PRICE‐TENURE REGRESSIONS WITH CENSORED DATA
| Date | 01 August 1987 |
| Author | Gary A. Zarkin,Robert C. Marshall |
| DOI | http://doi.org/10.1111/j.1468-0084.1987.mp49003006.x |
| Published date | 01 August 1987 |
OXFORD BULLETIN OF ECONOMICS AND STATISTICS, 49, 3 (1987)
0305-9049 $3.00
PRICE-TENURE REGRESSIONS WITH
CENSORED DATA
Robert C. Marshall and Gary A. Zarkin*
Hedonic regressions have been widely used in both labour and urban
economics to explain price dispersion. Tenure, the length of the
relationship between a given buyer and seller, has often appeared as an
explanatory variable in these regressions. In the labour market literature
much theoretical interest has focused on the relationship between a
worker's tenure and the wage offered by the firm. A fundamental result
in the human capital theory (Becker, 1975; Hashimoto, 1981) is that
employers reward workers for the accumulation of firm-specific human
capital. If job tenure is a good proxy for firm-specific human capital
accumulation then human capital theory predicts that finns offer
higher wages to workers with long job tenure than to identical workérs
with short job tenure. There also exist alternative theories which have
implications for the effect of job tenure on wage offers. Among this
latter group are models of job-matching (Jovanovic, I 979a, 1 979b),
shirking (Lazear, 1981), and self-selectivity (Salop and Salop, 1976).'
In the rental housing market literature models have been developed
which imply that a landlord's asking-rent is a decreasing function of a
given tenant's tenure in a unit. For example, Guasch and Marshall
(1987) have shown that landlords may offer a discounted rent to
sitting tenants in order to avoid the costs associated with marketing and
renovating a previously occupied unit?
A hedonic regression seems to be a natural way to obtain an estimate
of the parameter of interest - the effect of tenure on price. This note
points out that such a regression may not yield a consistent estimate
of the parameter of interest due to the data collection methodology of
nearly all microeconornic surveys.
* Robert C. Marshall acknowledges financial support from the NSF through Grant No.
SES-8509693.
'Within the job-matching literature the ability of a given employer and employee to assess
the quality of a match is an increasing function of tenure. Matches perceived as mutually 'good'
result in higher wage offers by the firm. The shirking literature analyses worker-firm interaction
within a principal-agent context. Firms implicitly offer an increasing wage-tenure profile in
order to induce workers to exert the optimal level of effort in the current period. In the self-
selectivity literature firms offer low initial wages and an increasing wage-tenure profile in order
to attract only those workers with a relatively low quit propensity.
2See also Merrill (1977), Downs (1980), Lowry (1980), Follain and Malpezzi (1980),
Goodman and Kawai (1981, 1982).
335
To continue reading
Request your trial