Printer Friendly

Funding pensions in Scotland: would independence matter?

Economic issues will be key determinants of the outcome of the Scottish referendum on independence. Pensions are a key element of the economic case for or against independence. The costs of funding pensions in an independent Scotland would be influenced by mortality risks, the costs of borrowing and the segmentation of costs and risks (i.e. pricing to Scotland's experience rather than pooled across UK experience). We compare the overall costs of providing pensions in an independent Scotland against the resources that are available to cover these costs. Scotland has worse mortality experience than the UK as a whole, and Scottish government debt is likely to attract a liquidity premium relative to UK government debt. An independent Scottish government would have to create a bond market for public debt. The liquidity premium would make pensions cheaper to buy, but taxpayers or the consumers of public services would have to pay the cost.

Keywords: pensions; bonds; independence; risk pooling; mortality; yields

JEL Classifications: H10; H11; H20; H21; H30; H50; H72; H73; H77


A principal argument in favour of maintaining political unions is the ability of the central government to implement stabilisation policies following asymmetric shocks. This is a form of risk pooling: if one part of a union suffers an adverse economic event, central government can transfer resources to the affected area, compensating agents for their losses and so maintaining the macroeconomic stability of the union as a whole. Thus, for the member nations, states or territories that make up a political union, and providing that mechanisms for appropriate fiscal transfers exist, such risk pooling may constitute an important reason for continued membership.

In his classic contribution to the literature on fiscal federalism, Oates (1972) argued that stabilisation policy, as described above, was one of the major roles that central government could fulfil. He also argued that centralised redistribution policies internalise spillovers across area boundaries and should therefore also remain part of the remit of central government. Thus he argued that the migration of poor households would undermine decentralised welfare programmes. However, other policies could be allocated to lower levels of government if the loss of benefits from an aggregate policy, such as economies of scale, is offset by the welfare gains from allowing local preferences to influence local public good provision.

The innovation of this paper is that it addresses some of these arguments, but not in relation to a public good. Specifically, it focusses on how pensions interact with the possibility of constitutional change in Scotland. Pension policy does not easily fit within the standard fiscal federalism literature. It is not immediately clear how the treatment of pensions might influence the economic case for or against remaining within a political union. In relation to the current constitutional debate, the Scottish government is in no doubt that pensioners would not be disadvantaged by Scotland breaking away from the rest of the UK (rUK).

"An independent Scotland will have the ability to protect and improve state pensions and ensure that private pensions are secure and saving for retirement is actively encouraged."

Scottish Government (2013a), p.139.

At first pass, one might conclude that risk pooling arguments relating to pensions would favour continued political union with rUK. But the risks associated with pensions are clearly different from those associated with macroeconomic shocks. Inter alia, they depend on the expected duration of the period between retirement and death. If this duration differs systematically between the area seeking independence and the rest of the political union, then, compared with a pooled insurance contract, separate pension arrangements may benefit the citizens of that part of the state that is leaving. We demonstrate that under certain conditions, this argument may hold in the case of Scotland.

Fiscal federalism was initially used to describe how local preferences for public goods might be accommodated within a larger political union. Pensions are clearly not public goods. Funded pensions are entirely private, existing simply to alter the time profile of consumption. The case for devolving control over such pensions within a unitary state is weak. Arguments relating to economies of scale, competition and the consistency of regulation would all seem to favour central government control over pension policy. However, from a political economy perspective in relation to potential independence, pensions have the unusual characteristic of being contracts held by many voters that extend beyond the date of independence: voting behaviour may be influenced by expectations of the way in which these contracts will be fulfilled post-independence. Equally, many voters contribute to the taxes that fund state and public sector pensions. If voters place a high weight on post-referendum income, those promoting independence then have the task of reassuring both taxpayers and pensioners. Considerations of loss-aversion may then lead, perhaps counter-intuitively, to proposals from those promoting independence to maintain the status quo as regards pension policy.

Debates on the economic effects of independence also focus on the division of public debt and the subsequent burdens that their shares of the debt impose on the successor and the new states. Rarely do these debates touch on contingent debt. Yet state and public pensions, which often comprise a major share of the contingent debt in developed countries, may constitute a significant threat to long-term fiscal sustainability if the political union breaks up. We discuss how this might play out in the case of Scotland and rUK.

In a UK context, some issues affecting contingent debt, such as population ageing, will occur irrespective of constitutional change. In the discussion of contingent debt, it is important to focus only on those aspects of the debt that may change as a result of the dissolution of the UK. This would include policy commitments relating to pensions that have consequences for future debt and that may be used to promote either the continuation or the break-up of the political union.

