An explanation, via buy-to-let analogy, of First Actuarial’s approach to the valuation of the USS pension scheme

Though on a much larger scale, the income and outgo of USS is structurally similar to that of a landlord with a buy-to-let mortgage, who is able and willing to hold onto his flat and rent it out for the next several decades. The landlord’s liabilities consist of regular and fairly invariant mortgage payments stretching over decades. USS’s liabilities consist of regular and fairly invariant pensions payments stretching over decades. The landlord’s income consists of the tenant’s rental payments, which are also regular and fairly invariant. USS’s income consists of employer and employee pensions contributions, plus investment income from the assets in the pension scheme, the combination of which is also regular and fairly invariant.

In order to determine whether his small business arrangement is solvent, the landlord should make a best estimate of his expected rental income and mortgage payments, and make sure that the income stream of the former is sufficient to cover his monthly mortgage liability.

Let us now suppose that our landlord is faced with regulations that impose the sort of funding requirements that private occupational pension schemes face. Such regulations would call for a triennial ‘mark to market’ valuation to determine whether or not this buy-to-let scheme is ‘underfunded’. Such a valuation takes the market value of his asset — his flat — at a given date once every three years and arrives at an estimate of whether the value of this asset is at least as great as the sum total value of his liabilities — the mortgage payments he owes in all future years. The value of these liabilities is discounted, to allow for the fact that there will be annual returns on the asset to help cover the cost of the liabilities. The estimated rate of return on this asset is known as the ‘discount rate’.

The most sensible way for the landlord to determine such a ‘discount rate’ is by reverse engineering a rate of the return on his asset from his independently derived expectation of rental income. Suppose, for simplicity, that he expects to receive £1,000 per month in rental income, and therefore £12,000 per year, for the next several decades. If his flat happens to be valued at £200,000 on the valuation date, he will divide £12,000 by £200,000 in order to derive a rate of return of 6%. If there is 7% housing inflation per annum between that valuation date and the next one three years later, so that his flat is worth £245,000 by then, then he will divide £12,000 by £245,000 in order to derive a new rate of return of 4.9%.

In estimating his rental income, it would make little sense, by contrast, for our landlord to adjust the expected amount of cash per month upward or downward significantly on the basis of fluctuations in the market value of his flat. This is because, as this graph shows, the market value of a flat fluctuates to a much greater extent than the rental income on that flat.

Image for post
Image for post
Annual percentage changes in housing prices versus rental returns in England

For the upcoming valuation of USS’s pension scheme, UCU’s actuary First Actuarial is advocating what I have described above as the most sensible way to determine the discount rate. On their approach, one takes the actual annual investment income on the assets in the pension scheme at date of valuation plus a best estimate of its rate of growth. Then one reverse engineers a discount rate by dividing this figure by whatever the market value of the assets happens to be at the date of valuation. Here, in more detail, is how First Actuarial describes their approach:

ONGOING VALUATION TECHNIQUE

6.11 In the >99% likely scenario of USS continuing as an open scheme sponsored by employers with a robust covenant, a good way to plan the contribution needs of the scheme is as follows.

6.12 The long run issue is the growth of asset income. A scheme which is open to new entrants has little need to buy assets (unless its membership is increasing) or to sell them (unless the membership is declining or the benefits have been cut). A valuation which plans for the >99% likely long term ongoing scenario would:

· Estimate the income expected from the assets (prudently, for a SFO [Statutory Funding Objective] valuation).

· Because the assets must be shown at market value, derive the rate of return which values the expected income at the assets’ market value (the internal rate of return).

· Use the internal rate of return to value the liability cash flows.

6.13 In this way, the ongoing planning valuation builds in a projection of asset income for comparison with the liability outgo. The prudent expected return on the actual assets of the scheme is incorporated. The cash flows of the scheme, on both sides of the balance sheet, are modelled and planned for.

In the first bullet point in 6.12 above, First Actuarial is proposing that USS provide a best estimate of future growth in current annual investment income, mainly from dividends from the stocks and coupon payments from the bonds they hold, and then prudently adjust this rate of growth downward.

In accordance with the second bullet point, that prudently adjusted estimate is then divided by whatever the market value of USS’s assets is at time of valuation.

This generates the discount rate (i.e., the third bullet point).

Some would resist the second bullet point because they embrace an ‘efficient market hypothesis’, which attempts to derive information about future income from the current market value of assets, where it is assumed that such income is a relatively fixed percentage of the market value.

As First Actuarial notes at 6.18 of the same document, this approach would be misguided when it comes to an estimation of income from bonds, where one takes the stated fixed coupon payment and divides it by the current market value of the bond to derive the yield:

6.18 Second, if the market movement is a move in the bond markets, then a corresponding change in the internal rate of return would not be disputed.

As I illustrated earlier, it would also be misguided for the owner of a buy-to-let flat to estimate his rental income as a relatively fixed percentage of the market value of the flat.

Is it sensible, by contrast, in the case of stocks, to suppose that dividend income will closely track the market value of the stock? This is what First Actuarial says at 6.19–6.22, which strikes me as sound:

6.19 It is in connection with a move in the equity markets that the question is raised. Does a 10% fall in market value signify an imminent and permanent 10% fall in dividends? If it does, the expected return on equities has not changed.

6.20 We cannot tell whether a short term market fluctuation signifies a move in dividends or not. We have to wait and see the dividends which emerge to distinguish noise, false expectations and irrational behaviour in the markets from a rational and precisely correct response to news affecting dividends.

6.21 The USS net cash flow position is not such that it need worry about short term market value volatility. It can afford to wait and see how dividends change. For a scheme like USS, which has a long term future with robust sponsoring employers, it is emerging fluctuations in asset income which matter most.

6.22 Turning the point around, a 10% rise in equity market value would not improve a cash flow based balance sheet by much either. On the other hand, a liability calculation using a gilt yield based discount rate would show an improvement in the balance sheet equal to the equity market rise. This is potentially imprudent, if the market rise is noise, false expectations or irrational behaviour in the markets rather than a rational and precisely correct response to news increasing dividends.

We can also observe, from graphs such as the following, that dividend income fails to track asset price, but is rather more predictable and less volatile:

Image for post
Image for post
Source: Robert Shiller, Irrational Exuberance, 3rd ed.

Robert Shiller offers the following commentary on the above graph:

The dividend value is extremely steady and trend-like, partly because the calculations for present value use data over a range far into the future and partly because dividends have not moved very dramatically. … [S]tock prices appear too volatile to be considered in accord with efficient markets. Assuming that stock prices are supposed to be an optimal predictor of the dividend present value, then they should not jump around erratically if the true fundamental value is growing along a smooth trend. Only if the public could predict the future perfectly should the price be as volatile as the present value, and in that case it should match up perfectly with the present value. If the public cannot predict well, then the forecast should move around a lot less than the present value. But that’s not what we see in Figure 11.2.

Written by

Professor, Dept. of Philosophy, Logic & Scientific Method, LSE

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store