ACCENTRO-IW Housing Cost Report

Transcription

1 ACCENTRO-IW Housing Cost Report Is the German Housing Market Threatening to Overheat? An Analysis of Rents and User Costs in 402 Counties Expert Opinion for ACCENTRO AG Jacopo Mingazzini CEO Uhlandstrasse 165 D Berlin Contact: Prof. Dr. Michael Voigtländer Björn Seipelt Cologne, 18 November 2016

2 Contact Details of the Editors Prof. Dr. Michael Voigtländer Phone: +49 (0) Fax: +49 (0) Björn Seipelt Phone: +49 (0) Fax: +49 (0) IW German Economy Institute, Cologne PO Box D Cologne Short Title Expert Opinion for ACCENTRO AG Page 2 of 35

3 Table of Contents Summary Background Data and Methodology Findings for Germany and Selected German Counties Findings for Germany Findings for German Metropolises and Selected German Counties Sensitivity Analyses Implications of the Findings for the Formation of Homeownership On the Interpretation of User Costs Alternative Approach to Determine the Sustainability of Financing Arrangements Search Behaviour and Speculative Bubble Evaluation of Google Data Evaluation of Real Estate Portal Data Conclusions Works Cited List of Figures Annex Expert Opinion for ACCENTRO AG Page 3 of 35

4 Summary The development of Germany's real estate markets in recent years has been paced by sometimes brisk price dynamics. A moderate supply growth coincides with increases in demand that have been quite drastic in some places most notably in major cities and campus towns. The widening gap between supply and demand, and the parallel increase in property rents and prices, have triggered a debate whether or not the housing market runs the risk of forming speculative bubbles. Overly optimistic expectations in the future price trend and with them a corresponding pricing divergence from fundamental property market factors can cause a market to overheat and thus intensify the looming threat of price corrections. This market anomaly, which can also be subsumed under the umbrella term of speculative bubbles, is the subject of this survey. Its purpose is to examine, on the basis of the user cost approach (to owner-occupied housing), whether the German housing market or its regional sub-markets show signs of overheating. On top of that, it seeks to derive the implications of increasingly affordable financing conditions for the formation of homeownership. These are subsequently compared with empirical evidence on the demand situation regarding homeownership in order to determine to what extent the improved financing conditions explain the behaviour on the demand side. To probe the significance of the findings and to derive potential reactions to a market environment of rising interest rates, the findings are subjected to in-depth sensitivity analyses. The analysis of the owner-occupied housing user costs or short user costs in juxtaposition to current rental costs suggests that living in an owner-occupied home is, broadly speaking, far more affordable than renting your home. When looking at Germany as a whole, the survey findings suggest that buying a home is roughly 41 percent more affordable on average than renting. The economic benefits of homeownership in German metropolises range from 34 percent in Munich to 47 percent in Hamburg. Within the framework of the regionalised analysis, a drilldown to the county level shows that the benefits can be anywhere between 13 and 67 percent. In addition, the analysis findings suggest that the benefits of homeownership would be sustained in the overwhelming majority of counties even if the interest load was to increase. In summary, it is safe to say that the way the costs of either housing format relate to each other does not suggest that the housing market is overheating. The economic benefits of owning your home actually indicate that home prices could continue to rise going forward. Another indicator that helps to identify speculative bubbles is the relative propensity to buy a home. In the wake of an overheating residential property market, optimistic price expectations and the imputed possibility to generate capital gains within a short period of time can inspire an increased propensity to buy. An initial screening of the empirical data underlying this survey, however, produced no evidence of such behaviour. The propensity to buy among demanders barely increased in recent years, and is currently flatlining. The favourable parameters would have made an upward trend likely, but regulations such as the Mortgage Credit Directive (WIKR) that hamper the acquisition of property have reduced the appeal of homeownership among prospective home buyers even further. All things considered, there are thus no signs suggesting that Germany's housing market is overheated. Expert Opinion for ACCENTRO AG Page 4 of 35

