Large-scale sales of public assets by the Greek government are the latest straw to which policymakers are trying to clutch, as the Greek debt saga threatens to drag down the European economy. However, privatisation is a mirage.
Privatisation can help when the problem is one of liquidity. If the problem is one of solvency, privatisation will only make matters worse, especially if it has to be done at distressed prices (Manasse 2011).
The government of Greece has officially promised to raise €50 billion (close to 20% of GDP) by 2015 through sales of public assets and privatisation receipts of €30 billion are the cornerstone of the second Greek rescue package from the EU/IMF/ECB that is being finalised these days. It is highly unlikely that the full sum will be realised, but the real question is whether privatisation is in principle a useful way to reduce public debt. The short answer is no.
That privatisation cannot solve a solvency problems should be clear from first principles. By selling assets the government can obtain funds to reduce its debt service burden, but it also loses future revenues. A priori the two should cancel out each other without a net gain or loss. But large-scale privatisation to reduce short-term debt (by paying maturing bonds in full) will actually increase the risk premium on longer-term debt.
When the government sells assets which would otherwise be at the disposal of Greece’s creditors in case of default the value of the remaining privately held (long term) public debt must surely decline because in the absence of default the upside for creditors is limited (by the face value) whereas the downside will become worse in case of default.
Privatisation thus devalues the value of the existing claims of long-term creditors.
On top of this we have to consider at what price Greek assets could be sold when the country is on the brink of bankruptcy. At present the market already discounts heavily the value of an unconditional promise of the government to repay a certain nominal amount in euro. Any investor is thus likely to discount even more the implicit promise not to undertake any measure to reduce the value of the assets the Greek government has to sell under duress. The price at which assets can be sold will reflect at least the same discount as the price of existing government bonds, which now average close to 50% to the face value for longer term bonds.
Moreover, when a government has to default it will be tempted to use regulatory and other measures to reclaim part of the returns from the assets it has previously sold. For example, it could lower regulated prices for utilities, port and airport fees, etc... This “time-inconsistency” problem implies that it would make more sense to privatise once the debt overhang has been dealt with (preferably through a market-based restructuring (see Gros and Mayer 2011).
Thinking about Greece as a company which is technically insolvent (i.e. that cannot pay its bills on time without official help) helps. If this company sells some of its assets to satisfy some privileged creditors who are closely linked to senior management, the other creditors, especially long-term bond holders, will surely be worse off. This is why bond covenants forbid this kind of operation.
There is no free lunch when a country has a solvency problem. Asset sales only reshuffle claims among different creditor groups. Any attempt by the official sector to reduce its own risk must ultimately increase the risk to private lenders and thus make market access more difficult. The simple model in the annex shows this formally.
The economic intuition behind this result can also be seen by a comparison with official lending which is senior to existing private creditors. Privatisation establishes a sort of de facto seniority.
Proponents of privatisation argue of course that privatisation will increase growth because private investors would put these assets to a more productive use. But most of the assets slated for the first round concern utilities and other natural monopolies. Transferring them into private hands might increase monopoly profits, but not growth. But even if this were not the case; if one could expect private monopolists to be much more productive and competitive, it is still the case that privatisation is highly unlikely to ignite growth on a significant scale. The existing capital stock in Greece (without land) is estimate to be worth around €600 billion. The full privatisation programme of €50 billion would thus not even affect 10% of the national capital stock.
It is interesting that in the case of Argentina very large scale privatisation programmes were implemented in the context of various reform programmes even years prior to the onset of the acute crisis. However, they could not stave of default because the government had lost the revenues from the assets which had been sold previously.
All in all it is thus likely that the risk premium on longer-term Greek government bonds will actually increase as the government privatises its most liquid assets, which are also those which yield solid revenues during normal times. In the end privatisation might make it actually even less likely that the Greek government will ever be able to return to the capital markets.
Gros, Daniel (2010) “The seniority conundrum: Bail out countries but bail in private, short-term creditors?”, VoxEU.org, 5 December.
Gros, Daniel and Thomas Mayer (2011), “Debt reduction without default?”, VoxEU.org, 11 February.
Manasse, Paolo (2011), “Why privatisation is not the panacea for Greece”, VoxEU.org, 30 May.
The following illustrative model makes the general point more explicit.
The model has a country with a (foreign) debt equal to one. Assume the future of the country in question is not certain. With probability (1-π) the sun will shine and the country will be able to repay its creditors in full in the next period. For simplicity we can assume that in this good state of the world the resources available to the government (and thus for creditors) are equal to (1+ρ), where ρ indicates the return on some government asset (stakes in state-owned enterprises, ports, airports, etc…). However, with probability π the country will encounter a bad state of the world in which the sum available for creditors will be only μ (some maximum tax revenues determined by the political system) plus the returns from the assets of the government, ρ.
If we consider the case of a country which privatises because it has financing difficulties we have to assume that μ+ρ <1, which is less than the full face amount of the debt. Clearly under the bad state of the world the country defaults.
The price (value) of the private long-term debt (as a proportion of its face value) will then be given by:
Where (1-( +ρ))>0 represents the hair cut private creditors have to accept in case of default.
Assume now that the government sells the assets which yield ρ. The proceeds from the official financing are used to reduce its short-term debt burden (i.e. fewer loans from its euro area partners). At what price can these assets be sold? We could argue that the returns from these assets are not directly affected by a default so that the expected return should not be discounted. However, this is not likely to be the case in reality since a default would certainly have a negative effect on the economy, thus reducing the returns from any real assets located in the country. Moreover, when a government has to default it will be tempted to use regulatory and other measures to reclaim part of the returns from the assets it has previously sold. For example, it could lower regulated prices for utilities, port and airport fees, etc... If the market discounts the value of an unconditional promise of the government to repay a certain nominal amount in euro it is likely to discount even more the implicit promise not to undertake any measure to reduce the value of the assets it has to sell under duress. This is why it is likely that the price at which the assets will be sold will reflect the same discount as the price of existing government bonds (given by PV in equation (1) above). Privatisation receipts will thus be equal to ρPV.
This implies that once privatisation is completed (it takes place within the current period) the outstanding amount of debt will be given by: (1- ρPV). (As shown below, the main result would not be affected even if the assets could be sold without a discount.)
After privatisation the amount available for foreign creditors of the country is only equal to µ if the bad state materialises. Keeping in mind that the (private) debt outstanding is reduced to (1-ρPV); this implies that the value of debt (as a proportion of its face value) is given by:
The impact of privatisation on the risk premium can be calculated by comparing the price after privatisation to the price before (given by equation (1)).
The difference between the two cases is given by:
This can be simplified:
This expression is clearly positive (under the maintained hypothesis that in the bad state total government revenues are insufficient to fully service the initial debt burden). It is also apparent that the drop in the price of long-term bonds is proportional to the probability of default, π. If it is close to zero privatisation has only a negligible impact.
Notice that in equation (6) the sign would not change even if the assets could be sold without a discount (in the model: if the debt burden could be reduced by ρ, instead of ρPV). The underlying reason is quite simple. In the good state creditors will at any rate obtain at most the face value of their claims. Privatisation cannot thus increase the upside for creditors in the good state, but privatisation will lower the downside in case the bad state materialises.
The general message from this little model is clear. A large privatisation programme (larger ρ relative to µ) will, ceteris paribus, increase the difference in the price, i.e. lower the price of existing long-term debt, or, equivalently increase the risk premium demanded by the private creditors.