One of the institutions that a new state has to create is a debt market. How this is structured and managed will have implications for domestic pensions. If the state floats its own currency, pension funds will have to invest heavily in this market to reduce scheme members' exposure to exchange rate risk. On the other hand, if a currency union is adopted, domestic bonds will have more competition from the bonds of other members of the currency union and pension funds may hold a more diversified bond portfolio. In turn, this will affect relative bond prices which will also influence the redistributive effects of pensions. For example, if Scottish bonds carry a liquidity premium and form a large part of Scottish pension fund portfolios, Scottish pensioners will gain from the higher yields, while Scottish taxpayers will have to fund higher debt charges, which in turn implies higher taxes. If Scottish bond yields are higher than those in rUK and Scottish occupational pensions are therefore cheaper to fund, the net effect is likely to be a redistribution of spending power from younger (poorer) taxpayers to older (richer) pensioners compared with rUK. We attempt to quantify this effect in the paper.

The structure of the paper is as follows: in the first section, we deal with the state pension, estimating the relative cost of its provision in an independent Scotland. We also deal with particular issues such as the recently implemented triple lock arrangement, the accumulation of contributions, and the payment of pensions to those living outside the UK. The second section deals with public sector pensions, and discusses the fiscal burden that they might impose on an independent Scotland. Next we consider the issue of occupational pensions and how these might be managed in an independent Scotland. The final section concludes.

The state pension in Scotland

This section considers the costs of the State Pension (SP) in Scotland: in the first part we compare its cost to that in the UK, both in relation to life expectancy, and in relation to tax revenues; in the second part we consider the cost implications of the 'triple lock' policy; and in the third we consider whether migration patterns between Scotland and rUK are likely to result in any geographical asymmetry between the location of individuals' contributions and where they may draw their pensions.

Costs of the state pension

Spending on the SP in Scotland was 6.3 billion [pounds sterling] in 2011/12 (Department of Work and Pensions, 2013), almost 10 per cent of total public sector expenditure in Scotland of 64.5 billion [pounds sterling] (Scottish Government, 2013b). The value of the average state pension payment is essentially the same in Scotland (130 [pounds sterling]

pw) as in England and Wales (129 [pounds sterling] pw).

These broad statistics do not capture the costs of providing SP at the individual level. It turns out that these differ substantially between Scotland and rUK. The reason is that life expectancy in Scotland is significantly less than that in the UK as a whole. One can relatively easily compare pension costs based on average Scottish mortality experience with those based on average UK mortality. This is an important comparison since the funding assumptions implicit in estimated UK pension costs are based on mortality experience across the whole of the UK.

The UK government plans to increase the State Pension Age (SPA) to 66 by 2020, and then to increase it further to 67 between 2026 and 2028. However, with different life expectancies in different locations, increases in the state pension age could affect Scotland but be caused by increases in life expectancy in the South of England. As part of the UK, Scotland would implicitly be part of a pension contract that would be actuarially unfair if Scotland was a separate state.

The SP is unfunded and therefore payments must be drawn from current tax revenues or from borrowing. The costs of funding pensions depend on the number of pensioners relative to total tax revenues. It may well be actuarially fair to defer the rise in the SPA in Scotland due to its heavier mortality, but if there are many pensioners and few workers then such a policy could still be costly to fund.

Consider the cost of providing an otherwise identical pension to an individual with the average mortality experience of Scotland compared to the cost to an individual with the average mortality experience of the UK as a whole. Assuming that the UK raises the SPA to 67 in 2027, one can calculate how long it would take for Scottish mortality experience to improve to that which the UK is expected to experience in 2027. This provides an estimate of the delay in the SPA which would give Scottish state pensioners the same expectation of life after the SPA as the average UK state pensioner in 2027. Heavier mortality, and the consequent lower cost of providing pensions on an individual basis, may provide the rationale for the Scottish government's proposal to consider the deferral of the rise in the SPA to age 67 in an independent Scotland. If we assume that the UK government plans to increase the SPA from 66 to 67 in 2027, then we can propose that a 'fair' basis for state pensions policy in an independent Scotland might be to consider raising the SPA from age 66 to age 67 when Scottish expectation of life at age 67 reached the UK expectation of life at age 67 in 2027. Figure 1 shows the difference between UK expectation of life at age 67 in 2027 and the Scottish expectation of life in year x. In this and in subsequent exercises, our estimates are based on the principle projections 2012 data from the Office for National Statistics (ONS).

Figure 1 shows that making policy on this basis would lead the Scottish government to raise the SPA to 67 in an independent Scotland in 2039 i.e. a delay of twelve years relative to the UK.

