Printer Friendly

Privatization in Russia: some micro-evidence based on housing markets.

1. Introduction

With the fall of Communism and the progress of capitalist institutions in the former Soviet Union as well as the Eastern European countries, there has been much interest in the privatization of the formerly state-owned enterprises and property. There is a vast amount of literature on this issue. See inter alia Alexander and Skapska (1994), Atkinson and Mickelwright (1992) and Wineicki (1993). There is also a lot of skepticism among the public in Russia about privatization. See, for instance, a recent article in the New York Times (Sunday 28 January 1996, Section 1, p. 1, col. 4) titled "Russian banking scandal poses threat to future of privatization". The author notes that privatization began as a revolutionary step to roll back 70 years of Communism, selling off state assets to create a society of property owners with a stake in the success of a free-market economy. But (because of inequities) it is deeply unpopular with voters who cynically refer to it as prikhvatizatsiya or "grabification", the giving away of government wealth to a few well-connected and unscrupulous businessmen and bankers. Under the deal, known as loans-for-shares, a few Kremlin-favored banks lent money to the government last year in return for a chance to buy shares in some of the state's most valuable assets at dirt cheap prices. Often, Soviet-era directors bought the companies themselves and continue to run them Soviet-style.

In order to examine the importance of the role of economic incentives in privatization, we decided to collect data at the micro-level. We collected data at auctions of state-owned apartments in Moscow. In this case there was no high level prearrangement as in the case of the auctions of big state enterprises. Two of us (Toda and Nozdrina) attended the auctions at which there were well over 20 people. The auctions were open to the public and newspaper reporters were also present. Sometimes two bidders fiercely competed, raising the price far above the starting price. We fit hedonic price functions to the data we collected and found that the prices were explained well by the characteristics of the apartment, particularly after the price liberalization program of January 1992. The data set we gathered is unique and there is no comparable micro data set in the communist countries.

In the following sections we shall describe the nature of the auctions, the data we collected and the hedonic price functions we estimated. Here we shall present merely a summary of the empirical findings. Details of the estimation can be found in Maddala et al. (1993, 1996). We are aware that Moscow does not represent Russia. However, the share of privatized apartments in Moscow was 37 percent at the end of February 1994, a figure much higher than the national average. Since the data had to be collected by us directly (they are not available from any public sources) concentrating on Moscow was the most practical thing to do. At some point in the future, we plan to extend our analysis to cover St Petersburg and Nyzhinyi Novgorod.

2. Development of auctions in Moscow: some institutional background

In the spring of 1991, the then Mayor of Moscow, G. Popov signed a resolution to hold three auctions in order to sell the municipal housing. Accordingly 30 apartments in three housing sites were auctioned. There were several restrictions on who could participate in the auctions (for instance, initially those with a residence permit to live in Moscow and were on the waiting list for municipal apartments were the only ones allowed to participate). These restrictions were later lifted and in 1992, real estate agents started participating in the auctions. In the first auction held on 5 April 1991, all the ten apartments put to auction were sold at prices which more or less corresponded to black market prices in dollars. At the second auction on 30 May 1991, when ten apartments were sold, the prices were much higher than in the first auction (for apartments with roughly the same characteristics). Subsequently several other auctions were held. But in December 1991, a rumor that the bank accounts of individuals would be frozen accelerated the price increase. For example, in the auction carried out by the realtor Banso on 24 December 1991, a comfortable three-room apartment in a prestigious residential area was sold for $670,000, a record price not surpassed till now. However, since the ruble became internally convertible with the liberalization of prices in January 1992, the difference between ruble and dollar prices diminished progressively and became practically zero in 1993.

In 1992, the Moscow city auctioned the apartments in the mass apartment sites in the city peripheries, but in 1993 the city also auctioned the units in the city center. The apartments in the city center were renovated and auctioned with the price around $1,000 per sq. m.

In 1992, the city started selling apartments in blocks rather than on an individual basis. With the restrictions on who could participate in the auctions removed, several real estate agents came into business in 1992. These agents were engaged in the sale of municipal housing and also private apartments. They would buy apartments in a block at the city auctions and then resell them on an individual basis. By 1993 there were about 250 real estate agents in Moscow but most of them were small, selling not more than five apartments a month. The largest firms (Reskom, Savva, Banso, S. I. Realty, Housing Initiative, the Moscow Central Exchange of Real Estate) are engaged in the auction of private apartments as well.

