|
|
|
Utilization of Contingent Workers and Firm Performance
Paula Alexander Becker
Stillman School of Business
Abstract
The nature of the employment relationship between employers
and employees is undergoing transformation in the new economy. Approximately
30 percent of the civilian work force works in "non-standard" work arrangements,
including agency temporary workers, direct-hire temporary workers, on-call
workers, leased employees, contract company workers, independent contractors,
self-employed workers, and regular part-time workers.1 Workers employed
in these non-standard work arrangements are referred to as contingent
workers. Employers have various reasons for utilizing contingent workers,
ranging from cost control and flexibility in staffing to screening for
qualified direct hires or finding specialized talent.2 This phenomenon
is the focus of a number of studies, and conditions under which externalization
of employment occurs have been identified.3 The question arises whether
the utilization of contingent workers affects firm performance.
Research Methodology
The American Management Association (AMA) conducted a survey
of its member organizations in June 1999 regarding the utilization of
contingent workers. The questionnaire included information on firm size,
whether the firm was engaged in manufacturing, extent of unionization,
and percent of the workforce which is contingent, as well as questions
relating to the reasons for using contingent workers, the departments
in which the contingent workers are used, and the type of contingent worker
employed. Over 1,200 human resource managers of the AMA member organizations
responded to the survey. This analysis was conducted among the publicly
traded firms that responded to the Contingent Worker Survey.4
Firm performance was measured in terms of traditional financial analysis:
earnings per share (EPS), dividends per share (DPS), and price-earnings
ratio (P/E ratio). The null hypothesis, that there are no differences
among firms that use more contingent workers, compared to firms that use
fewer contingent workers, was tested using two-tailed statistical tests,
since this author felt that there was no compelling theory to predict
the direction of observed differences in financial performance.
Results
Utilization of contingent workers IS related to firm performance.
Firms with a higher utilization of contingent workers (equal
to or more than 10 percent of their workforce) had significantly higher
price-earnings ratios in 1999 than firms with a lower utilization of contingent
workers (less than 10 percent of their workforce). Firms with a lower
utilization of contingent workers (less than 10 percent of their workforce)
had significantly lower price-earnings ratios in 1999 than firms with
a higher utilization of contingent workers (10 percent or more of their
workforce). There are no differences in either EPS or DPS between firms
with a higher utilization of contingent workers compared to firms with
a lower utilization of contingent workers. See Table 1.
Cost as a reason for using contingent workers. Companies
for which cost reduction was a more important motivation for using contingent
workers had significantly lower P/E ratios in 1999 and distribute significantly
higher DPS compared to firms for which cost reduction is less important.
See Table 2.
Firms that cited payroll cost reduction as a very important or somewhat
important reason for using contingent workers had significantly lower
P/E ratios than firms that state that payroll cost reduction is not at
all important. This result is particularly significant for non-manufacturing
firms. See Table 3.
Companies that cited healthcare cost reduction as a very important or
somewhat important factor had significantly lower P/E ratios in 1999 compared
to firms that stated that healthcare cost reduction is not at all important.
This difference is greater among non-manufacturing firms and among companies
whose contingent work force is comprised of more than 10 percent of its
total work force. See Table 4. Furthermore, companies that stated that
healthcare cost reduction is very important or somewhat important had
significantly higher DPS in 1999 compared to firms that stated that healthcare
cost reduction is not at all important.
Firms that stated that pension cost reduction is very important or somewhat
important had higher DPS than companies that stated that pension cost
reduction is not at all important, but there were no significant differences
in P/E ratios or earnings per share among firms that stated that pension
cost reduction was an important motivation for using contingent workers.
See Table 5.
