Public Infrastructure financing:
Submission to the Productivity Commission
David Richardson

Technical Brief

1

About TAI

The Australia Institute is an independent public policy think tank based in Canberra. It is funded by donations from philanthropic trusts and individuals, memberships and commissioned research. Since its launch in 1994, the Institute has carried out highly influential research on a broad range of economic, social and environmental issues.

Our philosophy

As we begin the 21st century, new dilemmas confront our society and our planet. Unprecedented levels of consumption co-exist with extreme poverty. Through new technology we are more connected than we have ever been, yet civic engagement is declining. Environmental neglect continues despite heightened ecological awareness. A better balance is urgently needed.

The Australia Institute’s directors, staff and supporters represent a broad range of views and priorities. What unites us is a belief that through a combination of research and creativity we can promote new solutions and ways of thinking.

Our purpose—‘Research that matters’

The Institute aims to foster informed debate about our culture, our economy and our environment and bring greater accountability to the democratic process. Our goal is to gather, interpret and communicate evidence in order to both diagnose the problems we face and propose new solutions to tackle them.

The Institute is wholly independent and not affiliated with any other organisation. Asan Approved Research Institute, donations to its Research Fund are tax deductible for the donor. Anyone wishing to donate can do so via the website at or by calling the Institute on 02 6206 8700. Our secure and user-friendly website allows donors to make either one-off or regular monthly donations and we encourage everyone who can to donate in this way as it assists our research in the most significant manner.

LPO Box 5096

University of Canberra, Bruce ACT 2617

Tel: (02) 6206 8700 Fax: (02) 6206 8708

Email:

Website:

Name of Brief

1

This submission is in two parts. Part A addresses the issues associated with the funding of infrastructure and Part B addresses the issues associated with the need for infrastructure in the first place.

Part A

Funding of infrastructure

Introduction

This part of the paper addresses issues associated with the question of the financing and provision of infrastructure. These tend to be economic considerations of a rather technical nature and simply take as given the need for infrastructure. However, this makes the decision as to whether or not to provide infrastructure a political decision and often reflects general predispositions on the merits of the public and private sector.

To borrow or fund out of consolidated revenue?

Government buildings and other structures are expected to last a very long time. The ABS for example works on the principle that non-dwelling construction assets will last an average 54 years while new government buildings are assumed to have the same average life as private commercial buildings of 65 years.[1] This raises the question of whether such long lived projects should be paid for by the present generation or by future generations who will also derive the benefits of the projects. Hence there are well-known public finance arguments to the effect that borrowing permits expenditures on long-lived projects to be shared across generations.

That approach seems useful when the question is one of financing a very large project relative to the population concerned. For example, a small community wishing to fund a large project may well wish to share the burden with future generations. However, in Australia it seems that there are numerous overlapping infrastructure projects that require not one-off approaches but a steady stream of expenditures. The profile of public investment in Australia looks much like any other category of government expenditure in that it is not particularly volatile as is shown in Figure 1.

Figure 1: General government fixed capital formation (% GDP)

Source: ABS (2013) Cat no 5204.0,

Figure 1 clearly shows that government infrastructure spending is fairly stable as a share of GDP with the exception of the stimulus following the global financial crisis and the temporarily high levels of spending in the mid-1960s possibly associated with peaks in spending on the Snowy Mountains scheme.

Given the actual profile of infrastructure spending we suggest a different approach is called for—one that takes account of the on-going nature of infrastructure spending.

Suppose the economy is growing at x per cent and we need to spend y per cent of GDP on infrastructure every year. Funding the infrastructure through borrowing is superior if the interest rate is lower than the annual growth in GDP. Conversely paying up front and taxing the present generation of taxpayers to raise y per cent of GDP is superior if interest rates exceed the rate of growth in the economy. That can be appreciated by using a numerical example.

For example suppose we need to finance projects worth 3 per cent of GDP every year. There are two extreme examples, first the government could raise all the money through general revenue with various taxes etc. In that case the revenue need would be 3 per cent per annum. Alternatively the government could borrow all its funding needs with a type of perpetual bond[2] but would need to raise sufficient revenue to cover the servicing costs. We can compare these two scenarios in the longer run.

Over the last 20 years GDP growth has averaged 6.35 per cent in nominal values. Interest rates on 10 year bonds have been around 3.5 per cent in recent times and averaged 5.1 per cent over the last decade. Over the last two decades they have averaged 5.96 per cent although that includes the period when interest rates were coming down from their very high levels of the early 1990s.