The large realtors are also involved in speculation. One of the largest companies, Reskom, for example, with loans from several banks, normally keeps 20 to 30 vacant apartments bought at low prices and sells them during periods of high speculative demand. Towards the end of 1992, Reskom bought (with 1.5 billion ruble bank loans) a 100 apartment block from the city and later sold them on an individual basis. According to expert opinion, these dealer operations give a monthly rate of return of 30 to 60 percent.

In spite of the lucrative dealer operations and regular sale of housing units as servicing agents, the large real estate companies are willing to spend money for auctions of private apartments for several reasons. First, it works as an advertisement. Second, they can choose a convenient date when clients have the need to spend unused money, for example the end of the year. Third, the apartments auctioned are not of the average class but with excellent layout and at the city center. Finally, the auction prices give the realtor information on prices it can charge for individual sales where the company acts as an intermediary. As the real estate market develops in Moscow, the auction of privately owned apartments will yield its place to individual sales where agents work as intermediaries, as in the USA. However, the auction of municipal housing units will continue.

The preceding discussion is a cursory review of the apartment auctions from which we collected the data. Since, as mentioned earlier, the data available from the real estate agents did not have information on characteristics necessary for hedonic estimation, we had to attend the auctions personally and gather the relevant information on the characteristics of the apartments. We have skipped a lot of detail on the changing structure of the housing market in Moscow, since that is going to distract from the main focus of this paper. More detailed discussion of the housing market itself without an econometric analysis can be found in Andrusz (1990), Khadduri (1992), Kosareva and Struyk (1992) and Renaud (1992).

3. The data

We gathered information on 582 apartments auctioned in Moscow between July 1991 and March 1993. Of these, 389 apartments were actually sold. For 179 apartments no buyers were found at the initial auction price. For the remaining 14 units the sellers withdrew them from the auctions even though there were several willing buyers. Thus, the data fall into three categories:

(1) For the 389 apartments sold, the final auction price is the equilibrium price.

(2) For the 179 apartments not sold, the implicit equilibrium price is lower than the observed initial auction price.

(3) For the 14 apartments withdrawn, the implicit equilibrium price is higher than the observed initial auction price. In this case the equilibrium price could not be observed because the auction was discontinued.

Most of the apartments put into auctions were individual units, but there were cases where a group of apartments, that shared the same entrance to a building or were located on the same floor of a building, were priced together and were sold together. Although the price could be divided to get the estimate of each single unit, the apartments were hardly distinguishable. They were not only in the same location, but were of the same size and quality. In these cases, we take the whole group as a single observation, though of course we acknowledge the fact that it contained multiple housing units. By this re-counting, the total number of observations was reduced to 466. They were divided into three groups mentioned earlier, each consisting of 348, 109 and nine observations respectively.

There are a number of "hedonic characteristics" attached to each apartment. The characteristics of the first kind are concerned with the location, e.g. the street address, the district of Moscow, the nearest subway station and the time needed from the station to the apartment either by walk or by public transportation. Second, there are the size factors such as the number of rooms in an apartment, the size of kitchen in square meters, the living space and the whole space (including corridors and bathrooms) also measured in square meters. The third set of variables are the quality and amenities, e.g. the building material (brick, panel or block), the year of construction, the need for repair work, the availability of elevator and telephone, the number of the floor where the apartment is located and the number of floors in the building. Fourth, the institutional factors must be taken into account. Whereas the empty apartments under the ownership of municipal government are sold for privatization, the private realtors are also selling private apartments. Furthermore, as mentioned earlier, while individual units are put into auctions separately, the groups of apartment units are also offered for bidding. They may have a different impact on the prices. Finally, the date at which the auction was held is also recorded.

We witnessed some of the auctions and collected the information by ourselves. We were also given by others some information on those we did not attend, but the information thus made available to us was not uniform. Many of the important hedonic characteristics were missing. By contrast, for many of the units on which we collected the data ourselves, we gathered detailed information. If one wants to see as many hedonic characteristics as possible, one is compelled to utilize only a few observations. On the other hand, in order to utilize as many observations as possible, one needs to restrict oneself to a limited number of characteristics. There is thus a trade-off between the number of characteristics and the number of observations.