One consistent way to calibrate the costs of this policy is to measure the ratio of SP costs to tax revenues. Suppose pensions are financed by issuing nominal government debt. Then if we assume that payments to pensioners grow in line with the nominal growth rate of the economy, and that the expected yield on this government debt is equal to the same nominal growth rate, then the real terms value of 1 [pounds sterling] per annum of pension in today's money depends only on demography. With a uniform retirement age across the UK, Scotland's heavier mortality clearly implies that its pensions will be cheaper. Figure 2 shows the percentage difference in cost between providing an average Scottish pensioner and an average UK pensioner with the same amount of pension. The outcomes are driven by the differences between Scottish and UK life tables. Details of the calculations performed are given in the Appendix. Note that age on the x-axis refers to the age of the relevant cohort in 2012.

Figure 2 shows that it is between 6 per cent and 8 per cent cheaper to provide a pension to an individual given Scottish mortality experience than it is given UK mortality experience. It also shows that the relative gap in life expectancy is predicted to be roughly constant for the current population aged between 30 and 60. Younger people in Scotland are predicted to have mortality experience that is slightly closer to the UK average so we can infer that ONS projections assume that the mortality differential between Scotland and the rest of the UK will narrow. We also see that the cost differential diminishes at very old ages: this is as expected as most of the unhealthy lives have died by this point and at these old ages, we are comparing the remaining healthy individuals of the two populations and, among this age group, life expectancy may be expected not to differ substantially. NB. The kink shown in figure 2 at age 65 is due to the change in calculation basis from (for ages less than 65) comparing the cost of pensions payable from age 65, to (for ages x greater than 65) comparing the cost of pensions payable from age x.

The SP may be cheaper in Scotland than in rUK, but affordability also depends on tax revenues in the two jurisdictions. Scottish Government (2013b) estimates that both Scotland and the UK as a whole pay tax at roughly the same level per person, of just over 9,000 [pounds sterling] p.a. Given the current age structure of the population, this means that revenues are roughly the same per worker. If all those aged between 18 and the SPA work, and those beyond state retirement age receive a pension of 120 [pounds sterling] per week, we can project the share of pension costs in total tax revenues in both Scotland and in the UK as a whole, using the ONS 2012 principal population projections. To keep the calculations simple, we again assume that pension payments and earnings both grow at the same rate as economic growth and that Scotland and the UK both raise the SPA to 67 in 2027. Figure 3 then shows the differential costs of state pensions as a share of tax revenues in Scotland and the UK as a whole between 2012 and 2062.

For Scotland, the cost of SP is projected to increase by more than 6 per cent of tax revenues over the period to 2062. Figure 3 also shows how the relatively faster ageing population in Scotland is much less important than the overall increase in SP costs: the differential between Scotland and UK peaks at less than 1 per cent of tax revenues. This is consistent with Lisenkova and Merette (2014), whose model also includes endogenous responses to the changing demographics and who also note that the extra cost of ageing in Scotland is small relative to the overall cost of ageing itself. It is worth noting that although Scotland is ageing faster than the UK as a whole, it is not ageing as fast as many other European countries.

Now suppose that Scotland delays the rise in the SPA to 67 for twelve years until 2039. Figure 4 shows that the cost of this policy is fairly constant on an annual basis, and so we can take simple averages. The cost of this policy is approximately 1.3 per cent of taxes for twelve years. In 2011-12, 1.3 per cent of Scottish taxes amounted to 750m [pounds sterling] which provides a rough estimate for the real terms annual cost of this policy. This cost is split into 550m [pounds sterling] additional pension costs and a 200m [pounds sterling] reduction in tax revenues. Note that this calculation does not take account of the extra benefits to which pensioners are currently entitled, such as Winter Fuel Payments. The 550m [pounds sterling] cost of paying extra pensions to those 66 year olds who would otherwise not receive a pension is a fairly certain cost. However, the 200m [pounds sterling] in lost tax revenues from those 66 year olds not working cannot be forecast with the same confidence since economic activity rates among older workers have changed significantly in recent years and their future path is highly uncertain.

The state pension is both cheaper on an individual basis in Scotland compared with the UK as a whole because of differential life expectancy, and slightly less affordable on an aggregate basis because of the pensioner dependency ratio. Current policy in the UK is to raise the state pension age, but there are implicitly regional transfers associated with this policy. For example, a SPA of 66, rather than 65, in 2020 represents a transfer from Scotland to rUK of nearly 50m [pounds sterling] per annum (based on the same calculations underlying figures 3 and 4). This is because Scotland has relatively fewer workers who benefit from lower taxes, it has relatively more people around SPA who lose out from not receiving their pension at age 65, and it does not have relatively more pensioners at the much older ages because of higher mortality. Whatever is the optimal policy for SPA in the UK as a whole, the optimal policy for the SPA considering only Scotland will be relatively lower. Given the consequence of this for affordability, it may be that an independent Scotland should choose a lower SPA and a lower state pension amount.

