A Model for Projecting the Number of Doctors in South Korea

Abstract

This paper develops a new model for forecasting potential shortfalls in the healthcare sector, providing an economically grounded framework for projections. The model is applied to assess potential doctor shortages in South Korea over the next decade under reasonable economic scenarios. Our analysis indicates that demand for healthcare, driven by aging-related factors, is projected to grow at an annual rate of 1.3% to 1.9%. In contrast, the supply of healthcare—bolstered by technological advancements, improved medical equipment, and natural growth in the doctor workforce—is expected to increase by 3.2% annually. These findings suggest that South Korea’s healthcare system is likely to meet future demand without necessitating an expansion of medical school admissions.

Graphical Abstract

INTRODUCTION

Since early 2024, South Korea has endured a prolonged political standoff between the government and the medical community over a proposed policy to increase medical school admissions by 2000, in addition to the existing quota of 3058. This conflict has led to significant patient suffering and increased public anxiety, as many doctors-in-training have left hospitals in protest of the government’s policy.

The justification for the government’s plan to expand medical school admissions hinges on whether South Korea currently faces, or is likely to face in the future, say within 10 years, a shortage of doctors. Although the government asserts that a shortage presently exists and will worsen over the next decade, it has not provided convincing evidence to support this claim, leading to strong opposition from the medical community. Conversely, doctors have also not presented a compelling argument against the assertion of a doctor shortage.

To resolve the healthcare crisis, an objective forecast is needed—one that the public can trust and agree upon regarding the projected doctor shortfall in 10 years. However, to date, there has hardly been any model available to provide an objective forecast of the shortfall of doctors.

To fill this gap, this paper develops a model for rationally forecasting potential shortfalls, offering a framework for an economically grounded projection that can serve as a basis for constructive dialogue between the government and the medical community in any country.

The model is then applied to assess potential doctor shortages in South Korea over the next decade under economically reasonable scenarios. Our findings indicate that, with plausible assumptions regarding increased demand due to aging, the annual growth in healthcare demand is projected to range between 1.3% and 1.9%. In contrast, the supply of healthcare—driven by technological advancements, increased medical equipment, and the natural growth of the doctor workforce—is expected to grow at an annual rate of 3.2%. Therefore, it appears that South Korea’s medical service supply will likely meet future demand without the need to expand medical school admissions.

MODEL FOR PROJECTING SHORTFALLS OF DOCTORS

Economically, whether there will be a shortage of doctors in 10 years depends first on whether there is currently a shortage of doctors and, second, on whether the rate of increase in demand for healthcare services will exceed the growth in supply.

This paper develops a model for projecting doctor shortages. The core idea of the model is straightforward: it compares the future growth rates of demand and supply of medical services over the next decade. If the current supply meets demand but demand grows faster than supply, an increase in medical school admissions will be necessary to address the shortfall. Conversely, if supply grows faster than demand, the existing supply will suffice, making expansion of medical school admissions unnecessary.

Our model for projecting doctor shortages consists of four stages. The first stage assesses whether the current demand for medical services is adequately met by the existing supply. The second stage projects the growth rate of the demand for medical services, while the third stage projects the growth rate of the supply of medical services. The final stage compares the two rates to determine whether there will be future shortages of doctors.

Predicting the growth rates of healthcare demand and supply precisely is challenging. However, we can still make reasoned predictions by applying established economic theory and utilizing any available information effectively.

Another distinctive feature of our model is its use of the standard production function approach in economics, along with the growth accounting method proposed by Nobel laureate Robert Solow in his 1957 paper, to predict the growth rate of medical supply.¹

On the supply side, the output of all goods and services, including healthcare, depends on technology, capital, and labor. Consequently, the production relationship for medical service is represented by the following production function:

Y=A F (K, L)

where Y denotes the output of healthcare service, A technology, K capital input (such as medical equipment), and L labor input (such as doctors).

The production function, a fundamental concept in economics, is widely used as it provides a structured way to analyze how inputs are transformed into outputs. The production function represents the relationship between input factors (labor and capital) and the quantity of output produced. If one knows the amount of output that a country needs to produce, the production function enables the calculation of the required combination of labor and capital.

This production function is a general concept applicable to the production of any goods or services, including medical services. Therefore, it is an extremely useful tool to evaluate or project the number of medical doctors needed by a country. For example, consider a country where one million people develop diabetes. In this case, the country needs to produce medical treatment services for one million diabetes patients. If the production function for medical treatment services is known, it becomes possible to calculate the required combination of medical doctors (labor) and medical equipment (capital). Furthermore, if the amount of medical equipment remains constant, one can determine the necessary number of medical doctors to treat one million patients.

This standard production function approach implies that labor and capital are substitutable. This means that various combinations of labor and capital can be used to produce the same level of output. For example, to treat 100 patients, one doctor with 10 medical devices could be utilized, or alternatively, 10 doctors with one medical device. Some countries may adopt a more capital-intensive approach to producing medical services, while others may rely on a more labor-intensive method.

Hence, the production function suggests that if the quantity of output to be produced is fixed, for example, at 100 patients an increase in capital (such as medical equipment) will necessarily reduce the number of doctors required. However, if the output is not fixed and can vary, an increase in capital raises the marginal productivity of labor (doctors), leading to higher wages for doctors. This, in turn, may result in an increase in the supply of doctors.

This production function also suggests that healthcare service output increases with better technology, more capital, and more labor. Consequently, the growth rate of healthcare supply will ultimately depend on the rates of technological progress, capital growth, and doctor growth.

Following the growth accounting approach proposed originally by Robert Solow, we can be more specific about the relationship among the growth rates by using the most popular Cobb-Douglas production function among possible production functions. The Cobb-Douglas production function, which was originally proposed by Cobb and Douglas’ 1928 paper, takes the following form:²

Y=A K^α L^1-α (1)

where α is capital income share, and 1-α is labor income share, which is equivalent to 1 minus capital income share.

By taking the logarithm of the Cobb-Douglas production function and taking a time derivative, as done in the growth accounting approach, we can express healthcare supply growth (denoted by g) as follows:

g=λ+αg_K+(1-α)g_L (2)

where λ represents the rate of medical technological progress, g_K the growth rate of capital, and g_L the rate of doctor growth.

Growth accounting, originally proposed by Solow¹, decomposes the growth rate of aggregate output of an economy into contributions from aggregate capital, labor, and technology, similar to equation (2), which applies to the medical industry. Since then, this approach has been widely used to measure an economy’s rate of technological progress—an otherwise difficult-to-measure factor—using data on aggregate output and input growth rates. For instance, Young applied this method to assess the role of technological progress and factor accumulation in the rapid economic growth of East Asian countries during the 1966-91 period.³ Barro further explored key issues associated with this growth accounting framework.⁴

Our model adapts this growth accounting approach, originally developed for the entire economy, to the medical services industry. The resulting growth accounting relationship in equation (2) means that healthcare supply growth rate is the sum of the rate of technological progress, capital income share times capital growth rate, and labor income share times the growth rate of doctors.

Using this relationship, we can estimate future changes in the key factors that determine healthcare service supply: technology, capital, and labor. Hence, the growth rate of healthcare supply will ultimately depend on the rates of technological progress, capital growth, and doctor growth.

APPLICATION TO THE KOREAN CASE

We apply this model, based on standard production function approach, to evaluate whether South Korea will face a shortage of doctors in 10 years, following four steps. First, we assess whether the current medical demand is adequately met by the existing supply. Next, we project future demand growth for medical services and the future supply growth of medical services. Finally, we compare the projected demand and supply growth.

Current demand and supply

To address the issue of current physician shortages, we first examine the number of physicians per capita. Economically, physicians are workers who produce healthcare services, such as diagnosing and treating illnesses. According to OECD statistics, as of 2021, South Korea has 130000 physicians, each responsible for the health of approximately 400 citizens.⁵ This translates to 2.6 physicians per 1000 people, significantly lower than the OECD average of 3.7.

If healthcare services are produced in direct proportion to the number of physicians, which implies Y=AL instead of equation (1), South Korea’s relatively low number of doctors would suggest an undersupply of healthcare services. In that case, the claim of a physician shortage would appear valid. This perception likely motivated the government’s efforts to increase the number of physicians.

The next step is to verify whether the lower number of physicians actually results in an undersupply of healthcare services, using data as evidence. One way to measure the quantity of healthcare services is by examining how many times citizens visit doctors for treatment in a given year.

In OECD countries, the average citizen visits a doctor eight times annually. If healthcare services were strictly proportional to the number of physicians (with implied production function Y=AL and the same A across countries), South Korean citizens, with fewer doctors per capita, would be expected to visit doctors fewer than six times per year.

However, the data shows that South Korean citizens visit doctors an average of 16 times per year—twice the OECD average. Despite having fewer physicians, the number of doctor visits per person, a measure of healthcare service production, is remarkably high in South Korea, ranking first among OECD countries.

In terms of healthcare service production per physician, or labor productivity, South Korean physicians perform approximately 6000 patient consultations annually. When measured by the number of consultations, South Korean doctors exhibit labor productivity three times higher than the OECD average. With such high productivity, it is difficult to argue that the country is unable to meet its healthcare service needs solely due to a low number of physicians.

Of course, the quality of care must also be evaluated to determine whether the high number of consultations per physician translates into successful patient outcomes. If an increased volume of consultations leads to a significant decline in quality, it could still support the argument for a physician shortage. Therefore, it is essential to assess whether the rise in consultation volume negatively impacts service quality. This includes evaluating not only consultation frequency but also treatment outcomes, such as success rates.

According to OECD data, South Korea ranks among the top countries in 5-year survival rates for cancer patients. It is ranked first for 5-year survival rates of stomach and colorectal cancer and third for lung cancer. These outcomes suggest that the quality-adjusted healthcare service production in South Korea is not low compared to other countries.

In conclusion, while South Korea has a lower number of physicians per capita, the combination of high productivity and high-quality care suggests that the country currently does not face a shortage of physicians.

The reason South Korea can provide top-tier healthcare services among OECD countries, despite having relatively fewer physicians, lies in the nature of its production function. The function follows a Cobb-Douglas form (1), where labor and capital are substitutable, rather than a simple linear form Y=AL. In addition, South Korea uses a capital-intensive production method that relies heavily on medical equipment rather than physicians. This is similar to how farmers in the United States achieve high productivity by using fewer workers and more machinery, such as tractors, in agricultural production.

Adopting a capital-intensive approach increases the healthcare output per physician. For example, increasing the use of medical devices and diagnostic equipment can significantly reduce the time physicians spend with patients. Consider a patient with stomach pain: instead of conducting a medical interview and prescribing acid reducers, followed by waiting 3–4 weeks to observe the medication’s effectiveness, a physician could perform a gastroscopy and arrive at a precise diagnosis within an hour. By using such equipment to shorten consultation times, a single physician can see more patients, enabling fewer number of doctors to treat a larger population.

In fact, South Korea ranks among the highest in the OECD in terms of medical equipment usage. For example, diagnostic tests using CT scans are conducted 280 times per 1000 people annually in South Korea, nearly double the OECD average of 160. Additionally, the number of CT, MRI, and mammography machines per million people is also among the highest in the OECD, comparable to the United States. This indicates that South Korea’s healthcare system employs a highly capital-intensive and time-efficient production model.

Upon examining the data and considering all available evidence, it is challenging to find objective support for the claim that South Korea currently faces a severe shortage of physicians. While the country has fewer physicians per capita compared to the OECD average, it compensates by using a capital-intensive, time-saving approach, producing and delivering more healthcare services per person than any other OECD country.

Demand growth

There are various factors that affect the future demand for medical services. Among these, the most significant socio-economic determinant likely to change in the next 10 years is the rapid increase in the elderly population, combined with a significant decline in the total population, as projected by Statistics Korea. As a result, the sharp rise in the aging population is expected to be the most significant factor influencing healthcare service demand in South Korea. In our benchmark case, we focus on the impact of this aging population trend.

While the decline in the total population is expected to reduce healthcare demand, the simultaneous increase in the elderly population will likely drive the demand upward. The net effect of these changes on the overall demand for medical services depends on which trend dominates. Even with a shrinking total population, healthcare demand may increase if the growth of the elderly population outpaces the decline in other demographic groups.

Therefore, it is crucial to carefully consider the impact of an aging population on increased healthcare demand over the next decade. For this analysis, we rely on projections of the non-elderly and elderly population provided by Statistics Korea. If we know the per capita medical demand (e.g., the average number of doctor visits per year) for both non-elderly and elderly groups, we can calculate the total national demand for medical services for each of the next 10 years. This is done by summing the product of the population size and per capita consultation rate for each group. A key factor in this projection is the elderly-to-non-elderly healthcare demand ratio (β), which represents how many more times an elderly person consults a doctor compared to a non-elderly individual.

According to Statistics Korea, from 2008 to 2016, the healthcare demand of one elderly person (aged 65 years and older) was 2.3 to 3.2 times that of a non-elderly person (aged under 65 years). Assuming this trend continues, the elderly-to-non-elderly healthcare demand ratio (β) is expected to reach 4.1 in 2025 and 5.1 in 2034, averaging 4.6 over the next decade. Thus, for every one hospital visit by a non-elderly person, an elderly person would visit 4.6 times.

Using this elderly-to-non-elderly healthcare demand ratio of 4.6 and Statistics Korea’s future demographic projections for the non-elderly and elderly population, healthcare demand is projected to grow by 1.7% annually over the next decade. If we assume the elderly-to-non-elderly healthcare demand ratio will be between 3.2 (the maximum for the 2008–2016 period) and 5.1 (the maximum for the next decade), then the annual growth in healthcare demand is expected to range between 1.3% and 1.9% (Fig. 1).

Fig. 1. Demand for medical service in the future.

Demand for healthcare services may also depend on the patients’ income, particularly that of elderly individuals. Many elderly people likely lack a stable income source after retirement, which could lead some to forgo seeking care due to financial constraints. This would reduce the elderly-to-non-elderly healthcare demand ratio. However, to take a conservative approach, we proceed with an estimate of 4.6.

Demand for medical services may also be influenced by prices set by the government. It is more likely that the government would raise prices, especially for essential services, rather than reduce them. If prices were to increase, the demand for medical services would likely decrease. However, to maintain a conservative stance, we assume no change in the price of medical services in our benchmark case.

There are several other factors, whichcould either reduce or increase the demand for medical services. For example, younger doctors may prefer not to work as intensively as their senior counterparts and may seek reduced working hours, or the public may desire more consultation time with doctors. If either of these occurs, it could increase the demand for medical services. Conversely, elderly people may experience improvements in health, or advancements in AI and robotics may reduce the need for doctor consultations. Both of these factors would decrease the demand for healthcare services. For our benchmark case, we assume that these countervailing effects cancel each other out, resulting in no net change in demand.

Supply growth

On the supply side, we project future changes in three key factors, which are technology, capital, and labor, to determine the growth rate of medical service supply. For this projection, we first consider a benchmark case with conservative assumptions.

To project technological progress in medical services, it is essential to consider the rapid advancements in AI and medical technology. However, no rigorous existing study currently provides an estimate of the growth rate for these advancements. In light of this, our strategy adopts a conservative approach for the benchmark case, assuming that the rate of technological progress in healthcare aligns with the Bank of Korea’s estimated national average of 0.4% per year over the past decade (see Cho, 2023).⁶ Based on this conservative projection, we assess whether supply growth can outpace demand growth. We then conduct robustness checks to explore alternative scenarios, including a zero-growth rate (e.g., if one believes that 0.4% is still too optimistic) or a significantly higher growth rate, to account for the potential impact of advancements in AI.

Medical equipment (capital) can be quantified by the number of devices, such as CT scanners, MRIs, and other diagnostic tools, as reported in the OECD Health Statistics 2024. In South Korea, the availability of such equipment grew at an average annual rate of 3.5% between 2016 and 2022, as shown in Fig. 2. Moreover, this growth rate exhibited a strong upward trend over the period. Assuming that the upward trend in the capital growth rate observed during this period continues, the capital growth rate over the next 10 years is estimated to be 5.4% per year. However, to adopt a more conservative assumption, let us instead assume that the past average growth rate of 3.5% per year will persist over the next decade.

Fig. 2. Growth rate of capital (=medical equipment).

To predict doctor growth, we note that the medical school quota has been fixed at 3058 students since 2006, but the number of active doctors has increased by an average of 2670 per year due to extended work years. However, since the growth trend gradually declined during the 2010–2020 period, it is reasonable to assume that this declining trend will continue when estimating the growth rate of physicians (Fig. 3). Under this assumption, the doctor growth rate over the next 10 years is estimated to be 1.8% per year.

Fig. 3. Growth rate of active doctors: average and trend.

For the labor income share (or alternatively, 1 minus the capital income share) in the healthcare industry, we can use data from the U.S. Bureau of Labor Statistics, given the similar capital-labor ratios between the U.S. and South Korean healthcare sectors. According to the U.S. Bureau of Labor Statistics, the labor income share is estimated to be 0.4.

With these conservative estimates and equation (2), we calculate the healthcare supply growth rate over the next decade as follows:

Healthcare supply growth rate=0.4%+0.6×3.5%+0.4×1.8%=3.2%.

Comparison of demand and supply growth

Thus, healthcare supply growth over the next decade will reach an annual average of 3.2%, exceeding the demand growth rate range of 1.3%–1.9% by at least 1.3 percentage points.

This suggests that the supply of medical service will sufficiently meet healthcare demand in 10 years. In conclusion, even with the increase in demand due to aging and other factors, the growth in supply—driven by technological advances, increased medical equipment, and the natural increase in the doctor workforce—can likely meet future demand without expanding medical school admissions.

Robustness check

We now perform a robustness test of the benchmark case results to explore alternative scenarios. What would happen if we apply even more conservative assumptions to the above scenario?

First, consider the possibility that the benchmark rate of 0.4% is still overly optimistic. Suppose there is no technological progress at all in the next decade. Under this assumption, the supply growth rate would decrease from 3.2% to 2.8%, which would still exceed the demand growth rate.

Additionally, suppose that the capital growth rate were further reduced to 2%. In this case, the supply growth would still be 2.3%, exceeding demand. Even when we assume that a doctor growth rate is 1% lower than estimated, supply growth would still reach 2.8%, exceeding demand growth. In the extreme case of no increase in the number of doctors, supply growth would still reach 2.5%, driven by capital growth and technological progress.

Suppose that factors other than the aging population, which increase the demand for medical services (e.g., younger doctors’ preference for reduced working hours), outweigh those that reduce it (e.g., the possibility that elderly people may experience improvements in health), resulting in a 0.5% increase in demand. In this case, the demand growth could range from 1.8% to 2.4%, still lower than the supply growth of 3.2%.

On the other hand, we might adopt less conservative assumptions to evaluate a broader range of possible outcomes. On the demand side, we could consider the possibility that the government might raise the prices of medical services for necessary treatments, leading to a decline in demand for medical services. Additionally, we might consider a scenario in which a significant number of elderly patients without labor income would limit their visits to doctors, resulting in a lower demand for medical services compared to the benchmark case.

On the supply side, we would factor in the potential for rapid development and adoption of medical AI, which could substitute for human physicians. In this scenario, the rate of technological progress would be much higher than in our very conservative benchmark case. This would lead to a more rapid growth of the medical supply compared to the benchmark.

Following the growth accounting tradition, our benchmark assumes that the growth of technology, capital, and labor are independent and are linearly added to calculate supply growth, as shown in equation (2). However, these elements can, of course, influence one another. For instance, an increase in capital raises the marginal productivity of labor, which in turn increases the supply of labor. This positive influence of capital on labor would result in a higher supply growth rate compared to our benchmark case. Considering this possibility, the projected gap between supply and demand growth would be even larger, further strengthening our argument.

These considerations suggest that, under less conservative assumptions, the rate of demand growth could fall short of the benchmark, while the rate of supply growth could exceed it, making it highly unlikely that South Korea will face a shortage of over 10000 doctors in the next 10 years.

A recent study by Shin was cited by the Korean government to support its claim that there will be a shortage of 10000 physicians in the next 10 years.⁷ This study projects a wide range of the number of doctors needed by 2035, based on various assumptions, including doctors’ working days and hours. It uses medical utilization data from 2010 to 2018 to evaluate the potential future shortage of doctors. While it is possible to construct scenarios that predict a shortage exceeding 10000 doctors within a decade, such projections would likely rely on unrealistic assumptions, particularly regarding the production function for medical services.

According to Lee,⁸ most existing studies on doctor shortages in South Korea, including his own, assume that the demand for doctors increases at the same rate as the demand for medical services. For instance, Shin⁷ also assumes that the demand for medical services directly translates into a proportional demand for doctors.

Such an approach disregards the roles of technology and capital in healthcare production, effectively presuming that only an increase in the number of physicians can address the growing demand. In terms of the production function, it is equivalent to assuming the production function for medical services is

Y=AL,

which is unrealistic, instead of the standard equation (1) for the production function. If one uses this unrealistic production function with an identical technology parameter A across countries, it would lead to the mistaken conclusion that South Korea, with a lower physician-to-population ratio than the OECD average, is currently supplying fewer medical services. This argument has been used by the Korean government to support its claims of physician shortages in the country.

The dynamic version of this unrealistic production function is

g=g_L,

which leads to the equally unrealistic conclusion that demand growth must be met solely by the equivalent rate of physician growth. The Korean government relied on reports that implicitly adopt this assumption to justify its claim of a physician shortage in 10 years.

In contrast, our approach considers that medical service demand is met by medical service supply, which depends not only on the number of doctors but also on capital and technology. This framework suggests that increases in medical demand can be addressed through growth in any of these factors—capital, technology, or physicians.

CONCLUSION

In this paper, we develop a novel framework for projecting doctor shortages in a rational manner. We then apply the model to assess potential doctor shortfalls in South Korea under economically reasonable scenarios, providing a basis for constructive dialogue between the government and the medical community.

The scenarios considered in this paper cover a wide range of possibilities but do not encompass every possible case. Before making arguments about the necessity of adding 2000 medical school seats, the government and medical community should clearly outline the specific assumptions behind their figures on doctor demand and supply, using the framework presented in this paper. Such transparency will allow the public to assess which assumptions are more reasonable and realistic, ultimately facilitating a resolution to the healthcare crisis through objective, data-driven projections.