A CRITIQUE OF “AN ECONOMIC ASSESSMENT OF MANDATED SICK LEAVE LEGISLATION (#080474) IN THE CITY OF PHILADELPHIA” Lonnie Golden, Professor of Economics and Labor Studies, Penn State-Abington And Dr. Stephen Herzenberg, Keystone Research Center ([email protected]; 717-805-2318) William Dunkelberg, the Chief Economist of the National Federation of Independent Business (NIFB), has just released a study of the cost of implementing paid sick days in Philadelphia.1 This policy brief analyzes the Dunkelberg study and finds that it has a combination of limitations and errors that makes it of little value for policymakers considering enacting paid sick days.  It contains a mistake that doubles the estimated potential maximum cost of the bill.  It makes an implausible assumption about how many paid sick days workers will actually take and fails to use readily available data on how much workers actually use paid sick time. (Many workers don’t take all their permitted sick days because they view them as insurance available in cases of more serious illness.)  It fails to capitalize on arguably the best source of information—actual experience following the paid sick days ordinance implemented in San Francisco in February 2007.  It overstates the adverse effect on jobs by using a questionable methodology that exaggerates the sensitivity of employer demand for labor to the implementation of paid sick days.  It does not consider potential cost-saving benefits for employers that would result from paid sick days. These include reduced turnover because paid sick days leads to fewer discharges and quits, and greater worker effort and productivity due to an improved employment relationship.  It neglects the potential costs of “presenteeism”—low-productivity performance by sick workers on the job—and the cost of the workplace spread of communicable illnesses.  It includes implausibly high estimates of the cost of compliance with a paid sick days ordinance. Dunkelberg also analyzes the cost of an ordinance which requires five days of paid sick days for employers with 10 or fewer employees and nine days for employers with 11 or more employees. The current Philadelphia proposal, based on proposed amendments to be voted on in City Council session, provides for a maximum of four days of paid sick days for employers with 10 or fewer employees and seven days for employers with 11                                                              1

 For perspectives on the National Federation of Independent Business, see.  http://bostonreview.net/BR33.4/kazee.php; http://boss.blogs.nytimes.com/2009/08/26/whom‐does‐the‐nfib‐ represent‐besides‐its‐members; and http://boss.blogs.nytimes.com/2009/09/03/how‐the‐nfib‐represents‐its‐ members‐part‐ii .   

Page 1 

or more employees. With these amendments, even if one accepted Dunkelberg’s analysis, his cost estimates would thus be inflated by more than 25%. The text below explains the limitations and errors in Dunkelberg’s analysis in more detail. (The text discussion does not map one-to-one with the list above because information from San Francisco is weaved in as appropriate not presented in a standalone sub-section.) The last section of the paper includes a broader discussion of Dunkelberg’s static, short-term approach, which is the same basic perspective that has been used to argue against all advancements in labor standards starting with prohibitions on child labor. The last section offers a longer-term and dynamic perspective on the relationship between advanced labor standards, such as paid sick days, and economic growth. Of course, the impact of paid sick days on the economy should not be exaggerated on the positive side just as it should not be exaggerated on the negative side. Properly understood and implemented, nonetheless, paid sick days can be part of Philadelphia’s rebranding as a center of innovation, a magnet for attracting and retaining young and skilled labor, fueled in part by 21st century rather than 19th century employment policies and practices. 1. Double Counting the Cost of Paid Sick Days The most basic mistake in the Dunkelberg study is its double counting of the maximum cost of paid sick days. Dunkelberg first estimates the cost of paid sick days assuming all workers who are not currently provided paid sick days by their employer use the maximum amount available under the law. He estimates that this would cost $350 million. Dunkelberg then adds to this the cost of substitute workers, an additional $350 million if all employers hire substitute workers (see Table 4, p. 7, left-hand column). The workers, however, would have already earned the first $350 million if they were working. So the maximum additional cost isn’t $350 million plus $350 million but just $350 million. In addition to this mistake, the assumption that all firms—or even 25% of firms—would hire replacement workers is inconsistent with actual experience. In San Francisco, 66% of all firms never hired a replacement and 90% hired replacements rarely or never.2 Only 1.2% hired replacement always. Small firms never hired replacements 68% of the time. Across all employers, these figures indicate that workers are replaced well under 10% of the time.3

                                                             2

 Data below on San Francisco’s experience with paid leave come from Robert Drago, Ph.D. and Vicky Lovell, Ph.D.,  San Francisco’s Paid sick days Ordinance: Outcomes for Employers and Employees (Washington DC: Institute for  Women’s Policy Research, February 2011), see Appendix Table 3, p. 33.  3  If we assume that “rarely” means workers are replaced 5% of the time and “frequently” means workers are  replaced one third of the time, then the weighted average replacement rate is 6%. 

