Cycles of the Housing Market in Hungary from the Economic Crisis until Today Laszlo Harnos PhD student at the LamfaLussy Sandor Department of Economics, University of Sopron, Hungary easy.harnos@gmail.com Abstract The main aim of this paper is to identify the underlying reasons for the cyclical nature of the Hungarian housing market, in particular the business cycles, the construction, and market participants' expectations. Our research was conducted based on analysis of statistical data and of the housing market indices. As a result, it can be stated that cyclic behaviour of the housing market may be explained primarily with business cycles, but state subsidies and mortgages also affect the variations. Accordingly, the increasing Lending and the high amount of subsidies can generate a price bubble. The supply of second-hand dwellings Looks more flexible compared with that of new ones. However, the expectations of market operators do not have a demonstrable effect on the housing market. Keywords: property market, housing market cycles, asset price bubble Introduction The requirement of a detailed assessment of the global crisis in 2008 also highlighted the importance of the research in the cyclic behaviour of the economy and the individual sectors, including the property market. The property sector is one of the most important sectors of the national economy. The aggregate value of the Hungarian residential properties is estimated to be about half the total estate of households, which is approximately 30-40 trillion Hungarian forints (Bekes, Horvath, & Sapi, 2016). Construction of new buildings provides 2-3% of the GDP, out of which, the production value of new residential properties is only responsible for 0.3-0.8% of the GDP (Table 1). The pro-cyclical behaviour of investments has long been treated as a fact in a macroeconomy. In the case of property investments, the short-term inflexibility of stock can be seen as the main reason for delays. The inflexible market stock of land is due to the limited quantity and the immovable nature of real properties. In the case of buildings, the relatively long time required for planning, authorisation, and construction is the reason why the only way the market can react to the variation of demand on the short term is a price change. The dominance of residential property ownership over rental has always been significant in Hungary, which was also amplified by the large-scale privatisation of state and council-owned properties at the beginning of the 1990s as well as ORIGINAL SCIENTIFIC PAPER RECEIVED: MARCH 2018 REVISED: MAY 2018 ACCEPTED: MAY 2018 DOI: 10.2478/ngoe-2018-0007 UDK: 332.72:338.12(439) JEL: E32, R21, R31 Citation: Harnos, L. (2018). Cycles of the Housing Market in Hungary from the Economic Crisis until Today. Nase gospodarstvo/Our Economy, 64(2), 3-14. DOI: 10.2478/ngoe-2018-0007 NG NASE GOSPODARSTVO OUR ECONOMY Vol. . 64 No. 2 2018 pp .3-14 3 NASE GOSPODARSTVO / OUR ECONOMY Vol. 64 No. 2 / June 2018 Year 2000 2001 2005 2010 2015 2016 Value of construction activities (million HUF) 59.694 124.566 170.266 101.158 96.369 190.479 397.109 471.129 669.614 591.231 684.585 779.628 Value of construction activities (percent of GDP) 0,45 0,81 0,75 0,37 0,28 0,54 2,97 3,06 2,97 2,17 1,99 2,20 GDP at current prices (billion HUF) 13.350,1 15.419,1 22.559,9 27.224,6 34.324,1 35.420,3 Table 1. Value of Construction Activities in Hungary (2000-2016) Residential buildings Buildings total Residential buildings Buildings total Source: KSH (2017a) the state subsidies of residential properties that have been introduced since then. Based on the data of the latest census in 2011, 92.1% of all residential properties were used by the owner or his relative or the person with a right of usufruct. The proportion of rented residential properties was only 6.8%, and uses based on other titles (lodging, service flat) constituted 1.1%. Such an overwhelming proportion of property ownership makes the analysis of market cycles more difficult because these cannot be assessed by the changes in rental costs. When assessing the economic and social effects of Hungarian property support, Hegedus (2006), Kiss and Vadas (2006), and Madi (2008) highlighted that the national property policy in Hungary has always favoured purchasing new properties over used ones, and that, in addition to nonreimbursable aids, mortgages also play a key role in financing. The latest type of subsidy, the family home creation support introduced in 2016, has not changed this either. The main purpose of this paper is to identify the underlying reasons for the cyclical nature of the Hungarian housing market, in particular the business cycles, the constructions, and market participants' expectations. The following key questions were mainly addressed by the research we conducted: • How state grants and housing loans granted influence the general housing market cycles. • Whether the general expectations of market players have any meaningful effect, influencing the Hungarian housing market. • Whether the time period required for new housing construction projects and the seasonality factor alter the cycles seen on the housing market. Our research was conducted based on a statistical analysis of the data published