«CHINA'S LAND MARKET AUCTIONS: EVIDENCE OF CORRUPTION Hongbin Cai J. Vernon Henderson Qinghua Zhang Working Paper 15067 ...»
NBER WORKING PAPER SERIES
CHINA'S LAND MARKET AUCTIONS:
EVIDENCE OF CORRUPTION
J. Vernon Henderson
Working Paper 15067
NATIONAL BUREAU OF ECONOMIC RESEARCH
1050 Massachusetts Avenue Cambridge, MA 02138 June 2009 ¸˛The authors thank Yona Rubinstein for helpful comments and advice, as well as workshop participants at Brown University and Hong Kong University. The views expressed herein are those of the author(s) and do not necessarily reflect the views of the National Bureau of Economic Research.
NBER working papers are circulated for discussion and comment purposes. They have not been peerreviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.
© 2009 by Hongbin Cai, J. Vernon Henderson, and Qinghua Zhang. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided that full credit, including © notice, is given to the source.
China's Land Market Auctions: Evidence of Corruption Hongbin Cai, J. Vernon Henderson, and Qinghua Zhang NBER Working Paper No. 15067 June 2009 JEL No. D44,H71,O38,O53,R14,R31,R52
ABSTRACTThis paper studies the urban land market in China in 2003—2007. In China, all urban land is owned by the state. Leasehold use rights for land for (re)development are sold by city governments and are a key source of city revenue. Leasehold sales are viewed as a major venue for corruption, prompting a number of reforms over the years. Reforms now require all leasehold rights be sold at public auction.
There are two main types of auction: regular English auction and an unusual type which we call a “two stage auction”. The latter type of auction seems more subject to corruption, and to side deals between potential bidders and the auctioneer. Absent corruption, theory suggests that two stage auctions would most likely maximize sales revenue for properties which are likely to have relatively few bidders, or are “cold”, which would suggest negative selection on property unobservables into such auctions.
However, if such auctions are more corruptible, that could involve positive selection as city officials divert hotter properties to a more corruptible auction form. The paper finds that, overall, sales prices are lower for two stage auctions, and there is strong evidence of positive selection. The price difference is explained primarily by the fact that two stage auctions typically have just one bidder, or no competition despite the vibrant land market in Chinese cities.
J. Vernon Henderson Department of Economics Box B Brown University Providence, RI 02912 and NBER firstname.lastname@example.org This paper studies the urban land market in China in 2003-2007. Urban land is owned “by the people” and its allocation done by the state. 1 In most cities, the local land bureau is responsible for the vast majority of allocations of land, allocated through auction sales of leasehold rights. 2 In China, land markets have been viewed as very corrupt, prompting a number of reforms over the years. We will provide some institutional context below but a couple of quotes illustrate that corruption is an on-going issue. In 2004, the China Daily wrote “China’s Ministry of Lands and Resources announced new measures to crack down on corruption and inefficiency in the land sector. The new rules forbid officials to receive personal benefits from parties under their administration [italics added]. It is estimated that in 2003, the country faced 168,000 violations of its Land Law.”
Yet in June 2008, the Asian Times reported
“Chinese government efforts to clean up land sales, a major source of official corruption…, face a rethink.
…Illegal transfers, corruption in land deals,…are rampant in major cities, according to an investigation published by the National Audit Office (NAO) last week.
…Governments in the 11 cities [studied by the NAO], including Beijing and Shanghai, were also found to have misused 8.4 billion yuan from land-grant fees, Zhai Aicai, of the NAO, said in the report.
….Some cities have given a flexible interpretation to the rules and the auction system has often existed in name only, resulting in a lack of competition among developers and the winning developer being able to secure the land at below its true market value.” Today, after considerable reform, leaseholds are, in principle, all sold at public auction. There are two main types of auction in most cities: regular English auction and an unusual type of auction which we call a “two stage auction.” The raw data suggest that leaseholds sold at two stage auctions sell at steep price discounts, relative to English auctions. Why are there such sales price differentials; and, related, how do city officials choose auction type for any particular property? Absent corruption, in terms of maximizing expected sales revenue, the paper argues that two stage auctions would be best for properties which are likely to have relatively few bidders, or are “cold.” Rural land is owned by the village and allocations done by the village leadership.
The central government (national asset committee) and the military may control portions of city land in particular cases, as for example the national capital Beijing.