Page 2 

2. How Many Paid Sick Days Will Workers Actually Take? The Dunkelberg study itself says that there is no way to know how many paid sick days workers will actually take—but then goes ahead and generates estimates assuming they’ll take it all. In fact, experience in San Francisco and national data provide two readily available data sources that indicate how much paid sick days workers are likely to take. Recall that the San Francisco ordinance permits nine paid sick days for workers at firms with more than 10 employees and five for those with 10 or fewer.  In a survey of 624 San Francisco workers4, 25% of workers did not take any paid sick days in the previous 12 months.  San Francisco workers took a median of only three paid sick days.  National survey data indicate that, with a nine-day cap on paid leave, workers will take an average of 3.1 days leave.5 The same source indicates that workers at small employers (with 10 or fewer employees) take only an average of 2.2 days per year.  Bureau of Labor Statistics data indicates that workers can take up to eight days paid leave each year but take only 27 hours—i.e., 3.4 days assume an eight-hour day. Research in San Francisco indicates that the reason many workers don’t take all their permitted sick days is that view the benefit as insurance available in cases of more serious illness. Additional indirect evidence that employees will not take the maximum allowed number of sick days comes from the fact Americans leave much of their annual paid vacation time on the table, unused (and not cashed out). More than one-third of employees (36 percent) did not plan to use their full vacation time allotment.6 (In addition, the proposed Philadelphia bill states explicitly that no cash out of unused time is required of employers.) 3. Lost Productivity During Paid Sick Days Is Compensated for When Workers Return or By Colleagues Many workers on paid sick leave who are not replaced make up for the lost productivity by performing more intensively, or slightly longer hours, when they return. They may also perform some work while out sick just as they often do on “vacation.” Over 20                                                              4

 Table 2, page 9, Robert Drago, Ph.D. and Vicky Lovell, Ph.D., San Francisco’s Paid sick days Ordinance: Outcomes  for Employers and Employees (Washington DC: Institute for Women’s Policy Research, February 2011)  5  Personal communications with Kevin Miller of the Institute for Women’s Policy Research (IWPR).To estimate paid  leave for workers’ own illnesses and for visits to doctors, IWPR relies on the 2008 National Health Interview Survey  conducted by the Centers for Disease Control and Prevention. Information on the survey is available at  http://www.cdc.gov/nchs/nhis/about_nhis.htm. To estimate time off for care of other family members, IWPR  relies on the 2000 update of the U.S. Department of Labor, Family and Medical Leave Surveys, made available on  April 12, 2005.  6  Families and Work Institute, 2006, conducting a survey in 2004 of 1,003 adults employed full or part time on  payroll 

Page 3 

percent7 to upwards of 50 percent8 of workers acknowledge planning to do or doing some work while on “vacation.” To the extent that employees make up for lost productivity when they return or are at home, then the actual cost to the employer shrinks to the cost of replacements—less than $35 million. In addition, the colleagues of an employee out sick may pick up the slack. 4. Reduced Turnover Costs and Increased Worker Loyalty9 Dunkelberg also fails to consider possible offsetting benefits to employers. One benefit is reduced turnover. A lack of paid sick days can lead to workers who don’t show up being discharged or lead employees to quit a job. A 1993 study found that having paid sick days reduces voluntary job mobility (quits) by 5.6% for married men and 3.6% for married woman.10 Worker turnover costs are high for entry-level workers relative to labor costs because the human resource and management staff who oversee recruitment are paid more than the entry-level employees. Vicky Lovell estimates that paid sick days nationally could reduce employer hiring and training costs by about $25 billion annually.11 Based on the city’s share of national employment, Philadelphia employers hiring and training costs would fall about $100 million.12 Lovell also notes that paid sick time not only reduces hiring and training costs but also increases worker loyalty, which generates additional productivity benefits. In San Francisco, roughly two thirds of employers in every firm-size range were “somewhat” or “very” “supportive” of paid sick days three years after its implementation.13 Even in the hotel, food services and retail and wholesale trade industries, more than 60 percent of employers were somewhat or very supportive.

