Why the rush in selling the Floating Rate Notes?

Over the past 12 months €2.5 billion of bonds issued as part of the IBRC liquidation in February 2013 have been purchased by the National Treasury Management Agency from the Central Bank of Ireland.

• 22 December 2014: €500 million of the 2038 bond
• 26 June 2015: €500 million of the 2038 bond
• 18 August 2015: €500 million of the 2038 bond
• 23 October 2015: €500 million of the 2038 bond
• 21 December 2015: €500 million of the 2041 bond

The IBRC was the entity formed when the carcasses of Anglo Irish Bank and Irish Nationwide Building Society were merged in 2011.  A year earlier a massive creditor bailout was engineered using €31 billion of Promissory Notes.  Under the structure the money to repay the creditors of these failed institutions (mainly deposits) was borrowed from the Central Bank of Ireland under the guise of Emergency Liquidity Assistance (ELA).

Under the terms of the Promissory Notes this ELA was to repaid in tranches with payments of €3.1 billion to be made in March of each year beginning in 2011.  There was a lot of confusion about this structure but as shown by Prof. Karl Whelan it would have taken payments of €3.1 billion each year up to 2022 to repay the ELA used to bail out the creditors of Anglo and INBS availed of using the Promissory Notes.

Under the Promissory Note structure, the key interest rates were:

1. The ECB’s main refinancing rate as this was the external cost to the broad government sector of the funding accessed through ELA, and
2. The rate on new government borrowing as this became the financing cost as each annual payment of €3.1 billion was made.

There were many issues with the Promissory Note structure but chief among them was the pace at which relatively cheap funding through the Central Bank was to be converted to more expensive funding through the NTMA.   That is, each €3.1 billion payment meant that less debt was being financed at the lower ECB refinancing rate and more debt was being financed at the higher government borrowing rate.

The €3.1 billion payment was made in 2011 and was fudged in 2012.  In February 2013 the Promissory Note structure was cancelled and in its stead the Central Bank was given €25 billion of long-dated, floating-rate Irish government bonds and could use the proceeds from these to settle the remaining liabilities outstanding created when the ELA that was issued in 2010.

The new structure was an improvement on what preceded it.  The main reason for these was that the pace at which debt was moved from the ECB refinancing rate to the government borrowing rate was slowed down.  Under the Promissory Note structure the debt spent an average of c.7 years being financed at the ECB rate.  Under the initial terms of the Long-Term Bond arrangement this was increased to c.15 years. 

And as the ECB rate has since reduced to 0.05% the external cost to the State of the debt created under the original Promissory Notes is relatively miniscule (0.05% of €25 billion is €12.5 million).

When the new structure was put in place it was said that:

The Central Bank of Ireland will sell the bonds but only where such a sale is not disruptive to financial stability. They have however undertaken that a minimum of bonds will be sold in accordance with the following schedule: to end 2014 (€0.5bn), 2015-2018 (€0.5bn p.a.), 2019-2023 (€1bn p.a.), 2024 and after (€2bn p.a.).

These terms were used to estimate the interest savings that the new structure would generate.  Instead of having the debt financed at the ECB rate paid off by 2022 (and replaced by higher costing debt funded by borrowing undertaken by the NTMA) this was not supposed to happen until 2032 when the last of the long-term bonds was to be sold by the Central Bank.

Under the original terms, a minimum of  €1 billion of the bonds were to be sold by the end of 2015.  The only update we get from the Central Bank on the bonds is in its annual report published each May.  However, as the press release linked above show the NTMA has bought €2.5 billion of the bonds in the past 12 months and subsequently cancelled them, with €2 billion in 2015 alone.  The Central Bank may have sold more of them this year but that is probably unlikely.

From the Central Bank’s 2014 annual report we do know that the only sale in 2014 was the €0.5 billion bought by the NTMA.  This is extracted from Note 15 to the report.

During 2014 the Central Bank’s holding of the bonds reduced from €25.0 billion to €24.5 billion because a €0.5 billion sale. The same note tells us that:

During 2014, the Bank sold €500 million nominal of the FRNs (2038 FRN) realising gains amounting to €180.3 million.