The question arises whether firms that invoke cost reduction as a reason
for using contingent workers are in worse "financial health" than firms
for which cost reduction is not an important reason for using contingent
workers. To test this hypothesis, firms were compared on measures of firm
financial health, independent of the outcomes measures for firm financial
performance, for the periods 1997, 1999, 2001, and 2002; that is, periods
before and after, as well as during, the time reference period for the
utilization of contingent workers. The measures of firm financial health
included cash flow from operations relative to common shares outstanding;
total debt relative to common shares outstanding; extraordinary items;
extraordinary items and discontinued operations; Earnings Before Interest
and Taxes and Depreciation (EBITDA),5 beta calculated on a calendar year
basis; and Z scores, a predictor of bankruptcy.6 Comparisons were made
between companies for which cost control, including payroll cost reduction,
health care cost reduction, and pension cost reduction, was an important
reason for using contingent workers, and firms for which cost control
was not an important reason for using contingent workers. For the years
1997, 1999, or 2001, there were no significant differences in cash flow
from operations relative to common shares outstanding between companies
for whom cost control was important compared to companies for which cost
control was not important, but in 2002 the companies for which cost control
was important had higher cash flow from operations relative to shares
outstanding (ANOVA p ¾ .035). For the years 1997, 1999, 2001,
or 2002, there were no significant differences in total debt relative
to common shares outstanding between companies for whom cost control was
important compared to companies for which cost control was not important,
using analysis of variance, but in 1997 the companies for which payroll
cost reduction was important had higher total debt relative to shares
outstanding (T test p ¾ .007) and the companies for which
health care cost reduction was important had higher total debt relative
to shares outstanding in 1997 (T test p ¾ .026). There were
no significant differences for extraordinary items or extraordinary items
including discontinued operations between companies for which cost reduction
was an important reason for using contingent workers compared to companies
for which cost reduction was unimportant in any period. There were no
significant differences in any year in EBITDAM, Earnings Before Interest,
Taxes, and Depreciation on an Annual Basis, between companies for which
cost reduction was an important reason for using contingent workers compared
to companies for which reduction was not important. For the years 1997
and 1999 there were no significant differences in beta between companies
for which cost reduction was important compared to companies for which
cost reduction was not an important reason for using contingent workers.
But in the years 2001 and 2002, companies for which cost reduction was
an important reason for using contingent workers had lower betas (p
¾ .05 and p ¾ .02, respectively). A Z score analysis
showed that companies for which cost reduction was not an important reason
for using contingent workers had significantly higher Z scores than companies
for which cost reduction was important for all years.7 Moreover, companies
for which cost reduction was not an important reason for using contingent
workers had Z scores that were significantly higher than 3.0 for all years
1997, 1999, 2001, 2002, 2003. In addition, companies for which cost reduction
was an important reason for using contingent workers had Z scores in 2002
and 2003 that are not significantly different than a Z score of 1.81,
and that are significantly lower than a Z score of 3.0. It is not likely
that cost control reasons are directly related to company poor financial
health several years later, but rather that the managerial decision-making
processes themselves are causally related to a poorer long term consequence.
Overall there are no significant differences between firms in "financial
health" that would be predictive of differences in firm performance.
Finding specialized talent as a reason for using contingent
workers. Firms that stated that finding specialized talent
is very important had higher P/E ratios than companies that stated that
finding specialized talent is somewhat important or not at all important.
Furthermore, firms that stated that finding specialized talent is very
important or somewhat important give higher DPS than companies that stated
that finding specialized talent is not at all important. These results
suggest that firms that use contingent workers for strategic reasons relating
to the acquisition of human capital gain a competitive advantage. See
Table 6.
The temp to perm track: screening for qualified
direct hires as a reason for using contingent workers. Companies
that stated that screening for qualified direct hires is very important
or somewhat important had significantly lower P/E ratios in 1999 compared
to firms that stated that screening for qualified direct hires is not
at all important. In addition, companies that stated that screening for
qualified direct hires is very important or somewhat important distributed
significantly more DPS in 1999 compared to firms that stated that screening
for qualified direct hires is not at all important. These results suggest
that more effective firms have other mechanisms for hiring qualified employees
than converting temps to permanent employees. See Table 7.
Earnings per share. Earnings per share,
an outcomes variable, was not significantly related to the independent
variables measured in the questionnaire.