by the Central Statistical Office for the various periods (the mean price per sqm of new and second-hand dwellings, the amount of housing loans and state grants provided, the total number of housing market transactions completed, and the total number of new constructions) as well as of the various housing market indices published by GKI Economic Research Co. The global crisis commenced in 2008 has caused a notable change in the Hungarian housing market, so the period from the commencement of the crisis will be evaluated primarily. In the case of some of the indicators used, statistical data are only available from 2007, so amongst the years before the crisis, the year 2007 was used as the starting year. The data regarding price and value used in the research are provided in Hungarian forints. At the time this manuscript was completed, the official exchange rate of the European Central Bank was EUR 1 = HUF 317.68 The study has four main parts. First, the theoretical bases and research results of the topic will be reviewed. The next part will show the changes of the Hungarian housing market from the political transformation until the economic crisis in 2008. The third part will analyse the behaviour of the market of residential properties over time, from the crisis until today. Finally, the conclusions will be drawn at the end of the study. Factors Affecting the Short- and Long-Term Behavior of the Housing Market The theoretical basis of the cyclic behaviour of the housing market, including the residential property, is the so-called four-quadrant model of DiPasquale and Wheaton (1992) (Figure 1). Quadrant I represents the demand for real estate stock (D), which primarily depends on the economic circumstances, mainly the income and the real estate stock (S). Market equilibrium is achieved at the intersection of the two curves. Due to the short-term inflexibility of stock, the 4 Laszlo Harnos: Cycles of the Housing Market in Hungary from the Economic Crisis until Today Figure 1. Relationship Between the Real Estate Market and Real Estate Development balance rent charge mainly depends on the variations in demand. Quadrant II shows the variations of market prices (value) of real estates, which depend on the money flow expected from the property investment, where the main element is the rent charge (R), and these are also affected by the expected income from the investment (i). The expected income depends on the long-term interest rates as well as on the risks of the investment. If the long-term market interest rate or the risk increases, the slope of the curve also increases. Quadrant III represents new construction. For the construction of new real estate, the market prices of real estate must exceed the construction costs, so the curve does not start from the origin. Finally, quadrant IV shows the variation of stock (e.g., quantity of residential properties), which is the difference of the newly constructed quantity (C) and the abandoned quantity (A). In the case of constant demand, these two quantities must be the same for the market equilibrium. If the income increases, the demand curve (D) moves upward, so the rent charges also increase in the market, accordingly; if the construction costs remain the same, more rentable properties are built, thereby increasing the supply, which ultimately leads to the decrease of rent charges. According to the Friedman's income hypothesis, households spend a constant proportion of their permanent income on housing; thus, according to DiPasquale and Wheaton (1992), if a family lives in a property owned by them, the housing costs, as with the installments of the mortgage and the utility costs, can be treated as rent charges payable by the household. If the income increases, households can afford to pay more on their housing, which also moves the demand curve upward, so the model is also suitable to assess housing markets, where ownership is predominant over rental. When long-term interest rates decrease, property prices increase, so, presuming that all other conditions remain 5 NASE GOSPODARSTVO / OUR ECONOMY Vol. 64 No. 2 / June 2018 unchanged, the stock will also increase. It is similarly clear that any potential increase in construction costs will result in fewer new properties being constructed, which will increase property prices as well as rent charges. When assessing the crisis of Houston's housing market in the 1980s and the prices of residential properties, Smith and Tesarek (1991) proved, using the so-called two-dimensional price index, that the cyclic behaviour of the housing market is also affected by regional factors, e.g., employment conditions of the relevant region and migration. They also established that even a drastic decrease in property prices does not result in an immediate increase of demand, but achieving the market equilibrium again takes several, even three to four years. By the beginning of the 1980s, nearly 200,000 jobs were lost in Houston due to the oil and gas industry crisis; thus, the number of households started decreasing rapidly. This was party due to the migration it started and partly because grown-up children moved away from the parents owing to the decreasing income and the lack of employment prospects, so the