                                                             7

 Families and Work Institute, 2006   American Management Institute  9  This section is adapted from Elise Gould, Kai Filion, and Andrew Green, Paid Sick Days: Doing Well by Workers in  Also Good Business, Economic Policy Institute, Washington, D.C., forthcoming.  10  Philip Cooper and Alan Monheit, “Does Employment‐Related Health Insurance Inhibit Job Mobility?” Inquiry, Vol.  30, No. 4, pp. 400‐416.  11  Vicky Lovell, Valuing Good Health: An Estimate of Costs and Savings for the Healthy Families   12  In 2010, employment in Philadelphia equaled 577,000 and unemployment national equaled 139,064,000.  Multiplying the Philadelphia share of national employment times $25 billion yields $104 million.  13  Robert Drago, Ph.D. and Vicky Lovell, Ph.D., San Francisco’s Paid sick days Ordinance: Outcomes for Employers  and Employees (Washington DC: Institute for Women’s Policy Research, February 2011)  8

Page 4 

5. The Productivity and Public Health Costs of Sick Workers Coming to Work Dunkelberg ignores the cost of "presenteeism"—the productivity loss when sick workers come to their job—and the health costs that result when attendance contributes to the spread of contagious illnesses and to outbreaks.14 The productivity decline when ill workers come to work has been estimated at 7% to 28%.15 The total cost of workers coming in to work while ill has been estimated to be 1.8 times the cost to employers of absenteeism.16 By contrast, staying home promotes a potentially quicker and fuller recovery. Analysis of the 2007 National Health Interview Survey found that providing paid sick days would have positive health results for employees and the public, by controlling the spread of pandemic flu, mainly because workers without paid sick days are more likely to go to work sick with an illness like the flu and are more likely to send a sick child to school than workers who have paid sick days. Providing paid sick days would also reduce the spread of food borne illnesses from restaurants, healthcare and daycare facilities.17 6. Implausible Compliance Costs On page 5 and 6, Dunkelberg discusses compliance costs for firms that do not currently offer paid sick days and firms that do. The presentation is not very clear and thus open to misinterpretation.18 If we interpret his tables and text currently, however, he presents unrealistically high estimates of compliance costs for firms that already have paid sick day benefits. In Table 3, Dunkelberg shows estimated compliance costs for the roughly half of small firms (with less than 50 employees) that already offer sick leave plans to be 38 cents per hour for small firms. This is almost exactly equal to his direct compliance cost for small firms without sick leave “…about $200 million per year.” This appears to mean that modifying payroll reporting to ensure documentation of compliance with the new law will cost as much as the full cost of paid sick days for firms that do not comply. That is simply not believable. With modern payroll reporting, compliance will be no more difficult that paying overtime or Social Security, and amount to at most a penny or two per hour. 7. Exaggerated Employment Effects To project the employment effects of paid sick days, Dunkelberg uses a single estimate (from 1998, 13 years ago) of the employment effects of the minimum wage increases.                                                              14

 Widera, Eric, Chang, Anna and Chen, Helen. Presenteeism: A Public Health Hazard Journal of General Internal  Medicine. Volume: 25 Issue: 11, 2010‐11‐01, 1244‐1247.  15  Paul Hemp, “Presenteeism at Work: But Out of It,” Harvard Business Review, October 2004, reprint R0410B.    16  Cooper and Dewe, 2008.  17  Rutgers, 2010.  18  It is also unclear how Table 3 relates to the compliance costs listed in Table 4. Also unclear is why compliance  costs table 4 are $52 million if workers take all available paid sick days, $103 million if they take 4 days, and $258  million if they take 4 (at small firms, presumably) and 6 (at large). 

Page 5 

That is, he takes his estimate of the cost increase resulting from paid sick days, then he uses one 1998 estimate of the sensitivity of the employment of minimum wage workers to cost, to justify the claim that paid sick days would cost 4,000 jobs. There are two problems with this procedure. The first is his use of an inflated cost estimate. Once the cost estimate is adjusted for how much workers actually take paid sick days, the compensating productivity increases when workers return and from peers, and the offsetting positive productivity benefits of worker loyalty and improved workplace relations, even with Dunkelberg’s employment elasticity, the job loss would be measured in hundreds, if that, not thousands. The second problem is that Dunkelberg has picked one estimate in a large literature. Based on extensive study of actual natural experiments in which state’s increased their minimum wage, the consensus among economists has shifted towards the view that modest increases in the minimum have little or no impact on employment.19 With regard to employment impact, San Francisco again provides real-world evidence. Since 2007, the number of jobs the city in both small and large businesses has increased relative to other counties. National comparisons find no relationship between the availability of paid sick days and either unemployment or competitiveness, nor is there a statistically significant effect of mandated paid sick days or leave on national unemployment rates, when considering the mandated duration of sick leave.20 An Alternative View of Labor Standards and Growth: Towards an Innovative City Dunkelberg takes a static view of labor standards focused on short-term costs. His perspective is similar to the view that has fueled opposition to labor standards going back to the first fights about prohibitions on child labor. An alternative view of labor standards was articulated by MIT economist Michael Piore in a well-known article some two decades ago.21 Piore points to the role of labor standards in shaping employer “business strategy” and influencing employers to embrace business strategies that lead to higher levels of service, productivity, and innovation. In today’s Philadelphia, lowcost, low-standard business strategies are likely to translate into low rates of service, quality, and innovation. They are also likely to perpetuate the large numbers of povertywage, dead-end jobs in the city. By contrast, higher-standard business strategies offer greater opportunities for increasing economic vitality and improving job opportunities in the city.                                                              19

