FX Intervention QnA

Last week’s FX related posts (1, 2, and 3) raised a lot of questions in the Comments section. Here are my answers to some of these questions:

1. The RBA has intervened many times, and did so during the financial crisis to arrest the fall in the AUD when it reached USD$0.60. Was that a mistake?

The canonical position on currency intervention is the ‘profit test’. It was advocated by Friedman in The Case for Flexible Exchange Rates (1953), and used by the RBA to assess their own interventions in their 2004 and 1994 papers.

The 2008 intervention was extremely profitable, so with the benefit of hindsight it seems that their judgement was very good. At the time I did not support their intervention, as I felt that the outlook was very bad, and that the economy needed both lower rates and a weaker currency. As it turns out, I was too pessimistic.

2. If the RBA sells the AUD high, and buys low, doesn’t it actually make a profit?

The profitability of an intervention depends on both the change in the foreign exchange value of the AUD and the cost of holding that position (the carry). Historically, carry has been about half the total return.

3. Do you expect the AUD to stay at this level or higher for an extended period (say >5 years?)

Yes, I do expect the AUD to remain at an historically high level for the forseeable future. We have a good export outlook, and our strong aggregate balance sheet, and high return on capital, make Australia an attractive investment destination. I think this is a medium term change. Additionally, now that our trading partners are faster inflating nations, I expect ongoing nominal effective exchange rate appreciation over the medium term – as it will be required to stabilise the real effective exchange rate.

4. You confuse money costs with opportunity cost. The balance sheet of the RBA, and hence the government must balance, but like money itself, it is a mere tool for implementing policy goals. The costs involved, in terms of actual resources and opportunity cost, arise from any inflation – which is at historic lows.

These matters are clear in my mind. I agree that profitability is not the sole basis on which a central bank ought to act, and i agree that the way they tax us to fund these operations is typically via inflation (which includes the AUD being weaker than it otherwise would be). I see the RBA’s target (2.5% inflation over the cycle) as being the key reason that intervening to weaken the AUD is poor policy – they will lose, as the FX market is bigger than they are, and the FX target is subordinate to the inflation target. If inflation is too low, they ought to cut the cash rate – not play bingo in the FX market.

5. Which risk is greater – intervening in FX markets, or selling off more domestic assets?

Intervening in FX markets is the greater risk. Capital markets are clearly imperfect, however i am more confident that the markets can price our assets than I am that the RBA (or anyone else) can identify the appropriate level for the foreign exchange value of the AUD. Also, I think that history is very clear that a policy of FX intervention bleeds into a policy of distorting trade and assistance — I strongly believe that we have been better served by freer markets. This is my main argument against intervention: I conceded that it may be optimal to intervene in the presence of a portfolio shift, but I do not believe that we will be about to keep the political classes from further tinkering.

In any case, most of the flow just now is bond buying — are you unhappy that they are lending us all that money so cheap? Would you sooner have that duration risk in your portfolio?

6. If the AUD was USD$2.00 would you still be happy with the “market” outcome? Or are we just arguing about the degree of market failure?

IF AUDUSD was 2.00 right now, and all other things were the same, i would probably say that the AUDUSD was fair at this level — as we would have low unemployment, around-trend GDP growth, and around-target inflation. If the AUD raced from ~1.05 to ~1.50 over the next few weeks, I’d be calling for rapid rate cuts. Once we were at 0bps, I’d be okay with FX intervention.

7. if FX intervention is so costly, why is everyone (Swiss, China, US etc) trying to devalue their currency? Surely external account imbalances need to be considered.

The answers are different for each nation. The Swiss have deflation and have no other tools; the Chinese were subsidising their export industry; the US is okay with USD weakness given the inflation and unemployment outlook, but i think they have been trying to deal with a destructive USD squeeze which had the unpleasant side effect of increasing the foreign exchange value of the USD. I do not agree that capital flowing toward our superior risk adjusted returns is undesirable.

8. If the RBA makes a loss, can they write-off the debts to themselves? Is there a law/regulation preventing this?

The RBA does not need positive capital to operate, however most academics and central bankers believe that positive capital is desirable as it improves the independence and hence credibility of monetary policy (this ECB paper contains a review). I believe that we’ll have better aggregate economic performance if we have an independent and credible central bank that is credible and focused on fighting inflation.

