Getting low: the effects of a weaker Euro

“Insanity: doing the same thing over and over again and expecting different results.” Albert Einstein

Following last week’s European Central Bank (ECB) monetary policy meeting, Mario Draghi was very clear in stating that the Governing Council stood ready to embark on further monetary easing should the Eurozone growth and inflation outlook deteriorate further. More specifically, he said that the QE programme could be adjusted in size, composition and duration while deposit rates charged to banks be lowered into negative territory, if necessary. As he spoke, financial markets moved sharply. Euro-USD exchange rate fell sharply, from just above 1.135 to a low of 1.10. (For those who aren’t familiar with FX magnitudes, that’s a large move in a short space of time!). 

Draghi for the past two or three years has quite often made a point of ‘talking the Euro lower’, with phrases like ‘the Euro is increasingly relevant in our assessment of price stability’ and ‘a strong Euro is a cause for serious concern’, so this new signal of monetary easing could easily push the Euro lower still. The question then remains: what does a weaker Euro actually mean for growth and inflation? This post tries to address this question.

(OK - so I must confess, this topic is now significantly more relevant than when I first started thinking about this issue, but hey, I’ll take it!)

A good way to think about exchange rate effects on economic variables is by looking at exchange rate pass through (PT). Pass-through is normally associated with inflation, but choosing which inflation measure to look at is clearly paramount. In this case, the exchange rate should work primarily through the trade channel (affecting import and exports), and as such I define PT as the effect of a change in the exchange rate on import and export prices. To get a reasonable idea of the Euro area as a whole, I look at the four largest Euro economies - Germany, France, Italy and Spain - which account for 78% of Euro area GDP.

The theory

Basic economic theory suggests that for exchange rates to matter for the economy, there has to be large or 'full' PT. This is the idea that import prices move one-for-one with the exchange rate while export prices (denominated in the home currency) do not move at all. 

For example, if the EUR depreciates by 5%, full PT would require German import prices (in Euros) to rise by 5% and export prices to remain unchanged in Euros and hence fall by 5% in foreign currency terms. This rise by 5% in import prices would mean foreign goods are now more expensive, and domestic German goods relatively cheaper. Both foreign and German consumers then switch to buying German goods, leading to higher German growth. This, in economic parlance, is called the expenditure-switching effect.

In addition, those goods that were previously imported are now more expensive, so prices of goods in the ‘consumer basket’ are higher, pushing inflation higher. Note that this is based on the assumption that all exported goods are priced in the currency of the producer. But let us work with this simplifying assumption for now. 

What does the data say?

I collect and calculate monthly changes in import price (IPI) and export price (EPI) indices for each country (both in home currency and using country sources), Brent crude oil price (OIL) and the trade-weighted euro exchange rate (EUR). The data are monthly in frequency, beginning in January 1990 and are collected using Thomson Reuters Datastream. 

-------------------------------For the more technical reader--------------------------------

To analyse PT, I run two autoregressions, similar to those published in Gopinath & Burstein (2013) and Campa and Goldberg (2005). For imports, I regress import prices at time t on its own 6 lags, 6 lags of EUR and 6 lags of OIL. Similarly, for export prices, I regress export prices at time t on its own 6 lags, 6 lags of EUR and 6 lags of OIL. I then sum across the betas of 6 lagged EUR variables in both regressions (a common technique in PT empirical studies) to get sensitivities over a 6 months time frame.

------------------------------------------------------------------------------------------------------

The charts above show the sensitivities (or ‘betas’) of the IPI and EPI in the 6 months following a change in the exchange rate. For illustrative purposes, France’s import price PT is -0.85, meaning that a 10% depreciation in the EUR would, on average, lead to a 8.5% rise in import prices over the next 6 months.

A few observations:

1) Export prices (in Euros) rise when the currency falls. This means that exporters raise the local currency price of their goods as the Euro depreciates. 

This is not what the theory says, and hence is a failing of the simple theory - export prices do to not show full PT. A simple example to drive this point home: Say Euros depreciated by 5%. Goods that would be normally 5% cheaper for foreign buyers are now only 1% or 2% cheaper because producers raise their price in Euros. This therefore means that import price PT is much higher than export price PT. 

2) The change in import prices is higher than export prices for each country. 

3) There are significant differences in the price sensitivities across the 4 countries.

Implications on the euro area:

Note that the ECB’s inflation mandate says that consumer price inflation (CPI) should be close to, but below 2% and that changes in the CPI can be broken down in domestic producer prices + import prices - export prices.

