The topic of prescribed assets featured at the ANC 2019 election manifesto has put many investment professionals and asset consultants on edge. They were first raised as a potential policy tool in 2017 when the topic returned to the national dialogue.
The definition of what prescribed assets would involve (if implemented) is open to interpretation and unspecific as to whether it will prescribe state-owned enterprises (SOE) funding such as Eskom, Transnet, SANRAL and others. To help inform investors on this topic, we set out to provide some history of the effect prescribed assets had on the South African financial sector, to draw lessons from the past and identify what some of the unintended consequences may be of forced or directive investments.
Prescribed assets are not a new concept to South Africans. In the lead up to the 1980s, the Nationalist government legislated prescribed assets in pension funds, peaking in 1977 when a fund had to include at least 77.5% of its assets in a combination of state-owned companies (SOCs) and government bonds. Prescribed assets were abolished in the National Budget of 1989.
The discussions around prescribed assets have once again raised suspicion that considering the introduction of such a framework is symptomatic of an underlying funding challenge that SOCs face. In recent years, investors have been unwilling to fund many of them any further due to governance and financial health concerns at these entities. As an example, SANRAL last issued public debt in 2016 and is eager to come to market again in 2019 for funding while appetite remains limited. In 2017, SOCs had negative net issuance, redeeming more outstanding debt (R13.7 billion) versus new debt issued (R12.6 billion). In 2018, net SOE issuance was a mere R5.7 billion, which is likely much less than required to achieve their objectives. Net issuance forecast by Standard Bank for the coming year penciled in for SOCs are only fractionally higher compared to last year.
On the one hand, the question is whether it may be prudent to introduce such a policy tool, but also an important consideration is the potential impact on market infrastructure. One of the key concerns is that retirement funds and other potential financial or savings vehicles e.g. unit trusts become an instrument of state, by avoiding the discipline of financial markets and responsibility imposed by asset managers and trustees to act in the best interests of its members if it yields below-market returns. The preamble of Regulation 28 states that funds have the fiduciary duty to act in the best interests of its members. Its duty of care is to the ultimate beneficiaries of the fund.
When we look at the large asset managers in the asset management industry who are custodians of retirement savings of members, we start with the PIC that is the largest. The PIC is a defined benefit fund with liabilities of R2.1 trillion. Long-term insurers (LTI) have liabilities totaling of R2.9 trillion, with low amount of free assets. When it comes to the official public sector self-administered pension and provident funds (e.g. public sector pension funds such as Transnet, Post Office, etc.) are also defined benefit schemes with liabilities totaling R2 trillion.
The privately administered pension and provident funds have assets of around R1.5 trillion, most of which are defined contribution funds. Finally, we have unit trusts which collectively hold around R2.3 trillion in savings. Readers should note that the ANC manifesto also made mention of unit trusts in the document, which means that non-pension/provident funds are also in scope for consideration. Of these, R350 billion are sitting in money market funds. For money market funds to invest in long-term developmental assets leads to a huge maturity mismatch and also creates issues with respect to Board Notice 90 for FSCA. Also, most unit trusts are equity oriented and must allow investors daily liquidity which is not suited towards a buy and hold strategy of long-term assets that may have low liquidity.
In total, we have over R10.8 trillion worth of assets that make up the various categories of funds in the industry. This is stock and not flow. Recall that stock of savings is accumulated over a long timeframe with a percentage of the member’s monthly salary being deducted and invested into these funds on a pre-tax basis. Any changes on asset allocation which impacts on stocks have got potential affect financial stability.
If there was a large-scale sale of equities to free up money to move towards these prescribed developmental assets this is likely to have a rather big impact on the equity market. It takes a long time to build up these pension fund assets. The Pension Fund Act was first promulgated more than 60 years ago in 1956. The current contribution rates that we face now are much lower than in the early days because there has been an increased cost of death and disability risk cover on the large group schemes. These retirement assets are very valuable national assets and how they are deployed should be very carefully considered.
So, as we go back to history (1958-1989) when the prescribed assets regime held, it originally started with the intention of protection for policyholders and pension fund members, but over time evolved and began to be regarded as a reliable and assured source of public sector funding. What this meant in practice was that 53% of all new money going into these funds had to be deployed in approved and government securities. A 33% ratio applied to all LTI (Life Companies) funds and a 75% ratio to all PIC funds.