 One of the major publications that began to shift the consensus was David Card and Alan Krueger, Myth and  Measurement: the New Economics of the Minimum Wage (Princeton: Princeton University Press, 1995).   20  Schmitt, Rho, Earlie and Heymann, 2009.  21  Michael Piore, “Labor Standards and Business Strategies,” in Stephen A. Herzenberg and Jorge F. Perez‐ Lopez (eds), Labor Standards and Development in the Global Economy (Washington, D.C.: Bureau of  International Labor Affairs, U.S. Department of Labor, 1990 (reprinted 1994)). Similar arguments are  incorporated into more recent writing about environmental standards by Harvard Business School Professor  Michael Porter.   

Page 6 

Clearly, the conditions the will produce more high-standard businesses are complex. Moreover, support for businesses to adopt high-standard approaches will be at least as important as limits on their ability to perpetuate low-standard ones. Nonetheless, it seems likely that implemented advanced labor standards, including paid sick days, is at least compatible with efforts to promote innovation in the city and, at best, a way to distinguish the city to forward-looking entrepreneurs and high-skill workers as a 21st century city not a 19th century one. In this regard, it is worth that PriceWaterhouseCoopers just named San Francisco, with its paid sick days law on the books, the 3rd best city globally for business and innovation in the world.22 To sum up, when you do the numbers carefully, it becomes clear that the sky will not fall if Philadelphia adopts a paid sick days bill. In fact, if the bill is implemented well, it could just make the Philadelphia sky just a little bit bluer and sunnier. For the workers and families who directly benefit, of course, the bill could make their quality of life more dramatically better.

                                                             22

 online at http://www.pwc.com/us/en/cities‐of‐ opportunity/2011/pdfdownload.jhtml?checked=B1FD1BB20BD87A9BA6C17171618B79F8.  

Page 7 

References Boots, S.W. Martinson, K., Danziger, A. Mandating Leave: Employers’ Perspectives on San Francisco’s Paid sick days Policy, 2008. Clemans-Cope, L., CD Perry, GM Kenney, JE Pelletier and M. S. Pantell (2008),“Access to and Use of Paid sick days among Low- Income Families with Children”, Pediatrics 122,480–86. Drago, Robert and Vicky Lovell, San Francisco’s Paid Sick Leave Ordinance: Outcomes for Employers and Employees, IWPR A138, The Institute for Women’s Policy Research, Washington, DC. February 2011. Families and Work Institute, 2005. Overwork in America: When the Way We Work Becomes Too Much. New York, NY. Gruber, Jonathan. The Incidence of Mandated Maternity Benefits. The American Economic Review, Vol. 84, No. 3, (Jun., 1994), pp. 622-641. Heymann, Jody, Hye Jin Rho , John Schmitt , Alison Earle, Ensuring a Healthy and Productive Workforce: Comparing the Generosity of Paid Sick Day and Sick Leave Policies in 22 Countries, International Journal of Health Services, Volume 40, Number 1 / 2010, Pages: 1 – 22. Ruhm, Christopher J., 1998. "The Economic Consequences Of Parental Leave Mandates: Lessons From Europe," The Quarterly Journal of Economics, vol. 113(1), pages 285-317, February. Rutgers University Center for Women and Work, 2011, A Health Impact Assessment of Paid Sick Days Policy in New Jersey, May. Scheil-Adlung Xenia & Lydia Sandner The case for Paid Sick Leave, World Health Report (2010) Background Paper, 9. Weil, David, Improving Workplace Conditions through Strategic Enforcement: AReport to the Wage and Hour Division, May 2010, Boston University, Boston MA. Widera, Eric, Chang, Anna and Chen, Helen. Presenteeism: A Public Health Hazard Journal of General Internal Medicine. Volume: 25 Issue: 11, 2010-11-01, 1244-1247.

Page 8