Triple lock

One of the policies to which the Scottish government has committed is the maintenance of the so-called 'triple lock' in SP after independence. This ensures that the SP will increase by the maximum of 2.5 per cent, the rate of growth of earnings, or the rate of growth of prices. This is a strong commitment which will eventually be unaffordable. The real costs of funding SP will rise particularly during periods of stagflation when high rates of inflation are combined with low economic growth.

However, the commitment to this policy by the Scottish government suggests a need to reassure voters that they will not experience significant losses as a result of Scottish independence, as discussed in the introduction. Anticipated loss aversion on the part of current and future pensioners may have a powerful influence on policy formation by supporters of independence.

Division of state pension liabilities

Part of the settlement of contingent debt will be determined by how pension liabilities should be assigned following independence. Currently, the UK government pays SP to some 1.2 million claimants living outside the UK (i.e. individuals who have accrued rights to the SP while living in the UK, but have retired abroad). Although SP rates are different for those living abroad, this raises an interesting hypothetical question: to what extent have the SP liabilities of Scottish residents been accrued whilst working in rUK, and vice versa? In other words, are taxpayers in Scotland (rUK) systematically exposed to a disproportionate level of liabilities that have been accrued in rUK (Scotland)?

Eleven per cent of Scotland's working age population were born in rUK and 11 per cent of Scotland's population of over 65s were born in rUK. For rUK, 1.3 per cent of the working age population were born in Scotland, against 2 per cent of the population of over 65s. These stylised facts, whilst not suggesting a significant difference in the make-up of the working age and pensionable age populations, tell us little about whether there are systematic differences between where people work and retire.

Since 2005, the Labour Force Survey (LFS) has asked respondents about their economic activity and residence 12 months previously. These data enable us to map migration flows between Scotland and rUK. Figure 5 shows total migration between Scotland and rUK by age group, using pooled LFS data from 2005-12. Migration clearly declines with age. Over the 2005-12 period, there was net migration from Scotland to rUK among the under 35s, and net migration from rUK to Scotland for those aged 40-59. There is net migration out of Scotland among those in their 60s, but little real evidence of a major increase in migration around retirement age. The data is pooled over eight years, so Figure 5 suggests net annual out-migration from Scotland of just 700 individuals aged 60-69. These data do not suggest that there are large flows between Scotland and rUK around the time of retirement.

Looking more specifically at the behaviour of retirees in the 12-month period around retirement, the LFS data suggests that of the 1.2m individuals who retired in Scotland between 2005 and 2012, only 0.2 per cent relocated to rUK in the year following retirement and just 0.02 per cent of rUK retirees relocated to Scotland in the year following retirement. These data cannot show the proportion of retirees' working lives that have been spent in rUK (Scotland) among those now living in Scotland (rUK). But it does not suggest that there is an obvious asymmetry of flows of working-age versus pensionable-age individuals between Scotland and rUK. The inference would be that net cross-border SP liabilities would be relatively small should Scotland become independent.

Public service pensions in Scotland

Another component of the contingent liabilities that an independent Scotland might face are those related to the payment of public sector pensions. Table 1 shows some stylised facts on public sector employment in Scotland and rUK. Scotland has a slightly higher proportion of public sector workers, and those workers tend to have had slightly longer tenure on average than public sector workers in rUK. Average public sector wages are slightly higher in rUK. However, the total public sector wage bill per capita is higher in Scotland than rUK, implying that the public pension liabilities are relatively higher in Scotland.

There are six main public service pension schemes in Scotland. Around one million people in Scotland currently have a direct interest in one of these six schemes, either as members or as pensioners and dependants. In 2009/10, the six schemes paid out 2.8 billion [pounds sterling] to pensioners while public bodies contributed 2.2 billion [pounds sterling] and employees paid 814 million [pounds sterling] to meet their expected long-term costs (Audit Scotland, 2011).

