Forecasting Insurance Demand
For this purpose, we use (i) a population projection for Germany, (ii) a macroeconomic simulation model and (iii) an empirical estimation of premiums for various insurance products, based on data from the German sample survey of income and expenditure (EVS). Our results show in the baseline scenario (i) a sharp decline in the labor force participation rate from 60% to 53% and an increase in the old-age dependency ratio from 35% to over 50% by 2040; (ii) a decline in returns to capital and income growth rates and, indexed to today's real terms, an increase in GDP per capita until approx. 2025, declining thereafter; (iii) a decline in indexed disposable income of about 12% until 2040; (iv) pronounced age and cohort effects in the demand for insurance products; (v) a decline in life insurance premiums by 8% in today's real terms until 2030; (vi) in the case of an increase in the retirement age, a decline in this demand by only 5% due to increased incomes; (vii) an increase in premiums for private health, long-term care, disability and accident insurance by 60%; (viii) all insurance products analyzed together, an increase in demand by 10% by 2040, which corresponds to an increase from the current approx. 4.2% to 4.7% of GDP.
Point of Departure
Demographic change will have a serious impact on many demographically aging economies in the next two decades. Within the European Union, Germany is one of the countries most affected by a decline in the labor force and an increase in the old-age dependency ratio. This development has received the attention it deserves in economic research for almost 40 years and is becoming omnipresent due to the current shortage of skilled workers. Its impact on the financing of social security systems and on the growth and distribution of resources in an economy has been studied in numerous academic papers. Here, we analyze the effects of demographic change on the demand for insurance products. In our analysis, we focus on the period up to 2040 and develop a quantitative forecasting model to answer the question. We focus on the demand for life insurance products measuring the "demand" for these products through the corresponding current premiums. Our analysis covers about 86% of the regular premiums of the German life insurance market. As an extension of our core analysis focused on life insurance, we also look at private health, long-term care, disability and accident insurance, motor insurance and private liability insurance.
Approach and Main Findings
The quantitative forecast model we develop is composed of four modules, (A) a population forecast for the German population, (B) a macroeconomic simulation model, (C) an empirical analysis of insurance demand based on the EVS and (D) a combination of the three aforementioned modules in a quantitative simulation based on a "shift-share" method.
Module A: Demographic Forecast
The demographic model extrapolates the population level of today into the future under various assumptions on the development of mortality, fertility and migration. In the baseline scenario, the results of the demographic forecast show a slight decline in the total population to about 80 million people by 2040. This slight shrinking process is accompanied by much-discussed dynamics of population ageing. The labour force participation rate ‑ the share of 20-64 year-olds in the total population in a given year ‑ will fall from currently about 60% to about 53%. Similarly, the old-age dependency ratio ‑ the ratio of the population aged 65 and older to the labour force (age 20-64) ‑ will rise from currently about 35% to about 50%.
Module B: Macroeconomic Framework
The macroeconometric modeling approach we use employs a so-called overlapping generations model, through which it is possible to represent demographic trends in the interaction between households and firms in as much detail as we calculate in the population model. The core of the model is the household sector. Model households, which differ in age, origin (native and non-native population) and education level, optimize their consumption and savings decisions over the life cycle. From these, together with assumptions about firm behavior, we can then calculate the amount of production factors capital and labor used in the production process in the economy, and finally the total income (gross domestic product). Due to the aging process, this per capita income-expressed in today's real purchasing power-will initially rise-the shrinking of the population means that a given income must be distributed among fewer people-and will already fall after 2025, due to the sharp decline in labor force potential. At the same time, the macroeconomic framework gives us information on the development of gross wages and returns to capital. The scarcity of labor in an aging population will cause real gross wages to rise and returns on capital to fall; thus, average returns will fall by about one percentage point by 2030. Real risk-free interest rates will remain negative over the next two decades in the baseline scenario.
By additionally modeling the government sector, in which we make projections of the expected trajectories of social security contributions and payments, we can also make projections of net wages and, ultimately, of real household disposable income. Our model predicts that average disposable income ‑ expressed in today's real purchasing power ‑ will fall by about 12% by 2040.