Assuming the change in exchange rate does not change domestic producer prices much, import prices will rise by more than export prices (observations 1 and 2) following a depreciation. This means that the change in import prices minus the change in export prices is positive, and so inflation rises. In other words: a euro depreciation should raise inflation across the Euro area.  

From observations 1 and 2, producers raise their export prices to offset the effective reduction in prices for foreign buyers arising from a currency depreciation. Consequently, demand will not increase as much as it would have done if they didn't raise the price (as the theory assumes). 

The end result?  The potential growth kicker from higher export volumes is dampened by firms’ pricing decisions. Moreover, even though import prices rise and hence domestic consumers buy more domestic goods, the boost to growth is likely to be lower than what people expect from a currency depreciation because of this dampened export PT.

Implications for country growth and inflation:

Now let’s combine this with observation 3. Import PT is the highest in Spain and France, while lowest in Germany and Italy. The mirror opposite applies to export PT. Let us take the two extremes, Germany and Spain, for illustration purposes.

On imports:

By only looking at PT rates, we can infer that a given depreciation in the Euro means higher growth in Spain than Germany - Spanish import are prices rise more than in Germany, so Spanish consumers switch to buying more domestic goods than German consumers, spurring growth.

Remember though, that Germany has the largest proportion of its trade with non-euro area members out of the big 4 countries. Interestingly, Germany seems well placed to take advantage of a depreciation from a trade perspective, but its poor import price PT means that it can’t. 

Conversely, Spain has the smallest proportion of its trade with non-euro area members, but has the largest PT. Though its import prices move a lot following a deprecation, it doesn’t actually matter for the economy because Spain predominantly trades with countries whose currency is also denominated in Euros.

On exports: 

Here, the story is the other way around. German firms’ decisions not to raise prices following a depreciation does mean cheaper goods for foreign buyers. Germany’s major trade links with non-euro area countries ends up being a major boon for its economy.

Spanish producers on the other hand choose to raise prices following a depreciation (more than Germany). In other words, the effective price of a good for a foreign consumer doesn’t change much even following a depreciation as Spanish producers raise its prices. Therefore, a depreciation doesn’t end up being a big deal for Spanish export growth.

The bottom line: 

There are two aspects that are required for a currency depreciation to have a meaningful effect. The first is a high pass through. The second is that the country must do a lot of trade with countries who have a different currency.

Draghi's message of a lower Euro has been a persistent narrative over the past few years and with more monetary easing on the horizon, a depreciating Euro is likely to continue for some time. When we take into account both trade links and the pass-through story, the analysis suggests a euro depreciation predominantly pushes German exports higher with little effect on much else. There are very little (first-round) effects on either the Spanish economy or German imports. It also follows that German growth is likely to be the major beneficiary of a euro depreciation, compared to Spain or France. 

In a world where euro-area growth has historically been very uneven as the Euro depreciates, this pass-through angle suggests that expecting anything different going forward would be nothing short of insane.

Changing market dynamics

“Noticing small changes early helps you adapt to the bigger changes that are to come.” Spencer Johnson

The Federal Reserve chose to keep interest rates on hold on their 18th September committee meeting. Despite improving household spending, business investment and labour market, the committee judged (correctly, in my opinion) to pay heed to economic and financial developments abroad. In response to the announcement, world equity markets fell. The S&P fell 1.6% on the day of the announcement and other equity markets followed suit.

Those who believe in market efficiency (i.e. that markets price in the ‘true value’ of the stocks at any point in time) would conclude that investors believed that the Fed made a mistake in its decision, and that the US economy was in fact strong enough to withstand a rate hike. In doing so, they would argue, the Fed has introduced more uncertainty into the economy, thereby warranting lower equity prices. The policy implication from this is that the Fed should ‘just get on with it’ and hike rates, as it should not cause too much of a disruption to the markets. A very simple bit of analysis by looking at what I call the ‘market reaction function’ to economic data suggests this isn’t quite the case.

First, a word on the data: I use the Citi US Economic Surprise index to represent ‘economic data’. It looks at every single US economic data release and compares it to what was forecasted by economists (the assumption to make it meaningful is that the economists’ forecasts closely represent what is expected by investors). Essentially, if the data release is better than expected, then a positive number is registered in the index. Likewise, if data is worse than expected, a negative number is registered. Using these elements, a cumulative index is created. The chart below shows the average (mean) return of the S&P on days when economic data releases ‘beat’ expectations (i.e. positive surprises), and days when it ‘missed’ expectations (i.e. negative surprises).

The period from 2005 onwards (which I think is a long enough sample set to think of as the long run/normality) shows how equities should react to economic surprises. Theoretically, equities should go up on days with positive releases as it signals to the market that the economy is healthier than people believe, which should translate to higher profits for companies (and therefore higher stock prices). Conversely, equities should go down on days with negative releases for exactly the opposite reason. 

Convention was turned on its head in the second half of 2013. With the Fed still conducting its QE programme and signs of its upcoming termination emerging, markets became jittery and appeared hooked on QE. Any piece of negative news that suggested the Fed would delay the ending of its QE programme was met with jubilation by the markets.  This is shown on the chart by S&P returns significantly higher on days where data ‘missed’ expectations than those where they ‘beat’. This was the very much a characteristic of the ‘taper tantrum’ that many commentators and media outlets continue to speak of. Essentially, the market would only go up if there were signs of a prolonging of QE, as there was little faith in the strength of the US recovery. The Fed subsequently ended its QE purchases in October 2014, and the market dynamics eventually returned to normal (the third pair of bars - Jan 2015 to August 2015). Until recently. 

Since the FOMC announcement in September, it seems the market has begun to be jittery again. Now, negative news is being met, on average, with positive returns in the S&P, and certainly much higher than the corresponding returns on a day with positive economic news. (Note that I have used mean returns here, but using medians produce the same results). From this, it doesn’t seem as if the market is sure about the strength of the economy at all - somewhat similar to the second half of 2013.

So what caused the S&P to fall after the announcement? Surely if the market wasn’t sure about US economic prospects, a delay of the rate hike should be push equities higher? In theory yes. In practice though, what I believe happened is that those investors who thought the Fed would not raise rates bet on this by buying equities BEFORE the announcement. The more and more people did this, the higher the price got. Once the announcement came out in favour of those investors, they cashed in and sold their stocks, having already made a profit running up to the event. And if a lot of people do this, the price falls. This is what, in ‘trader-talk’, is called ‘an unwinding of heavy positioning’. If we look at the price data, this is exactly what we see. From the 1st September to the 18th, the S&P went up 4%, as investors bought equities in anticipation of the Fed not hiking rates. Indeed this is corroborated with a various articles citing that the Fed funds rate rise expectations was being pushed out towards December/January 2016.  And then low and behold, as soon as the announcement was made, the S&P was down 1.6% the following day as those who made money on the way up sold their equities. 

The bottom line

The implication from all this is that the market reaction function has changed, and shows characteristics similar to that of H2 2013. Now, to be clear, I do not believe that we will get a ‘taper tantrum’-style scenario, with violent swings in all asset-classes, if the Fed does raise rates - the economy is certainly a lot healthier than it was back then, and I think everyone is aware of that. The point is that there again seem to be doubts creeping into markets as to how strong this recovery is. If the FOMC care at all about the market reaction to its decisions, it may well benefit from not normalising policy just yet. Indeed, noting this small change early on could save the Fed from a disastrous normalisation of policy that is to come.

Free money?

“There's No Such Thing as a Free Lunch.” Milton Friedman

My last post, ‘Too good to be true’, looked at Jeremy Corbyn’s most contentious economic proposal, People’s QE. I highlighted two issues: the threat to the BoE’s independence and the belief that the debt can be cancelled and then proceeded to discuss the first. In this post, I deal with the second.

Recapping from last week: 

"So what is People’s QE? Well, it begins the same as QE, with the BoE creating money. But this money is then only allowed to be used “under government direction and subject to government guarantees” to buy bonds of a new institution set up to promote investment and innovation the UK, called the National Investment Bank (NIB). Moreover, it is believed that the debt created by the NIB can be subsequently cancelled on a technical detail."

Earlier this year, Richard Murphy put forward his proposal for Green QE (the policy idea that spawned People’s QE). In his post, he writes that Government debt bought by the BoE as part of a QE programme is cancellable. This in and of itself, he argues, is no different to conventional QE:

"The government no longer pays interest on this £375 billion of its borrowing, because to do so would mean that it would simply be paying interest to itself since the government gilts now belong to the Bank of England, which is in turn owned by the Government. That’s one clear indication that the debt no longer existed, because if it did the interest would have been due. And that is quite obviously true: whilst one part of government can obviously owe money to another part, if no one outside government is due money as a result there can be no government debt owing, and that has been the consequence of QE."