There are both funded and unfunded public service pension schemes in Scotland. The Local Government Pension Scheme is the main funded scheme; the NHS, teachers, civil servants, police and fire-fighters' pensions are all unfunded. Of the unfunded schemes, payments to pensioners and their dependents increased by 32 per cent in real terms between 2005-2010, from 1,468 million [pounds sterling] to 1,936 million [pounds sterling]. This increase reflects the increase in the numbers of pensioners (up 13 per cent) combined with the fact that newly retired pensioners have higher pensions reflecting earlier increases in pay. In contrast, employers' contributions in the unfunded schemes have increased by just 15 per cent in real terms over the same period, from 1,167 million [pounds sterling] to 1,338 million [pounds sterling]. The Scottish Public Pensions Agency (SPPA) has indicated that contribution rates for the largest unfunded schemes (teachers and NHS) may need to increase by 2-4 per cent of pay.

The pension liabilities of Scotland's unfunded schemes increased by 45 per cent between 2006 and 2010. While some of this increase reflects an increase in members, an increase in pay, and an increase in life expectancy, it is largely driven by recent low interest rates which have reduced the level at which future liabilities are discounted (Audit Scotland, 2011). Similarly, the Local Government Pension Scheme (LGPS), which is a funded scheme, has been subject to a 26 per cent real terms increase in payments to pensioners and their dependents between 2006-10, from 667 million [pounds sterling] to 840 million [pounds sterling] a year. Over the same period, employers' contributions to the LGPS increased by 25 per cent in real terms, from 667 million [pounds sterling] to 836 million [pounds sterling] a year, whilst employees contributions increased by 11 per cent in real terms, from 243 million [pounds sterling] to 270 million [pounds sterling]. This reflects an increase of 10 per cent in scheme members and general increases in pay, but there were also increases in the employers' contribution rates to the LGPS, from 16.2 per cent to 19.3 per cent of pay between 2002-3 and 2008-9.

Table 2 shows net expenditure on public service pensions on individuals resident in Scotland. Prior to the recession in 2007/8, the Scottish schemes received more in contributions than they paid out to retirees. This position has deteriorated each year, and in 2011/12 the Scottish agency pension schemes (funded and unfunded) paid out over 300m [pounds sterling] more than was received. As alluded to above, this trend likely reflects a reduction in returns on investment, combined with real terms increases in payments which have not (for unfunded schemes) been equally matched by an increase in contributions.

For the most part, balancing payments to address annual deficits in public service pension schemes administered by devolved Scottish agencies (Scottish government, Scottish local authorities, etc.) are met by the UK government through Annually Managed Expenditure (AME). AME allocations make no direct claim on the Scottish government's core budget which comes through Departmental Expenditure Limits (DEL) funding. The exception to this is the Scottish police and fire-fighters' pensions which are paid from Scottish government DEL. Increased spending on balancing payments for police and fire-fighters' pensions implies less spending on other Scottish government DEL categories (including health and education, etc.). Other than the police and fire-fighters' schemes, direct spending on pensions does not immediately or directly affect the spending power of the Scottish budget, but changes in employers' pension contributions do (and cost implications of early retirement are met by the organisation granting the retirement).

In equilibrium, public service pension schemes should broadly balance, in other words the employee and employer contributions over a period of time should match payments to pensioners. This equilibrium can be disturbed if the public pensioner dependency ratio (i.e. the ratio of public sector pensioners to public sector workers) changes over time. Throughout the 2000s, the expansion of public sector employment helped keep public service pension schemes solvent. A shrinking public state could pose major challenges to the funding of public pensions, particularly those that are unfunded, for which current payments are met from current contributions. Shortfalls in such schemes could be met either by increasing current contributions (from employers, employees, or both), or through general government revenues. Filling a gap through general government revenues would imply an increase in taxation or reduction in spending on public services. Either case may be viewed as being inequitable, as it represents a redistribution of revenues to retirees from the public sector workers from the general population.

As it has a slightly higher proportion of public sector employment and slower growth in the population of working age, Scotland is likely to have a higher public sector pension cost per working age person than rUK. If Scotland became independent and its public sector increases in size relative to rUK, then maintaining the solvency of pay-as-you-go public sector pensions would be easier since there would be more contributing public sector employees, but the Scottish taxpayer would have to meet the additional costs of its larger public sector.

Occupational pensions in Scotland

Due to Scotland's lower life expectancy, we have shown that it costs less at an individual level to provide a given state pension in Scotland. The same level of benefits to a pensioner could be provided as in rUK but with a lower tax contribution history by that pensioner; or Scotland could choose to maintain tax contributions from prospective pensioners but defer proposed increases to the UK SPA.

However, it does not follow that private pensions in an independent Scotland would necessarily be cheaper. UK insurers already price annuities on the basis of highly disaggregated geographic (postcode) data and so higher mortality rates in Scotland are already reflected in the price that members of private pension schemes in Scotland are charged. In addition, actuarial valuations of occupational pension schemes are already conducted on the basis of scheme specific experience and so already reflect the geographic and socio-economic status of their members. Therefore, pension schemes in Scotland already benefit from lower funding costs due to higher mortality rates.

If mortality differences are already reflected in the price of Scottish occupational pensions it is less obvious that independence would affect their cost. There will likely be some short-term administrative costs as schemes determine which jurisdiction they operate under, but these are likely to be relatively small. There may also be additional costs if the market becomes less competitive due to its smaller size, though such effects are difficult to anticipate.

However, there may also be some issues on the asset side of providing pensions. The Institute of Chartered Accountants of Scotland (April 2013) argued that, under European Union regulations, pension schemes which operate across borders "would have to be fully funded at all times" and any initial underfunding following independence "would have to be rectified immediately rather than through a staged recovery plan". It is difficult to ascertain whether this issue involves any real costs: it may be that the EU would not enforce its rules for a transitional period commensurate with the staged recovery plan that these schemes would have had anyway. But even if this issue does involve costs, this is an area in which it is easy to overstate the costs by quoting some gross exposure to risk figure rather than calculating the real net costs.

Pensions World magazine estimated that "UK institutional pension fund assets hit an all-time high of 1.7 trillion [pounds sterling] in 2012". Allocating this by the shares of total employment income of taxpayers in 2009-10 from the Survey of Personal Incomes, suggests that the value of the aggregate 'Scottish pension fund' might be around 143bn [pounds sterling]. Pension consultants LCP estimate that "FTSE 100 member companies ... responsible for pension liabilities worth nearly 0.5 trillion pounds ... [have] pension scheme deficits ... [of] 43 billion pounds". Assuming these deficits are representative of all pension funds, and in particular, of Scottish pension funds, produces an estimated combined deficit for Scottish pension funds of 12bn [pounds sterling]. Assume that this entire 12bn [pounds sterling] is the relevant cross-border figure that the Scottish corporate sector needs to fund immediately (this will be a vast over-estimate). Employers are obliged to make up this deficit anyway, presumably from future profits. Firms now have to borrow money immediately to put into the pension fund, and they will meet the repayments on this borrowing through these profits.

This is a cost of capital issue. If we assume that the financing cost is the full margin between corporate borrowing and government borrowing (which is to vastly over-estimate the cost) of perhaps 4 per cent and that the moneys are borrowed as a 10-year loan schedule, then the present value of the extra payments under closing the deficits immediately is less than 3bn [pounds sterling]. This is very much an upper bound to the estimated cost to the corporate sector in Scotland.

Now consider the implications for Scottish pensions of the market in Scottish government bonds which will have to be created on independence. Pensions are a guaranteed income for pensioners. As such they must be funded by purchasing financial assets which provide a guaranteed income. Typically this is done using government bonds denominated in the currency in which the pensions are to be paid. To the extent that Scottish pension funds may view Scottish government bonds as the most appropriate matching assets for Scottish pension liabilities, there may be implications for Scottish pension savers.

The government debt of small nations attracts a liquidity premium i.e. the interest rates on the debt are higher since the price to buy this debt is lower. Armstrong and Ebell (2013) raise this issue as a cost to be faced by the government of an independent Scotland. This is true, but the counterparty to this transaction receives the benefit. This is not a risk premium due to extra default risk or extra revenue volatility risk, it is a margin that reflects the difficulty that active investors have in trading this debt given that its market is likely to be very thin. This reduces the demand for this debt from active investors, and so allows 'buy-to-hold investors', like pension funds, to buy these cashflows at a cheaper rate. Therefore, if Scottish pension funds wish to hold Scottish government debt, then they may be able to finance their pensions at a cheaper rate due to the liquidity premium on Scottish debt.

How much extra pension could Scottish pension funds buy, given this liquidity premium on Scottish government debt? Armstrong and Ebell (2013) estimate the liquidity premium on Scottish government debt relative to UK government debt at 0.63 per cent. To simplify how much extra pension can be bought taking this liquidity premium into account, assume that a pension fund will be used to purchase an inflation linked male annuity from age 65 in x years' time, under the '$1' series of mortality tables from the Continuous Mortality Investigation. Since the pension payments are inflation linked, the pensions funded by UK government debt should be discounted at the real rate of interest, whilst the pensions funded by Scottish government debt should be discounted at the real rate of interest + 0.63 per cent. The results are shown in table 3. We can see that the liquidity premium allows a fairly substantial extra amount of pension to be purchased, especially given a long investment horizon over which to accumulate this premium, and especially given an otherwise low interest rate environment. Scottish independence can be seen to help in the 'search for yield'.

