Tobacco Documents Online

TIMN 358269

z 14 WORKING PAPERSERIES = THE SOCIAL SECURITY COST OF SMOKING John B. Shoven Jeffrey 0. Sundberg John-P. Bunker Working Paper No..2234 J NATIONAL BUREAU OF ECONOMIC RESEARCH, INC. TIMN 358270

v The National Bureau of Economic Research is a private, non-profit, non-partisan organization engaged in quantitative analysis of the American economy. This paper has not undergone the re+riew accorded official NeER publications; in particular, it has not been submitted for approval by the Board of Directors. it is intended to make results of NBER research available to other economists in preliminary form to encourage discussion and suggestions for revision before final publication. ~ 1~-~ TIMN 358271 t

NBER WORKING PAPER SERIES THE SOCIAL SECURITY COST OF SMOKING John B. Shoven Jeffrey 0. Sundberg John P. Bunker Working Paper No. 2234 NATIONAL BUREAU OF ECONOMIC RESEARCH 1050 Massachusetts Avenue Cambridge, MA 02138 May 1987 Paper presented at the National Bureau of Economic Research Conference on the Economics of Aging, March 19-22, 1987, New Orleans, LA. This work originated when John Shoven and John Bunker were Fellows at the Center for Advanced Study in the Behavioral Sciences in 1984-85. The research reported here is part of the NBER's research program in Taxation. Any opinions expressed are those of the authors and not those of the National Bureau of Economic Research. TIMN 358272 ~

NBER Working Paper #2234 May 1987 The Social Security Cost of Smoking ABSTRACT Our paper is an examination of the Social Security cost of smoking from-an individual point of view. It is well known that smokers have a shorter life expectancy than nonsmokers. This meaas that by smoking they are giving up potential Social Security benefits. We estimate this cost and consider the effects on the system as a whole. We use mortality ratios, which relate the annuall death probabilities of smokers and nonsmokers, and the percentage of smokers in each age group to break down the life tables for men and women born in 1920 into the approximate life tables for smokers and nonsmokers. We then calculate expected Social Security taxes and benefits for each group, using median earnings as a base. We find that smoking costs men about $20,000 and women about $10,000 in expected net benefits. The implication of this for the system as a whole is that the prevalence of smoking has a direct effect on the financial viability of the system; every decrease in the number of smokers in society increases the system's liability. Changes in smoking behavior should be recognized as affecting the system. John B: Shoven Jeffrey 0. Sundberg John P. Bunker Department of Economics Department of Economics Stanford University Stanford University Stanford University School of Medicine Encina Hall, 4th Floor Stanford, CA 94305 HRP Building Room #7 Stanford, CA 94305 (415) 723-3712 (415) 725-0959 Stanford, CA 94305 (415) 723-6426 TIMN 358273

1. INTRODIICTION Smoking in the United States is associated with enormous costs to society. The Congressional Office of Technology Assessment has estimated the annual cost of medical care for smoking related illness at $15 to $30 billion, and that smoking relatad illness is responsible for an additional $49 to $70 billion in lost productivity. There are also substantial costs to tha individual who smokes in tarms of lost wages over a life time., primarily affecting those who - - die of smoking related disT -- --- -f ease while still active wage earners in the work force. Costs to tha individual also include approximately $500 to $1,000 per annum for pack and two pack-a-day smokers to purchase cigarettes. An additional cost to the individual is the loss of Social Security benefits as a result of smoking-induced loss of life expectancy. The data presented herein estimate the magnituda of this loss for single and married man and women born in 1920 and 1923, respectively. . Whila most of the previous literature on the costsof smoking and the benefits of quitting has overlooked the implications of smoking behavior on pension plans (see, for example, Oster, Colditz, and Kally (1984)), this is by no means universal. Gori at al. (1983) estimated that the savings realized by Ford Motor Company if the health of their employees improved (in terms of less expensive medical insurance, disability insurance, and life insurance) would be much smaller than the additional pension costs due to their increased longevity. Atkinson and Townsend (1971) noted that the financial benefits the British National Health Service would enjoy if there was a forty percent reduction in smoking in Britain would be more than offset by the increased cost of retirement pensions. In this paper, we examine the Social Security consequences of smoking from the individual or household perspective. From that vantage point, Social Security can be thought of as a prepaid life annuity. Contributions or taxes TIMN 358274

are collected during one's work life which entitle on. to an indexed life annuity beginning at age 65. The anauity can be commeaced at age 62 with a roughly fair actuarial adjustment, and can be started at an age beyond 65 with somewhat higher benefits reflecting the shorter expected remaining lifetime. In general, the system is not actuarially fair (favoring some cohorts relative to others, those with low incomes or short covered careers relative to others, and aarrieds, especially ons.eaxn.r couples, relative to singles). • , . . _.._ ..._,~ . . . ... _ , _ . . . .. .. Our point is that the system is unfair in a way very relevant to the decision of whether or not to smoke. Social Security does not have separate b.ae it structures for smakers aad nonsmokers even though sackers have a much ry lower chance of reaahing retirement age and a shorter expected length of retirement conditional on reaching that age. The Office of Technology Assessment (1985) estimated that 273,000 people died in the United States in 1982 of smoking related disease. Of those, 44 percent, or 121,000 died before they reached their 65th birthday. They may have never collected anything from Social Security. If they were married, their spouse may collect widov's benefits, but it is clear that their premature deaths greatly reduce their return on their participation in Social Security. Smoking also affects the Medicare portion of the Social Security system. While we concentrate on OASI, it is probably worth noting that the health insurance component is similarly affected. Many estimates of the effect of smoking on the total demand for health care services in the country find that it is small in the long run. Smokers certainly experience more health problems per year of life, but this is offset by the fact that they live fewer years. With a lower incidence of smoking, there would be more elderly who require additional health care services. The reduced demand caused by the improved health status of the former smokers is offset by extra care needed by the 2 TIMN 358275 ..~.~

additional elderly. There might be some initial reduction in the demand for health care if smoking was reduced. The improvements in health status would presumably occur before the age structure was significantly altered. However, in the long run the two effects offset each other. Despite the fact that total health care dsmand may be little affected by smoking, Medicare's finances are almost certainly affected. The reason is that it is a prepaid health insurance annuity for those aver 65. Medicare does not bear the higher health costs of nonelderly sac1cers, but benefits financially from the fewer numbers of elderly due to smoking. The other side of the coin is that smokers, as with their retirement benefits, pay while they work for old age health insurance which they are less likely to collect, or to collect for a shorter period than nosumokers. Wright (1986) estimates that each person who quits smoking increases the deficit faced by the HI component of Social Security for just these reasons. • Our study is the QASI analog of Wright's HI research. We assemble separate life tables for smokers and nonsmokers and then estimate the Social Security taxes, benefits, and transfers for members of the 1920 birth cohort. We do this for those who earn median wages for their age and cohort and for those who eara 60 percent of the median, in each case beginning at age 20. The results can be previewed by saying that we find the expected loss in net Social Security benefits accompanying smoking to be very large relative to the other costs of smoking. The loss exceeds the lifetime costs of cigarettes, is large relative to the estimates of the medical costs and lost wages due to excess morbidity and mortality, and is perhaps tan times greater than the corresponding Medicare figures of Wright. The next section of the oaper briefly reviews what is laiown about the effect of smoking on mortality. It discusses disease specific effects and also our technique of using mortality ratios to yield approximate separate life 3 TIMN 358276