5 1 Background Prices for condominiums and houses in Germany have surged since Especially major cities and certain mid-size towns have experienced price hikes that substantially exceed available empirical evidence. The reason for the price hikes is the insufficient building activity in combination with a strong demand for housing that is driven by high incoming migration, among other reasons (Deschmeier et al., 2016). Berlin alone has grown by nearly 115,000 residents over the past three years, but fewer than 30,000 flats were completed. Between now and 2020, the annual completions rate in Berlin would have to be 25,000 flats to meet demand. In Germany as a whole, at least 310,000 flats would have to be raised per year, but so far the annual rate has been roughly 250,000 flats. Despite the conspicuous housing shortage, however, the formation of a speculative bubble is not impossible. A speculative bubble means that a runaway price trend calls for correction that is, a drastic drop in prices. Speculative bubbles of this kind can form as a result of overly optimistic expectations concerning future price upsides, for instance (cf. Shiller, 2003). Experience from the financial crisis moreover suggests that particularly the combination of a robust economic development and a low-interest cycle can cause the housing market to overheat (Reinhart/Rogoff, 2009). This explains why the Deutsche Bundesbank has been warning for some time against the possibility of an adverse development on Germany's housing market. Stiglitz (1990) defined a speculative bubble as follows: If the reason that the price is high today is only because investors believe that the selling price will be high tomorrow when fundamental factors do not seem to justify such a price then a bubble exists. The definition implies an approach on how to identify speculative bubbles. The owner-occupier housing costs approach includes all the fundamental factors governing the real estate market, including particularly rents and interest rates. If price trend, interest rate development and rental growth cease to evolve in sync but create an imbalanced situation that is essentially resolved by a markdown in property prices, the occurrence of a speculative bubble becomes likely. The owner-occupier housing costs approach will be discussed in detail below along with the significance of the survey findings for Germany in general and the counties in particular. On top of that, we will also take a closer look at the search behaviour of private households. During a speculative bubble, the focus of web search queries shifts clearly toward the buying side because the property occupancy takes the back seat to reselling for the sake of quick capital gains. Data available from the ImmobilienScout24 real estate portal and the Google search engine are well suited for analysing the search behaviour of households, as a prior survey by Voigtländer, Pomogajko and Henger (2012) demonstrated. The survey wraps up with several conclusions. 2 Data and Methodology The basic idea underlying the user cost approach to owner-occupied housing is to make the costs of tenants and owner-occupiers comparable. Tenants, of course, pay a monthly amount (rent) whereas owners pay a one-off amount (purchase price). But the purchase price may be converted into a monthly or yearly annuity that, rather than the purchase price proper, takes into account the financing costs and the loss in returns on the equity employed. Additional costs are generated by real estate transfer taxation, maintenance and depreciation. On top of that, value changes may come into play, because any real property can appreciate or depreciate, as the case may be. Expert Opinion for ACCENTRO AG Page 5 of 35

6 Once you convert the purchase price into a recurrent payment of this sort, the respective costs of rented versus owner-occupied housing become comparable. Since households do compare these cost types with each other and respond to imbalances with changes in demand, rental costs and owner-occupier housing costs should be on a level in the long run. At least this is the basic logic underlying the user cost concept that goes back to Poterba (1984). If the costs of owneroccupied housing substantially exceed the rental costs, it would make sense to expect a correction, which would most likely take the form of falling prices. Whenever user costs far exceed rents, it can usually be blamed on an overly optimistic outlook regarding future price trends, and this is a situation necessitating a price correction. If, by contrast, the user costs undercut the rental costs, further price growth is to be expected. The discussion below will use the following model of owner-occupied housing user costs. Tenant costs as used here represent the annual rental costs, whereas the costs of owner-occupied housing are determined as follows: SNK ti = (1 + g ti ) P ti (b i F,t + [1 b t ] i A,t [1 t t ] + s + a P t t 4 ) The annual costs of owner-occupancy during period t (SNK ti ) are initially defined by the purchase price P ti and the real estate transfer tax g ti during t. This purchase price is financed with a debt capital b share and with an equity [1 b ] share, with interest at a certain interest rate i F,t paid on the debt and interest at an assumed alternative interest rate i A,t lost on the equity. Since these investments would be taxable, an average income tax rate of t t must also be taken into account. s represents the repair costs as part of the purchase price whereas a represents the asset erosion as part of the purchase price. By contrast, delta P t t 4 equals the capital growth of the property within the prior-year period, based essentially on increased land prices. The rent and price data was sourced from F+B, a market analyst that gathers asking data from internet portals and applies a hedonic method to calculate quality-adjusted square-metre prices. The interest rate data were obtained from Deutsche Bundesbank. The borrowing costs represent the average interest rate of a housing loan with a fixed-interest period between five and ten years, while the interest of a domestic corporate bond was chosen to represent the alternative interest rate. The average tax rate is based on data released by the Federal Ministry of Finance while the interest rate on debt capital was taken from the database of Dr. Klein, with an average established for simplicity sake. For the repair costs and the real-term asset erosion, a flat rate of three percent overall was assumed, as derived by Clamor (2013) among others. The average price increase reflects the mean annual rise in prices between 2005 and During this time period, prices largely manifested a moderate upward trend, so that the housing market may be considered to have been in equilibrium. Accordingly, the annual price increases during this period can safely be considered sustainable. Most alternative surveys (Himmelberg et al., 2005) consider the price increase of the previous year as proxy for future price increases. But doing so could distort results in the event that the expectations are actually too optimistic. So the way the owner-occupancy costs are determined can be said to be conservative. 3 Findings for Germany and Selected German Counties The section below discusses the rents and owner-occupancy costs for Germany as a whole and for selected counties, including without being limited to the main German metropolises. It will also Expert Opinion for ACCENTRO AG Page 6 of 35

