The basic goal behind the creation of New York City's rent stabilization system was the establishment of "fair" rents. Although imprecise, the policy of "fairness" has directed the rent stabilization system towards pursuing three main objectives. The first is the establishment of rents that protect tenants from price gouging in the face of low vacancy rates. The second encompasses the preservation of the returns reasonably expected by good faith (non-speculative) investors in the city's rental housing market. The final objective is to establish rent adjustments in an even handed way guided by reference to legitimate public policy concerns.

A common criticism of rent regulation is that such regulation increases rent inequities among tenants. Controls arguably provide "bargain rents" that induce tenants to occupy regulated units as long as possible, causing older, "empty nest" households to "over-consume" large dwellings which are better suited to the needs of younger, child rearing households seeking entry into the regulated sector. Regulations that do allow special rent increases for vacant units, such as New York's stabilization system, not only encourage long term occupancy, but also promote "rent skewing", in which identical apartments become differently priced over time due to variations in their turnover rates. Through this process, long standing regulated tenants pay much less rent than newcomers to the regulated sector, who in effect "subsidize" their counterparts. According to critics, these drawbacks seriously undermine the utility of controls as a means of making housing more affordable.

The presence of "rent skewing" in New York's stabilized housing would seem to contradict the "fairness" policy upon which the system was founded. However, this observation is based on the assumption that skewing is a unique by-product of rent regulations, and is not common to all types of housing markets. Objective assessment of the equity of skewing in New York's stabilized housing requires knowledge of the presence and degree of skewing in the city's non-regulated housing markets. The purpose of this analysis is to determine: a) the existence and extent of "skewing" within the city's stabilized sector as well as in the private market, b) the effects of "skewing" upon stabilized tenants as well as the city, and c) what actions the Rent Guidelines Board should undertake to either minimize or justify "skewing" within the stabilized system.

These questions will be explored first through a review of existing literature on rent skewing in both regulated and unregulated housing markets. Analysis of two hundred and twenty rent stabilized buildings for the presence of skewing within and between buildings will then be conducted. This analysis will be followed by a more detailed examination of the factors which contribute to any skewing which exists within the stabilized system.


Review of existing literature provides theoretical and empirical evidence of skewing in both regulated and unregulated housing markets. Statistical analysis of 1993 data from two hundred and twenty rent stabilized buildings as well as the cross-sectional data set of the 1991 HVS revealed statistically significant (non-random) skewing of rents for comparable apartments in both stabilized and unregulated rental buildings.

Further examination revealed similar average annual "length of occupancy" discounts (one measure of skewing) for sitting tenants in both sectors, but generally higher average total discounts for tenants in stabilized units than for those in unregulated rentals. This was due to the tendency of tenants to occupy stabilized dwellings for longer periods than other rental units, particularly in Queens and the part of Manhattan lying below 110th Street. Statistical tests undertaken on the length of tenure in stabilized and other rental units revealed a significant (non-random) positive association between stabilized status and length of occupancy.

Literature Review

Literature on rent skewing is often associated with arguments against rent regulations, which claim that controls distort market mechanisms for efficiently allocating housing supply and demand. Traditionally, urban theory attributed variations in rents among identical dwellings to differentials in location or public utilities.(1) However, there is a growing body of evidence for the existence of skewing in unregulated housing markets, whereby variations in lengths of tenure explain variations in rents between identical apartments.

Anthony Downs was among the first authors to provide a theoretical explanation for rent skewing in the private market. According to Downs, ownership of America's rental stock is dispersed among thousands of small property owners, with more than 60% of all rental units situated in buildings with less than five dwellings. Small landlords, due to the limited size of their holdings, tend to be much more sensitive to the costs incurred from vacancies than to the rents they receive from their property. In effect, Downs believes that small building owners are usually "turnover minimizers", who prefer to keep their units continually occupied, instead of "rent maximizers", who are willing to constantly refurbish vacant units to attract the highest paying tenants.(2) Given these attitudes, small landlords are willing to offer discounts to "good" (i.e., responsible, well behaved) tenants on a continual basis rather than to risk the expense of vacancies.