Dividends per share. Distribution of quarterly
dividends (dividends per share) is generally higher among firms with lower
P/E ratios. This suggests that distribution of dividends is used as a
means of holding investors. If a firm has a lower P/E ratio, it tends
to offer higher dividends instead. This finding may indicate the strategic
use of dividends distributions by firms with lower P/E ratios; namely,
in order to serve as an attractive opportunity for investors, those firms
with lower P/E ratios distribute more dividends than firms which are more
attractive investments in terms of their price-earnings ratios.
Unionization. Unionized firms have
lower price-earnings ratios and pay higher dividends per share than non-unionized
firms. See Table 8.
Manufacturing. Manufacturing firms
had significantly lower price-earnings ratios in 1999 than non-manufacturing
firms, and paid higher dividends than non-manufacturing firms. See Table
9.
Discussion and Conclusions
This analysis relates the utilization of contingent workers
to firm financial performance among publicly traded companies which participated
in the AMA 1999 Survey on Contingent Workers, concluding that greater
utilization of contingent workers is related to better firm performance,
particular P/E ratios. Relatively few investigations relate firm performance
to the utilization of contingent workers. However, Nayar and Willinger
(2001) recently examined the relationship between firm performance and
increasing use of contingent workers. Their study shows that firms which
increased their use of contingent workers have higher stock return measures
than firms which do not use more than 10 percent contingent workers.8
Their study used a different time period (1978-1991), a different group
of companies or databases, and different measures of firm performance.
Their conclusion, that increased reliance on contingent labor increases
firm profitability, however, is consistent with the findings of this study.
Shulamit Kahn, at Boston University, and her colleagues have also investigated
the relationship between the utilization of contingent workers and firm
financial performance. Although they interpret their findings as "mixed,"
9 firms that used more contingent workers had higher EPS and higher stock
prices than firms which used a lower percentage of contingent workers
(Kahn et al. 2001). In a follow up study, Kahn (2000) found that greater
utilization of contingent workers was related to higher firm productivity.10
The work of Professors Nayar, Willinger, Kahn, and Kahn's colleagues concurs
with the results reported herein.11 The convergence of these four different
studies provides confidence in the conclusion that companies that use
more contingent workers have better financial performance than firms that
use fewer contingent workers.
Factors relating to use of temporary workers have been identified by Davis-Blake
and Uzzi (1993). Davis-Blake and Uzzi found that independent contractors12
were used in jobs requiring "firm-specific or complex technical skills,"
whereas (other) temporary workers were used in situations where there
were "variations in employment needs," requiring flexibility in employment
on the part of the firm. Kahn (2000, 242) also found that the human resource
managers interviewed in her study appreciate the flexibility that the
utilization of temp workers provides to a company. The use of contingent
workers, both as independent contractors and other temporary workers,
provides a firm with the ability to meet variations in employment needs,
without incurring higher fixed costs associated with regular employees,
thereby adjusting to variations in the firm's market, and creating a competitive
advantage to such firms. Matusik and Hill (1998) theorize about the reasons
that a competitive advantage might be created by firms' use of contingent
workers. Their rationale, that contingent workers bring more up-to-date
market-based knowledge to the firms employing them, comport with our finding
that it is particularly the strategic use of contingent workers that is
related to more effective firm performance.
Cost cutting as a reason for the use of contingent workers was not related
to better firm performance in this study.13 Indeed the human resource
managers interviewed by Kahn (2000) acknowledge that the costs of using
contingent workers may be the same as, or greater than, the costs of regular
employees. Although firms may not pay their contingent workers medical
or pension benefits, temp agencies supplying the contingent workers charge
a fee that may be equal to, or greater than, the costs of the benefits.
This study found that firms which state that the reasons for their use
of contingent workers is to lower costs, payroll and healthcare costs
in particular, have lower P/E ratios. On the other hand, firms that use
contingent workers for the purpose of finding specialized talent, a strategic
approach to the use of contingent workers, have higher P/E ratios. The
utilization of contingent workers likely creates a competitive advantage
between firms; hence industry-wide measures might mask differences in
competitive advantage between firms within the same industry.
