VoxEU Column Labour Markets Macroeconomic policy

Risks to the individual in defined contribution pension schemes

The UK is about to make a massive change to its pension system. From October 2012, employers will be obliged to automatically enrol employees into a pension scheme – though individuals can opt-out. This column explores what this might mean for pension funding and argues that the risks are to the downside.

The UK is about to make a massive change to its pension system. Auto-enrolment into a pension scheme comes into effect in Autumn 2012, as a quasi-compulsory defined contribution national pension scheme is introduced called NEST (National Employment Savings Trust). From October 2012, employers will be obliged to automatically enrol employees into a qualifying pension scheme (though individuals can opt-out). These obligations initially affect only large employers, requiring them to pay a minimum of 2% of employees’ earnings into the scheme, but this will roll out to all employers by 2018, by which time contributions must be at least 8% of earnings.

This new UK scheme is part of a wider trend, as around the globe pension systems are moving towards funded defined contribution schemes. In a defined contribution scheme the individual is responsible for building up their own pension fund to provide an income during retirement. Important questions that this move to defined contribution raises are: how well do such schemes perform and since pension risks are now transferred to the individual, how much risk will individuals bear from participating in these schemes? Further, what explicit or implicit guarantees are offered? If in the future these schemes underperform so that the accumulated funds are not sufficient to generate an adequate pension in retirement, will individuals be able to fall back on the state to provide a pension?

One approach to assessing the risks in defined contribution pensions is to use historical data to calculate hypothetical retirement incomes based on regular contributions throughout a working life, and pose the counter-factual conditional: what pension would someone have achieved if they had been able to invest a defined contribution pension scheme and earned the investment returns which actually occurred over the 20th century?

Burtless (2003) addressed this question using data from the period 1927-2001 for six countries, but without using actual wage data. Building on work in previous research (Cannon and Tonks 2004), we examine pension risks using international historical data on stock and bond returns and wage growth over the period 1901-2007, to calculate hypothetical accumulated defined contribution pension wealth in sixteen major economies (Cannon and Tonks 2012). A single historical dataset suffers from the criticism that results depend on one realised set of returns. The advantage of a cross-country comparison is that we obtain a more general picture of the distribution of defined contribution pensions from these 'independent' realisations. We use the updated returns data in Dimson, Marsh and Staunton (2002) for 16 countries, and simulate pension fund ratios (size of pension fund to final labour income) and replacement ratios (size of pension income to final labour income). Labour incomes were taken from Mitchell’s (1988) comparison of international historical data, updated using data from the OECD and Maddison (2007). Standardising pension measures by labour income allows for a meaningful cross-country comparison.

The optimal value of the replacement ratio is unclear, although the UK’s Pension Commission (2004) suggests a range of benchmark replacement ratios of between 80 and 50%. Since annuity rates for a 65-year-old male would typically be in the range of 5-10%, then a target replacement ratio of 0.8 would suggest an optimal pension fund ratio of between 8 to 16.

In our scenarios, we assume that an individual saves from ages 25 to 65 and is continuously in employment during that time, making annual contributions of 10% of labour earnings, into a tax-privileged fund. The rate of return on investments depends upon that country’s bond and equity yields, and the assumed asset allocation; to allow for annual investment charges we deduct 2% for equity and 1% for bond investments. We consider three different asset allocation rules: (a) invest entirely in equity for the whole forty years; (b) invest entirely in bonds; and (c) a “lifestyle” scheme consisting of all-equity for the first 28 years with a gradual transition into all-bonds over the next nine years.

As well as reporting hypothetical pension fund ratios, we also construct synthetic annuity rates for each country in each year, using the relevant population mortality projections. The replacement ratio is the fund ratio multiplied by the annuity rate. To calculate the synthetic annuity rates we use an interest rate based on government bond yield data from the IMF’s International Financial Statistics. Mortality information was taken from the Human Mortality Database.

Some of our results are presented in Table 1. Panel A reports for the alternative investment strategies, the median fund ratio for six countries (US, UK, France, Germany, Japan and Australia) averaged over the time series for that country. All countries have median fund ratios considerably lower than sixteen for each asset allocation strategy. The average value across all countries of the median fund ratio for the all-equity strategy is only 10.22, and the all-equity strategy dominates the all bond strategy for each country. Except for Germany and Japan the equity strategy also has a higher average than the alternative lifestyle investments. Among the countries with average fund ratios greater than ten are the Anglo-Saxon countries (Australia, UK and USA).

