Actuaries and Investment Liquidity (1862-1933)

A second paper on life office investments followed Samuel Brown’s survey a few years later in 1862. Arthur Hutcheson Bailey’s ‘On the Principles on which the Funds of Life Assurance Societies should be Invested’ was the first actuarial paper that analysed investment strategy in the context of life assurance liabilities.[1] At five pages in length, it also has the distinction of being one of the shortest ever actuarial papers. It became a classic paper and was core reading in British actuarial examinations for at least 50 years following its publication. Bailey was the actuary of London Assurance, a significant life office of the period (and one of the very oldest, having been established by Royal Charter back in 1720). Like Samuel Brown, Bailey too went on to become President of the Institute of Actuaries, in Bailey’s case from 1878 to 1882.

Bailey considered investment strategy in the context of the characteristics of life assurers’ liabilities. The nature of these liabilities would inevitably place some constraints on the features of an appropriate strategy. Bailey, however, also recognised that these liabilities could create some unique comparative advantages for life assurers relative to other financial institutions:

They (life assurers) engage to pay fixed sums of money at periods generally long distance from the time when the contracts are entered into ... the probable amount of demands on their resources can be calculated from time to time within not very wide limits. Life assurance Societies, unlike banks and commercial enterprises generally, are not exposed to sudden or unusual demands on their resources in times of panic and financial difficulty.[2]

Put simply, life assurance liabilities are long-term, predictable and illiquid, especially when considered relative to those of banks. Bailey took it as a given that illiquid assets offered additional compensation to investors for sacrificing liquidity. He hence argued that life assurers should take full advantage of their unusual capacity to bear asset illiquidity:

The much larger proportion [of life office assets] may safely be invested in securities that are not readily convertible; and it is desirable...that it should be so invested, because such securities, being unsuited for private individuals and trustees, command a higher rate of interest in consequence [emphasis added].[3]

Mortgages, which Bailey defined as ‘every species of loan secure on tangible property’, were the main asset class he identified for investment in ‘securities that are not readily convertible’. He recognised that the size of these asset classes may be exhausted by the growing funds of life assurers and hence that alternative illiquid asset classes would need to be found. He identified loans for land development under recent Land Drainage Acts as a natural candidate—essentially, nineteenth century infrastructure debt.

Bailey viewed the illiquidity premium as the way in which life assurers should look to enhance their investment yield. He did not believe that life assurers should attempt to increase investment returns by taking equity risk or substantial credit risk. His starting investment principle was ‘That the first consideration should invariably be the security of the capital’,[2] and the notion of life assurers investing in equity investment was anathema to Bailey: ‘It is, I think, generally admitted that the ordinary stock and shares of trading and other companies are not eligible for Assurance Societies’ investments, as being too speculative.’[5]

Bailey’s thinking stands the test of time strikingly well. To the modern actuarial reader there are clear similarities with Bailey’s arguments and contemporary ideas on investment policy for long-term non-profit life business such as fixed annuities. The concepts of competitive advantage through unique longterm illiquid funding and the availability of asset illiquidity premia are increasingly central to modern actuarial thinking on investment strategy for illiquid liabilities such as these. Asset classes such as mortgages and loans remain the natural asset classes for life assurers intending to access illiquidity premia.

Bailey’s investment doctrine was largely consistent with the prevailing midnineteenth-century asset allocation practices described above. By 1890, life office allocations to mortgages were around 52 % and approximately 22 % was invested in various forms of illiquid loans. Another 6 % was invested directly in property. So by 1890, fully 80 % of funds were invested in illiquid asset classes. The remaining assets were invested in a range of asset classes such as British, foreign and colonial government securities (8 %), debentures (5 %) and cash and deposits (2 %).[6]

The final quarter of the nineteenth century was the high watermark for illiquid asset investment by British life assurers. In subsequent years, British life offices started to tilt their asset allocation towards more liquid assets. The final decade of the nineteenth century marked a decisive shift away from mortgages. Allocations to more liquid stock exchange securities such as long-term debentures and government bonds correspondingly increased. Between 1890 and 1900, the percentage of life assurance assets invested in stock exchange securities increased from around 18 % to 39 %. By 1910, this percentage had reached 49 %; by 1920 it was 60 %.[7] A similar trend was observed amongst the US life offices, and indeed started a decade earlier following the US banking crisis of the early 1870s. The US offices, however, retained a larger mortgage allocation than the British offices after the switch. Between 1874 and 1920, US life offices’ investments in mortgages had been reduced from 54 % of assets to 34 %.[8]

This great rotation from illiquid to liquid assets that occurred between 1890 and 1920 was not solely due to a desire for greater asset liquidity, though that was a factor. Other drivers were also important. There was a substantial fall in the value of British agricultural land during this era (perhaps as much as 30 %). This may have awakened life offices to the potential risks inherent in the mortgage asset class. It also reduced the value of the land upon which new mortgages could be offered. Increasing competition amongst mortgage lenders emerged, with legislative changes making the provision of mortgages more accessible to other investors such as trust funds. Mortgage yields fell.[9] In summary, there was a reduction in the illiquidity premium on offer, a reduction in the size of the market, and perhaps a re-appraisal of the risks associated with the asset class. All whilst the size of funds available for investment by life offices was increasing substantially.

The First World War also had a direct and significant impact on asset allocations. In particular, it resulted in a very substantial increase in life offices’ allocations to British government securities. Life offices were not legally compelled to buy government debt, but they reached a form of patriotic gentleman’s agreement with the government where they agreed to buy gilt issues during the war.[10] As a result, investment in British government bonds increased from 1 % to 35 % of life office assets between 1915 and 1920. However, this was not the main driver of the move from illiquid to liquid assets: much of the increase in gilt allocations was financed by reductions in liquid debenture securities (the asset allocation of which was reduced from 30 % to 12 %). Nonetheless, the scale of the gilt allocation necessarily further reduced illiquid asset holdings. Mortgage allocations fell from 21 % to 15 % between 1915 and 1920.[11]

Changes in the liquidity profile of life assurance liabilities played a significant part in motivating this fundamental shift in asset allocation and liquidity profile. The liquidity requirements of the liabilities were growing by the end of the nineteenth century. Unlike 100 years earlier, these liquidity requirements were not driven by the risk of mortality spikes arising from epidemics (whilst the influenza epidemic of 1918/19 did generate an unexpected concentration of claims, its impact was still relatively small in comparison to premium incomes).[12] Instead, the liabilities’ liquidity profile was being changed by surrender practices. Surrender rates had increased. In the fast-growing industrial assurance sector of the 1920s, double-digit surrender rates were the norm.[13] Historical anecdotes from the USA also highlighted how life liabilities may unexpectedly experience sudden liquidity demands. In 1873, a US banking liquidity crisis had created a surge of demand for the surrender of life assurance policies as these policies became the policyholders’ most direct route to cash. In British life assurance, there was an increasing practice, driven by competitive forces, of offering guaranteed surrender values (legislation passed in 1929 may also have encouraged guarantees in surrender values on some types of business).[14] A 1933 paper by William Penman noted that ‘if very large sums can be withdrawn, virtually without penalty and at very short notice, it becomes necessary to revise materially our views of Life Assurance finance ... it seems to me that the investment position, to the extent that guaranteed surrender values are made excessively liberal, approximates to that of a bank.’[15]

These changes in surrender practices materially impacted on the desired liquidity profile of the assets. With-profit life policies were never again thought of as illiquid. Today, actuaries’ illiquid investment strategies are largely focused on liability types such as non-profit fixed annuities that do not offer surrender values.

  • [1] Bailey (1862).
  • [2] Bailey (1862), p. 143.
  • [3] Bailey (1862), p. 144.
  • [4] Bailey (1862), p. 143.
  • [5] Bailey (1862), p. 145.
  • [6] Mackenzie (1891), p. 195.
  • [7] Murray (1937), p. 259.
  • [8] C.M. Gulland, in Discussion, Murray (1937), p. 274.
  • [9] Mackenzie (1891), p. 198.
  • [10] Whyte (1947), p. 223.
  • [11] Dodds (1979), Table A.3.
  • [12] Clayton and Osborn (1965), p. 65.
  • [13] Clayton and Osborn (1965), p. 64.
  • [14] Laing in Discussion, Penman (1933), p. 421.
  • [15] Penman (1933), p. 391.
< Prev   CONTENTS   Source   Next >