Given the positive effects of the utilization of contingent workers on
firm performance, it is unlikely that this trend will be reversed. Therefore,
the negative effects of contingent work should be managed.14 Specifically,
policies and programs for contingent workers whereby benefits and pensions
adhere to the individual worker and are portable across employers or even
agencies that place temporary workers should be developed, with a particular
emphasis toward "low end" contingent workers.15
In conclusion, the strategic use of contingent workers is associated with
better firm performance, whereas cost cutting approaches to the use of
contingent workers is not associated with firm competitive advantage.
Since many contingent workers are relatively disadvantaged in terms of
access to healthcare and pension benefits, policies and programs should
be developed to address these relative disadvantages, rather than attempt
to reverse the trend toward the increased utilization of contingent workers.
Acknowledgments
The author thanks the University Research Council, the Institute
for Work, the Institute of International Business, and the Stillman School
of Business for research support of this project.
Notes
1. The number of jobs in the temporary-help supply industry
rose 577 percent, while the total number of jobs rose 44 percent between
1982 and 1998 (U.S. General Accounting Office 2000).
2. See for example American
Management Association 1999.
3. See, for example, U.S. Department
of Labor, Bureau of Labor Statistics 1995; Washington Senate Democratic
Caucus 1999; U.S. Department of Labor, Bureau of Labor Statistics 2001;
and Carre et al. 2000.
4. Of the 1,248 employers responding
to the AMA survey, 220 were publicly traded companies.
5. The COMPUSTAT definition
of EBITDA is:
The annual concept is Earnings Before
Interest and Taxes and Depreciation Annual, Quarterly, or twelve-month moving,
divided by Sales<->Net. This total is then multiplied by 100.
The quarterly concept is Operating Income Before Depreciation divided by
Sales<->Net<->Quarterly. This total is then multiplied by
100. If no value is available for the current quarter, the value for the
previous quarter will be calculated.
The 12 Month Moving concept is Operating Income Before Depreciation 12MM
divided by Sales<->Net<->12MM. This total is then multiplied
by 100. If no value is available for the current quarter, the value for
the previous quarter will be calculated.
6. Z score is a bankruptcy prediction model developed by Edward Altman
at New York University. Data were obtained from Standard and Poor's COMPUSTAT
database, which defines Z score as:
If a value less than 1.81 is returned, than there is a high probability
of bankruptcy. If a value greater than 3.0 is returned, than there is
a low probability of bankruptcy. This item is designed to forecast failure
in the short-term (up to two years).
7. A higher Z score indicates
a lower likelihood of bankruptcy.
8. Nayar and Willinger define
the use of contingent workers by a notation in the COMPUSTAT database at
note 25 that "a firm has 10 percent or more seasonal or part-time workers
in that particular fiscal year" (2001, 666). They measured profitability
in terms of stock returns comparing "buy-and-hold excess returns in a 250-day
period subsequent to the fiscal year in which the reliance on contingent
labor is revealed . . . [with] a prior 250-day period" (2001, 678). Buy
and hold excess returns are defined on pp. 673-74.
9. Kahn, in a more intensive
case study of two companies in the South, found that in the textile industry
the companies which had made "sudden, radical shifts" to the use of temp
workers had declining financial performance, or financial performance that
was no different, in comparison with control companies which had not made
such shifts to temp workers (2000, 235-36). She interprets the case studies
to indicate that her findings reveal a mixed relationship between the use
of contingent workers and firm financial performance. However, extraneous
variables at work in the single company which made the "sudden, radical"
shift to the use of contingent workers and which subsequently experienced
a significant decline in financial performance, may also have accounted
for the decline in its financial performance; for instance, the shift in
employment policy may have been triggered by a need for sudden cost reductions,
which factor also accounted for the reduction in financial performance.
As discussed infra, cost as a motivator in the utilization of contingent
workers was not related to better financial performance in the present study.