The price paid for the bonds was around €680 million, with €500 million going to reduce the Central Bank’s liabilities to the Eurosystem (the process that saw the money “created” in 2010 is reversed with the money now “burned”) while most of the €180 million will be returned to The Exchequer as part of the Central Bank Surplus.

We know that the NTMA bought a further €2 billion of these bonds in 2015 but we don’t yet know the price paid.   From the table above we can see that, at the end 2014, the Central Bank valued the €24.5 billion of the bonds it held at €33.6 billion. 

Given that the yields on Irish government bonds were relatively stable in 2015 it would seem that the price paid by the NTMA for the €2 billion of the bonds it bought this year was probably around €2.7 billion.  Thus, there will have been a further €2 billion reduction in the Eurosystem liabilities of the Central Bank while most of the other €700 million will return to the Exchequer.

But the key question is: if only €1 billion of these bonds had to be sold by Central Bank by the of this year why have €2.5 billion of them being sold – and why has the NTMA bought them?

There are a couple of ways to approach this.  First, let’s also not that in 2014 the Central Bank sold €2.3 billion of the 2025 government bond that was created as part of the fudged 2012 Promissory Note payment.  Going back to Note 15 in the 2014 Annual Report of the Central Bank we are told:

In 2013 the Bank acquired €3.5 billion nominal of the 5.4% Irish 2025 Government Bond following the IBRC liquidation. During 2014, the Bank sold €2,300 million (2013: €350 million) nominal of the bond realising gains amounting to €537.6 million (2013: €24.8 million). As at 31 December 2014, the 5.4% Irish 2025 Government Bond was valued at €1.2 billion giving rise to an unrealised gain of €290.5 million (2013: €292.7 million) as at that date.

The Central Bank received around €2.8 billion for selling €2.3 billion of the bond.  But why didn’t the NTMA buy this bond as it has done with the others.  The reason is that this is a fixed coupon bond.  Although the annual coupon rate is 5.4 per cent the yield for the purchaser was much lower (probably around 1 per cent) as the purchase price was significantly greater than the face value. 

There was little to be gained by the NTMA buying this bond.  Most of the 5.4 per cent coupon will be covered from the gain realised by the Central Bank by selling to a third party. 

So why is the NTMA willing to buy the other bonds? Because they are Floating Rate Notes.  Unlike the 2025 bond above there is an element of interest rate risk associated with these bonds.  The Central Bank could sell these bonds to a third party but the NTMA risks paying higher interest to this third party in the future if interest rates rise.

As this useful note tell us:

The FRNs pay interest at a rate equal to the 6-month euribor rate plus a fixed spread which differs between the bonds of different maturities but has a weighted average of 2.62%.

Here is the history of the 6-month euribor:

It can be seen how exceptionally low the current rates are.  But if the rate was to rise to say 3 per cent then the interest cost on the floating rate notes would rise to an average of 5.6 per cent and would rise even more with further increases in the 6-month euribor.

However, all this tells us is that if the Central Bank is to sell the bonds then it is probably in our interest for the NTMA to purchase them (unless you think the 6-month euribor will stay below 1 per cent for the next decade).  It doesn’t answer the key question as to why the bonds are being sold ahead of the agreed schedule. 

In its own 2014 report the NTMA said:

Following discussions with the Central Bank of Ireland, €500 million nominal of the 2038 bond was bought back and cancelled by the NTMA in December 2014.

It could be that the accelerated sales are at the behest of the NTMA rather than the Central Bank (which faces pressure from the ECB to sell the bonds – monetary financing and all that).  Of course, we don’t really know but we do know that the current low borrowing costs for the government can be used to justify the sales. 

Back in September the NTMA issued 15-year debt at a yield of 1.8%.  If the ECB rate was to rise above 1.8% then the NTMA buying the bonds could lead to savings as the 1.8% financing cost locked in now would be below the financing cost if the Central Bank still held the bonds in circumstances where the ECB rate was higher.

And it is this justification that is offered by the Central Bank:

What implications do such sales have for the overall position of State’s finances?