What were the consequences for the savings industry during this period? The outcomes varied. During the 1960s inflation was under control averaging around 3%. During this period there was positive real returns on prescribed assets. During the 1970s things changed a lot when inflation averaged 11% and prescribed assets produced real returns of -2%. During this period the opportunity cost of not investing in equities (which returned 24% per annum on average during this decade) and in prescribed assets was -9%. During the 1980s inflation averaged 14.5%, resulting in negative 0.4% return in prescribed assets, with an opportunity cost relative to equities of -3%. This resulted in a destruction of wealth.
Moreover, besides the factor of returns and opportunity cost to equities one should also consider the impact of market liquidity. The Jacobs Committee of 1988 was appointed to investigate prescribed assets and investigate several distortions. Ultimately, the committee recommended this policy be abolished unless they involved positive real returns or under the policy that these available for investment on a purely voluntary basis. Due to their recommendation, the establishment of Primary Dealers and Financial Markets Advisory Board occurred. It marked the beginning of STRATE and BMA.
When we look at how this all impacted on changes in asset allocation we see a roughly 40% allocation to prescribed assets in the mid-to-late 1980s. After abolition, this fell as one would expect and is currently near 20% when we look at pension funds. For life companies this also fell from near 22% to close to 9% in 2016. A shift to equities occurred as asset allocation to that asset class increased.
Third party independent asset managers such as Coronation, Cadiz and the like started to emerge as provident fund money was no longer better off managed purely by life companies who had to only hold 33% prescribed assets. One of the Margo Commission outcomes was a change in taxation of life companies that created less of an advantage for them to manage pension/provident fund assets. Along with this for the first time came performance benchmarking when before no comparisons existed.
The consequence for liquidity is also a factor. In the era of prescribed assets, bonds were valued at the lower of cost or redemption value. That resulted in many bonds issued at the time being issued at a discount to par value. For example, the Transnet T4 had a 7.5% coupon. As it pulled to par you had the situation of instantaneously owning more prescribed assets. Trading, however, was inhibited as one would never sell a bond below book or par or nominal, since if you did so you would have to top-up your prescribed asset holdings. This meant huge disincentives to do any trading. The preference therefore was to buy to hold to maturity.
Is it a lack of demand or a lack of supply when it comes to feasible and attractive infrastructure and SOC investment opportunities that would necessitate the introduction of prescribed assets?
Whichever one it is, unintended consequences will need to be thoroughly assessed and understood. According to Gill Raine from ASISA who recently presented on this topic, their asset allocation numbers show that as of end 2016 the PIC had only 1.16% exposure to pure infrastructure or development assets. Their retirement funds and life companies had about 2.5% exposure respectively.
Prescription of investments in developmental projects yielding below commercial returns will not be welfare enhancing if there are not positive real returns. In a poor outcome it could even be destructive if foreigners and locals become forced sellers of bonds or if funds were forced to sell equities to meet the requirement. The risks associated with prescribed asset allocation are well known. Forcing savings into debt at the expense of other assets (such as equities) is suboptimal. A big risk, in our view, is that prescribed assets will adversely affect foreign holdings of local debt.
It is well known that foreign investors are the single largest holder of fixed-rate government debt; they held 47% (or R711 billion) of outstanding bonds in the fourth quarter of 2018. In comparison, local pension funds held R275 billion of outstanding fixed-rate debt. A big risk, in our view, is that prescribed assets will adversely affect foreign holdings of local debt. Prescribed investment in local debt (whether in government, SOE or municipal debt) would create artificial demand for bonds and result in lower (real) yields than would otherwise have been the case. While the initial decline in yields would be positive for incumbent bond holders, there will be a point where foreigners are not compensated, in real terms, for the risk they are taking. This is likely to lead to foreign selling of local bonds. This selling would have to be absorbed by local savings.
Reduced foreign holdings of local debt, combined with local savings channeled towards government and quasi-government debt, would lead to less savings being available for private sector borrowing. Ultimately, it may lead to a situation where most of the domestic savings is absorbed by government and semi-government debt (perhaps similar but not to the same extent as, what South Africa experienced in the 1980s).
Periods where the government crowds out the private sector tend to coincide with lower growth rates and lower fixed investment. Lastly, prescribed assets may not only crowd out private-sector borrowing but will likely also suppress demand for other assets such as listed equities and corporate debt.
Michael Grobler is a strategy analyst at Ashburton Investments.