tables for smokers and nonsmokers. The third section of the paper describes our simulation procedure for calculating the Social Security costs of smoking. It presents separate results for single individuals, one-earner and two-earner couples because of their separate treatment by Social Security. We conclude the paper with an interpretation of what our findings imply about thu private and social incentives to quit smoking. I 2: EFFECT OF SMOKING ON MOSTALITI There can be no statistical doubt that smoking is associated with increased mortality hazard rates. The overall finding of the 1979 Surgeon General's report on tiae subject was that ths mortality of all male cigarette smokers is about 170 percent of that of male nonsmokers. For two-pac3r a-day smokers, the average mortality ratio is 200 percent. For particular diseases the relative hazard is even greater~. For-exampls, two separate studies find smokers are between 9 and 15 times more likely than nonsmokers to die of lung cancer (Lubin at al. (1984) and Cowell and Hirst (1980)). The risk of dying of arterosclerotic and degenerative heart disease and myocardial insufficiencies has been estimated at 2.7 times as great for smokers as nonsmokers (Cowell and Hirst (1980)). There is further evidence that there is a significant interai~cion between smoking and other environmental factors such as exposure to asbestos. The finding is that while smoking is a major cause of lung cancer, smoking combined with other assaults (such as industrial exposure) greatly incra-Uses the mortality hazards (Schneiderman and Lsvin (1974). Our development of separate life tables for smokers and nonsmokers utxflizes the findings<of E.C. Hammond (1966) regarding the effect of smoking on mortality. Working for tize American Cancer Sociery, he conducted a comprehensive four-year study tracking a population of over one million subjects. He determines the death rates and the prevalence of certain causes of death for 4 - ---~" TIM~ 358277

smokers and nonsmokers of many different characteristics. The technique was to examine death certificates for the cause of death and to request information from the attending doctor whenever cancer was mentioned on the death certificate. Hammond's results are a very detailed set of mortality ratios1 for different types of smokers and for several different causes of death. In 1959 and 1960, Fiammond enrolled over one million volunteers from twenty-five different states to provide data on mortality. Subjects were classified by sex, age, type of tobacco smoked (cigarette, cigar, pipe, or none), age at which subject began smoking, daily amount of smoking, and degree of smoke inhalation. Each subject was contacted annually for four years to track the number and timing of fatalities in each group. Death certificates were received for over 97% of reported deaths to provide better information as to causes of death. IIsing the accumulated data, Hammond combined subjects with similar characteristics into five-year and ten-year age cohorts, and divided the number of deaths in each cohort during the study period by the number of "person- years" experienced in each cohort. This provided cohort death rates over the period for groups of similar age and sex, and varying smoking habits. This allowed Hammond to calculate mortality ratios for different groups. A sample r, of his findings is shown in Table 1. The separate mortality tables that we have produced are contained in the Appendix to this paper. The basic life tables used are the cohort life tables for men and women born in 1920, as estimated by the Social Security Administration. The mortality hazards are shown in column 8 of the appendix table for men and women. If we let Q~(4) represent the one-year death probability for males as a function of age (similarly Qw(a) for females), TIMN 358278

Table 1 Hortality Ratios for Smokers as Detesmined by E. C. Ha-ond Current Number Per Day Ag. 35-44 45-54 55-64 63-74 75-84 lien vi.th History of Only Cigarette Smoking 1-9 * 1.84 1.33 1.50 1.36 10-19 1.36 2.26 1.92 1.65 1.55 20-39 1.91 2.41 2.05 1.71 1.26 40+. 2.59 2.76 2.26 1.81 * Qom.n with History of Only Cigarette Smoking 1-9 0.90 0.95 0.99 1.09 1.07 10-19 0.97 1.22 1.31 1.18 1.21 20-39 1.35 1.54 1.46 1.51 * 40+ * 1.96 * * * * Signifies a very low number of expected deaths (small sampl• or low death rate). SOURCE: Hammond, p.133. 6 ~ TIMN 358279

let e(a) represent the fraction of inen who smoke as a function of age, and iSm(a) represent the mortality ratio of male smoker to nonsmoker as a function of age, then Qa (a) QNS(a) - 1 - fa(a) (1 - 2e(a) ) and QS(a) - Ie(a) x QNS(a) where (a) is the snnual death probability of male nonsmokers as a function of age and QS(a) is the annual death probability for male smokers. The formulas for women are identical with all the superscripts changed to w's. The appendix tables displaq_tlze assumptions for km(a) and Hw(a), derived from Fi-- ond (1966), and for fa(a) and fw(a) for the 1920 cohort, derived from Harris (1983). They also show the results for QNS(a), QS(a), QNS(a), and QS(a). Table 2 offers some summar,q statistics based on these derived life tables. Our life tables for the 1920 birth cohort show that 85,758 males and 88,787 females out of 100,000 births live to age 20. It is well known that smoking affects the mortality of women less than men. That is partially due to the fact that women smokers smoke less, inhale less, and are more likely to smoke filter cigarettes. Again, out of 100,000 births, 53,051 male smokers (who began smoking at age 20) survive until age 65, whereas 67,464 male nonsmokers survive until that traditional retirement age. Conditional on living to age 20, almost 79 percent of nonsmokers make it to 65, whereas slightly less than 7 TIMN 358280

Table 2 Life Ezpectaacy, Median Age at Death, And Surviving Population at Ages 20 and 65 For 1920 Birth Cohort Survivors Surrivors Xediaa Median at age 20 at age 65 Life Life age of age of out of out of expectancy expectancy death death 100,000 100,000 conditional conditionall conditional conditional births births on age 20 on age 65 on age 20 on age 65 man Smokers 85,758 53,051 68.7 78.8 70 77 Noasmokars 85,758 67,465 75.1 81.5 77 81 Qomea Smokers 88,787 ' 69,303 77.2 84.6 79 84 Nonsmokers 88,787 74,461 80.5 86.6 84 86 8 . .._, .r TIMN 358281 .

62 percent of smokers do so. At age 20, male smokers have a life expectancy 6.4 years shorter than male nonsmokers, and a median age of death 7 years yotmger. Conditional on surviving to 65, male smokers have a remaining life expectancy which is 2.7 years less than their nonsmoking cohort members. The figures for women show that th. life expectancy difference at age 20 is 3.3 years, while the difference at age 65 is 2 years. Our life tables cannot sort out causality and correlation. It is certainly true that smokers would not become identical to nonsmokers if they stopped smoking. Smokers drink more alcohol than nonsmokers, have a higher incidence of suicide, and, in general, may face higher mortality risks than nonsmokers for reasons correlated with smoking but not caused by smoking per se. We have not been able to separate these effects, although it is our belief that most of the higher mortality risks faced by smokers are due directly to the cigarette consumption. However, it should be kept in mind when interpreting our results that we have attributed all of the mortality difference to the cigarette consumption. 3. SIMULATION OF THE SOCIAL SECIJRIIY COSTS. OF SM0KING. We examine the Social Security consequences of smoking for 100,000 men born in 1920 and 100,000 women born in 1923. The three year difference approximates the average age gap in marriage for this cohort. The 1920 cohort life tables compiled by Social Security are taken to be applicable to the men and women in our study. We calculate the Social Security outcomes separately for single men, single women, and one and two earner couples. We assume that each person's probability of death is given by the life tables, and therefore' is independent of the status of their spouse. The number of women who become widowed in each year until the husbands.retire is noted, and each "widow cohort" is then tracked as a separate population. This is necessary, because 9 TIMN 358282

at retirement widows must choose between a benefit based on their own work record and one based on that of their spouse. In the case of the one earner couples, we assume that the husband is employed until retirement or death. If the husband should die before retirement, the widow is assumed to work until retirement or death. Qe have not been very sophisticated in developing our earnings profiles. The earnings series used are median earnings for men and women working full- time year-round, taken from the Census Bureau Current Population Report P-60, No. 142 (1982). Earaings before 1955 and 1982 are estiaated using a related series from the Department of Labor's Employment, Hours, and Earnings report. The earnings series are then adjusted to reflect a life-cycle pattern of lifetime earnings, using Cansvs data on mean incomes for different age groups from Report P-60, No. 137. Our analysis for low wage earners examines those who earn 60 percent of the median earnings profiles. There are several factors which could be added to our earnings series. First, we do not take into account the effect of th. increased morbidity of smokers on their earnings pattern. This is probably a relatively minor adjustment, but one which is conceptually desirable. Second, nonworking wives entering the workforce when widowed are assumed to immediately earn the median (or 60 percent of the median) amount for their age. This is certainly optimistic regarding their prospects. Finally, we do not take into account spells of unemployment, employment in the uncovered sector, or disability. The surviving members of the cohort are assumed to retire at age 65 and begin to receive benefits based on the 1985 Social Security law. We assume that the initial benefit received is fully indexed for inflation for their remaining life. The women in the simulations recire three-years later, simply reflecting that they are three years younger than the men. Because the median earnings of men exceed those for women, the Social Security OASI benefit based on a man's 10 TIMN 358Zg

earnings history exceeds the benefit based on a woman's work record. As a result, husbands and widowers will always elect to take their own benefit. tJives choose between their own benefit and one-half of their husband's, while widows may elect to receive their own benefit or the benefit which their husband would receive were he alive and had not worked since the year he actually died. In other words, a woman whose husband died in 1965 could take the ben,efit he would be receiving had he stopped wor3cing in 1965 and lived to receive his benefit, or she could take her own benefit. In the two-earner cohort, her benefit is based on her earnings from 1940 to her retirement in 1988, while in the one-earner cohort her benefit is based on a shorter work history, 1965 - 1988, since we assume she only begins work upon her husband's death. This means that a vidow's benefit may depend on when her husband died (and in the one-earner case must depend on it), necessitating our keeping track of the •widow cohorts' mentioned above. Wives over the age of 65 whose husbands are still slive will always receive one-half of their husband's benefit in the one-earner family, since they have no earnings history of their own. In the two-earner case, wives-will take their own benefit since their benefit exceeds half their husband's, given our earnings series. All benefits are calculated in real dollars, so compar:&on of 1985 and 1988 benefits is valid. The results for singles are shown in Table 3. All figures are stated in 1985 dollars, and the real discount rate used for cash flows occurring at other times is 3 percent. With those assumptions, the figures in the upper portion of the table for single men with median wage profiles in this cohort show that nonsmokers can expect to receive a net transfar from Social Security of $3,436, while the expected benefits received by smokers fall $17,782 short of the expected contributions. ?,11 of these figures are conditional on having survived to age 20. The Social Security cost of smoking for single men with median Il -- _ --- ~ TIMN 358284 ';

Table 3 Present Value of Social Sacuritp Benefits and Taxes (in 1985 $) And Intaraal Rata of Asturn to the Social Sacurity Program For Members of the 1920 Birth Cohort Real N.t Internal Expected Expected Present Rste of Present Present. N.t value Real Return value value Expected ' Conditional Internsl Conditional of OASI of OASI Present On Surv,iving Rats of On Surviving Benefits Taxes value IInti1 Ags 65 8sturn Until Age 65 Median Earnings Profile Men Smokers 53,497 71,279 -17,782 3,721 1.87 3.18 Nonsmokers 79,436 76,000 3,436 18,218 3.17 3.78 Women Smokers 65,512 57,386 8,126 21,843 3.45 4.03 Nonsmokers 75,788 58,395 17,394 28,283 3.87 4.27 • - - - - - - . - - - - • - - - - - • - - - - - - - - 'rfen Low (60% Median) Earnings Profile Smokers 41,378 50,342 .8,964 8,918 2.25 3.57 Nonsmokers 61,441 53,433 8,008 20,130 3.53 4.15 Women Smokers 47,159 34,431 12,728 23,166 4.06 4.65 Nonsmokers 54,556 35,036 19,520 27,801 4.47 4.88 12 T~MN 358~85

earnings patterns thus exceeds $21,000. The internal rate of return, which equates the expected value of payouts and payins, is 1.87 percent real for smokers and 3.17 percent for nonsmokers. If one only looked at those who survived until 65, the rates of return would naturally be higher. In that case, the real internsl rate of return for median wage male smokers in this cohort is,3.18 percent, whereas the rate for nonsmokers is 3.78 percent. The ' . .~ . . , dollar difference in the net transfer between male smokers and nonsmokers, conditional on surviving to 65, is still about $14,500. Table 3 indicates that the Social Security cost of smoking is smaller for single women than for single men. In general single women get a higher rate of . , .~ I , . . . return from Social Security for two reasons. First, they have longer life expectancies, and, second, they have lower earnings and the system is progressive. Conditional on age.20, the difference in the net transfer to median wage women nonsmokers and smokers is slightly more than $9,000. The real internal rate of return for smoking women is 3.45 percent, while the figure is 3.87 percent for nonsmokers. Conditional on reaching age 65, the dollar difference between smoking and not is about $6500.for median wage single women. The lower portion of Table 3 shows the results for single individuals with earnings 60 percent of the median for their age and cohort. The loss due to smoking in the expected transfer from Social Security is almost $17,000 for men and $7,000 for women at this earnings level. We conclude that the Social , Security cost of smoking is not terribly sensitive to earnings levels. The corresponding results for one and two earner married couples with median earnings profiles are shown in Table 4. One earner co uples receive larger transfers and a higher rate of return from Social Security because of . . -». . .... _ . .. - . . _ the benefits received by the nonworking spouse. The Social Security expected cost of smoking is similar for couples in either circumstance. The net TIIVIN 35828

Table 4 Present Value of Social Security Benefits and Taxes (in T985 $) And Internal Rate of Return to the Social Security Program For Members of the 1920 Birth Cohort Median Earnings Profile Real Net Internal Expected Expected Present Rate of Present Present Net Dalue Real Return Value value Expected Conditional Internal Conditional of OASI of OASI Present On Surviving Rzte of On Surviving Benefits Taxes value Until Age 65 Return Until Age 65 Married (one earner) Both Smoke 118,223 79,466 38,757 81,270 4.40 5.41 Nonsmokers 149,229 81,004 68,225 95,872 5.14 5.63 Male smoker 128,748 79,722 49,026 92,123 4.67 5.67 female nonsmoker Female smoker 139,353 male nonsmoker 80,860 58,493 87,010 4.93 . 5:43 2Ssrried (tvo earaers) Both Smoke 126,687 128,664 -1,977 38,639 2.95 3.87 Nonsmokers 162,985 134,395 28,590 56,371 3.68 4.18 Hale smoker 138,313 129,673 ' 8,640 48,103 3.22 4.04 female nonsmoker Female smoker 151,494 133,386 18,108 48,402 3.46 4.05 male nonsmoker ' YOTE: Figures are per household. , 14 TIMN 358287

expected present value of participation in Socia7. Security is $29,467 lower for one earner couples who both smoke relative to one earner couples where neither spouse smokes. If only the man smokes, the loss in the expected transfer from the system is $19,199, whereas if only the vife smokes the loss, is $9,732 relative to a one .arner couple in which neither smokes. To put these figures in perspective, one might note that the median earnings o£ 64 year old men in this cohort were $20,315. Thus, the Social Security loss for both smoking amounts to almost 1.5 years labor income. In fact, the loss in slightly greater than that given that Social Security benefits are taxed more favorably than labor income. The numbers for tsro earaer couples are that their expected net Social Security transfer is $30,567 less if both spouses smoke than if neither does. The real inteznall rate of return for two earner couples in which both smoke is 44 -.n 2.95 percent, whereas it is 3.68 percent if neither smokes. The cost of the husband only smoking is $19,950, and the cost of the wife only smoking is $10,482. For reference, the median annual earnings of_women is-abo_ut_$12,500, so the loss if they both smoke is roughly equivalent to 2.4 years of the wife's earnings. Table 5 contains the results for low wage one and two-earner couples. For t one earner couples, we find that the cost of both smoking is roughly $22,500. For two earner couples, the cost of both smoking is $23,500. Once again, the cost is roughly twice as large for men as it is for women. The dollar costs to smoking are greater relative to earnings for low wage households than for median earners. the The gain in Social Security benefits that accrue to the nonsmoker, or to smoker who quits, rearesents an equall and opposite drain on Social Security funds. This drain is only partially offset by the.increase in preretirement ?5 TI.___M.N 358288

Table 5 Present Value of Social Security Benefits and Taxes (in 1985 $) And Internal Rate of Return to the Social Security Program For Members of the 1920 Birth Cohort Lov (60% Median) Easnings Profile Real Nat Internal Expected Expected Present Rate of Present Present Net Vsltu Real Return Valu. Value Expected Conditional Internal Conditional of OASI of,OASI Present On Surviving Bats of On Surviving Benefits Taxes Valus Until Age 65 Return Until Age 65 Married (one sarner) Both Smoke 91,761 55,146 36,614 70,896 4.78 5.81 Nonsmokers 115,531 56,327 59,204 81,313 5.50 5.99 Male smoker 99,952 55,300 44,652 79,101 5.05 6.07 femala nonsmoker Female smoker 107,878 male nonsmoker 56,241 51,637 74,672 5.29 5.79 Married (tvo earners) Eo:.h Smoke 96,843 84,772 12,071 45,609 3.47 4.43 Nonsmokers 124,085 88,469 35,616 58,131 4.18 4.70 Male smoker 105,689 85,378 20,311 52,959 3.74 4.61 female nonsmoker Female smoker 115,429 male nonsmoker 87,864 27,566 52,003 3.97 4.56 NOTE: Figures are per household. lb TIMN 358289

taxes paid by nonsmokers and'ex-smokars in comparison with smokers, a substantially greater number of whom die prematurely. The potential cost to society, including the goverxsment, of a successful anti-smoking program has not gone unnoticed. In 1971, the Britishh government, in response to a recommendation by the Royal College of Physicians to mossnt an anti-smoking . .. campaign, estimated that such a campaign would save money in the short run. However, by the year 2000, they forecasted that a 40 percent reduction in cigarette smoking would result in a net loss to the government of 29 million pounds due to additional benefits received by surviving ex-smokers. The Congressional Office of Technology Assessment, in its recent report "Smoking- Related Death and Financial Costs," indicated that, in the event of a reduction T ~ - - a - '- • - of smoking, "there will be some increase in revenues to the government and the Social Security and Medicare trust funds because people will be working more years. The increas_ e in these revenues, however, may not equal the additional costs borne by these programs for the additional retirees." Limiting her , attention to Medicare's hospital insurance fund, Wright (1986) has estimated that the individual forty-five year old male who quits smoking will cost the fund between $204 and $2,745. We have emphasized the extent to which smoking reduces the expected value of Social Security payments below that of nonsmokers. TJe need to remember that because of the nature of the system, a drop in the number of smokers will provide a cost; every person who begins to smoke implicitly decreases the future liability of the system. The prevalence of smoking is an important factor in determining the financial viability of the system. The percentage of U.S. adults who smoke has fallen drastically in the last 20 years. In 1965, 52$ of men and 35% of women age 20 and older smoked; by 1983 the numbers fell to 35% and 30% respectively (Health United States, 1984). 17 TIMN 358290

This should result in an increase in the average lifespan; since the majority of these people are below retirement age, we should expect retirees in the Jfuture to live longer on average than current retirees, who are already living longer than previous retirees due to the reduced use of cigarettes. A higher percentage of all workers will live to retirement; those who do will collect benefits for a longer period. This should be reflected in the demographic projections upon which Social Security taxes and benefits are based. The trend toward fewer smokers has been a long one, .specially among men. Unless that trend was adequataly projected, we expect Social Security demographic projections to be too low. Ghile the 1958 and 1966 Actuarial Studies by the Social Security Administration do a good job of predicting 1980 totAl population levels, they predict too high a number of young people and too low a number of retired persons. This implies offsetting errors, perhaps forecasting a longer "baby boom" than actually occurred and underestimating the additions to life expectancy, some of which can be attributed to Iower smoking levels. This hurts the system twice; more retirees are currently drawing benefits than projected, and fe.rer workers will be paying-taaes in the future than pro j ected. Our simulations suggest that each median wage male smoker in the 1920 birth cohort roughly "saves' the Social Security system $20,000 and each median wage female smoker saves $10,000. To get an approximate idea of the aggregate effect of smoking,by members of this cohort on Social Security, we can multiply a these saving figures by the number of smokers born in 1920. The result indicates that if no one had smoked in this cohort, the net transfer to this population from Social Security would have been $14.5 billion greater. As this reflects only the change for those born in oise year, one can easily see that the total impact of smoking on the financial circumstances of Social Security amounts to hundreds of billions of dollars. 18 TIMN 358291 t~; ~P~

While we by no means claim that the reduction in smoking is responsible for all the gains in life expectancy achieved in recent years, we have demonstrated the enormous potential impact on the system of reductions in smoking rates. Changes in the prevalence of smoking should be included in the system's attempts to model future populations. 4. CONCLIISION The body of literature discussing the economic costs of smoking has largely ignored private and social costs with regard to Social Security. Our analysis is a first step in estimating these costs, both in terms of net benefits to smokers and reduced payments by the system. SJe find that the expected loss caused by smoking from participation in Social Security is a large one from the individual's perspective. The loss for a median wage male smoker is about $20,000, or about 11 months of earnings. The loss_for median wage women is•approximately $10,000 or about 10 months of earnings. These losses are quite significant even compared to the health cost consequences of smoking. Ws also found that these losses are not very different for workers with lower_wages. The aggregate implications of our results are that smokers "save" the Social Security system hundreds of billions of dollars. Certainly this does not mean that decreased smoking would not be socially beneficial. In fact, it is probably one of the most cost effective ways of increasing average longevity. It does indicate, however, that if people alter their behavior in a manner which extends life expectancy, then this must be recognized by our national retirement program. Looked at in this way, it is not surprising that the large pocential for increasing lifespans that reduced smoking offers, has sizable consequences for Social Security. :9 TIMN 358292 '`

Appeadiz Life Tables, by Sex and Smoking Status, for the 1920 Birth Cohort 1. Males Probability of Death withia Oae Year Survivors from 100,000 Births Fraction That Mortality Age Total Pop. Nonsmokers Smoker s Nonsmokers Smokers Saoke Ritio (1) (2) (3) (4) (5) (6) (7) (8) 20 .00244 .00230 .00253 85758 85758 0.60 1.10 21 .00265 .00250 .00275 85560 85540 0.60 1.10 22 .00287 .00271 .00298 85346 85305 0.60 1.10 23 .00353 .00333 .00366 85115 85051 0.60 1.10 24 .00369 .00348 .00383 84832 84739 0.60 1.10 25 .00360 .00340 .00374 84536 94415 0.60 1.10 26 .00231 .00218 .00240 84249 84100 0.60 1.10 27 .00217 .00205 .00223 84066 83898 0.60 1.10 28 .00210 .00198 .00218 83894 83709 0.60 1.10 29 .00206 .00194, .00214 83727 83527 0.60 1.10 30 .00213 .00164 .00246 83565 83348 0.60 1.50 31 .00222 .00171 .00256 83428 83143 0.60 1.50 32 .00227 .00175 .00262 83285 82930 0.60 1.50 33 .00232 .00178 .00268 83140 82713 0.60 1.50 34 .00230 .00177 .00265 82991 82492 0.60 1.50 35 .00243 .00157- .00300 82845 82273 0.60 1.91 36 .00256 .00166 .00316 82714 82026 0.60 1.91 37 .00288 .00186 .00356 82577 81766 0.60 1.91 38 .00311 .00201 .00384 82424 81475_ 0.60 1.91 39 .00337 .00218 .00416 82258 81162 0.60 1.91 40 .00375 .00243 .00463 82078 80824 0.60 1.91 41 .00411 .00266 .00508 81879 80450 0.60 1.91 42 .00451 .00292 .00557 81662 80041 0.60 1.91 43 .00501 .00324 .00619 81423 79595 0.60 1.91 44 .00557 .00360 .00688 81160 79103 0.60 1.91 45 .00608 .00329 .00794 80867 78558 0.60 2.41 46 .00681 .00369 .00889 80601 77935 0.60 2.41 47 .00746 .00404 .00974 80303 77242 0.60 2.41 48 .00842 .00456 .01099 79979 76490 0.60 2.41 49 .00904 .00490 .01181 79614 75649 0.60 2.41 50 .00972 .00527 .01269 79224 74756 0.60 2.41 51 .01033 .00568 .01370 78807 73807 0.58 2.41 52 .01125 .00629 .01515 78359 72797 0.56 2.41 53 .01196 .00679 .01636 77867 71694 0.54 2.41 54 .01272 .00734 .01769 77338 70521 0.52 2.41 55 .01340 .00879 .01801 76770 69273_ 0.50 _ 2.05 56 .01422 .00945 .01938 76096 68025 0.48 2.05 57 .01495 .01008 .02067 75376 66707 0.46 2.05 58 .01598 :01093 .02241 74616 65328 0.44 2.05 59 .01702 .01181 .02421 73801 63865 0.42 2.05 60 .01845 .01299 .02664 72929 62318 0.40 2.05 61 .01967 .01406 .02882 71982 60658 0.38 2.05 62 .02094 .01520 .03115 70970 58910 0.36 2.05 20 TIMN 358293

TZ 00'1 00'0 I T TL9ZIi' TL9Z*7' IL9Z+7 ' LTI 00'T 00'0 T £ ETZI*V' £tZt*7' £tZT*7 ' 9tT 00'T 00'0 £ S 9086£' 9086£' 9086E' StT 00'I 00•0 S 6 8+Iv8£' 8"8£' 8*rVBE' +7TT 00'T 00'0 R +7T S£IL£' SEILE' SEILE' £tT 00'T 00'0 ZT ZZ L98SE' L985£' L98SE' ZTT 00'T 00'0 6t h£ £+~9*7E' £+79*7f ' TTT 00'T 00'0 6Z ZS t99££' T9*Y££' t9VE£' OTT 00'T 00'0 £+7 LL OZ£Z£' OZ£ZE' 0Z£Z£' 60T 00't 00'0 £9 ZTT 8IZt£' 8TZTE' StZT£• 80t 00'T 00'0 16 09T IVSTO£' VSTO£' VSTO£' LOT 00'T 00'0 SZT LZZ SZT6Z' SZI6Z' SZI6Z' 90T 00'T 00'0 6LT ST£ Z£T8Z' Z£t8Z' ZETSZ' SOT 00't 00'0 SvZ ££9 9LTLZ' +7LTLZ' *ILTLZ' *70I 00'T 00'0 £££ 885. 8+7Z9Z' S+7Z9Z' 8*9Z9Z' £OT 00'T 00'0 9" 88L '75£SZ' *7S£SZ' h5£SZ' ZOI 00'T 00'0 169 £*l0T 06*7VZ' 06*1*iZ' 06"Z' , T0T 00'T 00•0 SLL 99ET 959£Z' 9S9£Z' 9S9£Z' 00I 00'T 00•0 +700I tLLt 058ZZ' 058ZZ' OSSZZ' 66 00'I 00'0 68ZT £LZZ ZLOZZ' ZLOZZ' ZLOZZ' 86 00'I 00'0 9E9T LSSZ OSZtZ' 0SZtZ' OSZtZ' L6 00'T 00'0 950Z 9Z9£ 06£0Z' 06£0Z' 06£0Z' 96 00'T 00'0 +/SSZ 50S*7 66*76t' 66V6T' 66'r/6T' 56 00'I 00'0 L£tE ££55 S858I' SBSST' S8S8t' +76 00'T 00'0 OIS£ 6119 *759LT' +759LT' *759LT' C6 00'T 00•0 LL5+7 *1L08 *7LL9T' *7LL9t' '4LL9t' Z6 00'T 00'0 S"5 *7096 V£65I' V£65T' V£6ST' 16 00'T 00'0 ST79 ST£TT TZTSt' TZIST' TZIST' 06 00'I 00'0 88*tL LOZET ZZ£*7t' ZZ£*iT' ZZE'7T' 68 00'T 00'0 6598 ZLZST TZS£T' tZS£t' tZ5£T' 88 00'I 00'0 6166 S6*!Lt 80LZT' 80LZt' 80LZT' L8 00'I 00'0 99'ZTt £586T LLSTT' LLBTT' LLStT' 98 00'T 00'0 TS9ZT £T£ZZ '4ZOTT' VZOtT' VZOtT' s8 00'T 00'0 s807T Z*!8*!Z t8tOT' TSTOT' t8I0T' 478 00'T 00'0 £+tSST £t*1LZ 08£60' 08£60' 08£60' £S 00'T 00'0 LIOLI Vt00£ 99980' 99980' 99980' Z8 00'T 00'0 TtSBT 8*79Z£ 99080' 99080' 99080' T8 00'T 00'0 SZOOZ 8T£5£ T9SL0' I95L0' T95L0' 08 9Z'T Z0'0 £86TZ 9008£ 0T680' tLOLO' 80140, 6L 9Z't +70'0 8L6£Z £690*7 0ZE80' £0990' ZL990' 8L 9Z't 90'0 966SZ • t9££'4 £SLLO' fiSI90' 6'4Z90' LL 9Z'I 80'0 L008Z *18659 68TL0' SOLSO' vZ850' 9L 9Z'I 0t'0 8666Z T*!S8*! 9£990' L9Z50' V0*tS0' 6L IL'T ZT'0 69SZ£ 1680; V68L0' LT9*l0' 0t0S0' *1L 1L'T *7T'0 0IT5£ 0*!T£S L£ZL0' Z£Z70' £59*70' £L tL'T 9T'0 909L£ 68ZSS L£990' t88£0' ZZ£*l0' ZL TL'T 8T'0 S*7OOh LZ£LS t6090' Z9S£0' LT0V0' TL IL'I 0Z'0 8IvZ*! 99Z65 V6550' ZLZ£0' 9£L£0' 01 TL'I ZZ'0 LTL" £0Tt9 t+7IS0' 900£0' 9Lii£0' 69 IL'I vZ'0 S£69*1 0+78Z9 SZLV0' £9LZ0' 'hEZEO' 89 IL'T 9Z'0 '17906*t 9L*ni9 T+/£'170' 8£SZO' L00£0' L9 IL't SZ'0 £OTtS LT099 066£0' £EEZO' L6LZ0' 99 IL'T 0£'0 TSOES 59VL9 ZL9£0' 8*7IZO' S09Z0' 59 50'Z Z£'0 60199 LTL89 ££L£0' tZ8t0' ££VZ0' '79 50'Z +7£' 0 SLOLS 16969 '4"EO' 09910' 08ZZ0' £9

118 .44180 .44180 .44180 1 0 0.00 1.00 119 .45743 .45743 .45743 0 0 0.00 1.00 120 .50000 .50000. .50000 0 0 0.00 1.00 2. Females Probability of Death withia One Year Survivors from 100,000 Births Fraction That Mortality Age Total Pop. Nonsmolcars Smokers NorLSmok.rs Smokers Smoka Ratio (1) (2) (3) (4) (5). (6) (7) (8) 17 .00195 .00191 .00201 88787 88787 0.40 1.05 18 .00191 .00187 .001.97 88617 88608 0.40 1.05 19 .00187 .00183 .00193 88451 88434 0.40 1.05 20 .00191 .00187 .00197 88289 88264 0.40 1.05 21 .00192 .00188 .00198 88123 88090 0.40 1.05 22 .00190 .00186 .00196 87957 87916 0.40 1.05 23 . .00190 .00186 .00196 87794 87744 0.40 1.05 24 .00182 .00178 .00187 87630 87573 0.40 1.05 25 .00171 .00168 .00176 87474 87409 0.40 1.05 26 .00163 .00160 .00168 87327 87255 0.40 1.05 27 .00156 .00153 . .00161 87188 87108 0.40 1.05 28 .00145 .00142 .00149 87054 86968 0.40 1.05 29 .00143 .00140 .00147 86930 86839 0.40 1.05 30 .00143 .00130 .00163 86809 86711 0.40 1.25 31 .00146 .00133 .00166 86696 86570 0.40 1.25 32 .00148 .00135 .00168 86581 86426 0.40 1.25 33 .00149 .00135 .00169 86464 86281 0.40 1.25 34 .00149 .00135 .00169 86347 86135 0.40 1.25 35 .00158 .00139 .00187 86230 85989 0.40 1.35 36 .00166 .00146 .00197 86111 85828 0.40 1.35 37 .00188 .00165 .00223 85985 85659 0.40 1.35 38 .00196 .00172 .00232 85843 85469 0.40 1.35 39 '.00208 .00182 .00246 85696 85270 0.40 1.35 40 .00235 .00206 .00278 85539 85060 0.40 1.35 41 .00246 .00216 .00291 85363 84823 0.40 1.35 !:2 .00273 .00239 .00323 85179 84576 0.40 1.35 43 .00301 .00264 .00356 84975 84303 0.40 1.35 44 .00324 .00284 .00384 84750 84002 0.40 1.35 45 .00359 .00295 .00455 84510 83680 0.40 1.54 46 .00393 .00323 .00498 84260 83300 0.40 1.54 47 .00422 .00347 .00534 83988 82885 0.40 1.54 48 .00467 .00384 .00591 83696 82442 0.40 1.54 49 .00487 .00400 .00617 83375 81954 0.40 1.54 50 .00528 .00434 .00669 83041 81449 0.40 1.54 51 .00556 .00459 .00707 82680 80904 0.39 1.54 52 .00577 .00481 .00741 82301 80332 0.37 1.54 53 .00624 .00522 .00805 81905 79737 ' 0.36 1.54 54 .00652 .00548 .00844 81477 79096 0.35 1.54 55 .00686 .00596 .00870 31030 78428 0.33 1.e66 56 .00726 .00633 .00924 80548 77746 0.32 1.46 57 .00760 .00665 .00971 80038 77027 0.31 1.46 58 .00813 .00717 .01047 79505 76279 0.29 1.46 22 TIMN 358295

59 .00862 .00764 .01115 78935 75481 0.28 1.46 60 .00954 .00849 .01239 78332 74639 0.27 1.46 61 .01030 .00924 .01349 77668 73714 0.25 46 1 62 .01099 .00990 .01445 76950 72720 0.24 . 1.46 63 .01214 .01098 .01603 76189 71669 0.23 1.46 64 .01297 .01183 .01727 75352 70520 0.21 1.46 65 .01393 .01264 .01909 74461 69303. 0.20 1.51 66 .01499 .01367 • .02064 73520 67980 0.19 1.51 67 .01609 .01481 .02236 72515 66577 0.17 1.51 68 .01719 .01589 .02400 71441 65089 0.16 1.51 69 .01829 .01699 .02566 70306 63527 0.15 1.51 70 .01948 .01827 .02759 69111 61897 0.13 1.51 71 .02079 .01959 .02958 67849 60189 0.12 1.51 72 .02219 .02101 .03173 66519 58409 0.11 1.51 73 .02368 .02264 .03419 65122 56556 0.09 1.51 74 .02532 .02433 .03673 63647 54622 0.08 1.51 75 .02723 .02676 .03345 62099 52615 0.07 1.25 76 .02938 .02902 .03627 60437 50855 0.05 1.25 77 .03167 .03136 .03920 58683 49011 0.04 1.25 78 .03407 .03390 .04238 56843 47090 0.02 1.25 79 .03672 .03663 .04579 54916 45094 0.01 1.25 80 .03969 .03969 .03969 52905 43030 0.00 1.00 81 .04321 .04321 .04321 50805 41322 0.00 1.00 82 .04750 .04750 .04750 48610 39536 0.00 1.00 83 .05268 .05268 .05268 46301 37658 0.00 1.00 84 .05866 .05866 .05866 43861 35674 0.00 1.00 85 .06522 .06522 .06522 41289 _ 33582 0.00 1.00 86 .07222 .07222 .07222 38596 31391 0.00 1.00 87 .07956 .07956 .07956 35808 29124 0.00 1.00 88 .08723 .08723 .08723 32959 26807 0.00 1.00 89 .09529 .09529 .09529 30084 24469 0.00 1.00 90 .10380 .10380 .10380 27217 22137 0.00 1.00 91 •.11285 .11285 .11285 24392 19839 0.00 1.00 92 .12251 .12251 .12251 21640 17600 0.00 1.00 93 .13285 .13285' .13285 18988 15444 0.00 1.00 94 .14392 .14392 .14392 16466 13392 0.00 1.00 95 .15480 .15480 .15480 14096 11465 0.00 1.00 96 .16527 .16527 .16527 11914 9690 0.00 1.00 97 .17517 .17517 .17517 9945 8088 0.00 1.00 98 .18429 .18429 .18429 8203 6671 0.00 1.00 99 .19243 .19243 .19243 6691 5442 0.00 1.00 100 .20095 .20095 .20095 5403 4395 0.00 1.00 101 .20984 .20984 .20984 4317 3511 0.00 1.00 102 .21912 .21912 .21912 3411 2774 0.00 1.00 103 .22881 .22881 .22881 2664 2166 0.00 1.00 104 .23893 .23893 .23893 2054 1671 0.00 1.00 105 .24949 .24949 .24949 1563 1271 0.00 1.00 106 .26054 .26054 .26054 1173 954 0.00 1.00 107 .27207 .27207 .27207 867 705 0.00 1.00 108 .28411 .28411 .28411 631 513 0.00 1.00 109 .29668 .29668 .29668 452 367 0.00 1.00 110 .30981 .30981 .30981 318 258 0.00 1.00 111 .32353 .32353 .32353 219 178 0.00 1.00 112 .33785 .33785 .33785 148 120 0.00 1.00 113 .35281 .35281 .35281 98 79 0.00 1.00 23

114 .36842 .36842 .36842 63 51 0.00 1.00 115 .38474 .38474 .38474 40 32 0.00 1.00 116 .40178 .40178 .40178 24 20 0.00 1.00 .117 .41958 .41958 •.41958 14 12 0.00 1.00 118 .43816 .43816 .43816 8 6 0.00 1.00 119 .45743 .45743 .45743 4 3 0.00 1.00 120 .50000 .50000 .50000. 2 2 0.00 1.00 24 TIMN 358297

Footnotes i A mortality ratio is the death rate of smokers divided by the death rate of nonsmokers of similar age and sex. 25

References Atkinson, A. and Townsend, J. 1977. "Economic Aspects of Reduced Smoking," Lancet, 492-495 (September 3). Census Bureau 1982. Current Population Report P-60 No. 137. Census Bureau 1982. Current Population Report P-60 No. 142. Cowell, M. and Hirst, B. 1980. "Mortality Digfarencss Between Smokers and Nonsmokers," Soc. Actuaries Transact. 32:185-213. Gori, G. and B. Richter 1978. "23acroaconoaics of Disease Prevention in the United States," Sciencs 200:1124-1134, Jun. 9. Gori, G., B. Richter, and SJ. Yu 1984. "Economics and Extended Longevity - A Case Study," Preventive Medicine 13(4):396-410. Hammond, E. C. 1966. "Smoking in Relation to the Death Rates of One Million Men and Women," In Epidemioloaical Study of Cancer and Other Chronic Diseases. National Cancer Institute Monograph 19, 127-204. Harris, Jeffrey 1983. "Cigarette Smoking Among Successive Birth Cohorts of Men and iJomen in the United States During 1900-80,_Journal of the National Cancer Institute, 71(3):473-479 (September). Kronebusch, Karl 1985. Smoking-Related Death and Financial Costs, Preliminary Draft, Office of Technology Assessment. Lubin, Blot, Berrino, Flamant, Gillis, Kunze, Scizmahl, and Visco 1984. "Patterns of Lung Cancer Risk According to Type of Cigarette Smoked," International Journal of Cancer 33:569-576. Oster, Colditz, and Kelly 1984. The Economic Costs of Smokina and Benefits of Quittina, Lexington Books, D. C. Heath & Co., Lexington, MA. Rogot, Eugene 1974. "Smoking and Mortality Among U.S. Veterans, Journal of Chronic Disabilitv 27:189-203. Scheiderman, M. and D. Lavin 1972. "Trends in Lung Cancer," Cancer 30(5):1320-1325 (November). TIMN 358299

Warner, K. and H. Murt 1984. "Economic Incentives for Health," Ann. Rev. Public Health 5:107-133. Wright, Virginia 1986. "Will Quitting Smoking Help Medicare Solve Its Financial Problema?" Inauiry 23:76-82 (Spring). ` TIMN 358300

Number Author Title Date 2188 Bruce Greenwald and Joseph E. Stiglitz Money, Imperfect Information and Economic Fluctuations 3/87 2189 Joseph E. Stiglitz Pareto Efficient and Optimal Taxation and the New New Welfare Economics 3/87 2190 Barry Eichengreen International Competition in the Products of U.S. Basic Industries 3/87 2191 Bronwyn H. Hall The Effect of Takeover Activity on Corporate Research and Development 3/87 2192 Alan J. Auerbach and David Reishus The Effects of Taxation on the Merger Decision 3/87 2193 Barry Eichengreen Hegemonic Stability Theories of the International Monetary System 3/87 2194 Andrew Weiss Incentives and Worker Behavior: Some Evidence 3/87 2195 Anne 0. Krueger The Importance of Economic Policy in Development: Contrasts Between Korea and Turkey 3/87 2196 David E.M. Sappington Joseph E. Stiglitz Privatization, Information and Incentives 3/87 2197 Katharine L. Bradbury and Helen F. Ladd City Taxes and Property Tax Bases 3/87 2198 Barry Eichengreen The Gold-Exchange Standard and the Great Depression 3/87 2199 Kevin Lang Why Was There Mandatory Retirement? Or The Impossibility of Efficient Bonding Contracts 3/87 2200 Martin Feldstein Should Social Security Benefits Increase with Age? 3/87 2201 Laurence Ball and Stephen G. Cecchetti Imperfect Information and Staggered Price Setting 4/87 2202 Joshua Aizenman Successful Adjustment in a Multi- Sectoral Economy 4/87 2203 Maurice Obstfeld Peso Problems, Bubbles, and Risk in the Empirical Assessment of Exchange-Rate Behavior 4/87 T1ll1N 358301

Number Author Title Date 2204 Jonathan Eaton and Mark Gersovitz Country Risk and the Organization of International Capital Transfer 4/87 2205 Wesley M. Cohen and Richard C. Levin Firm Size and R&D Intensity: A Re-Examination 4/87 2206 Charles W. Calomiris and R. Glenn Hubbard International Adjustment Under the Classical Gold Standard: Evidence for the U.S. and Britain, 1879-1914 4/87 2207 Richard G. Frank Jonathan P. Weiner Donald M. Steinwachs and David S. Salkever Economic Rents Derived from Hospital Privileges in the Market for Podiatric Services 4/87 2208 Robert E. Baldwin U.S. and-Foreign Competition in the Developing Countries of the Asian Pacific Rim 4/87 2209 Laurence J. Christiano Martin Eichenbaum and hall id Ma D The Permanent Income Hypothesis Revisited 4/87 av rs ~ 2210 Jeffrey A. Frankel International Capital Flows and Domestic Economic Policies 4/87 2211 David E. Bloom and Christopher•L. Cavanagh Negotiator Behavior Under Arbitration 4/87 2212 Matthew 0. Shapiro Measuring Market Power in U.S. Industry 4/87 2213 Warwick McKibbin Nouriel Roubini and Jeffrey Sachs Dynamic Optimization in Two-Party Models 4/87 2214 Sebastian Edwards Anticipated Protectionist Policies, Real Exchange Rates and the Current Account 4/87 2215 Lawrence Lindsey Capital Gains Taxes Under the Tax Reform Act of 1986: Revenue Estimates Under Various Assumptions 4/87 2216 Jeffrey A. Frankel and Kenneth A. Froot Short-Term and Long-Term Expectations of the Yen/Dollar Exchange Rate: Evidence from Survey•Data 4/87 2217 Steven F. Venti and David A. Wise Have IRAs Increased U.S. Saving?: Evidence From Consumer Expenditure Surveys 4/87 358302

Number Author Title Date 2218 Sebastian Edwards The United States and Foreign Competition in Latin America 4/87 2219 Stewart C. Myers and Richard S. Ruback Discounting Rules for Risky Assets 4/87 2220 William H. Branson and Jorge Braga de Macedo Smuggler's Blues at the Central Bank: Lessons frflm Sudan 4/87 2221 Steven G. Allen Relative Wage Variability in the United 4/87 States, 1860-1983 2222 Ray C. Fair The Effect of Economic Events on Votes for President: 1984 Update 4/87 2223 John Y. Campbell and Richard H. Clarida Household Saving and Permanent. Income in Canada and the United Kingdom 4/87 2224 Douglas Holtz-Eakin and Harvey S. Rosen Tax Deductibility and Municipal Budget Structure 4/87 `2225 Michael J. Boskin and Douglas J. Puffert The Financial Impact of Social Security by Cohort Under Alternative Financing Assumptions 4/87 2226 Jeffrey 0. Sachs Trade and Exchange Rate Policies in Growth-Oriented Adjustment Programs 4/87 2227 Robert J. Gordon The Postwar Evolution of Computer Prices 4/87 2228 James H. Stock and Mark W. Watson Interpreting the Evidence on Money- Income Causality 4/87 2229 Robert King Charles Plosser James Stock and Mark Watson Stochastic Trends and Economic Fluctuations 4/87 2230 Sven W. Arndt and J. David Richardson Real-Financial Linkages Among Open Economies 5/87 2231 Harvey S. Rosen The Marriage Tax is Down But Not Out 5/87 2232 Paul Wachtel Loan'Commitments and Monetary Policy 5/87 2233 Ray C. Fair and Robert J. Shiller Econometric Modeling as Information Aggregation 5/87 T1?qN 358303 t

Number Author Title Date 2234 John B. Shoven The Social Security Cost of Smoking 5/87 Jeffrey 0. Sundberg and John P. Bunker NOTE: Copies of the_above working Papers can be obtained by sending 52.00 per copy to Working Papers, NBER, 1050 Massacnusezts Avenue, Cambridge, MA 02138. Advance payment is required on all orders tota•ling less than $10.00. Please make checks payable to the National Bureau of Economic Research. TIMN 358304