Für neue Autoren:
kostenlos, einfach und schnell
Für bereits registrierte Autoren
27 Seiten, Note: 1,3
1 Introduction ... 1
2 Demographic Transition in Germany and other Nations ... 2
3 The Life-Cycle Saving Model ... 4
4 Demography and Asset Prices ... 5
5 Empirical Analysis ... 7
5.1 Model Variables and Expectations ... 8
5.2 Probit Estimation for 2004 ... 10
5.3 Probit Results over Time ... 14
6 Discussion ... 18
6.1 Discussion of the Empirical Results ... 18
6.2 Implications of the Empirical Results for Asset Prices ... 19
7 Conclusion ... 21
A Appendix ... i
B References ... iii
C List of Figures ... v
D List of Tables ... v
Most developed countries will be going through a strong demographic transition over the coming decades. The large Baby Boomer cohort is approaching retirement and whereas it is often believed that they brought an economic dividend when they were working, it is assumed they will prove to be a drain on economic growth as they retire. Equally, financial markets are worried that the Baby Boomers will put the financial assets they accumulated over their life-cycle simultaneously on the market, thereby causing asset prices to tumble. However, looking at data about the life-cycle saving behaviour of households casts doubts that there is strong asset deccumulation even in the very old age. This work contributes on this issue by providing empirical evidence on individual stock market participation over the life-cycle in Germany between 2000 and 2011. The results give only weak support for strong asset deccumulation during retirement. In combination with the institutional design of the German pension system and other findings, demographic ageing is unlikely to lead to an asset price meltdown.
Between 2015 and 2050 the share of the elderly in relation to the working age population is projected to almost double in Germany. As the Baby Boomers retire and exit the labour market, a shrinking labour force will likely depress economic growth and the German Government will struggle to keep its budget balanced (European Commission 2015).Regarding the latter, there was – and still is – great concern that the German pension system is not sustainable, so that future social security contributions will not be enough to provide promised public pensions (Raffelhüschen 2002). For this reason, Germany introduced the so-called Riester-Rente in 2002, which promotes private retirement provision plans and is likely to increase the amount of privately held financial assets.
But the value of these assets may decrease as demographic ageing is likely to affect financial markets, ultimately leading to tumbling asset prices referred to as the asset price meltdown hypothesis. The hypothesis is derived from the life-cycle saving model (Modigliani and Brumberg 1954), which suggests that people accumulate assets during their working age, which they sell to finance consumption during retirement. Thus, if the Baby Boomers put their assets simultaneously on the market when retiring, it could trigger a price drop and result in a great loss of the cohort’s retirement provisions. However, in contrast to most theoretical models, empirical evidence for the asset price meltdown hypothesis is mixed. For example, findings seem to be affected by the empirical specification or vary across countries (Brooks 2006), whereby most research focuses on the US. More importantly though, household survey data for the US seems to partly reject the life-cycle hypothesis (Poterba 2001). Consequently, if there is no asset deccumulation when retiring, demographic ageing could hardly affect asset prices from a supply and demand perspective.
The contribution of this work is to provide empirical evidence for the life-cycle hypothesis in Germany. Based on data from the German Socio-Economic Panel (SOEP), I test whether there is an effect of age on individual stock-ownership between 2000 and 2011using a Probit model.
The remainder of this work is organised as follows. Section 2 looks briefly at developments in global demographic change. Section 3 describes the life-cycle hypothesis in more detail and checks if it holds for the German case. Section 4 reviews the empirical literature regarding household asset holdings and the asset price meltdown hypothesis. Drawing from this literature, Section 5 presents my own empirical analysis followed by a discussion of the results and its implications for asset prices in Section 6. Finally, Section 7 concludes with a short summary of the paper.
If preferences about financial assets change with age, then the logical conclusion would be that a demographic change also changes aggregated preferences about financial assets, which in turn should induce price changes. Therefore, the stronger demography changes, presumably the larger the effect on asset prices. Hence, it is reasonable to first look at the magnitude of demographic change, before turning to the analysis of the relationship with financial assets in the following chapters.
Figure 1: Elderly dependency ratio by selected countries and regions between 2015 and 2050.
[Images are not displayed in this preview]
Source: United Nations, Department of Economic and Social Affairs, Population Division (2015). World Population Prospects: The 2015 Revision.
Figure 1 shows the elderly dependency ratio over time for selected countries and regions. The blue bars represent the medium and the black error bars the low and the high fertility variant respectively. We can see from the graph that Germany is particularly affected by demographic ageing. According to the medium variant, the elderly dependency ratio will approximately double in Germany between 2015 and 2050. Put another way, instead of just one, there will be two retirees that depend on three workers in 2050. The reason for that is the large Baby Boomer cohort that is moving up the population pyramid towards retirement, see Figure 2.
As Figure 1 also depicts, demographic ageing is not only a German phenomenon. In fact, most developed countries will have to cope with an ageing population.
Among them, Japan faces the most severe demographic transition, but also emerging countries like China see their elderly population rising fast. Only the least developed countries do not seem to be affected by demographic ageing very much.
Although population prospects are associated with a certain degree of uncertainty, they are much more reliable than forecasts of other economic variables, e. g. GDP. Still, there are some aspects that should be mentioned. First, as illustrated by the error bars, population prospects are particularly affected by a change in the fertility rate. At first, an increase in fertility would not immediately change the elderly dependency ratio, but as the respective cohort joined the working age population, the ratio would decline. Second, the mortality rate and life expectancy play an important role. Considering life expectancy, the longer the elderly live, the more the dependency ratio would increase. Third, if the base year for the population prospects was inaccurate, so are the predictions that are based on it. Fourth, events like a war or an epidemic disease (e. g. Ebola) can have drastic effects on population prospects. Finally, the refugee crisis in 2015 is a reminder of the effects of net migration. Although most people perceive the huge influx of refugees and migrants to Germany in 2015 as a fiscal burden, migration could actually smooth the demographic transition. However, migration is unreliable because changes in net migration are difficult to predict (Woellert and Klingholz 2014).
Figure 2: Population Pyramids for Germany in 2015 and 2050.
[Images are not displayed in this preview]
Source: United Nations, Department of Economic and Social Affairs, Population Division (2015). World Population Prospects: The 2015 Revision.
Generally, population prospects are more reliable for large and developed countries (International Monetary Fund 2004), which makes them quite reliable for Germany. Nevertheless, the development of the elderly dependency ratio may not reflect the problems implied by demographic aging adequately. The definition of the ratio is static, therefore neither a change in retiring age nor a change in the age when people start working would affect the dependency ratio. However, such changes could affect the age when individuals start to accumulate and deccumulate their assets. For example a rise in the retiring age would presumably postpone the start of asset deccumulation.
To sum up, albeit there are some uncertainties, Germany is going through a demographic transition that is strong enough to have consequences on financial markets.
After having determined the magnitude of demographic ageing, this section starts examining its consequences on financial markets. Most of them are based on the life-cycle model of saving, first introduced by Modigliani and Brumberg (1954). Basically, the model suggests that people save during their working age and draw down their accumulated wealth to finance consumption during retirement. Yet, the model has been subject to many tests across countries since its introduction and particularly the magnitude to which people dissave is strongly debated. A reason for why we may not observe a typical hump-shaped life-cycle, is that people have other saving motives than just retirement provision (Browning and Lusardi 1996). For example, people save to buy durables like a house or a car, build up precautionary savings for changes in economic or health conditions, or save to leave a bequest to their children. These saving motives can interfere with the retirement saving motive and lead to a saving profile that is not hump-shaped over the life-cycle.
Based on data from the Einkommens- und Verbrauchsstichprobe (EVS), BörschSupan et al. (2001) investigated the saving behaviour of German households between 1978 and 1993. They did find a hump-shaped life-cycle for the saving rate, however, the observed households had no intention to dissave when getting older. Even for the very old, saving remained positive and financial wealth was not drawn down. Since mandatory contributions to the German pay-as-you-go (PAYG) pension system make up for the vast majority of retirement income and pensions were very generous at the time, one would rather have expected that the contributions would crowd-out private saving (Börsch-Supan et al. 2001). BörschSupan et al. (2001) explain this German Saving Puzzle with the prosperous years of economic growth up to the 1970s, which lead the generation born between 1910 and 1930 to unexpected high levels of wealth. Then, as people retired, both their habitualisation to expenditure habits and their deteriorating health conditions lead to consumption levels too low to reach negative saving rates, the authors argue.
Although the findings by Börsch-Supan et al. (2001) contradict the life-cycle saving model, the authors carefully observe that younger generations may have taken the generous pensions into account and are thus incentivised to save less for retirement. On the other hand, the implementation of the Riester-Rente and further policy reforms towards a more privately prefunded pension system, will put future generations under pressure to save more for retirement. Either way, the saving behaviour over the life-cyle may not be consistent among cohorts. Regarding the effects on financial markets, even if the elderly do not withdraw their assets from the market, they purchase fewer assets (save less) compared to when they were young. Put another way, a growing share of the elderly could reduce the fuel required for financial markets to flourish in the future. Furthermore, wealth is an aggregation of all assets in an individual’s portfolio. That wealth remains constant or even increases with age does not necessarily mean that there is no individual asset that is deccumulated when an individual retires. For example, real estate, which accounts for a substantial share of households’ overall wealth (Börsch-Supan et al. 2001), is far less liquid than stocks. Hence, if an individual sells some assets when retiring, she presumably sells the liquid ones first.
To sum up, for whatever reason, German households have not withdrawn their savings during retirement in the past. Still, they might have deccumulated certain assets.
Combining the life-cycle saving model with the severe demographic transition would result in a large rise in demand for financial assets as the Baby Boomers enter their high-saving years, but also in a large fall when the Baby Boomers enter retirement. Some economists have attributed the rise in stock prices during the 1990s to the former. Consequently, they have also reason to believe that the latter implies a bearish outlook for Wall Street. Besides the aggregated effect of a change in demand for financial assets on asset prices, each asset category is expected to be affected differently by demographic ageing. More precisely, when building up assets, the Baby Boomers should have a preference for risky assets like stocks, which offer them excess return in the long-run, but at a greater risk in the short-run. High labour income can cushion short-run stock market fluctuations. However, this cushion fades with retirement, so that the Baby Boomers are more likely to switch to safe assets to secure their wealth when retired.
Most theoretical models predict that capital market returns will be affected by demographic ageing, but their results vary in magnitude (Poterba 2004). In contrast to theory, empirical evidence is more mixed. For example, Poterba (2001) solely finds a negative effect of the share of the population aged 40 to 64 on returns of treasury bills and government bond yields in the US. Yet, even this weak effect varies with the sample period and even reverses when estimating the same cross-section for Canada and the UK.
Instead, Bergantino (1998) finds that demographic factors accounted for 59 up to 81 per cent in the observed annual increase in real house, bond and equity prices between 1966 and 1997 in the US. He compares his results with those found by Poterba and Samwick (1997) – which did not indicate such a relationship – and argues that the difference could arise from the empirical specification. Whereas Bergantino (1998) estimated the effect of demographic factors on price levels, Poterba and Samwick (1997) estimated the effect of demographic factors on returns.
Correspondingly, Brooks (2006) notes that, estimating the effects on price levels should overstate the link between demography and asset prices, whereas estimating the effects on returns should understate it. Estimating Pooled-OLS for datasets covering most advanced economies between 1900 and 2005, he first confirms the theory by finding a positive effect of the middle-aged on financial asset prices and a negative one for the elderly. However, the relationship disappears for returns. In addition, the relationship is not consistent among countries.
Using panel data for advanced as well as emerging economies between 1960 and
2002, Davis (2007) also controls for macroeconomic effects and links demography to equity and bond markets. He finds that equity markets should initially benefit from demographic ageing as the share of the 40 to 64 years old grows, but then, a growing share of the elderly with preferences towards less risky assets should benefit the bond market.
After all, besides the financial assets that are directly held by US households, there are also many assets held through pension funds that offer a defined benefit. When these assets are sold for paying out pensions, they could add to the downward pressure on prices arising from a sell of directly held financial assets (Poterba 2004).
One reason why most empirical studies fail to find an affect of demographic factors on asset prices is that the underlying age-wealth profile, according to the life-cycle saving model, is not present in real data. Poterba (2001) used data from the US Survey of Consumer Finances that showed only slight – if at all – wealth deccumulation in old age. It is his main explanation for not finding a significant relationship between population shares and asset returns that differs among age groups. Hence, it is reasonable to first check if the elderly actually deccumulate their assets before arguing that a fall in demand due to demographic ageing will lead to a future asset price meltdown.