We can calculate the interest servicing costs and compare them with the costs of servicing the debt if all the finance comes from perpetual borrowing. Those comparisons are presented in Table 1. The cost of infrastructure financed through bond issues is calculated as the long run equilibrium value.[3] The scenario called ‘optimistic outlook’ refers to a possible scenario in which nominal economic growth continues at the average of the last 20 years (6.35 per cent nominal growth) and interest rates stay at roughly current levels of 3.5 per cent on 10 year bonds. The next scenario ‘extrapolation from the last two decades’, uses the same economic growth but uses the average interest rate on 10 year bonds over that period (5.96 per cent). The last scenario, ‘pessimistic outlook’, assumes the same economic growth but uses an interest rate of 7.5 per cent which is larger than the assumed growth in the economy (6.35 per cent).

Table 2: Cost of providing infrastructure; tax versus borrowing

Cost of infrastructure: Share of GDP (%)
Financed through current taxation / 3.00
Financed through bond issues
optimistic outlook / 1.65
extrapolation from last two decades / 2.82
pessimistic outlook / 3.54

Source: TAI calculations

Table 1 clearly shows that the cost of providing infrastructure depends on whether it is financed through taxation at the time of the investment or through borrowings. If financed through borrowing the actual interest rate is the critical factor. Financing through current taxation can be thought of as the reference case. Under the optimistic scenario interest rates are assumed to remain at roughly current levels (3.5 per cent). In that case the long term costs of servicing the debt is lower than raising the funds up front through current taxation. With interest rates of 3.5 per cent the costs of borrowing fall by almost half of the costs of the up-front taxation scenario.

Subject to interest rates being below the rate of economic growth we have to conclude that borrowing is superior to pay-as-you-go or up-front financing. That has been the case in the past and there is no reason to believe that the relationship will not persist into the future.

The argument so far has only considered the long run equilibrium position. It is also the case that even if we only consider one project at a time, so long as interest rates are below the rate of economic growth then raising taxes to pay interest on the debt will result in lower per capita tax burdens than trying to fund the projects out of current revenue. Of course the converse is also true; if interest rates are expected to be higher than the growth in the economy then debt finance involves increasing the per capita burden the longer the debt remains unpaid. That means that in principle governments should be flexible. However a sudden switch from debt financing to up-front financing would see the government effectively paying twice; it would be servicing the debt on existing infrastructure while fully funding new infrastructure. Taxpayers would be experiencing a larger burden than necessary over a fairly long transition period.

These considerations suggest that debt financing is likely to be the best approach in Australia. However, that assumes that the government would be responsible for infrastructure in the future. We now turn to the question of the extent of government involvement as distinct from private involvement in infrastructure.

Public or private funding?

The pragmatic view suggests that where it is mainly financing costs that are the critical factor in infrastructure there is a strong role on the part of the government and indeed, the Commonwealth government can borrow more cheaply than other levels of government. As we show below, funds secured via corporate finance will require premiums of well over 100 basis points and could add another 50 per cent to the cost of a project.

Comments by one Australian banker put the clear view that with interest rates currently so low the government should in fact borrow to invest. Cameron Clyne, the CEO of the National Australia Bank, put the view that the government should be borrowing more and exploiting its good credit rating and access to cheap capital. In his view government can finance long-term roads, rail and ports ‘far more effectively’. He also said ‘we don't have enough [debt]. We have a lazy balance sheet…We have a unique window as a AAA nation with strong demand for AAA debt to issue that debt and divert it to productive infrastructure’.[4]

Mr Clyne is absolutely right. As already argued above, the government has been able to borrow at well below four per cent or well under two per cent in real terms. Periods of low interest rates are the perfect time for investing in capital intensive projects. The hurdle rates of return that projects need to generate is so much lower and the borrowing costs are easily serviced.

Renewable energy projects are the perfect example of capital intensive projects. Once built there is minimal expense required for their on-going operation. Routine maintenance is about the only on-going cost. Depreciation expenses are also minimal because the turbines, solar collectors and the installations tend to be long-lived.

Mr Clyne did not describe how we can distinguish what should be done by government and what should be done by the private sector. However, his was clearly a pragmatic view that suggests different types of infrastructure will be handled best by different sectors.

Reserve Bank figures give a series for 10 year bond rates for both the Australian government and the NSW Treasury Corporation, the borrowing arm of the NSW government. Over the last decade NSW has had to pay an average premium of 49 basis points over the Australian Government borrowing rate. Over the life of a 50 year, billion dollar project the difference could cost $245 million or just under a quarter of the cost of the project. It is likely that the smaller states would have to pay a premium even higher than that paid by the NSW government. Local government would be in an even worse position.

Although the Australian government enjoys a lower interest rate on its debt than NSW, the NSW government still enjoys lower rates than those available to private corporations. For example over the last decade the bank lending rate for large business was 118 basis points above the Australian government 10 year bond rate.[5] Likewise RBA data give capital market spreads that indicate AA rated corporate bonds are typically well over 100 basis points above the government bond rate. Hence any form of indirect borrowing via the corporate sector is going to require a premium of well over 100 basis points and that would mean the whole of life cost of the project could be well over $500 million more expensive in the example examined above.

Those considerations suggest that government borrowing should be centralised and undertaken by the Commonwealth Government. However, the government could invent a new vehicle to attract loans from the public and there are historic precedents in Australia and overseas with special bonds designed to appeal to mom-and-dad investors. There is certainly a case for making it easier for such investors who want to purchase safe government paper in convenient sizes. But the design of government bonds takes us a long way from the topics at hand.

We now turn to the high rates of return in the private sector and why that is an argument against leaving the projects to the private sector.

High hurdle rates of return in the private sector

This section makes the point that there is a role for government in the provision of infrastructure and other services which can be profitably financed at the government bond rate but fall short of the very high rate of return hurdles that the private sector often sets for itself.

The lowest cost finance for infrastructure funding is available through commonwealth government borrowing. Australian infrastructure would be best financed by the general borrowing program—the regular issuing of government bonds. At the other end of the spectrum is privatising infrastructure development by giving responsibility for project investment and management to the private sector. This submission contends that the private sector tends to demand very high rates of return which would end up costing taxpayers dearly. (Or in the case of the provision of infrastructure financed by fee-for-service the commercial charge will be higher than the equivalent government charge because of the higher return on equity demanded by the private sector.)

A dramatic example of the role of rates of return in the government versus private sector came to light following the release of the report on the proposed very fast train. The report suggested a benefit cost ratio of over two and a commercial return of 0.8 to 1 per cent. However, it pointed out that:

An expected return of at least 15 per cent would be required at this stage of project development to be attractive to commercial providers of debt and equity to major infrastructure projects.[6]

A 15 per cent hurdle is very high and it is therefore unsurprising that many investments are undertaken by government by default. The government can borrow 10 year money at less than 3.5 per cent so the opportunity cost of capital for the government is a much lower interest rate than the private sector. Press reports suggest that pension funds usually require a return of 8 to 12 per cent on infrastructure investments.[7] By contrast governments should be prepared to invest when the rate of return exceeds the long term bond rate.

Note for example that the published expected rate of return to the government for its investment in the broadband network, NBN Co, is around seven percent.[8] By contrast Telstra earns a return on equity of 31.7 per cent[9] which implies a pre-tax rate of return of 45 per cent. Given the way companies like to brag about their return on equity there is simply no way that Telstra would contemplate investing in a broadband network with a return of seven per cent or that order of magnitude.

Years ago we had the big debate about banks closing down bank branches including in remote and regional centres that were ill-served by banking. A former governor of the Reserve Bank, Ian Macfarlane, made the point that by aiming for very high rates of return the banks were not investing in many things that would have been profitable, but were not able to reach hurdles of 18 to 20 per cent, or 26 to 29 per cent before tax. One consequence is that the big banks closed branches that while still profitable were not profitable enough.[10] Banks have made it less convenient for customers to undertake their banking even though the branches in question remain profitable.

This is a serious problem and is not unique to banks. Investment is major contribution to Australia’s economic growth and the rise in living standards over time. In a competitive market we expect that business will invest so long as they make a reasonable return. However, the banks put such a high hurdle on investments that they ignore investment opportunities that would otherwise contribute to overall wellbeing in Australia. This illustrates one of the many ways in which the existence of monopolies and oligopolies acts as a burden on the economy as a whole. In that regard note that the Australian economy is dominated by big business to the extent that sales of just the top five non-financial companies by market capitalisation had operating revenue of $240.2 billion equal to 16 per cent of the Australian economy while the top 10 controlled $323.9 billion or 22 per cent of the Australian economy. The ABS defines a large business as having 200 employees or more. On that basis large businesses had sales and service income of $1,117.5 billion in 2010‐118 or 80 per cent of GDP for that year.[11]