A number of empirical studies corroborate the existence of rent skewing in the unregulated housing sector. One of the most important is an article written by Allen Goodman and Masahiro Kawai in 1985 for the journal "Land Economics". The authors found statistically significant levels of rent skewing in eighteen out of nineteen metropolitan housing markets across the United States. In this analysis, length-of-tenure discounts averaged 1.3% per year for units of similar quality and location (i.e., the rents of established tenants declined by an average of 1.3% per year of occupancy compared to those of new tenants in comparable dwellings). The value of these discounts averaged eight dollars per month, which equaled 3.7% of the rents charged to new tenants.(3)

Among other authors concerned with private market rent skewing, Ira Lowry mentions a study undertaken by the Rand Corporation of rental properties in Brown County, Wisconsin and St. Joseph's County, Indiana in 1976, which found that monthly gross rents decreased from estimated "market levels" by a mean value of 3.8% per year of occupancy.(4) Another analysis conducted by Arthur D. Little in 1987 examined rent levels in cities with and without rent control and concluded that rents in all cities "...decrease with each year of additional occupancy".(5)

Many studies of rent regulated housing markets allude to rent skewing without exploring the issue in detail. Most link controls with reductions in tenant mobility, by claiming that below market rents encourage sitting tenants to occupy controlled apartments for extended periods of time, dramatically reducing turnover for some units. Systems with special vacancy allowances promote skewing over a period of years between dwellings that turnover frequently and those that do not. Anthony Downs elaborates on this point by observing that controls, in reducing the mobility of the initial occupants of controlled units, create shortages which make it harder to move from one controlled unit to another.(6)

Analyses that empirically examine tenant mobility in rent regulated cities present mixed results on this issue. One study of Los Angeles' rent stabilization system, conducted in 1984, found that tenant mobility in both the city and its surrounding communities (which did not regulate rents at the time) substantially decreased between 1977 and 1984. On the other hand, the authors found that length-of-tenure discounts substantially increased in Los Angeles over the same period, (the city's rent stabilization system was enacted in 1979). Although discounts for tenants with less than two years of occupancy declined between 1977 and 1984, discounts for those with three to four years increased by 39% and those for renters with more than six years of residency nearly doubled. From these findings the authors concluded that, overall, long standing tenants reaped the greatest rewards from stabilization in Los Angeles, whereas recent movers and tenants with less than three years of residence received the least benefits.(7)

Another examination of rent regulation, written in 1976 by Franklin James and Monica Lett, analyzed New York's rent stabilized housing stock. Despite its lack of quantitative sophistication, James and Letts' analysis concluded that the average variation between the highest rents and the lowest rents charged for four different types of apartments was 20% higher in rent stabilized buildings than in rent controlled buildings. The authors attribute this disparity mainly to the use of variable length leases, as well as vacancy allowances for unoccupied units, which they argue will cause skewing to increase over time, particularly during periods of rampant inflation.

To ameliorate this situation, James and Lett propose the exclusive use of one year leases, which, in their view, would prevent skewing from worsening over time.(8) However, James and Lett ignore the crucial option of modifying or eliminating vacancy allowances as a means of either remedying existing skewing or dampening future rent disparities.

Joel Brenner and Herbert Franklin's analysis of European rent controls also examines rent skewing, and provides a useful summary of strategies used to counteract it. Brenner and Franklin's discussion of skewing goes into greatest depth concerning Holland, where rapid inflation during the 1970's created great disparity in rents between new construction and older housing. They focus on two particular anti-skewing policies: a system of subsidies similar to HUD's Section 8 vouchers and a more recent attempt at "rent adaptation", whereby rents in older apartments are raised to levels comparable to new ones. According to the authors, the Dutch abandoned the former in the late 1970's, after rising inflation quickly devalued the subsidies paid out, opting instead to concentrate on "rent adaptation". This refers to a gradual re-alignment of rents in both old and new units, achieved through a series of slight rent increases in older dwellings combined with a new mortgage instrument, in which interest is spread evenly throughout the life of the loan to produce smaller monthly payments.(9)

Statistical Analysis

Examination of the presence and extent of skewing in New York's unregulated and stabilized rental stock required statistical analysis of two data samples. First, rents within "apartment lines" of 220 rent stabilized buildings obtained from the state Department of Housing and Community Renewal (DHCR) were analyzed to determine whether variation could be attributed to random chance or deeper structural causes. An "apartment line" represents a series of units which are vertically adjacent to each other, forming a vertical section of the property. These were chosen for analysis on the assumption that dwellings in a "line" are roughly similar in terms of size, number of bedrooms and internal layout, thus delivering roughly equal levels of "housing services" to consumers (tenants). Theoretically, in markets without skewing, apartments within a line should have equal rents. In the DHCR sample, considerable variation was observed in rents within "lines", with the difference between the highest and lowest rents (the "range") within a line averaging $242. Statistical tests on the range of rents within lines from the DHCR data revealed that the observed variation was not random, but emblematic of structural forces within the rent stabilization system. Thus, skewing between roughly equivalent stabilized units throughout New York is not a quirk of chance. The chart Difference Between Highest and Lowest Rent Within Apartment "Lines", Rent Stabilized Buildings, 1993 shows the range of rents within the lines in the DHCR sample for Manhattan, the Bronx, Brooklyn and Queens.

The simplicity of the DHCR data set precluded detailed insight into the factors behind skewing of rent stabilized rents beyond the fact that variance in rents between similar units (i.e., units within the same line) was positively correlated with location in Manhattan and negatively correlated with building size. Thus, Manhattan addresses were linked to increased differences in rents within apartment lines while larger buildings tended to have greater observed variation within apartment lines.

For non-Netscape table, click here or see end of current page.

"Length of Occupancy Discounts" in Rent Stabilized
and Unregulated Rental Buildings, 1991
Rent Stabilized BuildingsUnregulated Rental Buildings
Location:Annual DiscountTotal DiscountAnnual DiscountTotal Discount
Manhattan Core2.8%24.6%2.2%8.9%

Note: The number of stabilized properties in Staten Island was too low to be statistically reliable.
Source: NYC 1991 HVS

Use of the 1991 HVS cross-sectional data set, featuring detailed information on large numbers of stabilized and unregulated units, allowed us to follow Goodman and Kawai's footsteps to estimate another measure of skewing, "length of occupancy discounts", within both the rent stabilized and private rental sectors. In this instance, "unregulated rental dwellings" included buildings of all sizes, as well as co-op units which were rented by their owners. Regression analysis revealed that contract rents for units of similar size and quality in both markets were significantly associated with both location and the length of occupancy of the sitting tenants. Further analysis allowed us to calculate the mean discount received for each year of tenure in both stabilized and other apartments, as well as the mean total discount received by tenants in each sector, both of which are shown in the table above ("'Length of Occupancy Discounts' in Rent Stabilized and Unregulated Rental Buildings, 1991"). The term "Manhattan Core" refers to addresses below 110th Street on the West Side and below 96th Street on the East Side.

The figures illuminate an interesting relationship between rent skewing and regulation status. It appears that the main factor behind the total level of skewing experienced by tenants in both markets is the length of tenure rather than the annual discounts offered to renters. This is starkly demonstrated in Manhattan and Queens, where mean annual discounts were roughly equal across the sectors, while mean total discounts were much higher in the stabilized market due to mean lengths of occupancy double those in the unregulated rental market. In both the Bronx and Brooklyn, mean annual discounts are higher in unregulated units while the mean total discounts are approximately similar due to longer tenure patterns among stabilized tenants.

See chart Mean Length of Occupancy in Stabilized and Unregulated Apartments, 1991

These findings prompted further analysis of HVS data, to determine whether rent stabilization status significantly influenced length of occupancy, and thus the total amount of skewing observed in the 1991 HVS sample. Additional regression analysis revealed a strong positive correlation between length of occupancy and rent stabilization status among households of similar income in units of similar location, size and quality. Thus, tenants in rent stabilized units have strong incentives to occupy their units longer than their counterparts in unregulated rental dwellings, particularly in Manhattan, and especially in the high rent "Manhattan Core". In turn, this significantly influences the total amount of skewing observed in the stabilized sector.


  1. Goodman and Kawai, Land Economics, May 1985, p. 93.
  2. Weicher, John, et al., Rental Housing: Is There A Crisis?, p. 88-89.
  3. Goodman and Kawai, p. 94-95. The authors go on to show that, in the presence of rent skewing, recent movers "consume" less housing for every marginal increase in rents than do long term tenants, although their consumption patterns are similar in all other respects.
  4. Weicher, et al., p. 30-31.
  5. Arthur D. Little, Inc., Housing Gridlock in New York. May 1987, p.2.
  6. Downs, Residential Rent Controls: An Evaluation, p.21-25.
  7. Hamilton, et al., The Los Angeles Rent Stabilization System: Impacts and Alternatives, p. 19-49. Average length-of-tenure discounts for tenants with 1-2 years of occupancy declined from 4.2% in 1977 to 0.6% in 1984, those for tenants with 3-4 years of occupancy rose from 8.5% to 11.8% and those for tenants with six or more years of occupancy grew from 15.6% in 1977 to 30.1% in 1984.
  8. James and Lett, The Economics of Rental Housing in NYC, p. 80-85.
  9. Joel Brenner and Herbert Franklin, Rent Control in North America and Four European Countries, p. 28-37.

     Length of Occupancy Discounts" in Rent Stabilized and
               Unregulated Rental Buildings, 1991

                  Rent Stabilized           Unregulated Rental
                     Buildings                  Buildings
                --------------------       ---------------------
                 Annual      Total          Annual       Total
Location:       Discount    Discount       Discount     Discount

Manhattan         2.6%       23.4%           2.6%         10.7%
Manhattan Core    2.8%       24.6%           2.2%          8.9%
Bronx             1.3%       11.4%           2.0%         12.1%
Brooklyn          1.7%       14.8%           2.2%         14.2%
Queens            1.6%       16.2%           2.0%         11.0%

Note: The number of stabilized properties in Staten Island was too low to be statistically reliable.
Source: NYC 1991 HVS
To return, click here