• Selected
• Entire List

Abstract:
The nation's spending for prescription drugs has grown dramatically in recent years. Previous studies have shown that the replacement of older drugs by newer, more expensive, drugs is the single most important reason for this increase, but they did not measure how much of the difference between new and old drug prices reflects changes in quality as better, newer drugs replace older, less effective medications.

In this paper we analyzed prescribed medicine event-level data (linked to person- and condition-level data) from the 1996 Medical Expenditure Panel Survey (MEPS) to provide evidence about the effect of drug age on mortality morbidity, and total medical expenditure, controlling for a number of characteristics of the individual and the event. (Previous researchers have hypothesized that differences in treatment patterns across individuals and areas may occur because of physicians' uncertainty and ignorance over the best medical practice.) The MEPS data enable us to control for many important attributes of the individual, condition, and prescription that influence outcomes and non-drug expenditures and that may be correlated with drug age. These include sex, age, education, race, income, insurance status, who paid for the drug, the condition for which the drug was prescribed, how long the person has had the condition, and the number of medical conditions reported by the person. Indeed, the fact that many individuals in the sample have both multiple medical conditions and multiple prescriptions means that we can control for all individual characteristics both observed and unobserved by including "individual effects".

Abstract:
The nation's spending for prescription drugs has grown dramatically in recent years. Previous studies have shown that the replacement of older drugs by newer, more expensive, drugs is the single most important reason for this increase, but they did not measure how much of the difference between new and old drug prices reflects changes in quality as better, newer drugs replace older, less effective medications. In this paper we analyze data from the 1996 Medical Expenditure Panel Survey (MEPS) to provide evidence about the effect of drug age on mortality, morbidity, and total medical expenditure, controlling for sex, age, education, race, income, insurance status, who paid for the drug, the condition for which the drug was prescribed, how long the person has had the condition, and the number of medical conditions reported by the person. The results provide strong support for the hypothesis that the replacement of older by newer drugs results in reductions in mortality, morbidity, and total medical expenditure. People consuming new drugs were significantly less likely to experience work-loss days and to die by the end of the survey than people consuming older drugs. The estimates indicate that reductions in drug age tend to reduce all types of non-drug medical expenditure, although the reduction in inpatient expenditure is by far the largest. Reducing the age of the drug results in a substantial net reduction in the total cost of treating the condition. Allowing people to use only generic drugs would increase total treatment costs, not reduce them, and would lead to worse outcomes.

Abstract:
No abstract is available for this paper.

Abstract:
We analyze the effect of mergers on various aspects of airline performance during the period 1970-84, using a panel data set constructed by Caves et al. Estimates derived from a simple "matched pairs" statistical model indicate that these mergers were associated with reductions in unit cost. The average annual rate of unit cost growth of carriers undergoing merger was 1.1 percentage points lower, during the five-year period centered on the merger, than that of carriers not involved in merger. Almost all of this cost reduction appears to have been passed on to consumers. Part of the cost reduction is attributable to merger-related declines in the prices of inputs, particularly labor, but about two-thirds of it is due to increased total factor productivity. One source of the productivity improvement is an increase in capacity utilization (load factor).

Abstract:
Between 1960 and 1997, life expectancy at birth of Americans increased approximately 10% from 69.7 to 76.5 years and it has been estimated that the value of life extension during this period nearly equaled the gains in tangible consumption. While life expectancy has tended to increase, there have been substantial fluctuations in the rate of increase. In this paper we investigate whether an aggregate health production function can help to explain the annual time-series behavior of U.S. longevity since 1960. We view longevity as the output of the health production function, and output fluctuations as the consequence of fluctuations in medical inputs (expenditure) and technology. We estimate longevity models using annual U.S. time-series data on life expectancy, health expenditure, and medical innovation. Reliable annual data are available for only one type of innovation new drugs but pharmaceutical R&D accounts for a significant fraction of total biomedical research. The empirical analysis provides strong support for the hypothesis that both medical innovation (in the form of new drug approvals) and expenditure on medical care (especially public expenditure) contributed to longevity increase during the period 1960-1997. Increased drug approvals and health expenditure per person jointly explain just about 100% of the observed long-run longevity increase. The estimates provide strong evidence against the null hypothesis that public health expenditure has no effect on longevity, but not against the null hypothesis that private health expenditure has no effect on longevity. This is at least partly attributable to the fact that public health expenditure exhibited much greater variability during the sample period than private health expenditure. The estimates imply that the medical expenditure needed to gain one life-year is about $11,000, and that the pharmaceutical R&D expenditure needed to gain one life-year is about $1,345. This suggests that increased development of new drugs may be a more cost-effective way of increasing life expectancy than increased medical expenditure in general. Previous researchers have estimated that the average value of a life-year is approximately $150,000. This figure implies that the benefit-cost ratio of general medical expenditure is 13.6, and that the ratio for pharmaceutical R&D exceeds 100.

Abstract:
We update and extend our previous study of the effect of drug age - years since FDA approval - on total medical expenditure, in several respects. The estimates indicate that, in the entire population, a reduction in the age of drugs utilized reduces non-drug expenditure 7.2 times as much as it increases drug expenditure. In the Medicare population, a reduction in the age of drugs utilized reduces non-drug expenditure by all payers 8.3 times as much as it increases drug expenditure; it reduces Medicare non-drug expenditure 6.0 times as much as it increases drug expenditure. About two-thirds of the non-drug Medicare cost reduction is due to reduced hospital costs. The remaining third is approximately evenly divided between reduced Medicare home health care cost and reduced Medicare office-visit cost. We also found that the mean age of drugs used by Medicare enrollees with private Rx insurance is about 9% lower than the mean age of drugs used by Medicare enrollees without either private or public Rx insurance.

Abstract:
No abstract is available for this paper.

Abstract:
Due in large part to intense takeover activity during the 1980s, the extent of American firms' industrial diversification declined significantly during the second half of the decade. The mean number of industries in which firms operated declined 14 percent, and the fraction of single-industry firms increased 54 percent. Firms that were "born" during the period were much less diversified than those that "died", and "continuing" firms reduced the number of industries in which they operated. Using plant-level Census Bureau data, we show that productivity is inversely related to the degree of diversification: holding constant the number of the parent firm's plants, the greater the number of industries in which the parent operates, the lower the productivity of its plants. Hence de-diversification is one of the means by which recent takeovers have contributed to U.S. productivity growth. We also find that the effectiveness of regulations governing disclosure by companies of financial information for their industry segments was low when they were introduced in the 1970s and has been declining ever since.

Abstract:
We perform an econometric analysis of the effect of new drug launches on longevity, using data from the IMS Health Drug Launches database and the WHO Mortality Database. Our data cover virtually all of the diseases borne by people in 52 countries during the period 1982-2001, and enable us to control, to an unusually great extent, for the effects of other potential determinants of longevity, e.g. education, income, nutrition, the environment, and 'lifestyle'. We find that launches of new chemical entities (NCEs) have a strong positive impact on the probability of survival. Launches of (older) drugs that are not NCEs do not increase longevity. NCE launches account for a significant fraction of the long-run increase in longevity in the sample as a whole. Between 1986 and 2000, average life expectancy of the entire population of sample countries increased by almost two years. Our estimates imply that NCE launches accounted for 0.8 years (40%) of the 1986-2000 increase in longevity. The average annual increase in life expectancy of the entire population resulting from NCE launches is .056 years, or 2.93 weeks.

Abstract:
Only about one third of the approximately 80 drugs currently used to treat cancer had been approved when the war on cancer was declared in 1971. We assess the contribution of pharmaceutical innovation to the increase in cancer survival rates in a 'differences in differences' framework, by estimating models of cancer mortality rates using longitudinal, annual, cancer-site-level data based on records of 2.1 million people diagnosed with cancer during the period 1975-1995. We control for fixed cancer site effects, fixed year effects, incidence, stage distribution of diagnosed patients, mean age at diagnosis, and surgery and radiation treatment rates. Cancers for which the stock of drugs increased more rapidly tended to have greater increases in survival rates. The increase in the stock of drugs accounted for about 50-60% of the increase in age-adjusted survival rates in the first 6 years after diagnosis. New cancer drugs increased the life expectancy of people diagnosed with cancer by about one year from 1975 to 1995. The estimated cost to achieve the additional year of life per person diagnosed with cancer - below $3000 - is well below recent estimates of the value of a statistical life-year. Since the lifetime risk of being diagnosed with cancer is about 40%, the estimates imply that new cancer drugs accounted for 10.7% of the overall increase in U.S. life expectancy at birth.

Abstract:
We investigate the economic effects of leveraged buyouts (LBOs) using large longitudinal establishment and firm-level Census Bureau data sets linked to a list of LBOs compiled from public data sources. About 5 percent, or 1100, of the manufacturing plants in the sample were involved in LBOs during 1981-86. We find that plants involved in LBOs had significantly higher rates of total-factor productivity (TFP) growth than other plants in the same industry. The productivity impact of LBOs is much larger than our previous estimates of the productivity impact of ownership changes in general. Management buyouts appear to have a particularly strong positive effect on TFP.

Labor and capital employed tend to decline (relative to the industry average) after the buyout, but at a slower rate than they did before the buyout. The ration of nonproduction to production labor cost declines sharply, and production worker wage rates increase, following LBOs. LBOs are production-labor-using, nonproduction-labor-saving, organizational innovations. Plants involved in management buyouts (but not in other LBOs) are less likely to subsequently close than other plants. The average R&D-intensity of firms involved in LBOs increased at least as much from 1978 to 1986 as did the average R&D-intensity of all firms responding to the NSF/Census survey of industrial R&D.

Abstract:
We perform an econometric investigation of the contribution of pharmaceutical innovation to mortality reduction and growth in lifetime per capita income. In both of the periods studied (1970-80 and 1980-91), there is a highly significant positive relationship across diseases between the increase in mean age at death (which is closely related to life expectancy) and rates of introduction of new, priority' (as defined by the FDA) drugs. The estimates imply that in the absence of pharmaceutical innovation, there would have been no increase and perhaps even a small decrease in mean age at death, and that new drugs have increased life expectancy, and lifetime income, by about 0.75-1.0% per annum. The drug innovation measures are also strongly positively related to the reduction in life-years lost in both periods. Some of the more conservative estimates imply that a one-time R&D expenditure of about $15 billion subsequently saves 1.6 million life-years per year, whose annual value is about $27 billion. All age groups benefited from the arrival of new drugs in at least one of the two periods. Controlling for growth in inpatient and ambulatory care utilization either has no effect on the drug coefficient or significantly increases it.

Abstract:
Several econometric studies have concluded that technical progress embodied in equipment is a major source of manufacturing productivity growth. Other research has suggested that, over the long run, growth in the U.S. economy's 'health output' has been at least as large as the growth in non-health goods and services. One important input in the production of health - pharmaceuticals - is even more R&D-intensive than equipment.

In this paper we test the pharmaceutical-embodied technical progress hypothesis - the hypothesis that newer drugs increase the length and quality of life - and estimate the rate of progress. To do this, we estimate health production functions, in which the dependent variables are various indicators of post-treatment health status (such as survival, perceived health status, and presence of physical or cognitive limitations), and the regressors include drug vintage (the year in which the FDA first approved a drug's active ingredient(s)) and indicators of pre-treatment health status. We estimate these relationships using extremely disaggregated - prescription-level - cross-sectional data derived primarily from the 1997 Medical Expenditure Panel Survey.

We find that people who used newer drugs had better post-treatment health than people using older drugs for the same condition, controlling for pre-treatment health, age, sex, race, marital status, education, income, and insurance coverage: they were more likely to survive, their perceived health status was higher, and they experienced fewer activity, social, and physical limitations. The estimated cost of the increase in vintage required to keep a person alive is lower than some estimates of the value of remaining alive for one month. One estimate of the cost of preventing an activity limitation is $1745, and the annual rate of technical progress with respect to activity limitations is 8.4%. People consuming newer drugs tend to experience greater increases (or smaller declines) in physical ability than people consuming older drugs.

Most of the health measures indicate that the effect of drug vintage on health is higher for people with low initial health than it is for people with high initial health. Therefore in contrast to equipment-embodied technical progress, which tends to increase economic inequality, pharmaceutical-embodied technical progress has a tendency to reduce inequality as well as promote economic growth, broadly defined.

Abstract:
With substantial fixed costs of drug development, more common conditions can support more products. If additional pharmaceutical products are beneficial, they will attract greater consumption and promote better health, e.g. greater longevity. We ask how market size - measured by condition prevalence - affects consumption and longevity. We document in condition cross sections that both the tendency to use a drug and longevity are higher for individuals with more prevalent conditions. We also make use of the 1983 Orphan Drug Act (ODA), which promoted development of drugs for the treatment of rare conditions. Longevity and drug use have grown more quickly for persons with rare diseases and even more quickly for persons with conditions with substantial orphan drug use.

Abstract:
This paper presents an econometric analysis of the effect of changes in the quantity and type of pharmaceuticals prescribed by physicians in outpatient visits on rates of hospitalization, surgical procedure, mortality, and related variables. It examines the statistical relationship across diseases between changes in outpatient pharmaceutical utilization and changes in inpatient care utilization and mortality during the period 1980-92. The estimates indicate that the number of hospital stays, bed-days, and surgical procedures declined most rapidly for those diagnoses with the greatest increase in the total number of drugs prescribed and the greatest change in the distribution of drugs, by molecule. The estimates imply that an increase of 100 prescriptions is associated with 1.48 fewer hospital admissions, 16.3 fewer hospital days, and 3.36 fewer inpatient surgical procedures. A $1 increase in pharmaceutical expenditure is associated with a $3.65 reduction in hospital care expenditure.

Abstract:
We test several hypotheses regarding the relationship between ownership structure and corporate performance. Our findings support the proposition that equity ownership by financial institutions in Japan may effectively substitute for the missing external takeover market by resulting in monitoring and intervention which minimizes the danger of lapses in productivity. In contrast, we also find evidence that high levels of intercorporate shareholding insulate firms from their own problems, at the expense of firm performance. Further, we find a notable positive influence of insider ownership, but see no evidence that the influence of financial institutions has diminished in the globalization and prosperity of the 1980s.

Abstract:
We develop a simple theoretical model of the allocation of public biomedical research expenditure, and present some empirical evidence about the determinants of this allocation. The structure of expenditure should depend upon the relative costs as well as the relative benefits of different kinds of research. Analysts of technical change typically have data on neither of these, but the measures of disease burden we use are indicative of the benefit of achieving advances against different diseases. We calculate distributions of government-funded biomedical research expenditure, by disease, from records of all re-search projects supported by the United States Public Health Service; to obtain a reasonably complete accounting of disease burden, we utilize data on both the dying (from the Vital Statistics-Mortality Detail file) and the living (from the National Health Interview Survey). We find a very strong positive relationship across diseases between total life-years lost before age 65 and public R&D expenditure. But the amount of publicly-funded research on a disease decreases with the share of life-years before age 65 lost to the disease that are lost by non-whites, perhaps because lack of scientific knowledge is a less important cause of premature mortality among non-whites than it is among whites. The number of research grants mentioning a chronic condition is completely uncorrelated with the number of people with the condition but very strongly positively related to the number of people whose activities are limited by that condition. There tends to be more research about chronic conditions that are prevalent among people living in low-income households, and that are prevalent among the young (under age 18) and the old (above age 75).

Abstract:
We hypothesize that pharmaceutical-embodied technical progress increases per capita output via its effect on labor supply (the employment rate and hours worked per employed person). We examine the effect of changes in both the average quantity and average vintage (FDA approval year) of drugs consumed on labor supply, using longitudinal, condition-level data. The estimates indicate that conditions for which there were above-average increases in utilization of prescriptions during 1996-1998 tended to have above-average reductions in the probability of missed work days. The estimated value to employers of the reduction in missed work days appears to exceed the employer's increase in drug cost. The estimates are also consistent with the hypothesis that an increase in a condition's mean drug vintage reduces the probability that people with that condition will experience activity and work limitations, and reduces their average number of restricted-activity days. The estimates imply that activity limitations decline at the rate of about one percent per year of drug vintage, and that the rate of pharmaceutical-embodied technical progress with respect to activity limitations is about 18% per year. Estimates of the cost of the increase in drug vintage necessary to achieve reductions in activity limitations indicate that increases in drug vintage tend to be very 'cost-effective.'

Abstract:
Between 1960 and 1997, life expectancy at birth of Americans increased approximately 10% - from 69.7 to 76.5 years - and it has been estimated that the value of life extension during this period nearly equaled the gains in tangible consumption. We investigate whether an aggregate health production function can help to explain the substantial fluctuations in the rate of increase in longevity since 1960. We view longevity as the output of the health production function, and output fluctuations as the consequence of fluctuations in medical inputs (expenditure) and technology. We estimate longevity models using annual U.S. time-series data on life expectancy, health expenditure, and medical innovation. Reliable annual data are available for only one type of innovation - new drugs - but pharmaceutical R&D accounts for a significant fraction of total biomedical research. The empirical analysis provides strong support for the hypothesis that both medical innovation (in the form of new drug approvals) and expenditure on medical care (especially public expenditure) contributed to longevity increase during the period 1960-1997. The estimates imply that the medical expenditure needed to gain one life-year is about $11,000, and that the pharmaceutical R&D expenditure needed to gain one life-year is about $1,345. Previous researchers have estimated that the average value of a life-year is approximately $150,000.

Abstract:
This paper is a re-examination of the relationship between research and development (R&D) activity and total factor productivity (TFP) at the industry level during the period extending from the early 1960's to the mid-1970's. The data base consists of NSF data on applied R&D expenditures by product class, matched to TFP indices derived from the detailed Census-Penn-SRI manufacturing data file.

A hypothesis suggested by previous research on the R&D-productivity relationship is that, due, perhaps, to the depletion of scientific opportunities, the "potency" of R&D as a source of technological progress has declined in recent years. Our findings indicate, however, that the relationship between an industry's R&D-intensity and its productivity growth did not disappear; if anything, the relationship was stronger in recent years. The overall deceleration in productivity in recent years has affected R&D-intensive industries, but to a lesser extent than it has other industries. What cannot be found in the data is strong evidence of the differential effects of the slowdown in R&D itself. The time series appear to be too noisy and the period too short to detect what the major consequences of the retardation in the growth of R&D expenditures may yet turn out to be.

Abstract:
Coe and Helpman(1995) have measured the extent to which technology spills over between industrialized countries through the particular channel of trade flows. This paper re-examines two particular features of their study. First, we suggest that their functional form of how foreign R&D affects domestic productivity via imports is probably incorrect. We provide an alternative model which turns out to be more accurate, both theoretically and empirically. Second, we take into account two new potential channels of technology transfer: inward FDI and technology sourcing, as proxied by outward FDI. The empirical results show that outward FDI flows and imports flows are two simultaneous channels through which technology is internationally diffused. Inward FDI flows are not a significant channel of technology transfer. The hypothesis of technology sourcing associated with MNEs activities abroad is therefore confirmed while the widespread belief that inward FDI is a major channel of technology transfer is rejected.

Abstract:
This paper extends previous research on the effect of investment on labor productivity at the country level by accounting for investment in R&D, as well as for investment in fixed and human capital. Privately-funded R&D investment is found to have a significant positive effect on productivity. Moreover, this effect appears to be quite large. The estimated social (national) rate of return to private R&D investment is about seven times as large as the return to investment in equipment and structures. The elasticity of GNP with respect to the privately-funded research capital stock is about 7 %--about 1(3 as large as the physical-capital elasticity (whose estimate is substantially reduced when R&D is accounted for). These findings do not support the hypothesis that there arecomplete, or at least instantaneous, international R&D spillovers. The social marginal product of government-funded research capital appears to be much lower than that of private research capital.

Abstract:
Using micro data on virtually all of the drugs and diseases of over 500,000 people enrolled in Puerto Rico's Medicaid program, we examine the impact of the vintage (original FDA approval year) of drugs used to treat a patient on the patient's 3-year probability of survival, controlling for demographic characteristics (age, sex, and region), utilization of medical services, and the nature and complexity of illness. We find that people using newer drugs during January-June 2000 were less likely to die by the end of 2002, conditional on the covariates. The estimated mortality rates are strictly declining with respect to drug vintage. For pre-1970 drugs, the estimated mortality rate is 4.4%. The mortality rates for 1970s, 1980s, and 1990s drugs are 3.6%, 3.0%, and 2.5%, respectively. The actual mortality rate is about 16% (3.7% vs. 4.4%) lower than it would have been if all of the drugs utilized in 2000 had been pre-1970 drugs. Estimates for subgroups of people with specific diseases display the same general pattern.

Abstract:
This paper uses confidential Census longitudinal microdata to examine the association between R&D and productivity for the period 1972-1985. These data allow for significant improvements in measurement and model specification, yielding more precise estimates of the returns to R&D. Our results confirm the finding of existing studies: 1) positive returns to R&D investment 2) higher returns to company-financed research 3) a productivity "premium" on basic research

These results are robust to our attempts to adjust for "influential" outliers. Also, it appears that the return to company-financed R&D (but not total R&D) is an increasing function of firm size.

Abstract:
A patent only protects an innovator from others producing the same product, but it does not protect him from others producing better products under new patents. Therefore, one may divide up the source of competition facing an innovator into within-patent competition, which results from production of the same product, and betweenpatent competition, which results from production of products on other patents. Previous theoretical and empirical micro-based analyses have emphasized the effects of intellectual property regulations on within-patent competition by showing how protecting innovative returns from imitators raises R&D incentives. However, between-patent competition affects innovative returns, particularly through creative destruction in the many high-tech industries that seem central to overall economic progress. This suggests that a fuller understanding of IP-regulations take into account its effects on between-patent competition. We find that the total effects of intellectual property regulations depend heavily on whether these unexplored effects are present. We attempt to estimate the relative magnitudes of the two sources of competition in limiting innovative returns in the U.S. pharmaceuticals market. In this market within-patent competition from so-called generic producers has been analyzed relatively more compared to competition between-patents through so called therapeutic competition. We estimate that between-patent competition, most of which occurs while a drug is under patent, costs the innovator at least as much as within-patent competition, which cannot occur until a drug is off patent. The reduction in the present discounted value of the innovator's return from between-patent competition appears to be at least as large as the reduction from competition within-patents, and may be much larger.

Abstract:
A patent only protects an innovator from others producing the same product, but it does not protect him from others producing better products under new patents. Therefore, one may divide up the source of competition facing an innovator into within-patent competition, which results from production of the same product, and betweenpatent competition, which results from production of products on other patents. Previous theoretical and empirical micro-based analyses have emphasized the effects of intellectual property regulations on within-patent competition by showing how protecting innovative returns from imitators raises R&D incentives. However, between-patent competition affects innovative returns, particularly through creative destruction in the many high-tech industries that seem central to overall economic progress. This suggests that a fuller understanding of IP-regulations take into account its effects on between-patent competition. We find that the total effects of intellectual property regulations depend heavily on whether these unexplored effects are present. We attempt to estimate the relative magnitudes of the two sources of competition in limiting innovative returns in the U.S. pharmaceuticals market. In this market within-patent competition from so-called generic producers has been analyzed relatively more compared to competition between-patents through so called therapeutic competition. We estimate that between-patent competition, most of which occurs while a drug is under patent, costs the innovator at least as much as within-patent competition, which cannot occur until a drug is off patent. The reduction in the present discounted value of the innovator's return from between-patent competition appears to be at least as large as the reduction from competition within-patents, and may be much larger.

Abstract:
I investigate the effect of large increases in the number of drugs available to treat rare diseases and HIV on mortality associated with them. Mortality from both diseases declined dramatically following increases in drug approvals. Before the Orphan Drug Act went into effect (between 1979 and 1984), mortality from rare diseases grew at the same rate as mortality from other diseases. In contrast, during the next five years, mortality from rare diseases grew more slowly than mortality from other diseases. I estimate that one additional orphan drug approval in year t prevents 211 deaths in year t+1 and ultimately prevents 499 deaths, and that about 108 thousand deaths from rare diseases will ultimately be prevented by all of the 216 orphan drugs tha t have been approved since 1983. Deaths are more closely related to the number of orphan product designations (which include experimental drugs) than they are to the number of approvals. Consistent with previous patient- level studies of HIV, I find that new drugs played a key role in the post-1995 decline in HIV mortality. I estimate that one additional HIV drug approval in year t prevents 5986 HIV deaths in year t+1 and ultimately prevents 33,819 HIV deaths. HIV drug approvals have reduced mortality both directly and indirectly (via increased drug consumption). HIV mortality depends on both the quality and the quantity of medications consumed, and new drug approvals have a sizeable impact on drug consumption: one additional HIV drug approval in year t results in 1.2 million additional HIV drug units consumed in year t+1 and ultimately result in 3.6 million additional HIV drug units consumed.

Abstract:
This study examines the effect of the introduction of new laboratory procedures and other medical goods and services on the health of Americans during the period 1990-2003. We hypothesize that, the more medical innovation there is related to a medical condition, the greater the improvement in the average health of people with that condition. To test this hypothesis, we estimate models of health outcomes using longitudinal disease-level data. We measure innovation in five types of medical procedures or products: pathology & laboratory procedures, outpatient prescription drugs, inpatient prescription drugs, surgical procedures, and diagnostic radiology procedures. We examine two kinds of (inverse) indicators of health: mortality and disability. The mortality indicator we analyze is the mean age at death of people whose underlying cause of death is medical condition i. The disability measures we analyze are the fraction of people with medical condition i who (1) missed work, or (2) spent one or more days in bed, due to that condition. Our estimates indicate that conditions with higher rates of lab and outpatient drug innovation had larger increases in mean age at death, controlling for other medical innovation rates and initial mean age at death. The 1990-1998 increase in mean age at death attributable to use of new lab procedures is estimated to be about 6 months. This is 42% of the total increase in mean age at death (1.18 years) in our sample of diseases. New laboratory procedures introduced during 1990-1998 are estimated to have saved 1.13 million life-years in 1998. Expenditure per life-year gained from new lab procedures is estimated to be $6093. Treatments that cost this amount are generally considered to be quite cost-effective. In the analysis of disability, when we don%u2019t control for the initial level of disability, we find that conditions with higher rates of lab and outpatient innovation had greater declines in the probability of missing work during 1996-2003. This suggests that the use of new laboratory procedures reduced the number of work-loss days in 2003 by 42 million. When we control for initial disability, the inverse relationship between lab innovation and disability changes disappears. This is because there is a significant inverse relationship between initial health and the extent of laboratory innovation. But due to errors in measuring initial health, controlling for this variable may cause the impact of innovation on health to be underestimated.

Abstract:
This paper examines the output contributions of capital and labor deployed in information systems (IS) at the firm level during the period 1988-91 throughout the business sector, using two different sources of data on these inputs. Our production function estimates suggest that there are substantial excess returns to both IS capital and IS labor, although the size and significance of the excess returns to IS capital is larger. Computer capital and labor jointly contribute, or account for, about 21 percent of output, although only about 10% of both capital and labor income accrue to IS factors. Although IS employees accounted for a very small share of total employment by 1986, IS employment growth is estimated to have made a larger contribution to 1976-86 output growth than non-IS employment, due to the very rapid growth (16% per annum) of IS employment. The estimated marginal rate of substitution (MRS) between IS and non-IS employees, evaluated at the sample mean, is 6: one IS employee can be substituted for six non-IS employees without affecting output.

Abstract:
This paper examines trends in computer usage and the effect on productivity growth for a sample of federal government agencies over the period from 1987 to 1992. We link data from the Bureau of Labor Statistics (BLS) on the growth in real output per employee with data from a marketing research firm, Computer Intelligence (CI), on the growth in per capita computer assets for a sample of 44 federal agencies. The data show that computer usage increased dramatically and that there was a shift towards more powerful, lower cost, distributed systems and that usage diffused more extensively throughout the sampled agencies. These trends mirror, while perhaps lagging, those experienced by large private firms over the same period. From estimates of a Cobb-Douglas production function for government services, we derive an estimated output elasticity for computers of 0.06, which allows us to conclude that computers did contribute significantly to output growth, thereby refuting the Computer Productivity Paradox as it applies to the public sector. Computers do not appear to be responsible for the disappointing productivity performance of the service sector. Although the magnitude of our estimated elasticity suggests that the returns to computer investments exceeded those to other types of capital, our results are not conclusive. We also observe a positive correlation between increased computer usage and compensation growth which is consistent with skill-biased technical change.

Abstract:
The rate of increase of longevity has varied considerably across U.S. states since 1991. This paper examines the effect of the quality of medical care, behavioral risk factors (obesity, smoking, and AIDS incidence), and other variables (education, income, and health insurance coverage) on life expectancy and medical expenditure using longitudinal state-level data. We examine the effects of three different measures of the quality of medical care. The first is the average quality of diagnostic imaging procedures, defined as the fraction of procedures that are advanced procedures. The second is the average quality of practicing physicians, defined as the fraction of physicians that were trained at top-ranked medical schools. The third is the mean vintage (FDA approval year) of outpatient and inpatient prescription drugs. Life expectancy increased more rapidly in states where (1) the fraction of Medicare diagnostic imaging procedures that were advanced procedures increased more rapidly; (2) the vintage of self- and provider-administered drugs increased more rapidly; and (3) the quality of medical schools previously attended by physicians increased more rapidly. States with larger increases in the quality of diagnostic procedures, drugs, and physicians did not have larger increases in per capita medical expenditure.

Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Abstract:
In this paper, we address the issue of spurious correlation in the production of health in a systematic way. Spurious correlation entails the risk of linking health status to medical (and nonmedical) inputs when no links exist. This note first presents the bounds testing procedure as a method to detect and avoid spurious correlation. It then applies it to a recent contribution by Lichtenberg (2004), which relates longevity in the United States to pharmaceutical innovation and public health care expenditure. The results of the bounds testing procedure show longevity to be linearly related to these two factors. Therefore, the estimates reported by Lichtenber (2004) cannot be said to be result of spurious correlation, to the contrary, they very likely reflect an effective relationship, at least for the United States.

health, life expectancy, innovation, pharmaceuticals, health care expenditure, cointegration

Abstract:
There is a large body of work that documents a strong, positive correlation between education and measures of health, but little is known about the mechanisms by which education might affect health. One possibility is that more educated individuals are more likely to adopt new medical technologies. We investigate this theory by asking whether more educated people are more likely to use newer drugs, while controlling for other individual characteristics, such as income and insurance status. Using the 1997 MEPS, we find that more highly educated people are more likely to use drugs more recently approved by the FDA. We find that education only matters for individuals who repeatedly purchase drugs for a given condition, suggesting that the more educated are better able to learn from experience.

Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Abstract:
Importation of drugs into the U.S. would result in a decline in U.S. drug prices. The purpose of this paper is to assess the consequences of importation for new drug development. A simple theoretical model of drug development suggests that the elasticity of innovation with respect to the expected price of drugs should be at least as great as the elasticity of innovation with respect to expected market size (disease incidence). I examine the cross-sectional relationship between pharmaceutical innovation and market size among a set of diseases (different types of cancer) exhibiting substantial exogenous variation in expected market size. I analyze two different measures of pharmaceutical innovation: the number of distinct chemotherapy regimens for treating a cancer site, and the number of articles published in scientific journals pertaining to drug therapy for that cancer site. Both analyses indicate that the amount of pharmaceutical innovation increases with disease incidence. The elasticity of the number of chemotherapy regimens with respect to the number of cases is 0.53. The elasticity of MEDLINE drug cites with respect to cancer incidence throughout the world is 0.60. In the long run, a 10% decline in drug prices would therefore be likely to cause at least a 5-6% decline in pharmaceutical innovation.

Abstract:
We develop a simple theoretical model of the allocation of public biomedical research expenditure, and present some empirical evidence about the determinants of this allocation. The structure of expenditure should depend on the relative costs as well as the relative benefits of different kinds of research. Analysts of technical change typically have data on neither of these, but the measures of disease burden we use are indicative of the benefit of achieving advances against different diseases. We calculate distributions of government-funded biomedical research expenditure, by disease, from records of all research projects supported by the US Public Health Service: to obtain a reasonably complete accounting of disease burden, we utilize data on both the dying (from the Vital Statistics-Mortality Detail file) and the living (from the National Health Interview Survey). We find a very strong positive relationship across diseases between total life-years lost before age 65 and public R&D expenditure. But the amount of publicly-funded research on a disease decreases with the share of life-years before age 65 lost to the deases by non-whites, perhaps because lack of scientific knowledge is a less important cause of premature mortality among non-whites than it is among whites. The number of research grants mentioning a chronic condition is completely uncorrelated with the number of people with the condition but very strongly positively related to the number of people whose activities are limited by that condition. There tends to be more research about chronic conditions that are prevalent among people living in low-income households, and that are prevalent among the young (under age 18) and the old (above age 75).

Abstract:
No abstract is available for this paper.

Abstract:
There is substantial evidence from the literature on individual wage determination that length of service to the firm is an important determinant of earnings and thus of labor productivity, holding constant employee at-tributes such as age, sex, and education. Earnings growth associated with increased tenure is usually interpreted as a reflection of firm-specific on-the-job training (OJT). In this paper a model of producer technology consistent with the hypothesis of firm-specific OJT is formulated and estimated. Empirical implementation of the model on data for U.S. manufacturing provides the basis for estimation of the marginal productivity of workers classified by length of service to the firm, i.e., of the tenure-productivity profile. The parameter estimates also enable us to determine the effect of recent changes in the tenure distribution (due to changes in labor turnover behavior) on manufacturing productivity performance.

Abstract:
In this paper we estimate variants of a labor demand equation derived from a (restricted variable) cost function in which "experience"on a technology (proxied by the mean age of the capital stock) enters "non-neutrally." Our specification of the underlying cost function isbased on the hypothesis that highly educated workers have a comparative advantage with respect to the adjustment to and implementation of new technologies. Our empirical results are consistent with the implication of this hypothesis, that the relative demand for educated workers declines as the capital stock (and presumably the technology embodied therein) ages. According to our estimates, the education-distribution of employment depends more strongly on the age of equipment than on the age of plant, and the effect of changes in equipment age on labor demand is magnified in R&D-intensive industries.

Abstract:
Official government statistics on the "mission-distribution" of U.S. R&D investment are based on the assumption that only the government sponsors military R&D. In this paper we advance and test the alternative hypothesis, that a significant share of privately financed industrial R&D is military in orientation. We argue that in addition to (prior to) contracting with firms to perform military R&D, the government deliberately encourages firms to sponsor defense research at their own expense, to enable the government to identify the firms most capable of performing certain government contracts, particularly those for major weapons systems. To test the hypothesis of, and estimate the quantity of, private investment in 'signaling' R&D, we estimate variants of a model of company R&D expenditure on longitudinal, firm-level data, including detailed data on federal contracts. Our estimates imply that about 30 percent of U.S. private industrial R&D expenditure in 1984 was procurement- (largely defense-) related, and that almost half of the increase in private R&D between 1979 and 1984 was stimulated by the increase in Federal demand.

Abstract:
We estimate the medical cost per life-year gained from increased utilization of HIV drugs by estimating the impact of increased drug utilization on the life expectancy and drug and hospital expenditure of HIV/AIDS patients, using aggregate (U.S. national-level) data for the period 1982-2001. We use IMS Health data on the aggregate number of and expenditure on HIV drug prescriptions, the CDC's AIDS Public Information Data Set, and data from AHRQ's Nationwide Inpatient Sample. Estimates of mortality models imply that actual life expectancy of HIV/AIDS patients in 2001 was 13.4 years higher than it would have been if the drug utilization rate had not increased from its 1993 level. Estimates of a model of hospital discharges imply that increased utilization of HIV drugs caused hospital utilization to decline by .25 to .29 discharges per person per year during the period 1993-2001. Medical cost per additional life-year is estimated to have been $17,175. Treatments that cost this amount are widely considered to be cost-effective. The consistency of this estimate with those from previous studies suggests that analysis of aggregate data may be a useful alternative or additional approach to evaluating the cost-effectiveness of new treatments.

Abstract:
No abstract is available for this paper.

Abstract:
This paper examines R&D activities in the European Community using the Community R&D Information Service (CORDIS) databases. We find that a country's private companies tend to be specialized in the same scientific fields as its universities and public organizations. In addition, we construct indicators of the degree of R&D tacitness and find that greater expected ability to communicate research outcomes encourages less centralized R&D programs. Programs that yield tangible results are less geographically and administratively centralized. The more that research leads to codifiable knowledge, the less centralized R&D activity needs to be.

Abstract:
There have been a number of econometric studies of the effect of changes in management and control on the productivity and employment of private,but not but not of public, enterprises. This paper examines the impact of changes in political administration on the productivity and employment of the entire executive branch of the U.S. government using data compiled under the Bureau of Labor Statistics' Federal Productivity Measurement Program. The estimates Measurement Program. The estimates indicate that the mean rate of productivity growth in the first year of administrations is 2.6 times as high as the mean growth in subsequent years. Also, employment growth is strictly increasing with respect to the administration's tenure: 95% of federal employment growth during the period 1967-94 occurred in the fourth or later years of political administrations, although administrations were that old only 36% of the time. These findings are broadly consistent with evidence about the private sector. They suggest that the inauguration of a new administration initially purges the executive branch, but as an administration's tenure increases, fat and inefficiency tend to accumulate.

Abstract:
In this paper we investigate the incidence of measurement errors in two independent estimates of long-term price change, within the framework of "multiple indicators" models of price measurement. We develop estimates of the measurement-error variances associated with both the Producer Price Index (PPI) and the Census Unit Value Relative (UVR). Our estimates provide support for the generally accepted view that the PPI is a far more reliable indicator of long-term price change: the estimated signal-to-noise ratios for the PPI and UVR are 2.72 and 0.53, respectively. Our estimates should be useful for both constructing an optimal indicator of price change, and for identifying econometric models including error-ridden price- or output-growth terms as regressors. Our analysis suggests that "scores" assigned to product deflators provide useful information about their reliability. By extending our model to explicitly incorporate product-quality change, we are able to assess the importance of the problem posed by quality change for price and productivity measurement. Less than half of quality change, which we estimate to occur at an average annual rate of 1.3 percent, appears to be adjusted for in the PPI. Consequently, estimates of productivity growth based on the PPI underestimate "quality-adjusted" productivity growth by an estimated 43 percent.

Abstract:
Dean Baker and Adriane Fugh-Berman have published a critique of a study I performed in 2007, entitled “Why has longevity increased more in some states than in others?” One of the conclusions I drew from that study was that medical innovation accounts for a substantial portion of recent increases in U.S. life expectancy. Baker and Fugh-Berman claim that my study was subject to a number of major methodological flaws. Many of their claims pertain to the role of infant mortality; the definition of drug vintage; the issue of age adjustment; and the appropriateness of controlling for AIDS, obesity, and smoking in the analysis of longevity.

In this article, I make the case that their claims about my study are largely incorrect. I show that infant mortality was not an important determinant of the growth in U.S. life expectancy during the period that I studied, and that my estimates are completely insensitive to the inclusion or exclusion of infant mortality. Controlling for the age distribution of the population also has essentially no effect on the longevity equation estimates.

I argue that my definition of drug vintage, based on the initial FDA approval year of a drug’s active ingredient, is quite reasonable, and it is consistent with the FDA’s evaluation of the therapeutic potential of new drugs.

I argue that controlling for AIDS, obesity, and smoking in longevity analysis is entirely appropriate and consistent with the epidemiological literature.

Baker and Fugh-Berman express deep skepticism about my study’s conclusion that medical innovation has played a very important role in recent U.S. longevity growth, but they offer no explanation of why life expectancy increased by almost a year during 2000-2006, a period of increasing poverty and obesity and declining health insurance coverage.

longevity, innovation, pharmaceuticals, obesity

Abstract:
This study examines the effect of changes in the vintage distribution of cardiovascular system drugs on hospitalization and mortality due to cardiovascular disease using longitudinal country-level data. The vintage of a drug is the first year in which it was marketed anywhere in the world. We use annual data on the utilization of over 1100 cardiovascular drugs (active ingredients) in 20 OECD countries during the period 1995-2003. Countries with larger increases in the share of cardiovascular drug doses that contained post-1990 or post-1995 ingredients had smaller increases in the cardiovascular disease hospital discharge rate, controlling for the quantity of cardiovascular medications consumed per person, the use of other medical innovations (CT scanners & MRI units), potential risk factors (average consumption of calories, tobacco, and alcohol), and demographic variables (population size & age structure, income, and educational attainment). The estimates also indicate that use of newer cardiovascular drugs has reduced average length of stay and the age-adjusted cardiovascular mortality rate, but not the number of potential years of life lost due to cardiovascular disease before age 70 per 100,000 population. The estimates indicate that if drug vintage had not increased during 1995-2004, hospitalization and mortality would have been higher in 2004. We estimate that per capita expenditure on cardiovascular hospital stays would have been 70% ($89) higher in 2004 had drug vintage not increased during 1995-2004. Per capita expenditure on cardiovascular drugs would have been lower in 2004 had drug vintage not increased during 1995-2004. But our estimate of the increase in expenditure on cardiovascular hospital stays is about 3.7 times as large as our estimate of the reduction in per capita expenditure for cardiovascular drugs that would have occurred ($24).

Abstract:
Dean Baker and Adriane Fugh-Berman have published a critique of a study I performed in 2007, entitled “Why has longevity increased more in some states than in others?” One of the conclusions I drew from that study was that medical innovation accounts for a substantial portion of recent increases in U.S. life expectancy. Baker and Fugh-Berman claim that my study was subject to a number of major methodological flaws. Many of their claims pertain to the role of infant mortality; the definition of drug vintage; the issue of age adjustment; and the appropriateness of controlling for AIDS, obesity, and smoking in the analysis of longevity. In this article, I make the case that their claims about my study are largely incorrect. I show that infant mortality was not an important determinant of the growth in U.S. life expectancy during the period that I studied, and that my estimates are completely insensitive to the inclusion or exclusion of infant mortality. Controlling for the age distribution of the population also has essentially no effect on the longevity equation estimates. I argue that my definition of drug vintage, based on the initial FDA approval year of a drug’s active ingredient, is quite reasonable, and it is consistent with the FDA’s evaluation of the therapeutic potential of new drugs. I argue that controlling for AIDS, obesity, and smoking in longevity analysis is entirely appropriate and consistent with the epidemiological literature. Baker and Fugh-Berman express deep skepticism about my study’s conclusion that medical innovation has played a very important role in recent U.S. longevity growth, but they offer no explanation of why life expectancy increased by almost a year during 2000-2006, a period of increasing poverty and obesity and declining health insurance coverage.

Abstract:
No abstract is available for this paper.

Abstract:
We examine the impact of pharmaceutical innovation on the longevity of Australians during the period 1995-2003. Due to the government's Pharmaceutical Benefits Scheme, Australia has much better data on drug utilization than most other countries. We find that mean age at death increased more for diseases with larger increases in mean drug vintage. The estimates indicate that increasing the mean vintage of drugs by 5 years would increase mean age at death by almost 11 months. The estimates also indicate that using newer drugs reduced the number of years of potential life lost before the ages of 65 and 70 (but not before age 75). During the period 1995-2003, mean age at death increased by about 2.0 years, from 74.4 to 76.4. The estimates imply that, in the absence of any increase in drug vintage, mean age at death would have increased by only 0.7 years. The increase in drug vintage accounts for about 65% of the total increase in mean age at death. We obtain a rough estimate of the cost per life-year gained from using newer drugs. Under our assumptions, using newer drugs (increasing drug vintage) increased life expectancy by 1.23 years and increased lifetime drug expenditure by $12,976; the cost per life-year gained from using newer drugs is $10,585. An estimate made by other investigators of the value of a statistical Australian life-year ($70,618) is 6.7 times as large as our estimate of the cost per life-year gained from using newer drugs. We discuss several reasons why our estimate of the cost per life-year gained from using newer drugs could be too high or too low.

Abstract:
No abstract is available for this paper.

Abstract:
Public policy can impact innovation, eitherencouraging or inhibiting the creation of innovative technologies.Howfive public policies have impacted continued innovation within the U.S.pharmaceutical industry is examined.At issue are the 1962 Kefauver-HarrisAmendment; the 1992 Prescription Drug User Fee Act; Medicare; Medicaid; and the1984 Hatch-Waxman Act. The policies set forth in each act are described, as are their effects oninnovation in the industry.Using a theoretical approach, the impactthat the speed and the ease of pharmaceutical imitation has upon pharmaceuticalinnovation is analyzed, with the belief that it impedes the rapidity ofinnovation within the industry.This theory is assessed, noting that theresearch on the impact of rapid imitation is mixed. The predicted impact of President Clinton's proposed 1993 health care reformpolicies is evaluated by studying the impact of the policies upon the marketvalue of pharmaceutical corporations and upon R&D investment. The findingsindicate that the proposed health care reform policies of the ClintonAdministration threatened the pharmaceutical industry, resulting in a reductionin the growth rate of R&D expenditure. (AKP)

Imitation, Hatch-Waxman Act of 1984, U.S. Food & Drug Administration (FDA), Kefauver-Harris Amendment of 1962, Prescription Drug User Fee Act of 1992, Medicare, Medicaid, Industrial research, Federal regulations, Product development, Public policies, R&D expenditures, Healthcare industry, Policy analysis, Innovation process, Market value, Patents, Pharmaceutical industry

Abstract:
We present three types of evidence about the effect of cancer drug vintage-year of initial world launch - on cancer survival and mortality. All three are based on group-level data, and employ difference-in-difference research designs, which enable us to control for the influence of potentially confounding variables far better than cross-sectional or time-series models.

The first analysis uses data on cancer drug vintage, survival, and other variables, by primary cancer site and year, for U.S. cancer patients during the period 1992-2002. We control for cancer stage distribution, mean age at diagnosis, expected survival, number of people diagnosed, and innovation in diagnostic, radiation, and surgical procedures. We find that the cancer sites whose drug vintage (measured by the share of post-1990 treatments) increased the most during the 1990s tended to have larger increases in observed survival rates, ceteris paribus. Estimates of the fraction of the 1992-1999 change in the observed survival rate that is attributable to the increased utilization of post-1990 drugs range from 12% to 121%; the mean of these estimates was 44%.

The second analysis uses data by primary cancer site and country, for 5 large European countries. Drug vintage (in this case measured by the share of post-1985 treatments) has a positive and statistically significant effect on both 1-year and 5-year survival rates. The difference in the fraction of post-1985 cancer drugs accounts for 14-19% of the (probit of the) 5-year survival rate differential, adjusted for international differences in distribution of cancer sites. Since the data on survival and on drug utilization pertain to different time periods, this estimate is probably conservative.

The third analysis is based on data by country and year, for all cancer sites combined, for 20 countries during the period 1995-2003. We find that countries with larger increases in the mean launch year of cancer drugs had larger declines in the age-adjusted cancer mortality rate. A 10-year increase in drug vintage is estimated to reduce the cancer mortality rate by 5.9%, controlling for per capita GDP growth. The increase in cancer drug vintage - in other words, the use of newer cancer drugs accounts for about 30% of the GDP-growth-adjusted decline in the age-adjusted cancer mortality rate.

cancer, pharmaceutical, innovation

Abstract:
After increasing steadily from 1987 to 1995, the number of U.S. deaths caused by HIV declined sharply from 1995 to 1998. We analyse aggregate data to consider the hypothesis that this decline was due to a rapid increase in the number of drugs available to treat HIV. The evidence suggests that new drugs played a key role in the post-1995 decline in HIV mortality. The annual number of HIV deaths is estimated to have been reduced by over 6000, on average, by an additional HIV drug approval. The social return to HIV drug innovation appears to be extremely large.

Health, HIV, mortality, new drug approvals