Armstrong and Ebell's (2013) estimate of the size of the liquidity premium is calculated specifically assuming a common currency. Under a common currency it is less clear why Scottish pension funds would hold Scottish government debt rather than rUK government debt, and why schemes from the rUK would not hold Scottish debt if it offered such attractive features for pension funds. As already mentioned, the liquidity premium is a cost to the Scottish government which may be passed on to the upper part of the Scottish income distribution if it is paid by taxpayers through higher taxes. Or it may fall on the lower part of the Scottish income distribution if it is paid by Scottish public service recipients through reduced levels of public services. The liquidity premium is a benefit to the (hold to maturity) holders of Scottish government debt. If these are Scottish pension funds then the liquidity premium represents a transfer from one section of society in Scotland to another. If the debt is held by external investors, then it will be an overall loss to the Scottish economy. A separate Scottish currency will ensure that Scottish pension funds seek to buy this debt, since they will wish to match pension liabilities denominated in the Scottish currency.

Finally, current annuity holders in Scotland have very little to concern themselves with, with regard to Scottish independence. Their annuities will be backed by UK government debt and so they are guaranteed their income in sterling. The only risk that they face is under the case of a separate Scottish currency, and then in the direction that implies relative success for the Scottish economy. The rUK pound sterling would have to fall in value relative to the new Scottish currency for these pensioners to be out of pocket. This is likely to happen under, say, a scenario of very high oil prices and booming Scottish revenues. This is a risk, but a risk which comes with the hedge of a Scottish government flush with cash and able to make good on any losses to this powerful section of the electorate.


This paper has examined how pension policy might interact with the debate about Scottish independence. Pensions play an important role in the independence debate partly because they comprise contracts that extend well beyond the date when independence might occur. Voters are seeking some reassurance that they will be no worse off in respect of their future pensions if Scotland should become independent. State pensions and public sector pensions are also important components of the contingent debt that an independent Scotland would have to manage.

Our analysis has shown that the costs of the state pension would be lower in Scotland, due to Scotland's lower life expectancy. This could, in theory, be used to reduce contributions or to delay increases in the state retirement age in Scotland, relative to the rest of the UK. However, Scotland has a relatively smaller working age population than rUK and therefore the costs of the state pension measured as a share of tax revenues will be higher in Scotland as the population ages. Nevertheless, the difference in the tax costs of the state pension is small relative to the increased taxes that will have to be raised to pay for population ageing throughout the UK.

Meeting the costs of public sector pensions poses a challenge to the UK government. The same would be true for an independent Scotland. Scotland has a somewhat larger public sector than rUK. Many public sector employees belong to unfunded pension schemes, whose costs are funded from current employer and employee contributions and the taxpayer. If an independent Scotland wanted to keep a larger public sector than rUK, it would have relatively more employees and employer contributions to meet the costs of unfunded pensions. But the overall size of the public sector would have to be met from higher taxes or reductions in the components of public spending that do not incur labour costs.

Another effect of independence would be the creation of a Scottish bond market in which one might expect Scottish pension funds would invest. If Scotland has to pay a liquidity premium on its debt relative to rUK due to its smaller market size, then Scottish pensions will be less costly to fund. This is good for pensioners, but bad for taxpayers or for the consumers of public services, since they have to meet the higher costs of Scottish debt either through higher taxes or reduced services. If the debt is held outside Scotland, then these groups will lose, but the gains will accrue to pensioners or others living outside Scotland.

Scottish independence is just one of a number of factors which will influence the affordability of Scottish pensions--rising life expectancy and a growing pensioner dependency ratio will pose critical challenges for pension affordability regardless of whether Scotland becomes independent or remains part of the UK. In itself, independence is unlikely to materially alter the scale of the affordability issue. This paper has shown that the extent to which pensions are likely to be more or less affordable under independence will depend on factors including the size of the public sector, the value of tax revenues, and the performance of a Scottish bond market.

What can be said from the analysis in this paper is that there are important demographic differences between Scotland and the UK as a whole, and these differences suggest that an independent Scotland may, or perhaps should, make different policy choices with regard to pension eligibility. However, the importance of economic issues to the outcome of the referendum suggests that the political economy of pensions will play a key role in determining the attitudes of taxpayers and pensioners to the referendum. Considerations of loss aversion may then paradoxically result in those proposing a breakaway from the political union to argue that they will preserve existing contribution and benefit rates.