10. In the same study, Kahn
(2000, 256-57) found that the use of independent contractors, but not other
categories of temporary workers, was negatively related to firm profitability;
however, change in the use of independent contractors between 1995 and 1997
and change in profits between 1995 and 1997 were unrelated. Moreover, her
study was based on industry-wide measures. Thus, differences between companies
which are more or less competitive based on firm management practices may
be masked by the level of measurement.
11. The author feels that the
findings reported herein are particularly robust since two-tailed tests
of statistical significance were used. See, supra, the section on
research methodology.
12. Independent contractors
were included in the definition of contingent workers in the AMA Survey
used in this study.
13. Similarly, control of labor
costs by downsizing is negatively related to stock price. See Worrel, Davidson,
and Sharma (1991), who found that downsizing by itself was related to declines,
rather than increases, in stock price. Casio, Young, and Morris (1997) examined
the effect of employment downsizing on stock price for the period 1980-1994;
they conclude that employment downsizing alone, in the absence of asset
re-structuring, was related to lower average return on stock price compared
to stable employers in the short run, but related to better returns in the
years following the downsizing.
14. The U. S. General Accounting
Office (2000) study of contingent workers notes that medical benefits and
pensions are significantly lower for temporary workers, other than independent
contractors.
15. Many temp agencies offer
benefits packages to the temporary workers which the agencies contract out.
However temporary workers sometimes rotate between and among temp agencies,
often at the suggestion of the temp agency client organizations where the
temps work, so that portability of benefits is still important for agency
temp workers.
References
American Management Association. 1999. "Contingent
Worker Survey of 1248 Employers." New York: American Management Association.
Carre, F., M. Ferber, L. Golden, and S. Herzenberg, eds. 2000. Nonstandard
Work: The Nature and Challenges of Changing Employment Arrangements.
Champaign, Ill.: Industrial Relations Research Association.
Casio, Wayne, Clifford Young, and James Morris. 1997. "Financial Consequences
of Employment-Change Decision in Major U.S. Corporations." Academy of
Management Journal, Vol. 40, pp. 1175-89.
Davis-Blake, Alison, and Brian Uzzi. 1993. "Determinants of Employment Externalization:
A Study of Temporary Workers and Independent Contractors." Administrative
Science Quarterly, Vol. 38, pp. 195-223.
Kahn, Shulamit. 2000. "The Bottom-Line Impact of Nonstandard Jobs on Companies'
Profitability and Productivity." In Carre et al. 2000.
Kahn, Shulamit, Fred Foulkers, and Jeffrey Heisler. 2001. "Large Companies
and the Changing Use of Temporary Workers: Trends and Impacts on Financial
Measures of Performance." In Working Time in Comparative Perspective--Volume
II: Life-Cycle Working Time and Nonstandard Work, ed. Susan Houseman
and Alice Nakamura. Kalamazoo, Mich.: W.E. Upjohn Institute for Employment
Research.
Matusik, S. F., and C. W. L. Hill. 1998. "The Utilization of Contingent
Work, Knowledge Creation, and Competitive Advantage." Academy of Management
Review, Vol. 23, pp. 680-97.
Nayar, N., and G. L. Willinger. 2001. "Financial Implications of the Decision
to Increase Reliance on Contingent Labor." Decision Sciences, Vol.
32, pp. 661-81.
U.S. Department of Labor. Bureau of Labor Statistics. 1995. "Contingent
Worker and Alternative Employment Arrangements, Report 900."
———. 2001. Occupational Outlook 2001. http://www.bls.gov/oco/home.htm.
U.S. General Accounting Office. 2000. "Contingent Workers: Income and Benefits
Lag Behind Those of Rest of Workforce." Report to Senators Edward Kennedy
and Robert Torricelli. GAO/HEHS-00-76.
Washington Senate Democratic Caucus. 1999. "Short Term Work, Long Term Consequences:
How New Work Arrangements Are Transforming Washington Labor Markets."
Worrel, D. L., W. N. Davidson III, and V. M. Sharma. 1991. "Layoff Announcements
and Stockholder Wealth." Academy of Management Journal, Vol. 34,
pp. 662-78.
|