establishment of new households was delayed in time. Nevertheless, building residential properties continued until 1983, so the number of empty residential properties doubled by the middle of the decade. Due to over-construction, house prices decreased by an average of 30%, but there were regions where the decrease in value was nearly 40% in the case of detached houses and 60% in the case of flats. However, empty properties were mostly residential properties built a few years earlier, where, owing to the decrease in value, the property no longer provided sufficient security for the mortgage debt. The concentration of empty properties, however, was not stable, but it started to spread to other segments of the housing market because many people moved to a lower category property, while the new demand was mainly aimed at residential properties in good condition built a few years earlier, with prices significantly lower than before. As a result of such market correction, the proportion of empty residential properties effectively reached the pre-crisis level in all segments of the housing market by the end of the 1980s. Wheaton (1999) established that the cyclic behaviour of various real estate market segments might differ in various aspects. Although the main reason of the cyclic behaviour of all sectors is in the economic cycles, there are market segments (offices, stores), where this relationship is not as close as in other segments (residential properties, industrial properties). The shorter the transition period of property development is, and the more inflexible the stock is, the more effect the external factors have on the cyclic behaviour of the market. However, in the case of long transition periods, and more flexible stock, internal cycles are much more likely to develop, so the cycles of the relevant segment will be less aligned to the economic cycles. In the case of residential properties and industrial properties, the time period of property development is around one year, while in the case of offices and shopping centres, this may even take up to four to 10 years. The market stock of residences is much more flexible than the market of offices or stores, but, according to Wheaton (1999), there had not been sufficient information on industrial properties in the United States. The cyclic behaviour of the property market was assessed in Hungary by Horvath et al. (2016). Horvath et al. (2016) described the market of offices based on profit, rental turnover, number of empty properties, and the variations of new completion in time. Their research proved that changes in the profits of investments in the market of offices had the first effects on the number of empty properties and rental turnover (with a delay of three to four months), while the reaction of rent charges and office constructions was delayed significantly more, and it roughly took one year. From the second half of the 1990s, house prices increased rapidly on the American housing market; it stopped in the mid-2000s, and the prices started to decrease. Zhou (2010) was looking for an answer whether this effect is merely random, or the fall of the house prices are determined by market powers, thereby leading the property market toward the long-term equilibrium. During his research, he tested the long-term joint movement of the economic fundamentals (income, employment ratio, interest rates, etc.) and house prices. In many cases, earlier empiric research has not found any evidence on the long-term relationship of these factors and house prices, and Zhou believes that the reason for this is that these assessments presumed a linear relationship between the fundamental variables and house prices. In his research, he investigated the relationship of these fundamental macroeconomic variables and property prices in 10 major cities of the United States, and he could only prove a linear relationship only in one metropolis, while this relationship proved to be nonlinear in six other major cities. With his method, there were only three cities, where the relationship between house prices and the economic variables could not be proven. Neng Lai and Van Order (2017) assessed the long- and short-term relationship of house prices, rent charge indices and interest rates in 45 regions of major American cities. Based on the analysis of quarterly data between 1980 and 2013, the authors concluded that these factors affect house prices on the long-term regardless of the locality. On the short term, however, the market prices of residential properties are also affected by local factors of the cities, so the short-term effects of the assessed variables are not significant. 6 Laszlo Harnos: Cycles of the Housing Market in Hungary from the Economic Crisis until Today Due to the financial crisis started in August 2007, it was also proven that missed mortgage payments had a negative effect on property prices (Ascheberg, Jarrow, Kraft, & Yildirim, 2014). Having analysed the cyclic behaviour and the efficiency of the housing market of the United Kingdom, Meen (2000) highlighted the importance of the interactions among house prices, construction costs, and interest rates in the formation of high amplitude cycles. The operation of the property market is not efficient, so the excess supply of houses is a common characteristic of the market. Davidoff (2013) has also analysed the rapid price increase affecting the American housing market in the first half of the last decade. While the negative correlation between the expansion of the stock and the prices was treated fundamental, which was also supported by empiric research at the time of the price increase in 1980, the time period subject to Davidoff's (2010) research has proven a positive correlation between the stock of residential properties and their prices. Amongst others, the Hungarian housing market was studied by Hegedus (2006), Kiss and Vadas (2006), Madi (2008), Farkas et al. (2010), and Bekes et al. (2016). In their studies, Hegedus (2006) and Madi (2008) analysed the state policy regarding residential properties and the property subsidy systems in Hungary between 1990 and the first half of the 2000s. A main element of the subsidy system introduced in 2000 was the interest subsidy provided for mortgages. Some households brought forward their intention to purchase their property, which generated a dynamic increase in state-supported mortgages, so budgetary funding of the introduced property policy was not sustainable for the long term. In the years following the political transformation, the real prices on the housing market decreased by 35-45%. However, house prices increased by 100-150% between 1998 and 2001, which was slightly before the increasing number of mortgages after the 2000s. Subsequently, house prices increased even further until 2005, which was due to the increased demand. Hegedus (2006) explained the change of demand resulted partly by the wide stabilisation of household income and partly by the reallocation of savings in the money market to the property sector. Madi (2008) primarily analysed the positive effect of residential construction and the housing market on the economic growth, the structure of the economy, employment, the equilibrium of public finances, lending, and growth of imports. As he established, the expansion of the housing market in the first half of the 2000s was mainly the result of the announced state policy on housing. Kiss and Vadas (2006) studied the effect of the Hungarian monetary policy on the housing market in the first half of the 2000s. The authors found that the changes in interest rates did not have a large effect on the available income of households and thereby to the demand for residential properties. The reason for this is, first, the interest subsidy provided by the state, meaning that the interest rate risks were actually born by the state; second, mortgages with long interest periods were predominant, so their interest rates reacted to the changes of the key interest rate with a relatively long delay. The effect of interest rate changes on the prices, investment, and house sales is limited to the Hungarian market. In their agent-based simulation model, Farkas et al. (2010) investigated the role of social and individual learning in the adaptability of the housing market. Market operators on the housing market conclude transactions only rarely. The market is monitored before each purchase or sale, but market operators also strongly consider their previous experience as well. Farkas et al. (2010) concluded that learning based on individual information results in unstable adaptability, so price bubbles and crashes occur on the housing market from time to time. Using a hedonic regression method based on the Hungarian tax authority database, Bekes et al. (2016) assessed the relationship of each town-level factor affecting the demand for residential properties in Hungary (natural properties, availability, educational and healthcare services, official classification) with the housing prices. On average, natural properties (especially nearby natural waters), accessibility by road or public transport, distance from institutions, and the administrative classification of the town are responsible for approximately 15% of the house prices in total. The properties of the residential property explain approximately 30% of the price variation. The four factors analysed by Bekes et al. (2016) have a strong correlation with the size of the town and the average income of people living there, which are cruical factors for the prices in the housing market. The previous domestic research focused mainly on the period from the beginning of the 2000s until the crisis, which was exclusively about demand and price increase. In this study, the behaviour of the Hungarian housing market is evaluated in respect of time from 2007, which is the year directly before the beginning of the economic crisis. In accordance with the international trends, the crisis also caused a significant decrease in demand in Hungary concerning durable consumer products, including residential properties. However, housing prices started to increase from 2011, which significantly accelerated in the past three years. The volume of remitted mortgages and state support had been decreasing every year from the beginning of the crisis and only started to increase again from 2014, which corresponds 7 NASE GOSPODARSTVO / OUR ECONOMY Vol. 64 No. 2 / June 2018 to the beginning of the housing market expansion in progress today. The period passed since the beginning of the crisis is therefore suitable for analysing the cyclic behaviour of the Hungarian housing market. The Hungarian Housing Market from the Political Transformation until the Economic Crisis The events of the housing policy and housing market after the political transformation can be divided into three sections. In the period between 1990-1994, construction by the state (which was the main form of construction in the previous decades, even though it was hardly operational in the 1980s already) has completely ceased, and the division channels of the socialist housing system have been terminated. Before 1990, the Hungarian state provided massive amounts of mortgages with minimal interest rates through the only consumer bank operating at that time. This credit volume became extremely large by 1990. As a result of the sudden increase in unemployment in the years following the political transformation, households were unable to pay their mortgages, so the resolution of this credit volume could no longer be delayed. This was the period when the large-scale privatisation of state and council-owned rental properties started, and this has a significant effect on the housing structure even today. The number of constructed houses decreased year after year. Less than half as many houses were built in 1994 than in 1990. In the period from 1995 until the turn of the millennium, the privatisation of the construction industry has been completed, and the institutions of the housing market have been established. New institutions appeared in financing, e.g., housing associations and banks providing mortgages. The state subsidy provided for construction and purchase of houses has been increased substantially from 1995. The number of constructed houses had a slight increase until 1997; then it decreased significantly from 1998 (Hegedus, 2006; Madi, 2008). The turning point for the housing policy was the year 2000; new programmes started for supporting mortgages and for supporting the construction of rentable council houses. The construction of houses increased by 50% in one year, and such increase continued. By 2003, the interest rate subsidies by the state have effectively become impossible to finance from the national budget, so the government had to implement a significant correction of the housing support system (Hegedus, 2006). At this time, the growth in the number of house constructions became slightly slower, but the turnover of the housing market kept increasing during the economic boom until the crisis. The financing role of state-supported mortgages was gradually overtaken by currency-based mortgages, but these had an inherent risk concerning exchange rates. Once the recession started, the decrease in demand has occurred first in the case of durable consumer products, especially for real properties, including residential properties. Due to the uncertainty of their income prospects, households delayed their intention to buy their properties. The decrease of house prices only followed the lower demand with a delay because most sellers were waiting, the offer prices started to decrease only after the crisis deepened. Amongst the years preceding the crisis, the year 2007 was chosen for the starting year of further analysis because, in the case of certain indicators used for the analysis, statistical data are only available from this year. Behaviour of the Hungarian Housing Market in Respect of Time from the Beginning of the Crisis until Today Similarly to other durable consumer products, the behaviour of the Hungarian housing market is primarily aligned with the macroeconomic cycles. There is a strong relationship in the positive direction between the constant-price GDP and the specific price of newly constructed residential properties, which is indicated by the value of the correlation coefficient (r = 0.8725) (Figure 2). According to our calculations, the variations of the GDP have an approximately 76% effect on the variance of the square-meter price of new residential properties (the value of the determination coefficient is R2 = 0.7613). The average specific price of new residential properties was continuously decreasing, with a total decrease of around 9% from the beginning of the economic recession started in 2008 until the year 2011 constituting the lowest point in prices, even though the output of the national economy showed a slight increase from 2009. Due to the inflexibility of the property market, residential property prices primarily depend on the demand (DiPasquale & Wheaton, 1992); this behaviour can be explained by houses being durable consumer goods because, if households consider their future income to be unstable due to the stagnation or decrease of their salaries or because of the increasing unemployment, they decrease their consumption primarily in relation to such things. The demand in the housing market is significantly affected by the provision of mortgages and state subsidies, as a strong positive correlation 8 Laszlo Harnos: Cycles of the Housing Market in Hungary from the Economic Crisis until Today Figure 2. GDP and the Mean Price Per SOM of New Dwellings in Hungary (2007-2016) 330 - 320 310 "O ü 300 D O