Table 1 Summary statistics of hypothetical pension fund ratios and pension replacement ratios, by country

Figure 1 Hypothetical pension fund ratios obtained from all-equity asset allocation strategies

Figure 1 shows the pension fund ratios for the all-equity asset allocation strategy for the six countries. The two striking features about these graphs are that:

  • Each country’s historical pension fund ratio is volatile, implying that the timing of retirement has a large impact on pension income;
  • The pension fund ratio appears to be low for nearly all countries in nearly all time periods; so that there may be insufficient funds available to generate an adequate retirement income.

Panel B in Table 1 provides the lower decile of the fund ratio, and shows that there is considerable downside risk for all types of investment strategy. Except for Germany, the all-equity strategy though risky dominates the other asset allocations.

The replacement ratios obtained by matching the hypothetical annuity rates to the pension fund ratios, for each country are summarised in Table 1: Panel C. The median replacement ratio appears more than satisfactory, with an average value across countries of 1.38; substantially greater than unity: implying an income in retirement more than necessary for income smoothing. Even the numbers for the lowest decile, suggest relatively high replacement ratios, so that individuals in the worst-performing economy (Japan) would have a 10% probability of ending up with a replacement ratio of 0.29.

It might seem surprising that given the low fund ratios identified in Figure 1 and Panel A, the replacement ratios in Panel C are not lower. But bond and equity returns are less than perfectly correlated, and the annuity rates that we have simulated are determined by the bond rate. The historical low correlation between equities and bonds acts as a natural hedge in the annuitisation process: when fund values are high due to high equity returns, the annuity rate will be low, reducing the replacement ratio. Conversely when fund values at retirement are low, the annuity rate and hence the replacement ratio will be relatively high.

Recent speeches from the Bank of England on the impact of quantitative easing has emphasised the trade-off between decreases in long-term gilts and annuity rates being offset by increases in share prices.

In summary, it appears there is considerable variation in fund ratios across both time and country. There is some support for the idea that countries keenest on individual defined contribution pension accounts have the highest fund ratios (UK and USA), but the historical evidence is that all investors in all countries face considerable downside risk.


Basu, AK, and ME Drew (2010), “The appropriateness of default investment options in defined contribution plans: Australian evidence”, Pacific-Basin Finance Journal, 18(3):290-305.

Bean, C (2012), “Quantitative easing and the economic outlook”, speech given to the Scottish Council for Development and Industry

Burtless, G (2003), “What do we know about the risk of individual account pensions? Evidence from industrial countries”, American Economic Review, 93(2):354-359.

Burtless, G (2007), “International investment for retirement savers: historical evidence on risk and returns”, Center for Retirement Research at Boston College, Working Paper.

Cannon, E and I Tonks (2012), “The Value and Risk of Defined Contribution Pension Schemes: International Evidence”, forthcoming Journal of Risk and Insurance, DOI: 10.1111/j.1539-6975.2011.01456.x

Cannon, E and I Tonks (2004), “UK Annuity Rates and Pension Replacement Ratios 1957-2002”, Geneva Papers, 29(3):371-393.

Dimson, E, P Marsh, and M Staunton (2002), Triumph of the Optimists: 101 Years of Global Investment Returns, Princeton University Press.

Economic Policy Committee and the European Commission (DG ECFIN) (2006), “The impact of ageing on public expenditure”, Special Report 1/2006

Human Mortality Database, University of California, Berkeley (USA), and Max Planck Institute for Demographic Research (Germany). Available at www.mortality.org or www.humanmortality.de (data downloaded August, 2008).

International Labour Office Yearbook of Labour Statistics 1990-2002

Maddison Statistics on World Population, GDP and GDP Per Capita, 1-2006 AD. Available from the Groningen Growth & Development Centre (can be downloaded from http://ggdc.net/maddison/).

Miles, D (2012), “Asset prices, saving and the wider effects of monetary policy”, speech given at the pro. Manchester Business Conference, March 2012

Mitchell, BR (1998), International Historical Statistics: Europe; The Americas; Africa, Asiana and Oceania; 1750- (Basingstoke: Macmillan).

Pension Commission (2004), Pensions: Challenges and Choices: The First Report of the Pensions Commission, The Stationery Office.

Pension Commission (2006), Implementing an Integrated Package of Pension Reforms: The Final Report of the Pensions Commission, The Stationery Office.

Poterba, J, J Rauh, S Venti and D Wise (2007), “Defined contribution plans, defined benefit plans, and the accumulation of retirement wealth”, Journal of Public Economics, 91(10):2062-2086.

Samwick, A and J Skinner, (2004), “How will 401(k) plans affect retirement income?”, American Economic Review, 94:329-343.

Shiller, RJ (2006), “Life-cycle personal accounts proposal for Social Security: An evaluation of President Bush’s proposal”, Journal of Policy Modeling, 28:427-444.

2,940 Reads