I have posted this thread to Twitter:
Simon Clarke MP, who is Chief Secretary to the Treasury claimed in several broadcast interviews this morning that increasing interest rates by 1% would increase the cost of government borrowing by £23bn a year. He's wrong. A quick thread...
According to the ONS Public Sector Net Debt on 31/12/21 was £2,340 billion pounds. So to come up with his figure it's pretty clear that Simon Clarke multiplied this by 1%. That's how complicated his fag pocket calculation was. But that number is wrong.
First, according to the Debt Management Office of HM Treasury at least 33% of all UK government debt is owned by the Bank of England on behalf of HM Treasury - so the government owes this money to itself. That cancels one-third of the debt and so the interest cost.
Second, UK debt is issued with fixed interest rates. The only amount of debt with varying rates outside the Bank of England and National Savings and Investments, both of whom operate bank deposit accounts, are inflation-linked bonds. But the rate on those varies with inflation, and not interest rates.
The average repayment period over which the interest of government debt is fixed is, again according to the HM Treasury Debt Management Office, near enough 15 years.
In other words, an interest rate change now would on average take 15 years to have an impact on the cost of government borrowing as rates are fixed. That covers all the national debt excluding National Savings and Investments and Bank of England deposit accounts that they hold for UK clearing banks.
Allowing for the fact that one-third of the national debt is owned by the Bank of England that's about a net £1,200 billion (or thereabouts) where an interest rate change will make almost no difference to government spending for a long time to come.
We're left then with changes of interest due on NS&I and on accounts held with the Bank of England. Together these come to about £1,100 billion. So at a push you could claim this will cost £11 billion in interest if the government allowed the interest rate in full to savers.
But even if they did, most of this would be paid to companies paying tax at 19% and savers maybe paying rates of tax higher than that. So, about £2 billion would be recovered in tax on this interest paid, at least.
In that case the maximum cost of a 1% increase in interest rates looks to be £9 billion.
Now, I don't know about you, but if I was an examiner looking for an answer to the question 'what would a 1% interest rate rise cost the government?' and the right answer was £9 and the answer given was £23 billion I'd say the student failed their exam.
The Chief Secretary told Kay Burley this morning that he thinks the prime minister does not lie. We all know he does. And I think Simon Clarke did not tell the truth about the cost of a change in interest rates on government finances this morning.
Thanks for reading this post.
You can share this post on social media of your choice by clicking these icons:
You can subscribe to this blog's daily email here.
And if you would like to support this blog you can, here:
Oh dear.
Around £114bn is held in premium Bonds at NS&I. No tax is paid on what premium Bonds pay out.
The vast majority of those saving with NS&I are smaller investors. They can receive up to £1,000 a year in interest without paying any tax. Even 40% taxpayers receive the first £500 interest tax free and with interest rates on a typical account paying as low as 0.2% interest you’d need £500,000 saved to breach the £1,000 interest figure if you were a basic rate taxpayer. I doubt many do.
If I was an examiner looking for an answer to the question ‘what would a 1% interest rate rise cost the government?’ and was given an answer that didn’t take these two factors into account, I’d say the student failed their exam.
I over-egged it…
So you are saying I should state it as £9bn less tax
I can live with that. My assumption was cautious, not reckless
And if you knew the national debt it is more complicated than this
Richard,
You have made several errors in your article.
“That cancels one-third of the debt and so the interest cost.”
This is your main one. Even if you claim that QE debt owned by the BoE is cancelled (which it isn’t, but that is a topic for another time) the money created to purchase that debt still sits on reserve accounts, so still carries interest.
If interest rates rise 1% the £885bn of cash created for QE will cost the government £8.85bn in extra interest payments because of it.
“Second, UK debt is issued with fixed interest rates.”
In the narrow sense you are correct. Fixed coupon bonds pay a fixed amount of interest depending on their coupon.
Broadly speaking though, this is totally immaterial because as interest rates rise the value of bonds will fall, pushing yields higher. As bonds mature and are replaced this pushes the net total cost to government for financing to that of yields, not coupons.
So it is not correct to say that it takes on average 15 years to have an impact on the cost of government borrowing. Given the scale of new issuance and redemption, higher rates feed through to higher interest rates almost immediately.
So for a back of envelope calculation, Simon Clarke’s £23bn number is probably not far wrong. If anything, it is likely to be an understatement – with the inflation linked bonds as well as things like NS&I the increase in government interest costs is likely to be closer to £40bn a year.
Your £9bn a year number is miles out though, for the reasons I have provided above.
a) I have allowed for the reserve accounts – read what I wrote
b) I have allowed for £9bn of extra interest
So I have covered what you claim]
You are wrong
I am right
As I am on it taking 15 years for rates to get through to the market
If you are in finance you really are very, very bad at it
Ah sorry – I missed what you had written about reserve accounts because your article was poorly laid out. That said, trying to claim that tax paid on those interest payments can be offset is wrong – the vast majority of it will not be taxable.
“As I am on it taking 15 years for rates to get through to the market”
This is most definitely wrong though.
Fixed coupons does not mean fixed value, and an average maturity of 15 years dos not mean it takes 15 years for the effect of higher interest rates to be felt.
Your logic would suggest that governments should issue long dated zero coupon bonds – and therefore never pay interest at all.
The problem is that in real life bonds are discount instruments, which means as yields go up the effective interest costs increase as well. You can see this in practice:
For year ending 2020 interest payments were £30.5bn. For year ending 2021 they were £50.5bn.
In that period almost all debt issued was bought by the BoE, which increased QE by £250bn. Those reserves were only paying 0.1% at the time, so the effect on total interest payments would be very small. Inflation linked bonds will have had a contribution, but because the index is lagged the big increases are yet to filter through.
The big effect though, is of course from the increase in nominal bond yields. At end 2020 the 10y Gilt yielded about 0.3%. End 2021 it was around 1.2%. Which accounts for most of the £20bn increase in interest payments over the course of the last year.
QED.
You can see how strong and fast the pass through to higher interest costs from higher yields is just from this.
“If you are in finance you really are very, very bad at it”
Apart from you being insulting, I’d have hoped an economics expert who claims to be a Professor would understand basic bond maths. Which you clearly don’t.
It was a Twitter thread …. I said so.
And you are wrong re tax – why are banks not taxable? Your claim makes no sense
Nor does that on average maturity make sense
Some of this debt is not redeemed until 2070 and is at fixed rate until then. How can I be wrong?
All I am saying is it will take a long time on average for the implication of any interest rate change to change real cost, materially – and that is a fact.
And your analysis is bizarre. £250 million was issued at a lower average cost than the amount redeemed. And the 2021/22 cost per the OBR will not exceed £40bn. The excess over 2020 is due to inflation.
To claim that 0.0% of £250 billion explains £20 billion is very odd indeed…
I get the maths
You clearly do not
And you have, I also note, used seven identities on this blog
“And you are wrong re tax – why are banks not taxable? Your claim makes no sense”
They are, but only on profits. Not on the amount of reserves they hold.
“Nor does that on average maturity make sense”
“Some of this debt is not redeemed until 2070 and is at fixed rate until then. How can I be wrong?”
Easily it turns out. Thanks to the NPV + the constant turnover of bonds thanks to new issuance and redemption, the yield to maturity is the best estimate of government interest costs.
Which stands to reason really. The owner of such a bond earns the YTM, and the government paying it so…..
“All I am saying is it will take a long time on average for the implication of any interest rate change to change real cost, materially – and that is a fact.”
As the evidence shows, not a fact.
“And your analysis is bizarre. £250 million was issued at a lower average cost than the amount redeemed.”
Tell me. If two bonds have the same maturity date, but one has a higher coupon, which has a higher yield to maturity?
“The excess over 2020 is due to inflation.”
OBR says about £4bn is due to inflation in the figures I gave in my last post. So where is the other £16bn coming from? By your logic, reserves/QE cancels debt so the amount of bonds outstanding has barely changed. Yet interest payments have increased by £20bn over the last year, of which £4bn is from the RPI bonds.
So where is that £16bn coming from?
“I get the maths. You clearly do not”
Not sure you do. I get the impression you still think of bonds as totally fixed value.
“And you have, I also note, used seven identities on this blog”
Well, I haven’t but what does it matter to the points I am making.
Oh dear…
The government does not pay NPVs
It pays fixed yields
You really are out of your depth, there is so much wrong in this
“You really are out of your depth, there is so much wrong in this”
OK. When you buy a bond as an investment, is your rate of return the coupons or the yield to maturity?
And if I own a bond and receive a rate of return, is not the case that the bond issuer is paying all those cash flows?
Should be a simple enough question for an economics professor to answer. Doubt you can.
Wow…
Do you not understand the viewpoint from government and market are different?
Talk about worm’s eye view
The government is not worried about your NPV. Its payment is fixed.
I think you need to leant a lot
But please don’t call again until you have
You are hilarious.
If the bondholder is receiving payments, they come from the government.
Because bonds yield move, the price of the bond is not fixed. Which means that even though the coupon is fixed the government might effectively be paying far more.
For example, a 10y bond with a 1% coupon will have a price of less than 100. Lets say 95. But will still pay 1% on that 100. So the effective interest rate the government is paying is higher than 1%.
because the NPV is how the yield and price of the bond are calculated.
I think you need to learn basic bond maths. To think an economist doesn’t understand this is just downright crazy. Are you really a professor in the field, or are you just some wannabe pretend expert using his corner of the internet to fluff his ego?
The market can vary the rate by changing the price of the bond
The cost to the government does not alter over the life of the bond. That was fixed on day one and is all that matters here
If you do not understand that then it is deeply worrying as you obviously think you do understand this
So @Henry, when Alistair Darling wrote to the Bank of England in Jan 2009
https://ypfsresourcelibrary.blob.core.windows.net/fcic/YPFS/Darling_2009.01.29_APF.pdf
“The Government will indemnify the Bank and the fund specially created
by the Bank of England to implement the facility from any losses arising out of
or in connection with the facility.” [The Asset Purchase Facility]
You seem to be asking us to believe that he actually meant that QE could never be cancelled?
Really? And when the Treasury in any case owns the Bank of England?
I wonder how much Lynton Crosby will be paid to help Johnson to win the next election?