7 undertake sensitivity analyses to find out how great the potential for corrections might be once the interest rates are lifted. 3.1 Findings for Germany For Germany as a whole, it is safe to say that rents are steadily pushing up, and have been gathering momentum lately. According to the F+B data, the average net rent for existing flats in Germany was just below 7.00 euros/sqm in the third quarter of This means that rents have increased by around 11 percent since Q The costs of owner-occupied housing have been much more volatile by contrast. As Figure 1 shows, owner-occupancy costs persistently exceeded rents until 2010, but, began to be more affordable than renting as of Q To be sure, owneroccupancy costs briefly perked up twice in the years since, but the general tendency shows a widening gap between the costs of renting and the costs of buying. At the moment, buying is 41 percent more affordable than renting in Germany. Figure 1: Comparing rents and owner-occupancy costs in Germany (costs per sqm) 8,50 8,00 7,50 7,00 6,50 6,00 5,50 5,00 4,50 4,00 Owner-occupancy costs Rent Source: IW Köln, F+B A key driver of this development is mortgage interest. As recently as 2011, the interest rate for long-term loans was above 4 percent, whereas in 2016, the average interest rate was 1.6 percent. This interest rate development more than offset the price trend, making buying clearly more affordable than renting today. Accordingly, the risk that the German market could overheat is negligible. The market would only overheat if owner-occupancy costs were higher than rent rates and created a potential for adjustment. In truth, however, owner-occupancy costs are far below rents, and this suggests that prices will continue to go up. Expert Opinion for ACCENTRO AG Page 7 of 35

8 3.2 Findings for German Metropolises and Selected German Counties Taking a closer look at Germany's metropolises will confirm this picture. Rent rates in the socalled Big Seven (the seven Class A cities) cities are well ahead of owner-occupancy costs, as Figure 2 illustrates. Figure 2: Owner-occupancy costs and rents Class A cities Berlin Hamburg Munich Cologne Frankfurt Stuttgart Düsseldorf Owner-occupancy costs Rent Source: IW Köln, F+B In fact, the difference between owner-occupancy costs and rents in some of these cities is greater than the national average. In Hamburg and Berlin, it is more than 45 percent. The advantage of owner-occupancy is lowest in Munich and Stuttgart at 34 and 35 percent, respectively. One reason for the wide gap is the anticipated price growth, which also exceeded the national average between 2005 and The situation is compounded by the generally brisk rental growth. When combining these factors, the drop in interest rates not only makes up for the price hikes, but overcompensates for them, and makes homeownership even more attractive than it is nationwide. For a drilldown of the results on the county level, please see the Annex. But even a look at the overview map in Figure 3 shows, homeownership is more advantageous than buying in any county at the moment. The benefits can be quantified in a spectrum ranging from 16 percent in Kelkheim County to 67 percent in Wittenberg County. Expert Opinion for ACCENTRO AG Page 8 of 35

9 Figure 3: Overview map of the advantageousness of homeownership over renting Source: IW Köln Expert Opinion for ACCENTRO AG Page 9 of 35

10 With a view to this constellation, a sudden drop in prices seems relatively improbable. On the contrary, the drastic differences suggest that there is reason to expect further price hikes in the near future. The fact that the divergence between buying and renting is particularly great in economically undeveloped areas should not be read as a sign to invest, though. Rather, these areas are the ones where the long-term price expectations, being based on the evidence of the period 2006 through 2010, seem inflated. Conversely, the price expectations imputed for the major cities and many other fast-growing locations could well be too modest and the benefits generally be greater than assumed. 3.3 Sensitivity Analyses That being said, the advantageousness could reverse itself if the parameters for buying residential property were to change. To find out just how robust the present market situation is, the parameters are carefully reviewed below. The first step will assume that prices for residential property will not increase going forward. As elaborated above, the price upside is actually quite high, and nowhere higher than in the major cities. Figure 4 shows, however, that homeownership is more attractive even in that case. Although the benefits would admittedly diminish, they are positively there, ranging from 36 percent in Cologne to 6 percent in Munich. The calculation should be qualified as extremely conservative, for the assumption that condominium and property prices in a major city will see no long-term growth is rather far-fetched. This is particularly true for Munich: Figure 4: Owner-occupancy costs, without anticipated price hikes, and rents in the Class A cities Berlin Hamburg Munich Cologne Frankfurt Stuttgart Düsseldorf Owner-occupancy costs Rent Source: IW Köln, F+B In addition to the imputed price expectations, the interest rate development in particular is being watched with great concern. If interest rates were to go up, it would create potential for setbacks on the German market. Analyses by the Bundesbank, for one, suggest as much. The first step in determining just how robust the German market and the local markets are is to define a so-called Expert Opinion for ACCENTRO AG Page 10 of 35

