Wednesday, 6 December 2017

Answering Andrew Neill's Question

BBC's Andrew Neill has repeatedly asked Labour's financial spokespeople to answer the question of how much Labour's plans for renationalisation, and for infrastructure spending will cost. The same theme has been taken up by other presenters of TV politics programmes. As an economist, Neill should actually know that the basis of his question is fundamentally flawed. However, to be fair to him, his job is to ask the questions, so as to get the politicians to provide the answers. Again to be fair to him, the answers that Labour politicians have usually given have been pretty abysmal, so its no wonder he keeps asking the question. So, let me try to provide the simple answer to the two parts of his question. First, the cost of renationalisation, and second the cost of the infrastructure investment programme.

The basis of Neill's question in relation to renationalisation, or the taking back in house of PFI programmes, is fundamentally flawed, for the reason that the IFS themselves have given. That is that because any expenditure, or borrowing undertaken to buy up the shares of companies to be renationalised is simply a matter of exchanging one asset for a liability, it makes no difference to the government balance sheet. Issuing government bonds, to borrow money to buy the shares of a private company is rather like a reverse co-co. A co-co, is a contingent convertible bond. That is a bond, which offers the buyer the same fixed coupon, as any other bond, but which contingent upon certain future conditions, can be converted into a share, which then pays dividends, at a variable rate. If the government exchanges bonds for shares that is really just the same thing in reverse.

There are two ways that could be done. Either the government could convert existing shares in Company X into commercial bonds, with a fixed coupon, or else the government could issue Gilts to raise the funds required to be able to buy the shares in Company X, at their current market price, or some other price deemed appropriate. In the first case, the existing shareholders would obtain bonds in exchange, in the second, they would not, but would receive cash for their existing shares. Because, bonds have less risk than shares, and because they pay a fixed coupon each year, bond yields are generally lower than dividend yields on shares. For that reason, government spokespeople have been absolutely correct to say that such a programme of renationalisation would have no cost. If say, the average yield on shares in the utility sector is around 4.4%, the average for the FTSE 100, but the government, as it can, can borrow for ten years, at an interest rate today of only 1.3% on a 10 year Gilt, the net revenue for the government from such a swap amounts to 3.1% per share.

If we take, Severn Trent, for example, its dividend yield is 4.03%. In the last five years it has varied from around 3.5% up to around 4.75%. Severn Trent has a market capitalisation of £4.9 billion, consisting of approximately 235 million shares. To round things up, for ease of calculation, let's say the market capitalisation is £5 billion. So, the government would have to issue £5 billion of say, 10 Year Gilts. The annual interest it would have to pay on those gilts is currently 1.226%, or £61.3 million per year. But, now as the owner of £5 billion of Severn Trent shares, it rather than private shareholders will receive the dividends. The dividends on £5 billion of Severn Trent shares comes to £201.5 million per year. So, far from such a transaction representing a cost to the government, it would provide it with a net annual current income of £201.5 million - £61.3 million = £140.2 million. Over ten years, this net income amounts to £1.402 billion, or more than a third of the initial capital cost. The government could, however, raise the initial funds by issuing 30 year Gilts, which currently yield 1.792%. The annual interest on those bonds would be £89.6 million, giving a net annual revenue of £111.9 million, which over 30 years is equal to £3.36 billion.

In other words, the effect of such a transaction is to reduce the government's budget deficit by either £140 million per year, or £89.6 million per year, depending upon whether it is financed with 10 or 30 year bonds. Whilst, it results in an increase in total debt to the extent of the issuance of these bonds, that debt is cancelled by the increase in government assets in the form of shares, and is to eventually reduce government debt, as the income received from that asset is greater than the cost of debt issued to buy it. That is just the benefit that would accrue to the government from issuing debt to buy Severn Trent shares alone. That benefit is multiplied up for all of the other water companies, and other utility companies whose shares are bought by the issue of shares in this way.

Obviously, the net revenue that the government obtains from such transactions depends on the number of companies whose shares it buys up. The more companies whose shares it buys, and the higher the dividend yield on those shares, the more net revenue the government will obtain. This same calculation essentially applies to the taking back in house of PFI contracts. The government will have to issue bonds to buy back the contracts, but having done so, it will then obtain the revenue that private companies currently obtain each year from those contracts. The current yields on PFI contracts are often exorbitant, and so make the average yields on company shares look chicken feed. By taking these contracts back in house, the government would obtain significant net revenue. Once again, it represents not a cost, but an income.

In relation to the train companies, not even this argument applies. All that labour need do with the train companies is to wait until the current franchises run out, and then take them back in house. Its true that the government would then need to buy actual productive-capital in the shape of trains and rolling stock, but given that the current train companies can hardly use that equipment elsewhere, there would be a lot of such equipment available on the market at knock-down prices, to be bought up.

Now let's turn to the other issue, which is the cost of Labour's infrastructure investment programme. Labour has proposed to invest £500 billion in Britain's crumbling infrastructure, thereby mending Britain's leaking roof, whilst the sun of historically low interest rates is shining. Recent estimates, and analysis indicate that the multiplier effect for Britain is around 1.5. That is £500 billion put into the economy as government spending, results in an increase in GDP of £750 billion. Labour proposes to spread this programme over ten years, which thereby translates into £50 billion of investment per year, and an increase of £75 billion a year in GDP. But, let's deal with the total figure of £500 billion of investment.