He explains again, in another post:

"If the government buys its own debt then it cancels it. This, as  a matter of fact, is true. It is, of course, technically legally possible to argue that the debt still exists, but if I create a loan to myself, even if it is legally recorded, it has no economic consequence:  my repayments of my loan from me to me means that no money actually in net terms leaves my own pocket and that is exactly what  happens when, as a consequence of any form of quantitative easing, the government owns its own debt.   So, quantitative easing actually cancels government debt and does not increase it."

Now, this might not sound like a big deal but it has enormous ramifications for economic policy. First, it means that the motivation for embarking on a QE programme could have nothing to do with monetary policy, but rather fiscal issues (which bring us down the route of questioning central bank independence). Second, it means that the government could essentially be getting free money. 

So, it makes sense to figure out what is actually happening. In the rest of this post, I’ll explain why I think government debt bought in a QE programme CANNOT be cancelled under the current framework.

The evidence

A look at the correspondence between the BoE and the Treasury gives us a fair idea of the financial agreements in place between them. The first thing to remember though, is that the Government is indemnified to the Bank of England (BoE) and its subsidiaries, i.e. the Government covers losses as well as receives profits arising from the BoE’s activities. 

In the Monetary Policy Committee (MPC) meeting of 7-8 November 2012, committee members were briefed on (and subsequently agreed to) the Government’s request to receive excess cash, on a quarterly basis, from the entity set up to carry the QE transactions (the Asset Purchase Facility - APF). This excess cash was defined as the interest paid by the Government on the bonds the APF held as part of their QE programme, minus operating and other costs. 

When the APF was initially set up, it was agreed that the total profit/loss for the APF would be settled with the Government when QE was finished and the APF was closed. During that time, the Government would borrow money to finance the interest payments on the bonds that the APF owned (just as if the bonds were owned by private investors). 

George Osborne, in his letter correspondence with Sir Mervyn King (former Governor of the BoE), argued that with the UK recovery sluggish, it was apparent that the APF would be around for a while longer. This then meant that the APF’s cash balances from received interest payments (a sizeable £35 billion at the time) would continue to grow. Osborne (correctly) argued that the Government was borrowing too much to fund these payments, and the set-up was economically inefficient. Therefore, he said, it made sense to transfer the cash back on a quarterly basis.

Essentially, the Treasury believed that as the APF was just another governmental department, it made little sense for one part of the government borrowing to fund another, and so the interest payments were economically moot. So far, so good for Murphy and his theory.

The letters though, made clear that these payment will likely be reversed in the future (due to the Government’s indemnity to the BoE) as the APF incurs losses on its bond positions when the MPC decides to sell their bond holdings and increase the Bank rate in response to a strengthening economy.

What does this mean?

The OBR noted in its press notice that the fiscal effect of this transfer would be two fold. In the short term, Government net borrowing would be lower than it would have been - so QE provides a helping hand to the government finances. In the longer term, the OBR said, it is likely to be higher than it would have been as the Government has to borrow to fund any losses the APF incurs, just as the letters noted. It's critical to realise that these losses could amount to MORE than what the Treasury saved from the APF’s quarterly transfers to it (depending on how violently the UK bond market reacts to the MPC deciding to raise the Bank rate). In this case, QE doesn’t cancel any debt, but in fact raises it. This is where Murphy's theory breaks down.  

What’s more, Mervyn King was sure to spell out that principal (cash) received from the maturing bonds would not be transferred to the Treasury, and would  be used at the discretion of the MPC. In an example: if the APF bought £100 billion of government debt as part of QE at an interest rate of 3%, though government would get a £3 billion windfall from this transfer set up, it would still have to pay £100 billion back to the APF (and hence the BoE) once the bonds mature. 

Also bear in mind that once the MPC normalises monetary policy by selling its government bond holdings back to the market and raising the Bank rate, the APF will no longer be the owner of Government bonds - private investors will. The Government will then have no choice but to resume making interest payments and the interest costs become, in some sense, ‘live’ again. 

So, government debt cannot be cancelled; only interest payments can be cancelled (and that, only for a short period of time). But as I explained above, if the APF incurs substantial losses as part of the MPC raising rates, this doesn’t mean that QE will necessarily reduce government borrowing.

The bottom line

For People’s QE to be fiscally neutral (i.e. have no effect on government debt), the debt must be cancellable. This is, in my view, impossible under the current set up of QE. One way it could be done, is if the MPC is instructed to buy and hold government debt until maturity, and then create more money to finance any losses it might incur. The outcome? The loss of central bank independence.

In my first ever economics lesson, we were taught that economics is the theory of allocating finite resources to satisfy infinite wants. Richard Murphy seems to have forgotten the fundamental principle of economics; and in doing so has forgotten that there’s no such thing as a free lunch.