One way of thinking about the Exchequer’s position in this regard is to consider the disposal by the Central Bank of FRNs to the NTMA. For example, in December 2014, the NTMA bought €500m (nominal) of the 2038 FRN from the Central Bank for a cash price of €680m. The €680m was, of course, funded by the NTMA in the market, at the current low interest rates, but the additional €180m has augmented the 2014 profit of the Central Bank, the bulk of which will be distributed back to the Exchequer as surplus income. So the Exchequer is now servicing the €500m at current market yields with the bulk of the €180m being returned to it in cash. Thus, from the point of view of the Government's budget, the disposal can be seen as a refinancing of part of Government debt at current market rates, which are much lower than the rates which prevailed when the FRNs were issued.

As long as the Central Bank is the holder of the FRNs, and as long as the effective cost of funds to the Central Bank (in practice, the ECB Main Refinancing Rate or ‘MRO’ rate, currently 0.05 per cent.) is below the coupon on the bonds, there will be some net interest contribution to the Central Bank’s profits. Since most of the Central Bank’s profits are remitted as surplus income to the Exchequer, it might appear that a relatively slow pace of disposal would benefit the Exchequer. The eventual impact will, of course, depend on a range of factors, including future interest rates, which cannot be predicted with accuracy.

Nevertheless, when interest rates on new issues of Government debt were much higher than they are today, refinancing would have seemed relatively expensive and this would have supported the idea that slower sales might be beneficial to the Exchequer. However, the current low cost of refinancing by the NTMA, though still somewhat above the Central Bank's effective cost of funds, reflects the exceptionally low level of interest rates and it could, therefore, be below the cost of the same refinancing at some point in the future. As a result, any small short term gain to the Exchequer resulting from slower refinancing could be more than offset by a higher cost of refinancing in the future.

The same point was made here.  But the reality is we simply can’t tell with any certainty what, if any, the benefits would be.

The best situation would be if the debt didn’t exist at all. Next best would be for the Central Bank to hold the bonds until maturity ensuring that the net cost remains at the ECB refinancing rate which will always be below the government’s borrowing cost. 

But because the bonds are to be sold we are in a grey area as interest rates and the government’s borrowing cost are likely to be higher in the future.  We could save by refinancing the debt at the lower rates that currently prevail.  This is a justification for the accelerated sales but there are questions that remain:

1. If there are benefits from doing this now why move €1.5 billion ahead of the minimum sale schedule for the end of 2015? Why not more?
2. Will the accelerated pace of sales continue in 2016?
3. Why are the shortest dated of the bonds (the 2038s) being cancelled first rather than the longest dated (the 2053s) or a combination of bonds?
4. If there are accelerated sales now does that mean there can be a sales “holiday” in the future if there is a period of high interest rates and/or an abnormally high spread between the 6-month euribor and the ECB’s main refinancing rate?
5. What pace of interest rate increases are needed for a breakeven point of these transactions to be reached? And what profit is possible if interest rates go above that? And,
6. If the Promissory Notes still existed would we be making accelerated payments on them?!?

Tweet

Balance of Payments Mismatches

In 2014 Ireland imported €4.8 billion euro of services from Bermuda with the Netherlands importing €13.9 billion of services from Bermuda in the same period.  In 2014 Ireland and the Netherlands accounted for 86 per cent of the EU’s service imports from Bermuda.  Luxembourg had imports of €1.6 billion and the UK figure was €0.4 billion.  Total service imports from Bermuda for the remaining 24 countries of the EU was €1.0 billion.  This is a total of €21 billion of inflows to Bermuda from EU countries under the guise of services.

Bermuda is a small island in the North Atlantic. It is a British Overseas Territory and has a population of around 70,000.  It has a GDP of around US$5 billion which gives a pretty impressive per capita figure of around US$80,000.

But a quick question emerges.  If Bermuda is receiving nearly €18 billion in exports to just two countries (Ireland and the Netherlands) why isn’t this showing up in value-added in the Bermudan GDP figures?  Bermudan GDP is high but it should be off the charts.  Are there offsetting imports to Bermuda?

Here is a table from Bermuda’s Department of Statistics giving the payments and receipts of the Current Account.