4. The variables used in the hedonic regression equations

If we try to utilize all the observations in hand, we can test but a handful of characteristics. Among the location variables, we use only district. We divide the Moscow city and its environs into 15 districts and rank them (starting at one for the best area). Our ranking is based on the rates of land tax the Moscow city applies to different districts. Thus, in principle, the closer to the city center, the higher is the rank. However, the area with high

commercial value and hence with high land tax is not necessarily a good residential district because of its ecological deficiency. As the construction of single-family units for "the new rich" picks up its pace in ecologically cleaner peripheries, a simple, unimodal regional distribution of prices as it is currently seen might disappear.

The size variables such as the number of rooms, the size of the kitchen, the total space and the living space are available for all the observations. The size of the kitchen, not only being a factor by itself, is also considered as a surrogate for the quality of the apartment.

Not many units have the quality variables such as the building material, the age of the building, the need for repair, the number of floors where the apartment is located, etc. We were not able to utilize any of these variables.

The information whether the unit was put into auction by a government agency or by a private body is available. Whether an individual unit was sold separately or whether several of the units were sold together is also known. Thus we can test how powerful these differences may be in affecting the apartment price.

The date of auction is always available. We are thus able to test the effect on the auction price of the macro-economic variables such as the rate of inflation, the ruble exchange rate, the money emission, etc. which we gathered from Goskomstat (1994).

We are interested not in the absolute price of an apartment but in the relative price, because the apartment is one of many things people choose to buy and sell. So we calculate the ratio by dividing the auction price by the retail price index (with the March 1993 level as 100). Further, we standardize the size of a housing unit by dividing the price ratio by the whole apartment space. Thus the auction price whose variation we want to explain in estimating a regression is the constant price per square meter. We also used the size variable as an additional regressor in some of the regressions but found its coefficient to be insignificant.

In periods of inflation, the scale of macro-economic variables such as the price index, the exchange rate, the money emission, etc. are increasing fast. The larger the scale of these variables, the larger may well be the scale of the error term. If so, then the assumption of homoskedasticity is again violated. In order to avoid this problem, we transform all the variables into the logarithmic form before estimating the hedonic regression equation.

To measure institutional factors affecting the apartment prices, we use a dummy variable to distinguish an apartment sold collectively from the one sold individually. The variable assumes unity if the apartment in question is the former and takes zero if it is the latter. We also use a dummy variable to divide the sellers into two groups. The variable takes unity if it is sold from a governmental ownership and assumes zero if the source of supply is nongovernmental.

5. The empirical results

There are unusual problems with the data that necessitate the use of more elaborate techniques than ordinary least squares (OLS). There are 348 equilibrium observations and 109 observations for which the observed price is above the equilibrium price. There is also the problem of identification of supply and demand functions in the estimation of hedonic price functions. In our case, this problem does not arise because the supply can be considered as inelastic. At each time only one apartment is offered for auction and also the set of apartments auctioned is determined by administrative considerations. The stock of apartments under governmental control is fixed. Thus, we can safely conclude that the hedonic price function we are fitting represents a demand function.

In Maddala et. al. (1993) all the observations were treated as equilibrium observations and the hedonic price equations were estimated by ordinary least squares using the two groups separately. The main findings were (we are omitting the details):

(1) The explanatory variables (hedonic characteristics) were all significant at the 5 percent level and had the expected signs. Also the [R.sup.2]s were in the range of 0.47-0.54.

(2) Most importantly, there was evidence of a structural break in 1992, which coincides with the dating of the price liberalization program.

In Maddala et. al. (1996) the disequilibrium nature of the 109 observations on apartments withdrawn from the auction (because the initial auction price was too high and there were no bidders) is taken into account. The method used was one of estimation of disequilibrium models (see Maddala, 1983) but the likelihood function turned out to be similar to (but not exactly the same as) that of a tobit model. Since the modifications to the tobit program could not be made using the standard computer packages we used our own computer program (which is available on request). The results were (again we are omitting the details which can be found in our paper):

(1) All the explanatory variables (hedonic characteristics) were significant at the 5 percent level and had the expected signs.

(2) The results from the estimation of the disequilibrium model were not much different than those from the use of ordinary least squares. This suggests that the equilibrium price of the apartments withdrawn from the auction (because of high initial auction price) was perhaps not too far below the initial auction price. We do not know whether these apartments were sold in a subsequent auction. All we know is that they were not offered at the particular auctions we (Toda and Nozdrina) attended.

(3) There was clear evidence of a structural break in 1992. In fact the estimation results for set 1 (equilibrium observations) is not as good for the period January 1992 as for the post January 1992 period. The [R.sup.2] was 0.478 for the pre-January 1992 period compared to 0.836 for the post January 1992 period. Also some of the coefficients had the wrong sign in the former period. Thus, there is some evidence that there is a structural break in 1992. See, for instance, Feige (1994) for a discussion of the consequences of the price liberalization program of 1992.

6. Conclusions

We have estimated hedonic price equations for apartments auctioned in Moscow. We have analyzed the data for 457 such auction prices, of which 109 were disequilibrium observations in the sense that the apartments were not sold because the initial auction price was too high. These withdrawn apartments were not sold at the auctions that we attended subsequently They might have been sold at auctions that we did not attend.

We have estimated the hedonic price equations, both by OLS ignoring the fact that for the 109 apartments not sold, the auction price is not the equilibrium price and also using a disequilibrium model, where it is assumed that for the unsold apartments, the equilibrium price is below the auction price. Our results indicate that for these apartments, perhaps the equilibrium prices are not much below the initial auction prices.

Although initially we used price per unit area as our explained variable we also estimated equations where price is not proportional to area. Although the results showed that the proportionality factor was not significantly different from one, they indicated that price increases more than proportionately with area. The larger apartments are perhaps of better quality and the variable area proxies unmeasured quality differences.

We also found a structural change around January 1992, when prices were liberalized and inflation became overt. We also found that the hedonic price equation fits much better after the price liberalization.

We have been constrained in our analysis by data limitations. Using all the variables would have meant dropping a lot of observations. Hence we made use of only those explanatory variables for which we could use almost all the observations. However, with the variables that we used, our equations showed very good fits with [Mathematical Expression Omitted] around 0.84-0.93. We can conclude that the auction prices are explained very well by economic factors. This is particularly true after price liberalization. Before the price liberalization of 1992, the fits of our hedonic price equations were comparatively poor (in one equation we got an [Mathematical Expression Omitted] of 0.48). In our future work we plan to collect more detailed data on auction prices in Moscow, St Petersburg and Nyzhnii Novgorod and check how prices are determined by economic factors under the privatization programs. It is also important to see whether a system of real estate agents as in the USA develops in Russia as well.


Alexander, G.S. and Skapska, G. (Eds) (1993), A Fourth Way? Privatization, Property and the Emergence of New Market Economies, Routledge, New York, NY.

Andrusz, G. (1990), "Housing policy in the Soviet Union", Chapter 6 in John Silince (Ed.), Housing Policies in Eastern Europe and the Soviet Union, Routledge, New York, NY.

Atkinson, A.B. and Mickelwright, J. (1992), Economic Transformation in Eastern Europe and the Distribution of Income, Cambridge University Press, Cambridge, MA.

Feige, E.L. (1994), "The transition to a market economy in Russia: property rights, mass privatization and stabilization", in Alexander, G.S. and Skapska, G. (Eds), A Fourth Way? Privatization, Property and the Emergence of New Market Economies, Routledge, New York, NY, pp. 57-78.

Goskomstat (1994), Sotsial'no-ekonomicheskoe polozhenie Rossii, yanvar'-iyun', Moskva.

Khadduri, J. (1992), "Notes on the residential real estate market in Russia", The Urban Institute, September.

Kosareva, N. and Struyk, R. (1992), "Housing privatization in the Russian Federation", The Urban Institute, May.

Maddala, G.S. (1983), Limited. Dependent and Qualitative Variables in Econometrics, Cambridge University Press, Cambridge, MA.

Maddala, G.S., Toda, Y. and Nozdrina, N. (1993), "The price of apartments auctioned in Moscow", Voprosy Ekonomiki, July, pp. 99-110.

Maddala, G.S., Toda, Y. and Nozdrina, N. (1996), "Privatization in housing markets in Russia: a disequilibrium analysis of auction prices of apartments in Moscow", paper presented at the Royal Economic Society, Swansea.

Renaud, B. (1992), "The housing system of the Former Soviet Union: why do the Soviets need housing markets?", discussion Paper, World Bank, March.

Wineicki, J. (1993), Post-Soviet-Type Economies in Transition, Avebury, Brookfield.
COPYRIGHT 1998 Emerald Group Publishing, Ltd.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1998 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Maddala, G.S.; Toda, Yasushi; Nozdrina, Nadezhda
Publication:International Journal of Social Economics
Date:Feb 1, 1998
Previous Article:Jeremy Bentham and the New South Wales convicts.
Next Article:Privatization of public enterprises and its problems in Japan.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters