forecasts of output growth in a seven-country data set. Journal of Forecasting 23(6): 405–430.

26.

Tiebout

(1956) A pure theory of local expenditures. Journal of Political Economy 64(5): 416–424.