In the receipts panel we can see that total service receipts are BD$1,377 million. The Bermudan dollar is on a one-for-one peg with the US dollar.

So at one end we have €18 billion of payments flowing out of Ireland and the Netherlands with the data at that end showing this as flowing to Bermuda.  At the other end we have no sign of this being reported as an inflow in Bermuda.  It seems there is a Bermuda triangle in the Balance of Payments.

Is it difficult to know what is going on.  It could be due to coverage with the Bermudan authorities simply choosing to ignore these flows on the basis that the services receipts would be fully offset by income outflows.  The only significant income flow in the table above is receipts of Employee Compensation.

It could be a definitional issue with only ‘local’ companies included.  The Bermuda statistics release includes the following:

Resident: The concept of residency is very important in the BOP because the BOP is in fact a statement of transactions between residents and non-residents. A unit must have a centre of predominant economic interest within an economic territory for at least one year or more to be considered resident of that territory.

It could be that the entities receiving the payments from Ireland and the Netherlands are not considered residents of Bermuda.  The infamous double-irish scheme devised to avail of the ‘same-country exemption’ in US tax law is based on two Irish-incorporated companies – one which is resident in Ireland and one which is not.

A key feature is that the second company is deemed non-resident in Ireland as the test of management and control rather than the test of incorporation is applied.  Of course, there is nothing to say that territory in which management and control is exercised considers the company resident there.  What matters is that the company is non-resident in Ireland and, of course, residency or otherwise doesn’t really matter in Bermuda as the rate of corporate income tax is nil. 

It is likely the case that the payments are being made to entities that are essentially ‘stateless’.  Thus we have a situation where companies are reporting outgoing payments in Ireland and the Netherlands to Bermuda but are not having the equivalent inbound payments recorded in Bermuda.  And these are all likely to be intra-group transactions. 

Tweet

Ireland’s net external debt falls to just €23 billion (sort of)

Today’s International Investment Position and External Debt release from the CSO shows that Ireland’s net external debt (excluding the impact of the IFSC) has fallen to just €23 billion from €180 billion at the start of 2012.

This would be a remarkable improvement if it was true but as with lots of aggregate statistics on the Irish economy it is true, sort of. 

The first issue that needs a second glance is our gross external debt.  This actually rose €60 billion in the year to Q2 2015 before falling back slightly in Q3.  Here is the sectoral breakdown of our gross external debt (again excluding the impact of the IFSC as all the charts here will).

Some components of our gross external debt are coming down in particular through the monetary authority (central bank) and, to a lesser exent, monetary financial institutions.  Both of these are functions of the deleveraging of the domestic banking system. 

The sector going in the other direction is direct investment debt, the vast majority of which is likely originating from multinationals – both MNCs which have subsidiaries here and MNCs which have re-domiciled here.

This FDI effect is equally seen if we look at external assets in debt instruments with those associated with direct investment increasing very rapidly over the past two and half years.

If we subtract the external assets in debt instruments from the gross external debt we can get the net external debt to see to contribution of each sector to the aggregate totals in the first chart above.

This shows that three of the five sectors have net external debt positions of around zero.  The central bank has rapidly moved there while the banks and other sectors (financial intermediaries, pension funds etc.) have moved into small creditor positions (they are owed more debt from the rest of the world than they owe).

The two diverging sectors are the government sector (which borrowed hugely from abroad between 2009 and 2012) and the debt associated with direct investment.  We can see that there are far more external assets in debt instruments than there are gross external debt liabilities linked to direct investment.  This means companies here are owed much more from abroad than they owe externally themselves.  Again this is likely to be an MNC effect.

The impact of this is that our gross external debt is likely to be overstated (because many of the debt liabilities owed from Ireland originate from MNCs) and our net external debt is likely to be understated (because there is an even greater amount of external debt owed to MNC subsidiaries or headquarters in Ireland). 

So if we strip out FDI from the first chart we get the following:

This probably better reflects our position.  There still is improvement with both the gross and net external debt figures falling over the past four years.  Excluding direct investment (and the IFSC, of course) Ireland’s net external debt is €101 billion – with by far the largest contributor (only?) to that being the government sector.