Before doing that, let me deal with another point raised by Sky's Nial Paterson, last Sunday, who argued that putting additional fiscal stimulus into the economy at a time of near full employment would be likely to cause inflation. Firstly, although employment is certainly rising, and unemployment falling, the fact that a lot of the employment continues to be in part-time or zero hours employment, that Britain's appallingly low productivity levels indicate that a lot of labour is underemployed, typified by the large number of people in fake self-employment, shows that there continues to be a lot of slack in the UK labour market. With Brexit starting to tank the UK economy that slack is likely to increase. Moreover, if wages do rise, that will prompt firms to improve productivity, so that the net effect is likely to be a further increase in GDP, and potentially a greater multiplier effect on the back of it.

However, back to the substantive point. Labour has suggested that it would borrow £250 billion, and use it as part of a National Investment Bank to lever in additional private funds. Given the experience with PFI, I would have thought it better not to rely on such kindness from strangers. So, let us assume that Labour were to borrow the whole £500 billion for the ten years, by issuing 10 Year Gilts with a yield of 1.23%. That would mean the annual interest cost of such borrowing would be £6.15 billion per year. Assuming that such investment only resulted in a multiplier effect of 1.5, though as investment in productivity raising infrastructure it might well be expected the effect would be considerably greater, especially as much spending would suffer much less in leakages in imports etc., the overall increase in GDP would then be £750 billion. But, the immediate result of higher GDP, is that government tax revenues rise, and government expenditure on welfare tends to fall. 

For example, currently, £9 billion a year goes straight into the pockets of private landlords in Housing Benefit. If the government used some of the infrastructure spending to build council houses, so that people had houses to rent at reasonable rents, Housing Benefit could be abolished, and that alone would more than cover the annual interest cost of the borrowing for such spending. In addition, as the infrastructure spending put more people to work, in decent paying jobs, so the payments for Tax Credits, and other benefits would be slashed. As with the use of capital spending to buy shares through the issuing of bonds, the net effect would be a reduction in current spending, thereby reducing the budget deficit, and consequently facilitating the ultimate reduction in debt.

But, setting aside those current net revenue benefits of such capital spending, let's look at the immediate effect of the growth of GDP, and government revenues. According to the OECD, and other estimates, tax accounts for around 35% of current GDP. If the annual interest cost is £6.15 billion, over ten years that is £61.5 billion of interest payments. But, on the basis of a 1.5 multiplier, £500 billion of investment increases GDP by £750. In turn that means that government taxes rise by £262.5 billion. So, Andrew Neill is quite right to say, that such borrowing to cover infrastructure spending would not be self-financing in terms of the increase in tax revenues resulting from the rise in GDP, as a consequence of the multiplier. That might account for around half of the interest cost of such additional borrowing. However, as indicated earlier, this additional tax revenue is not the only net revenue effect of such investment.

As indicated, a saving of £9 billion a year on Housing Benefits alone would cover the annual interest cost; the saving on welfare benefits as workers are put to work, and wages rise, represents another significant saving on current government spending; if the infrastructure spending put Britain's productivity growth back to its historic 2% p.a. level that would add £40 billion a year to GDP, which means an increase in taxes of around £14 billion a year in additional revenues, which more than covers the £6.16 billion annual interest cost. A look at the queues of lorries and other traffic lined up for hours on Britain's motorways indicates just how much GDP could be increased by higher productivity from such infrastructure spending; a look at Britain's pathetic broadband infrastructure, and mobile phone network coverage shows how much GDP could be raised.

So, Labour's spokespeople are quite right to say that Neill is asking the wrong question. The basis of his question is flawed, because Labour's plans to borrow to invest, represent a reduction in current net spending, not an increase, and thereby also represent a reduction in the current deficit, along with a longer-term reduction in debt, and of the debt to GDP ratio. They need to do the maths to get the figures the next time he asks the questions. Its not as though they have not had time to do so, or that they shouldn't be expecting him to ask the question!

Finally, however, as I have pointed out many times in the past, the question of the cost of renationalisation is a bogus one anyway. What the government would be buying in any such renationalisation would not be the productive-capital of the business, but only the shares. As a corporate entity, Severn Trent owns its productive-capital, not the shareholders, be those shareholders private shareholders or the government. All that shareholders own is shares. There is no need for the government to renationalise companies like Severn Trent. All it needs to do, and this applies to all corporations, is to change the UK laws on corporate governance so as to limit the rights of shareholders to those of any other creditor, such as a bond holder. In other words, they should have no right to elect directors to the Board of Directors, or to otherwise determine corporate policy. Company boards should be elected from amongst the managers and workers of the company itself, as it is the company that owns the productive-capital. The shareholders should simply be entitled to receive a competitively determined rate of interest/dividend on their shares.

If Labour followed that policy, it would have no need to expend any money to buy company shares, dividend yields on shares would be competitively determined in the market as with any other rate of interest, freeing up profits for reinvestment, and companies would be run democratically by their workers and managers.

No comments: