Asset Valuation (1829-1952)
Nineteenth-century actuarial thinking on valuation was resolutely focused on the liability side of the balance sheet. Actuaries generally took asset values as they were given to them by the life office’s investment department. This started to change in the mid-twentieth century, most notably in the late 1940s and early 1950s, when a number of sophisticated actuarial papers were published discussing how assets should be valued for the purposes of measuring solvency and surplus. The increased holdings in risky asset classes such as equities and property, together with the substantial movements in interest rates that were experienced in the 1930s and 1940s, were both significant in motivating this new actuarial focus on the asset side of the balance sheet.
Early nineteenth-century actuaries were aware that asset valuation could play an important role in the assessment of surplus. William Morgan, in his address to the Equitable general court in 1829, stated:
‘That the Society may probably have the right to estimate its Stock in the Public Funds at the price it bears at the time of determining the surplus, I do not pretend to dispute ... The great fluctuation in the value of the funded property [long-term government bonds] will always render the surplus very precarious. It might perhaps be desirable that the Stock in the Funds should be invariably fixed at some mean price, rather than that the amount of the surplus should be suffered to depend on property perpetually changing its real value.’
This is an interesting early example of actuaries’ disdain for market prices. But it is important to understand the context of this statement. Morgan was specifically considering the determination of surplus that had emerged between two valuation dates for the purpose of setting the bonuses that would be distributed to with-profit policyholders. His comment is essentially a reflection on the amount of ‘bonus smoothing’ that he believed was desirable to have in with-profit distributions over time and between generations. This need not have any direct implication for his view of whether market values were a sound indicator of economic worth or whether or not they should be used when considering the absolute solvency of life offices.
The actuarial preference for the asset side of the balance sheet to use something more stable than market values was well-established by the early twentieth century. The standard practice was to hold the lower of an asset’s book value and market value, a practice summarised as ‘always write down, never write up’. A rationale for this approach can be found in R.K. Lochhead’s 1930s textbook for actuarial students:
The established practice of British Offices is to write down security values when prudence dictates such a course and only in very exceptional circumstances to take credit for capital appreciation apart from realized profits on maturity or sale. The justification for this procedure lies in the necessity for creating a reserve against adverse movements in the value of assets, so endeavouring to secure that the life assurance fund shall be represented by assets of equal value.
Lochhead’s justification is notably focused on solvency rather than the equitable distribution of surplus. It is also interesting to observe that his justification can be construed as a prudential solvency approach where assets are valued at market value less a reserve for risk. More generally, a number of possible justifications were forwarded in support of the ‘always write down, never write up’ approach, relating to both measuring surplus and ensuring future solvency: it produced a prudential valuation (any excess of market value over book value was a ‘hidden reserve’); it could achieve more consistency with the net premium valuation of liabilities (which tended to use ‘sticky’ interest rate assumptions that would only be revised when absolutely necessary); it smoothed the release of surplus to policyholders over time and produced a less volatile public balance sheet.
As mentioned in the liability valuation discussion above, a gross premium valuation became the standard actuarial practice for the internal assessment of the solvency position of with-profit funds. In that context, and in the context of surplus assessment using a bonus reserve valuation, the use of market values for assets was more typical, and was increasingly viewed as the natural valuation approach to take in such cases, as was argued by T.R. Suttie in his important Institute paper of 1944:
There does not seem to be any object in making a bonus reserve valuation unless it is based upon the true facts, i.e. the actual premiums to be received, the actual renewal expenses, the experience rate of interest, and the market value of assets. Hidden reserves seem entirely contrary to the principles of a bonus reserve valuation ... The position is quite different under a net premium valuation. The valuation basis being more or less fixed, hidden reserves are necessary, not only to avoid violent fluctuations in the rate of bonus if appreciation or depreciation occurs, but also to enable the results of the published net premium valuation to correspond with those of the unpublished bonus reserve valuation.
The net premium valuation was still used widely, both in measurement of surplus for bonus determination, and in the published regulatory returns of life offices. The ‘always write down, never write up’ asset valuation method was invariably used with net premium valuations in the first half of the twentieth century. As the asset and investment thinking of the profession developed in the 1940s, the justifications for its use came under inevitable scrutiny. It made the relative treatment of assets and liabilities somewhat opaque, particularly in the context of volatile long-term interest rates. In the event of a general fall in the level of interest rates, assets’ book values could substantially understate the market value of long bonds, creating a substantial hidden reserve. Meanwhile, the actuary would be under pressure to reduce the valuation interest rate used for the liabilities to reflect the reality of the new market prices. But the assumed valuation interest rate was at the discretion of the actuary and need not be directly related to the market yield. The net impact of the rate fall would therefore be confused or even arbitrary. Similarly, in the event of an interest rate rise, the assets would be written down to their new market values, and the actuary would have some choice over whether this was offset by a reduction in the size of the estate or by increasing the liability valuation interest rate. It was the role of actuaries to use their judgement in these choices so as to meet their twin objectives of ensuring the solvency of the fund and delivering an equitable distribution of surplus to different generations of policyholders (and holders of different types of with-profit policy). But it is easy to see how a combination of a liability valuation that assumes premiums that have no direct relationship to the actual premiums that will be received and that uses an interest rate that is not directly related to that which is expected to be earned, together with an asset valuation that bears no direct relationship to current asset market values, could make non-actuaries somewhat suspicious of these actuarial dark arts.
There was a more specific issue with the logic of the ‘always write down, never write up’ asset valuation approach: if a reserve against adverse asset movements was desirable, as per Lochhead’s textbook, why was the difference between market values and book values the right size for this reserve? If such a reserve was important, should a more risk-based and scientific approach not be used to determine its necessary size? This line of argument was developed in a couple of papers presented in 1947-1948—Pegler’s Institute paper that arose in the discussion of equity investment above; and a Faculty paper written by the investment actuary Lewis Whyte.
Pegler’s 1948 paper that advocated the maximising of expected investment returns also advocated what was, for the time, a novel approach to dealing with equity valuation and the distribution of surplus arising from equity investment. He argued that equities should always be valued at market value, dismissing book values as ‘illogical’ and based on ‘fallacies’. The recognition of the surplus created by an increase in market values could be partly deferred by creating an explicit reserve for depreciation. This reserve would be set based on an assumption about how far equity market values could fall. Pegler suggested an assumed fall of 50 % in equity market values might be a reasonable basis, though he conceded the choice would be subjective. He argued for a similar approach to be applied to fixed interest securities, and suggested a long-term interest rate rise of 0.5 % would be a reasonable assumption for determining the depreciation reserve. The sizes of these two ‘stresses’ are interesting: both were ‘plucked out the air’ rather than being the result of any exhaustive statistical study, but both are reasonably consistent with the range of stress sizes applied to equities and long-term interest rates in twenty-first-century insurance solvency assessments.
In his paper, Whyte argued for an asset valuation approach that had a similar effect to Pegler’s market value plus depreciation reserve approach. Whyte proposed that assets should be valued at a ‘notional value’, which was ‘defined to reflect its fair value ... less an appropriate margin for adverse price fluctuations’. In the context of setting a notional price for equity holdings he wrote ‘it would be helpful to have in mind a price below which the security in question would be considered as unlikely to fall’, though he did not venture any quantitative suggestions. For long-term government bond yields, he, like Pegler, suggested a margin of 0.5 % (Pegler’s paper came after Whyte’s and Pegler’s assumption was influenced by Whyte’s).
Both these approaches are notable for advocating a prospective risk-based asset reserve rather than an arbitrary ‘hidden reserve’ based on historical transaction prices. This concept was fairly revolutionary, and inevitably was resisted by the more traditional thinkers. The depreciation reserve approach also required difficult assumptions to be made about how much depreciation would be reasonable to reserve for. A parallel to the net premium valuation versus bonus reserve valuation debate can be observed here—actuaries were reticent to adopt new methods that required them to make new, difficult and explicit assumptions relative to the more opaque requirements of their established methods.
The application of the depreciation reserve concept to bonds was especially confusing. Actuaries understood that from an asset-liability perspective they were often more exposed to falls rather than rises in interest rates. So it did not make sense to hold a reserve based on how much bond values would fall if rates went up. These were teething issues that would be ironed out in the coming decades. Pegler and Whyte’s contributions should be recognised as important groundwork for the later development of solvency capital systems that were based on the assessed riskiness of assets’ market values.