If the RBA wanted to, it could print AUD to ‘pay’ its liabilities by issuing more high-powered money – and it could do this without limit. Who pays when the central bank does this? Everyone who presently holds AUD — via inflation.

Ultimately, this is a ‘quasi-fiscal’ operation, and i believe that the costs (and benefits) should be made explicit by passing it directly over to the taxpayer. How to do this?

The Treasury could issue additional bonds, put the funds on deposit with the RBA, and then ask the RBA to enter the market to purchase them some foreign exchange. From the perspective of the aggregate government balance sheet, this is the same as the RBA creating the reserves, selling them for foreign exchange, and then ‘sterilizing’ the additional AUD liquidity by lending the money back from the banking sector via other open market operations.

There seems to be a bit of controversy about who pays, so it’s worth noting that, on Friday, at the house, RBA Gov Stevens confirmed that the losses accrue to the taxpayer:

… the Reserve Bank has a very large open foreign currency position which has affected our earnings, as you know. It is our job to accept that. For policy reasons we cannot manage that away and we would not seek to, but they are not inconsiderable risks that the taxpayer would effectively take on with extremely large-scale intervention. In extremis you could make arguments that that should be done. But it is a big call and it is not a call that we have felt should be made to this point in time.

9. Are financial losses irrelevant for a monopoly currency issuer?

No, i do not think so. The open market operations add to the supply of the AUD, and an increase in the supply of AUDs will — all things equal — make their value lower than it would otherwise have been. The answer to Q8 (above) contains an example that shows the equivalence between intervention that’s pre funded by the Treasury, and intervention that’s post funded via sterilisation.

The Qantas / Frequent Flyer example given in the comments section is a perfect example case – if Qantas gave everyone 1000 points for every point they currently have, but didn’t change the point-price of their flights, they’d pretty soon find themselves with no excess capacity, and they’d have to put up their prices.

That’s exactly what would happen if the central bank tried to lower the currency via loss making interventions, and simply kept at it until the nominal value of the exchange rate fell. We’d have inflation.

The losses are the taxpayers – we can pay them explicitly by putting them on the balance sheet, or implicitly via inflation.

10. Can’t the central bank ‘always win’ if it’s a monopoly issuer of the currency? They cannot run out of AUD to sell, just like how Qantas cannot run out of frequent flyer miles!

Sure, they can put the nominal value of the exchange rate where ever they’d like it, if they are willing to let their inflation goals go (as well as any other general macro-economic performance aims).

However, at this point we need to remember why we were doing this — presumably we wanted a competitiveness boost, as we felt that the REAL effective exchange rate was too high. If the RBA shoves the nominal value of the AUD down by using their ‘monopoly privilege’, there is no guarantee that the real effective exchange rate will decline. If the appreciation the market wanted to push onto was appropriate, we’ll just take the higher real effective exchange rate as higher domestic inflation.

My guess is that we’d have a period of stag-flation if the RBA moved the nominal exchange rate lower by causing a domestic inflation.

11. Does the RBA mark to market?

It seems from their annual report that they do mark to market. I happen to think that all central banks should use mark-to-market accounting and a VAR framework — like investment banks and hedge funds.

The larger their balance sheets, the more important this is — for while they should not seek to maximise their own narrow profits (they should stabilise aggregate inflation), we should at all times be able to see the explicit costs and benefits of their activities. Good decision-making requires good information.

12. Will the SNB regret their EURCHF floor of 1.20? If so, why?

Yes, i do think that the SNB will regret their policy of putting a floor under EURCHF at 1.20.

The private sector is running away from financial claims on European Sovereigns, and the SNB is acquiring, credit risk, currency risk and duration risk at probably the worst imaginable time.

If the European project has a future it’s at a much lower exchange rate (as this will allow the broke periphery to close their twin deficits via increased net exports). Partly due to the fact that their lower inflation is causing the real effective exchange rate to decline just now, the Swiss economy does not require a much lower nominal exchange rate – nor is it likely to do so. If anything, Switzerland going to require a stronger exchange rate at some point, as the REER will decline to a ‘normal level’ in due course if present trends continue.