11 neutral interest rate on loans. The latter represents the interest rate at which rental costs and owner-occupancy costs are on a level. If the interest rate were to rise any higher, it could trigger an adverse price reaction. Figure 5 shows Germany's counties sorted by the neutral level of interest rates. There are 8 counties where a rise in interest from currently 1.48 percent to a hypothetical rate of 2.5 percent would shift the advantage from buying to renting. In another 35 counties, the advantageousness would reserve itself at an interest rate of 3 percent, but everywhere else, interest rates would have to more than double to restore the attractiveness of renting. Such a sudden and abrupt shift in interest rates is not to be expected. More likely is a slow growth in interest rates and a combination of continued rental growth and a lateral movement of prices that would keep the equilibrium intact. Figure 5: Neutral interest rate on the county level 8% 7% 6% 5% 4% 3% 2% 1% 0% Source: IW Köln Striking to note is the high number of counties in Bavaria where the critical interest rate level is very low. All of the 8 counties that show a critical interest rate of less than 2.5 percent are located in the metro region of Munich, the Bavarian state capital. This could indicate a potential for correction. Then again, it needs to be remembered that the demographic growth in these counties is particularly strong. So the assumed capital growth, which is based on the period 2005 through 2010, could actually be too modest. In spite of all this, the figures demonstrate that prices in many Bavarian counties are already rather high compared to the fundamental values, and that other regions might offer greater opportunities for price growth. Overall, it is safe to state that the formation of speculative bubbles is very improbable at this time. It is much more likely that many counties will see further price hikes, among them the metropolises of Berlin, Hamburg and Cologne. Expert Opinion for ACCENTRO AG Page 11 of 35

12 4 Implications of the Findings for the Formation of Homeownership The findings on owner-occupied housing user costs are primarily meaningful for answering the question of a speculative bubble. The findings contain no evidence of overheating markets, but suggest instead that further price hikes are likely. But there is yet another implication to be derived from the findings: Buying is currently much more affordable than renting. The low-interest cycle offers a splendid window of opportunity for the formation of homeownership. But it is precisely this diagnosis that is often questioned. One of the caveats voiced, for instance, is that the observation fails to take the payments of the principal into account, and that the debt load depends not so much on the financing arrangement but on the acquisition value. In response, the interpretation of owner-occupied housing user costs shall be more narrowly defined below. On top of that, an alternative approach shall be applied for the purpose of qualifying the refinancing risks. 4.1 On the Interpretation of User Costs The objective underlying the user cost approach to owner-occupied housing is to make the costs of tenants and owner-occupiers comparable. The acquisition of a property is principally a one-off transaction, whereas paying rent is a recurrent encumbrance. But taking the financing into account makes it possible to identify a monthly encumbrance for owner-occupiers as well. The capital invested in the property earns no interest on the capital market (this being the opportunity costs of the equity) while interest payments are made on the debt capital. But to factor in the repayment of the principal would mean bringing in an alien component into the user cost approach. The repayment is associated with a continuous asset accumulation that is without analogue on the tenant side. If the loan repayments were taken into account, it would distort the picture because in old age, for instance, a tenant does not benefit from the added relief of a debtfree property. A more apt comparison to tenancy would be a situation in which a property is acquired but the debt not repaid. This would mean that the property, while being used, does not contribute to the asset accumulation. Even in this scenario, though, an owner would be able to turn a profit, e.g. if the land appreciated. Such calculation models are common in some countries, but are virtually unknown in Germany. It is also of the essence for the user cost approach that the data relate to a key date. The comparison always references a certain point in time, and cannot simply be carried forward. Especially with variable interest rates, the burden on the owner can change quickly. The standard in Germany, however, is a fixed-interest period of ten years, which means that the user cost approach remains a constant at least for this decisive cost factor. So the user cost approach is quite suitable for comparing renting and buying. But many households do their own math, and factor in their payments above all without realising that homeownership is also a form of pension scheme that will substantially lower their housing costs in retirement. To do justice to this way of looking at things, an alternative approach shall be applied below that takes the risks of follow-up financing into account in addition to the amortisation. Expert Opinion for ACCENTRO AG Page 12 of 35

13 4.2 Alternative Approach to Determine the Sustainability of Financing Arrangements The concern of many is that while it may be very well to opt for homeownership now they won't be able to make their payments once the interest rates tighten. To address this concern, we shall do the following calculation: Let us assume that a household acquires a property at the present time. The recurrent financial burden from interest payments, repayments of the principal, and repair costs is supposed to match the rent level of a comparable flat. Neither lost income on equity nor changes in the property's value are taken into account. After ten years, a follow-up financing arrangement is taken out. The idea is to pay off the property within the next 25 years. The owner's annual debt load after ten years is again not supposed to exceed a comparable rent. The price growth of the rental evidence is assumed to equal 1.5 percent for half of the ten or 25 years each, and is retrospectively averaged for this period. Next, the maximum interest rate that would still make the financing arrangement work is identified. The basic idea is the following: During the remainder of the acquisition phase (35 years), the household will bear the same costs as a tenant household, and will earmark the benefits it reaps according to the user cost approach toward the repayment of the loan. In return, such a household will have a debt-free home by the end of the financing phase, and lower costs as a result. If you leave out the committed equity of 20 percent, the asset accumulation under this model will not expose a home buyer to a greater burden than a tenant. The interest level which should not be exceeded in ten years so as to make this financing model work will hereinafter be called neutral interest rate on loans. The model calculation below uses the fictional acquisition of a 100-square-metre condominium in the city of Cologne to illustrate the point. In a first step, the annual financial burden associable with the acquisition of a condominium is calculated and matched with the rental costs of a comparable unit. In the example at hand, the annual encumbrance through the property acquisition, which divides into interest payments and maintenance costs, add up to a total of 6,124 euros. The far greater costs of renting a flat of comparable size in the amount of 11,871 euros result in a cost difference of 5,747 euros between the two types of use. This difference in amount could be used by the condominium buyer to repay the loan taken out. Assuming a loan-to-value ratio of 78.3 percent, the borrowed amount that is needed to finance the above-described property in Cologne would add up to a total of 218,873 euros. Assuming a repayment in an amount equal to the difference between both types of use as calculated above, the owner-occupier could shoulder a maximum repayment rate of 2.63 percent to keep the costs of buying on a level with the costs of renting. According to this repayment rate, 26.3 percent of the loan will already be repaid after a ten-year period, bringing the debt total down to 161,405 euros. In order to repay the remaining loan amount over the period of another 25 years, the annual repayment rate should go up to 4 percent, which would imply an annual repayment total of 6,456 euros. Assuming a rental growth of 1.5 percent per year during the first half of the 25 years, the tenant would be facing costs in an average amount of 14,586 euros. To keep the sum of interest payments and repayments of the principal on level with the rental costs so that buying remains Expert Opinion for ACCENTRO AG Page 13 of 35