Allocating cold properties to two stage auctions would suggest negative selection on property unobservables, and would be one possible explanation for the price differential between the auction types. However estimation suggests that selection on unobservables into two stage auctions is in fact positive. The paper argues that two stage auctions are more corruptible, which would explain positive selection, as city officials divert “hot” properties to the more corruptible auction form. Consistent with how we think the corruption process works, much of the price differential between auction types seems to be explained by the fact that two stage auctions are much more likely to have just one bidder, or no competition, despite their benefiting from positive selection on unobservables and despite the vibrant land market in Chinese cities.
The paper is organized as follows. We start with essential background information on Chinese land markets and especially the two auction formats. We then present a conceptual framework to model the key differences between the auction formats. In section 3, we discuss the data and patterns in the data. Section 4 estimates a reduced form model of price differences between the two auction types, and discusses instruments for auction type used to estimate selection into auction type. In Sections 5 and 6 we split the analysis of price differences into its two key components, accounting for joint selection into auction type and competition: whether a property is likely to sell competitively or not depending on auction type; and whether there are price differences across auction types, conditional on a property selling competitively.
1. Background In the Maoist era and in the early reform years after 1978, land allocations were done by the state, with no market mechanisms involved. Starting in 1986, land administration changed with major reforms over the years (Ding and Knapp 2005, Valetta 2005). The first change was to charge new users for development rights and some incumbent users for use rights. After 1988, use rights for vacant land in the city were allocated through leaseholds, where, for a fixed sum, users obtained a long lease for a specified use (e.g., 70 years for residential land), subject to restrictions on intensity of development. In the 1990’s, many of these allocations were done by “negotiation” in a hidden process, where reportedly leaseholds were often sold for a tiny fraction of market value. This had two consequences.
First, leasehold sales are a major source of revenue for many cities, in essence potentially being a full Henry George tax, allocating all “surplus” land rents to the city.
For example, in 2004 and 2005 for Chengdu, Suzhou and Chongqing, leasehold sales revenues ranged from 2.6% to 5% of local GDP. Cities have an expenditure budget and on-the-book revenue sources. On-the-book revenues account for about 70% of total expenditures. Leasehold sales revenues are mostly off-the-book revenues, which are used to effectively close the on-the-book deficit. Negotiated sales at well below market prices deprived cities of major revenues.
Second, negotiated sales were reportedly inherently corrupt, resulting in some indictments of corrupt officials and a variety of reforms, one of which in 2004 was quoted above. Another reform in 2002 banned the secondary market for “land development rights,” which had allowed large traditional holders (e.g., state owned enterprises) to, in effect, privately sell off their own land use rights (Zhu, 2004, 2005).
Today the local land bureau is supposed to be in charge of almost all allocations of land for (re)development. Finally and most critically for this paper, a third recent reform was a 2002 law which banned negotiated sales by the land bureau, with the last date for any negotiated sales being August 31, 2004. For the last 4 years at least, all urban land leasehold sales are to be done through public auctions, with details of all transactions posted to the public on the internet.
How does the land market work? Local land use planning is done by an independent committee (albeit with 1-2 representatives from the land bureau on the committee). Given the overall land use plan for the city, at the beginning of each year, annual allocations are planned, based on existing urban land and converted rural land which should be ready for redevelopment during that year. Each plot of land is large with, in our sample, a median area of 22,300 square meters and a median sales price of USD $7 – 8 million. The committee decides the use and other constraints (like floor-area ratio) of each plot to be sold. Once the land becomes available during the year, the committee sets the reserve price, using a formula based on the appraised value submitted by, in principle, independent appraisal companies. Then the land is turned over to the land bureau which prepares it for sale (land for redevelopment in principle should be cleared), and chooses an auction type.
There are three types of auction used in China’s land market. About 97% of sales in major cities are accounted for by two auction types, with the third type generally appearing only in Beijing and Shanghai. We ignore this third type of auction and our econometric specifications exclude Beijing and Shanghai. 3 The two main types are guapai auction which we call two stage auction and paimai which is an English auction.
English auctions are standard ascending auctions, usually publicly announced 20 working days before the auction. At announcement, basic details (e.g., use restrictions, reserve price, location) are publicized; and potential bidders for a small fee can obtain more detailed information, as well as inspect the site. Participation requires a cash deposit, usually about 10% of the reserve price, which is a non-trivial requirement given the large sizes and sales prices of such properties. English auctions are quite public, often videotaped with the press present. Winning bidders in principle must develop the land themselves.