In all probability, equilibrium for Europe is a much lower currency and a higher rate of inflation — which also means higher interest rates. When rates rise, the SNB will make a massive loss on their duration.

So long as they hold the floor until the bonds mature, they can ‘resist’ this mark to market loss — however at just this point they are likely to see higher inflation at home, due to their lower exchange rate.

At a minimum they are likely to want to begin to sterilise some of that excess liquidity as inflation picks up — however as the rise in rates will have reduced the value of their portfolio, they will not have the assets required to collateralise the reverses required to drain the liquidity and therefore achieve their inflation goals.

The SNB may then use unsecured borrowing from the banking system (a deposit facility) or administrative measures (a tax on the banking system) to sterilise. In the former case, the interest costs could be funded by further money creation — however at this point we are deep into Ponzi-waters.

Most likely the Swiss Treasury would re-capitalise the SNB — probably by gifting the SNB interest bearing debt that’s sufficient to sterilise the monetary base (or selling it directly into the market themselves and putting the cash on deposit at the SNB).

In the medium term, this is a very risky policy. The Swiss economy is presently doing okay — partly because it has a low cost of capital, as it is a safe and stable place to invest. The SNB’s current policy creates a credible medium term risk to that prosperity: does a central bank really have a mandate to do that? I do not think so — it is a task for elected officials.

13. Is the market wrong?

I do not believe that the market always gets it right, however I do believe that markets are efficient, in the sense that there is no systematic way to make a profit. Thus, i do not agree that it’s so obvious that the AUD is over-valued. The ABS reports that international investment position in Australia is ~880bn as of Q1’12 – which is split between ~140bn of equity and ~740bn of debt.

The foreign sector owns ~170bn of our bonds, and if the other ~710bn of money invested in Australia felt that the AUD was over-cooked, they’d sell out and we’d pretty soon have a weaker AUD.

That’s not happening — which tells me that the weight of money does not think that the AUD is over-cooked. For that matter, neither does the RBA — the following is RBA Gov Stevens’ answers to AUD related questions in the QnA following his appearance at the House of Reps Standing Committee on Economics on 24 August 2012:

the fact that it is high does not mean on its face mean that you can say, ‘That must be a distortion.’ It could be, but let us put it this way: in the internal work that I have asked the staff for, I say: ‘We have had various comradeship relationships over the years. How does what is in the market now compare with what would be predicted?’ For some of those it is higher than would be predicted. There are one or two models you can come up with that say it is quite undervalued. I personally would not draw that conclusion but, as with any model, you can find various answers. I think it is probably the case that on balance it looks high relative to the mean prediction of models, though the size of that gap is not especially large in comparison to other forecasting errors or ‘misalignments’ we have seen in the past. So it is a bit on the high side, I would say, but probably not dramatically so …

… I would describe it as probably at present trading a bit above where I would have thought it would be, based on these past relationships and my instinct of the sorts of things that have been happening in, say, the past six months—terms of trade have declined, global growth outlook is not catastrophically bad but it is softer than it was and so on. I am probably, to be honest, a little surprised that it is not a little bit lower than it is. But we are not talking 20c worth or something—I do not want to give figures. I would say it is probably a little on the high side, but in terms of the statistical relationships you can fit you cannot actually say that error is all that significant. It is just my instinct that suggests that it is a bit on the high side …

… the big strategic question is whether you think that there is fundamental change in global relative prices, a lot of which is going to be persistent—not all of it but a fair bit of it—then we will have a larger resource sector and some other sectors will be smaller to be at full employment. If that is the case then you would think that the mean exchange rate we used to see will not be seen any more and instead on average we will see something higher. I cannot really tell you exactly where that would be. If that is what you think, you still would not be sure that the exchange rate is ‘correct’, but that would colour your thinking.

If you really think the whole thing is going to go away imminently—the mining boom is going to completely crash, the relative price shift is not there, the whole thing is not going to persist—then the Aussie dollar in that scenario presumably would need to be much lower than it is. It probably will go there if that scenario unfolds—that would be my guess. But in the interim, if that was your view, you would probably be worried that it is too high