Among others, Yoo (1994) and Ameriks and Zeldes (2004) provide empirical evidence on this issue for the US. Whereas Yoo (1994), using cross-sectional data, finds that retirees demand relatively less stocks and risky assets, Ameriks and Zeldes (2004), using panel-data, report no decrease in equity shares with increasing age.
Both the estimation of an age effect on asset holdings and the estimation of an age effect on asset prices share an identification problem, which Ameriks and Zeldes (2004) and Poterba (2004) point out respectively. When estimating the age effect at a specific time, one has to distinguish between the time, the cohort and the age effect (Poterba 2004). The time effect could be a period of prosperous economic growth that incentivised people to buy more assets, which leads to higher prices. Instead, cohorts that, for example, experienced the Dot-com bubble between 1997 and 2000 may be more cautious when it comes to stocks than future cohorts, resulting in a cohort effect. What remains is the age effect that most economists attempt to find in the data regarding the life-cycle hypothesis. The identification problem refers to the fact that at most two of the three effects can be estimated with repeated cross-sections and panel data, "[...] because the cohort effect is a linear combination of the age and time effect" (Poterba 2004).
To sum up, empirical evidence on the link between demography both asset prices and asset holdings is mixed. Results differ across countries and are sensitive to the sample period and the empirical specification.
The findings of Börsch-Supan et al. (2001) rejected the life-cycle hypothesis for Germany. But as I pointed out in Section 3, this does not necessarily mean that there is no asset deccumulation for specific assets. This section will provide some evidence on this issue. Based on this evidence, Section 6 discusses if the results suggest falling asset prices for an aging Germany.
I have data from the German Socio-Economic Panel (SOEP) in 2012, a longitudinal dataset from the German Institute of Economic Research (DIW) that is representative for the German population and samples about 10,000 households and 30,000 individuals annually since 1984. I am interested in individual stock holdings over the life cycle because stocks represent an asset that is (a) easy to liquidate and (b) more risky, which should make it appealing to the elderly to sell it before retirement. Specifically, I would be interested in the effect of age on the amount of privately held stocks, but unfortunately, the SOEP only has a binary outcome variable that equals 1 if someone holds securities with a floating rate of return and 0 if not. Therefore, the variable does not reveal if someone gradually reduces her stock holdings, only if she leaves the stock market completely. Furthermore, the SOEP survey did first not distinguish between fixed and other securities. For that reason, I can only observe since 2001 if an individual held stocks in the previous year, i.e. since 2000. I estimate a simple Probit model of the form
Pr(Y = 1|X) = Φ(X’β) , (1)
where Y is the binary outcome variable, stock-ownership, and X consists of the explanatory variable of interest, age, and the set of control variables, which are both described in Section 5.1.
 According to the medium variant of the United Nations, Population Division (2015).
 The Baby Boomer generation is the large cohort born between 1955 and 1969.
 The elderly dependency ratio is defined as the share of the over 65 years old (i.e. the retirees) divided by the share of the population aged 15 to 64 (i.e. the working age population).
 In 2015 the Baby Bomer cohort relates to individuals that are 45 to 60 years old.
 In 2010 people with migration background in Germany were relatively younger than natives (Woellert and Klingholz 2014).
 Henceforth, I will refer to these securities as stocks. However, note that other securities with a floating rate of return, e. g. bonds, also fall into this category.
Bachelorarbeit, 84 Seiten
Habilitationsschrift, 109 Seiten
Forschungsarbeit, 27 Seiten
Masterarbeit, 101 Seiten
Bachelorarbeit, 66 Seiten
Bachelorarbeit, 60 Seiten
Der GRIN Verlag hat sich seit 1998 auf die Veröffentlichung akademischer eBooks und Bücher spezialisiert. Der GRIN Verlag steht damit als erstes Unternehmen für User Generated Quality Content. Die Verlagsseiten GRIN.com, Hausarbeiten.de und Diplomarbeiten24 bieten für Hochschullehrer, Absolventen und Studenten die ideale Plattform, wissenschaftliche Texte wie Hausarbeiten, Referate, Bachelorarbeiten, Masterarbeiten, Diplomarbeiten, Dissertationen und wissenschaftliche Aufsätze einem breiten Publikum zu präsentieren.
Kostenfreie Veröffentlichung: Hausarbeit, Bachelorarbeit, Diplomarbeit, Dissertation, Masterarbeit, Interpretation oder Referat jetzt veröffentlichen!