14 as affordable as renting, the loan interest in this example must not rise above 5.04 percent. A loan interest rate of 5.04 percent would raise the nominal interest load up to 8,130 euros, which together with the repayments of the principal would in turn add up to 14,586 euros, that is, the same as the annual rental costs. The outcome of this calculation is quite surprising, for it suggests that interest rates could go up by a considerable margin in many German counties without putting the buyer at a disadvantage over the tenant. As Figure 6 illustrates, the neutral interest rate on loans exceeds 4 percent in 325 counties, and actually tops 5 percent in 280 counties. In many instances, an interest level of 4 percent is deemed normal (Schier/Voigtländer, 2015). Figure 6: Histogram of neutral interest rate on loans Source: IW Köln Even in the major cities, the neutral interest rate on loans is often at least 3 percent, as shown in Figure 7. Munich is the only German city where the interest rate is very low. But even this does not imply that buying property in Munich is unattractive. But unlike in many other cities, a household will indeed have to pay more than a tenant in order to be a debt-free home owner in old age. Then again, it should be taken into account that the rate of appreciation is steeper in Munich than elsewhere, and that the wealth built up is also bound to be higher here than in many other regions. Expert Opinion for ACCENTRO AG Page 14 of 35

15 Figure 7: Neutral interest rate on loans in Class A cities 0,0% 1,0% 2,0% 3,0% 4,0% 5,0% 6,0% Berlin 3,7% Hamburg 2,6% Munich 0,8% Cologne 5,0% Frankfurt Stuttgart 3,7% 3,5% Düsseldorf 4,2% Source: IW Köln The calculation of the neutral interest rate on loans underscores the great opportunity for the formation of homeownership that the present funding environment offers. It puts households in most counties in a position to buy residential property and to repay their debt without shouldering a heavier burden than they would by renting. Only if interest rates were to see an extreme increase in ten years' time would owner occupiers pay more than tenants, assuming the repayment of the principal is taken into account. From the present perspective, the global increase in savings, the declining propensity to invest, and the increasingly moderate inflation trend make it safe to say that a major hike in interest rates is not to be expected, not even in the 2020s (Summers, 2014). 5 Search Behaviour and Speculative Bubble A speculative bubble not only points to a pricing deviation from the fundamental values, but is also a psychological phenomenon. The expectation of continued price hikes can become the main motive for market players to step up their property acquisitions. As a result, transaction activities start speeding up. The swelling ranks of home buyers are complemented by a growing number of people just seeking to turn a profit. But the currency of transaction data in Germany is subject to a serious time lag. Most helpful for monitoring the latest price performance is actually the data available on the internet. Google search queries or real estate portal searches can be quite useful in tracking changes in behaviour. Particularly revealing are comparisons of looking-to-buy with looking-to-rent requests. A speculative bubble would typically be reflected in a surge in lookingto-buy queries as the length of holding periods declines and the number of households wishing to exploit the price trend rises. The trend in search queries is shown and rated below, based on data both from Google and the ImmobilienScout24 real estate portal. Expert Opinion for ACCENTRO AG Page 15 of 35