Module C: Empirical Analysis
The aim of our empirical analysis, which is the core of our work, is to identify and quantify the drivers of demand for insurance products. To this end, we estimate an econometric model using data from the EVS based on the survey waves of 2003, 2008, 2013, and 2018, which conditions current premiums on age, birth cohort, disposable income, and other covariates. As a result, we find pronounced age and cohort effects, which are summarized here using the two main products of our analysis ‑ the endowment insurance and private pension schemes ‑ as examples. According to our decomposition by these effects, demand for endowment insurance falls at old age while that for annuity insurance rises at old age, with both profiles flattening out from age 65. The identified cohort effects ‑ these describe the change in savings behavior in the products across individual birth cohorts ‑ indicate that the demand for endowment insurance is lower the later the household heads were born. The reverse is true for private pension insurance: Households with younger household heads demand more of these products. The effect of disposable income on premiums for both products is positive.
Module D: Simulation in a Shift-Share Analysis
Finally, we merge Modules A-C into a simulation model, which takes the demographic projections of Module A, the income and return projections of Module B, and the coefficients of Module C estimated in the empirical analysis as input factors. Holding the regression coefficients of Module C constant, we produces a quantitative evaluation of insurance demand over the next two decades through a "shift-share" analysis.
The quantitative analysis based on the simulation model ultimately leads us to our core results. The year 2018 serves as the starting year, as this is the last year of the EVS. According to our estimates, current contributions to endowment life insurance (ELI) will steadily decline, falling from around 14 billion euros in 2018 to zero in around 2035. By contrast, contributions to private pension schemes (PPS) will rise sharply, from €31.5 billion in 2018 by almost 50% by 2040. The reason for these opposing effects for the two products is in particular the age and cohort effects outlined earlier, which cause a drop in demand for ELI and an increase for PPS. Similar to the ELI, the age and cohort effects for term life insurance (TLI) also point to a decline. For example, premiums for this product will decline by about 90% by 2040, from 3.2 billion euros in 2018. If we combine ELI, PPS and TLI, we see a demand reduction by 2035 due to a 8% decline from 48 billion euros in 2018. It is important to emphasize that all these results are calculated in today's purchasing power. Relative to GDP, this means that the share of contributions to these three products (ELI+PPS+TLI) will decrease from 1.45% to 1.35%. Of the various demographic scenarios we consider as sensitivity analyses, we would like to mention in particular the effect on insurance demand of increasing the retirement age by indexing it to life expectancy. As this causes an increase of incomes, the decline in demand for ELI+PPS+TLI is reduced from 8% to 5% and the demand dip as a share of GDP is shifted forward ‑ the minimum demand is already reached in 2025 ‑ with a complete recovery by 2040.
Furthermore, from the non-life sector, we should mention the demand for private health, long-term care, occupational disability and accident insurance (HI). Demand for these products is positively affected by the aging process and will increase by around 60% from EUR 49 billion by 2040. This will stabilize the insurance industry along the demographic transition. If we summarize the entire insurance market we analyzed (i.e., ELI, PPS, TLI, HI and other insurance products), premiums will increase from 4.1% of GDP in 2018 to 4.7% in 2040.
Demographic change will likely lead to a decline in demand for life insurance products ‑ the sum of private pension schemes, endowment life insurance, and term life insurance. We show that demand for life insurance will fall by 8% relative to an economy with constant real growth ‑ in particular, constant demographics. In the process, demand for endowment life insurance falls sharply. By contrast, demand for private pension schemes will rise sharply-by about 40% by 2040. Raising the retirement age by indexing it to life expectancy would reduce some of the demographically induced demand effects. This negative trend in the life sector is counteracted by demand for private health insurance. As the latter is set to increase, this positive demand more than offsets the negative effects in the life sector. Thus, in the long-term view aimed at here, the demand for all insurance products analyzed rises from a current share of 4.1% of GDP to 4.7%.