Yes, this is a higher figure than the €23 billion figure we started with above but it probably isn’t too bad all things considering.  As a percentage of GDP it comes in at under 50 per cent – not exactly a headline grabbing number.

And if we look at the broader net international investment position of all financial assets we can see where this improvement has gotten us (with a negative figure indicating a net liability position).

It won’t be long before the explanation will be that a positive figure indicates a net asset position.  In fact, if we look at the sectoral positions we can see the components of the overall net position.

So if our pension funds (a large part of financial intermediaries) can earn more on their investments abroad than the government pays out in external interest costs we may even be able to turn a bob or bob from all this.

Tweet

Why isn’t Compensation of Employees increasing in the National Accounts?

One of the notable features of the national accounts in recent quarters has been the continued stagnation of compensation of employees.

Compensation of employees reached a peak of €21 billion per quarter at the end of 2008 but since the start of 2010 it has hovered within a few hundred million of €17.7 billion, apart from Q4 2013 when it was €18.7 billion from which it quickly fell away.  The most recent data is for Q2 2015 when it was €18.1 billion.

The pattern of compensation of employees seems to run counter to other indicators in the economy.  From the QNHS we know that employment is growing, the ELC shows that average weekly earnings are rising and the Exchequer Returns show rising receipts from Income Tax.

The following chart shows indices of total employment and compensation of employees with both set to 100 when employment bottomed out in early 2012.  To smooth out some of the volatility four-quarter moving averages are used.

The divergence during 2014 is evident.  The final chart shows the source sectors for compensation of employees which, in line with the fist chart above, shows very little happening in recent years.

Shouldn’t the navy part, at least, be showing some improvement? Tweet

Legal Proceedings and Repossessions in the Central Bank’s Mortgage Arrears Statistics

The release of the Q3 2015 update of the Central Bank’s arrears statistics garnered some attention for the continued reduction in the overall level of arrears.  The release also contains information on legal proceedings issued and concluded by lenders.

First we see that the number of legal proceedings issued declined in 2015.

The series jumped hugely after the lacuna in the law introduced by the Land and Conveyancing Law Reform Act 2009 was corrected by the Land and Conveyancing Law Reform Act 2013.  We can see the peak of over 3,000 legal proceedings issued per quarter was reached in the first half of 2014 while the figure for the most recent quarter is around 50 per cent of the peak.  Almost 22,000 legal proceedings for PDH properties have been issued since the middle of 2013.

Perhaps surprisingly, there has also been a slowdown in the number of court proceedings concluded.

The number of court orders granted for PDHs in the first three quarters of 2015 was 1,314.  Since the start of 2014, just over 5,000 legal proceedings have concluded and 2,300 (45 per cent) have concluded with a court order.  This tallies with observations of the proceedings in the Cork Circuit Court that more legal proceedings conclude by other means (mainly strike outs) than conclude with the granting of a possession order.

The outcome that is likely to be of most interest is the actual number of repossessions and this has increased significantly in 2015.

So far in 2015 there have been 564 court-ordered repossessions of PDHs carried out on behalf of lenders.  That is around 14 a week.  The figures suggest that around 40 per cent of possession orders are executed to give a court-ordered repossession.  It would be useful to know how many of these were vacant at the time the order was granted or were let out even though the original loan was for a PDH.


There are also occasions when lenders do not have to rely on a court order or executing a court order to take possession of a property.

While the number of court-ordered repossessions has increased in 2015 the number of voluntary surrenders has declined slightly.  In the first three quarters of 2014 there were 692 voluntary surrenders and this fell to 631 in 2015.   The statistics do not give the number of forced sales that occur though collecting data on “assisted voluntary sales” should be possible.

There are around 17,000 possession proceedings before the courts at the moment.  In rough terms, the aggregate data to date indicate that of these, around 45 per cent will conclude with the granting of a court order for possession and, of which again, around 40 per cent will lead to a court-ordered repossession.  That suggests there are around 7,500 orders for possession to come out of the courts with around 3,000 court-ordered repossessions following from those. Tweet