16 5.1 Evaluation of Google Data In the United States, Google data proved useful in tracing the real estate crisis, at least retrospectively (Hohenstatt et al., 2011). Google saves all of its web search queries and makes (anonymised) evaluations of them available under Google Insights for Search. Google data have been used for a variety of issues (e.g. on influenza cases) other than real estate (Ginsberg et al., 2009). For certain search queries, Google makes indexed data with a maximum score of 100 available. In this case, we had Google evaluate the frequency of the following queries (i.e. their German equivalents): flat to sell, flat to rent and condominium. Figure 8 shows the rolling weekly average for the aforementioned search queries. Figure 8: Google search queries for Germany Source: IW Köln, ImmobilienScout24 Condominium Flat to let Flat to sell All things considered, the search term Flat to sell (German: Wohnung kaufen ) was more frequently used, and the curve has essential been on the rise since Simultaneously and almost in sync, the number of queries for Flat to rent (German: Wohnung mieten ) went up. In a market characterised by strongly speculative demand for housing, you would expect the gap between web search queries to widen, and a surge in looking-to-buy requests to coincide with a drop in looking-to-rent requests. The picture that the actual data present, however, points to a generally increased occupier demand. The stable score for the term condominium (German: Wohneigentum ) does not suggest any change in behaviour or excessive optimism. In fact, it is rather odd to note that the favourable parameters have not triggered a serious increase in web search queries concerning condominiums. In addition to queries for Germany, the evaluation also covered queries for the country's three largest cities: Munich, Hamburg and Berlin. These cities display a pattern quite similar to that of Germany as a whole, with looking-to-buy and looking-to-rent search queries rising in tandem, whereas queries for condominiums remain relatively stable. The findings for Berlin, shown in Figure 9, illustrate the point. Expert Opinion for ACCENTRO AG Page 16 of 35

17 Figure 9: Google search queries for Berlin Condominium Flat to let Flat to sell Source: IW Köln, ImmobilienScout Evaluation of Real Estate Portal Data To analyse the search behaviour of private households it is helpful to evaluate, in addition to the queries run on Google and other search engines, the data of real estate portals, too. Germany's leading platform for private property transactions is ImmobilienScout24, covering both buyers and tenants. The portal's own data mining captures not only the property listings but also the search queries. These data were made available to the IW Economic Institute in Cologne for the purpose of this survey. Their analysis reveals whether or not the relation between looking-to-rent and looking-to-buy requests has shifted. In general, it is safe to say that the absolute demand for real estate has increased. However, the fact is to some extent explained by the platform's own expansion, too. But looking at the relationship of looking-to-buy requests to the total number of queries will still reveal a possible shift. As Figure 10 shows, the number of queries for property to sell has actually declined lately. In analogy to the declining owner-occupied housing user costs, available-for-sale property initially experienced a steady growth in demand between 2011 and 2015, but in 2016 the trend reversed itself. This is possibly explained by recent credit constraints arising from the new Mortgage Credit Directive (cf. Voigtländer/Hude, 2016), among other reasons, yet the change is not reflected by the Google data. It should be added that the stats for Germany as a whole are made available on a quarterly basis, whereas the drilldowns to the county level are provided on an annual basis only. Although the pro-rata share of looking-to-buy requests has increased in Germany's three largest cities, the increase is by no means significant, as Figure 11 shows. Especially in Munich, which is the market that seems most likely to overheat, the number of looking-to-buy requests is actually stagnant. A Expert Opinion for ACCENTRO AG Page 17 of 35

18 speculative bubble would manifest itself in a steep rise in search queries as demand is gathering momentum in general and the transaction volume expands as a result of shorter holding periods. There are no signs of this at the moment. Figure 10: The share in web search queries for property acquisitions in relation to all web search queries in Germany 21% 20% 19% 18% 17% 16% 15% Source: IW Köln, ImmobilienScout24 Figure 11: The share in web search queries for property acquisitions in relation to all web search queries in the three largest cities 35% 30% 25% 20% 15% 10% 5% 0% Source: IW Köln, ImmobilienScout24 Berlin Hamburg Munich Expert Opinion for ACCENTRO AG Page 18 of 35

19 6 Conclusions Condominium prices have grown at a brisk pace in Germany, but there is no risk of overheating markets at this time. Indeed, the survey findings based on the user cost approach show that there is a demonstrable price upside. This is true even for the major cities. Even if interest rates were to rise abruptly, the adjustments to be expected would be moderate, or so the sensitivity analyses suggest. Analogously, the evaluation of web search queries reveal no evidence for overvaluation. Striking to note is that in spite of the favourable funding terms there are no signs of an increased propensity to buy. One main reason for this could be the extremely tight buyer requirements. In addition to an equity stake of 10 to 20 percent expected by the bank, there are incidental acquisition costs and removal expenses to factor in, so that property buyers in Germany need to have about 30 percent of the purchase price on hand in capital (Voigtländer, 2016). Assuming an average acquisition value of 250,000 euros, this would equal approximately 75,000 euros. With the Mortgage Credit Directive having become effective in March 2016 and macro-prudential instruments possibly introduced in the future, there is a good chance that the requirements might actually tighten further. So while the window of opportunity for homeownership is particularly attractive at the moment, these hurdles pose a serious challenge for many households. According to the calculations, owners will have a debt-free property in their retirement age after an acquisition phase whose financial burden is no greater than that of a tenant. Only in the event of an extreme interest rate development would home buyers have to shoulder a greater burden, but in many scenarios they would pay less than comparable tenants even during the acquisition phase. The big difference between owner-occupancy costs and rents makes it reasonable to assume that property prices will keep going up in Germany, and the continued demographic growth especially in the metro regions suggests the same. Yet while there are no signs of overheating today, the market should be continuously monitored because the chance of sudden changes in behaviour is always there. This is why ACCENTRO AG has asked the IW German Economic Institute in Cologne to check the market every six months from now on, using the user cost approach. Expert Opinion for ACCENTRO AG Page 19 of 35

20 Works Cited Clamor, Tim / Brügelmann, Ralph / Voigtländer, Michael, 2013, Abschreibungsbedingungen für den Mietwohnungsneubau, in: IW-Trends, vol. 40, no. 2, pp Deschermeier, Philipp / Henger, Ralph / Seipelt, Björn / Voigtländer, Michael, 2016, Zuwanderung, Wohnungsnachfrage und Baubedarfe. Aktualisierte Ergebnisse des IW Wohnungsbedarfsmodells, in: IW-Report, no. 18/2016, Institut der deutschen Wirtschaft Köln, Cologne Ginsberg, Jeremy, et al. 2009, Detecting Influenza Epidemics Using Search Engine Query Data, in: Nature, vol. 457, no. 7232, pp Henger, Ralph / Pomogajko, Kirill / Voigtländer, Michael, 2012, Gibt es eine spekulative Blase am deutschen Wohnimmobilienmarkt?, in: IW-Trends, vol. 39, no. 3 Himmelberg, Charles / Mayer, Christopher / Sinai, Todd 2005, Assessing High House Prices. Bubbles, Fundamentals and Misperceptions, in: Journal of Economic Perspectives, vol. 19, no. 4, pp Hohenstatt, Ralf / Käsbauer, Manuel / Schäfers, Wolfgang, 2011, Geco and its Potential for Real Estate Research: Evidence from the US Housing Market, in: Journal of Real Estate Research, vol. 33, no. 4 Poterba, James M., 1984, Tax Subsidies to Owner-Occupied Housing: An Asset-Market Approach, in: The Quarterly Journal of Economics, vol. 99, no. 4, pp Reinhart, Carmen M. / Rogoff, Kenneth S., 2009, This Time Is Different. Eight Centuries of Financial Folly, Princeton Schier, Michael / Voigtländer, Michael, 2015, Immobilienpreise. Ist die Entwicklung am deutschen Wohnungsmarkt noch fundamental gerechtfertigt?, in: IW Trends, vol. 42, no. 1, pp Shiller, Robert J., 2003, From Efficient Markets Theory to Behavioral Finance, in: Journal of Economic Perspectives, vol. 17, no. 1, pp Stiglitz, Joseph E., 1990, Symposium on Bubbles, in: The Journal of Economic Perspectives, vol. 4, no. 2, pp Summers, Laurence H., 2014, Reflections on the New Secular Stagnation Hypothesis, in: CEPR (ed.), Secular Stagnation: Facts, Causes and Cures, London, pp Voigtländer, Michael, 2016, Transaction Costs for Housing Purchases. A High Financial Burden for German Home Buyers, IW Short Report, no. 72, Institut der deutschen Wirtschaft Köln, Cologne Voigtländer, Michael / Hude, Marcel, 2016, Trends in der Wohneigentumsbildung. Gutachten für die Schwäbisch Hall Stiftung bauen-leben-wohnen, IW expert opinion, Cologne Expert Opinion for ACCENTRO AG Page 20 of 35

21 List of Figures Figure 1: Comparing rents and owner-occupancy costs in Germany (costs per sqm)... 7 Figure 2: Owner-occupancy costs and rents Class A cities... 8 Figure 3: Overview map of the advantageousness of homeownership over renting... 9 Figure 4: Owner-occupancy costs, without anticipated price hikes, and rents in the Class A cities Figure 5: Neutral interest rate on the county level Figure 6: Histogram of neutral interest rate on loans Figure 7: Neutral interest rate on loans in Class A cities Figure 8: Google search queries for Germany Figure 9: Google search queries for Berlin Figure 10: The share in web search queries for property acquisitions in relation to all web search queries in Germany Figure 11: The share in web search queries for property acquisitions in relation to all web search queries in the three largest cities Expert Opinion for ACCENTRO AG Page 21 of 35

22 Annex County Code Area Schleswig-Holstein User costs in Q3/2016, in * Cost benefit over renting Flensburg % Kiel % Lübeck % Neumünster % Dithmarschen County % County of Herzogtum Lauenburg % Nordfriesland County % Ostholstein County % Pinneberg County % Plön County % Rendsburg-Eckernförde County % Schleswig-Flensburg County % Segeberg County % Steinburg County % Stormarn County % Hamburg % Lower Saxony Braunschweig % Salzgitter % Wolfsburg % Gifhorn County % Göttingen County % Goslar County % Helmstedt County % Northeim County % County of Osterode am Harz % Peine County % Wolfenbüttel County % Expert Opinion for ACCENTRO AG Page 22 of 35

23 03241 Hanover County % Diepholz County % Hameln-Pyrmont County % Hildesheim County % Holzminden County % County of Nienburg (Weser) % Schaumburg County % Celle County % Cuxhaven County % Harburg County % Lüchow-Dannenberg County % Lüneburg County % Osterholz County % County of Rotenburg (Wümme) % Soltau-Fallingbostel County % Stade County % Uelzen County % Verden County % Delmenhorst % Emden % Oldenburg % Osnabrück % Wilhelmshaven % Ammerland County % Aurich County % Cloppenburg County % Emsland County % Friesland County % County of Bentheim % Leer County % Oldenburg County % Osnabrück County % Expert Opinion for ACCENTRO AG Page 23 of 35

24 03460 Vechta County % Wesermarsch County % Wittmund County % Bremen Bremen % Bremerhaven % North Rhine-Westphalia Düsseldorf % Duisburg % Essen % Krefeld % Mönchengladbach % Mülheim an der Ruhr % Oberhausen % Remscheid % Solingen % Wuppertal % Kleve County % Mettmann County % County of Rhein-Kreis Neuss % Viersen County % Wesel County % Bonn % Cologne % Leverkusen % County of Städteregion Aachen % Düren County % Rhine-Erft County % Euskirchen County % Heinsberg County % County of Oberbergischer Kreis % County of Rheinisch-Bergischer Kreis % Expert Opinion for ACCENTRO AG Page 24 of 35

25 05382 Rhine-Sieg County % Bottrop % Gelsenkirchen % Münster % Borken County % Coesfeld County % Recklinghausen County % Steinfurt County % Warendorf County % Bielefeld % Gütersloh County % Herford County % Höxter County % Lippe County % Minden-Lübbecke County % Paderborn County % Bochum % Dortmund % Hagen % Hamm % Herne % Ennepe-Ruhr County % Hochsauerland County % County of Märkischer Kreis % Olpe County % Siegen-Wittgenstein County % Soest County % Unna County % Hesse Darmstadt % Frankfurt am Main % Offenbach am Main % Expert Opinion for ACCENTRO AG Page 25 of 35

26 06414 Wiesbaden % Bergstrasse County % Darmstadt-Dieburg County % Gross-Gerau County % County of Hochtaunus % Main-Kinzig County % Main-Taunus County % Odenwald County % Offenbach County % Rheingau-Taunus County % Wetterau County % Giessen County % Lahn-Dill County % Limburg-Weilburg County % Marburg-Biedenkopf County % Vogelsberg County % Kassel % Fulda County % Hersfeld-Rotenburg County % Kassel County % Schwalm-Eder County % Waldeck-Frankenberg County % Werra-Meissner County % Rhineland-Palatinate Koblenz % Ahrweiler County % County of Altenkirchen (Westerwald) % Bad Kreuznach County % Birkenfeld County % Cochem-Zell County % Mayen-Koblenz County % Neuwied County % Expert Opinion for ACCENTRO AG Page 26 of 35

27 07140 Rhein-Hunsrück County % Rhein-Lahn County % Westerwald County % Trier % Bernkastel-Wittlich County % County of Eifelkreis Bitburg-Prüm % Vulkaneifel County % Trier-Saarburg County % County of Frankenthal (Pfalz) % Kaiserslautern % Landau in der Pfalz % Ludwigshafen am Rhein % Mainz % Neustadt a. d. Weinstrasse % Pirmasens % Speyer % Worms % Zweibrücken % Alzey-Worms County % Bad Dürkheim County % Donnersberg County % Germersheim County % Kaiserslautern County % Kusel County % County of Südliche Weinstrasse % Rhein-Pfalz County % Mainz-Bingen County % Südwestpfalz County % Baden-Württemberg Stuttgart % Böblingen County % Esslingen County % Expert Opinion for ACCENTRO AG Page 27 of 35

28 08117 Göppingen County % Ludwigsburg County % Rems-Murr County % Heilbronn % Heilbronn County % Hohenlohe County % Schwäbisch Hall County % Main-Tauber County % Heidenheim County % County of Ostalbkreis % Baden Baden % Karlsruhe % Karlsruhe County % Rastatt County % Heidelberg % Mannheim % Neckar-Odenwald County % Rhein-Neckar County % Pforzheim % Calw County % Enz County % Freudenstadt County % Freiburg im Breisgau % County of Breisgau-Hochschwarzwald % Emmendingen County % Ortenau County % Rottweil County % Schwarzwald-Baar County % Tuttlingen County % Constance County % Lörrach County % Waldshut County % Expert Opinion for ACCENTRO AG Page 28 of 35

29 08415 Reutlingen County % Tübingen County % County of Zollernalbkreis % Ulm % Alb-Donau County % Biberach County % County of Bodenseekreis % Ravensburg County % Sigmaringen County % Bavaria Ingolstadt % Munich % Rosenheim % Altötting County % County of Berchtesgadener Land % County of Bad Tölz-Wolfratshausen % Dachau County % Ebersberg County % Eichstätt County % Erding County % Freising County % Fürstenfeldbruck County % Garmisch-Partenkirchen County % County of Landsberg a. Lech % Miesbach County % County of Mühldorf a. Inn % Munich County % Neuburg-Schrobenhausen County % County of Pfaffenhofen a.d. Ilm % Rosenheim County % Starnberg County % Traunstein County % Expert Opinion for ACCENTRO AG Page 29 of 35