Pensions provide an interesting case study for the economics of independence and fiscal federalism. Arguments for independence have typically emphasised the benefits of cultural or societal homogeneity: they have not focussed on whether the citizens of the new state belong to a different risk class from that which is appropriate for the entire political union. Similarly, though the early versions of fiscal federalism focussed on differences in preferences for public good provision, one might argue that permitting local areas to determine state pension and contribution rates would allow them to match more closely local mortality conditions, which would be welfare enhancing. However, these benefits might be lost if migration renders contribution and pension rates inappropriate. At least in this respect pensions do share some of the characteristics of the public goods generally associated with fiscal federalism.


ONS provides q_{x, t + 1/2} i.e. deaths rates at age x last birthday, and at time t (where t = 0 is 1st Jan 2012).

Want probability of death at each future lifespan in order to calculate expectation of life.

Assume the number of lives aged x last birthday at time 0, l_{x, t = 0} = 100,000 for all x (i.e. initial conditions are unimportant).

Have q_{x, t + 1/2} = (l_{x, t} - l_{x + 1, t + 1}) / l_{x, t}

So l_{x + 1, t + 1} = l_{x, t} (1 - q_{x, t + 1/2})

Then the probability of dying in year t (which you would have started with age last birthday of x + t, and so average lifespan from time 0 of t + 1/2) is pd_{x + t, t} = (l_{x + t, t} - l_{x + t + 1, t + 1})/l_{x, 0}. NB these probabilities sum to 1.

The expectation of life at 2012 for a life aged x last birthday is the sum over all t from 0 upwards of (t + 1/2) * pd_{x + t, t}.

The probability of survival to draw 1 [pounds sterling] of pension in year t is ps_{x + t, t} = l_{x + t, t}/l_{x, 0}. NB These probabilities sum to much greater than 1.

The value of 1 [pounds sterling] of pension (assuming economics cancel so no discounting) is the sum over all t from 0 upwards of delta(x + t)*ps_{x + t, t} where delta(x + t) = 0 if x + t < SPA, 1 otherwise.


Armstrong, A. and Ebell, M. (2013), 'Scotland's currency options', Centre for Macroeconomics Discussion Paper 2013-2.

Audit Scotland (2011), The Cost of Public Sector Pensions in Scotland, prepared for the Auditor General and the Accounts Commission, February, Edinburgh, Audit Scotland.

Department of Work and Pensions (2013), Benefit and Caseload Tables, Department of Work and Pensions, London, accessed at:, Dec 2013.

Lisenkova, K. and Merette, M. (2014), 'Can an ageing Scotland afford independence', National Institute Economic Review, 227, pp. R32-9.

Oates, W.E. (1972), Fiscal Federalism, New York, Harcourt-Brace Jovanovich.

Scottish Government (2013a), Scotland's Future: Your guide to an independent Scotland, Edinburgh, Scottish Government.

--(2013b), Government Expenditure & Revenue in Scotland 2011-2012, Edinburgh, Scottish Government.

David Bell, David Comerford and David Eiser *

* University of Stirling. E-mail:

Table 1. Public sector employment in Scotland and rUK,

                                    Scotland            rUK

Total population                    5,313,600        58,391,400

Public sector employees              631,845          6,166,347

Public sector employees               25.6%             23.1%
as per cent total employees

Average tenure with current           11.4              10.1
public sector employer (yrs)

Average weekly public sector       452 [pounds       481 [pounds
wage                                sterling]         sterling]

Total public sector wage bill    14,897 [pounds    154,701 [pounds
(m [pound sterling])                sterling]         sterling]

Public sector wage per capita     2,804 [pounds     2,649 [pounds
                                    sterling]         sterling]

Source: Quarterly Labour Force Survey, April June 2012.

Table 2. Net public service pension spending on
individuals resident in Scotland (fm)

                   2007-8    2008-9    2009-10   2010-11   2011-12

Scottish depts/
  agencies          -358      -287      -205       -44       337
rUK depts/
  agencies           337       558       586       694       785

Source: Scottish government personal communication.

Table 3. The increase in pension amount purchased due
to the liquidity premium (per cent)

                                     X (years)

Real interest rate        0         5        10        20

0                        6.4       9.8      13.3      20.6
2                        5.6       8.9      12.3      19.4
4                        4.9       8.1      11.5      18.4

Source: Author's calculations.
COPYRIGHT 2014 National Institute of Economic and Social Research
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2014 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Bell, David; Comerford, David; Eiser, David
Publication:National Institute Economic Review
Geographic Code:4EUUS
Date:Feb 1, 2014
Previous Article:Scotland: currency options and public debt.
Next Article:Can an ageing Scotland afford independence?

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters