Please note the content on this website is for Investment Professionals only and should be shared responsibly. No other persons should rely on the information contained within this website.
Logo of Bond Vigilantes
Friday 29 March 2024

One of the major factors that has enabled inflation to stay low despite the economic strength in major western economies has been the fall in the price of oil. Given the huge price volatility over the past 18 months it is interesting to depict the falling influence of oil on actual end inflation.

In the UK, the most direct way that changes in the oil price affect inflation is through petrol prices. Falls in the input cost translate into a fall in the price at the pump. The price at the pump is however not just a function of the price of oil, but includes costs of transportation, retail margin, and most significantly tax and duty. According to petrolprices.com, the cost of a litre of petrol can be broken down as follows:

Duty: 57.95p

Product: variable

Retailer/delivery: 5p

VAT: 20%

If we use the above as a guide we can examine how changes in the price of oil (product) can affect headline inflation. The chart below shows how changes in the oil price feed through to the price of petrol at the pump.

The effect of changes in oil prices on the price of one litre of petrol in the UK

As you might expect, it is not a one for one relationship. As the price of oil falls, its effect on the price of petrol becomes less significant because it makes up a smaller and smaller portion of the total price at the pump. So a 50% fall in the price of oil from $160 to $80 creates approximately a 35% fall in the price of petrol from 267p to 172p. A further 50% fall in the price of oil from $80 to $40 creates a 28% fall in the petrol price and a 50% fall from $40 to $20 translates to only a 19% fall in the petrol price. Although the GBP/USD exchange rate is another factor in determining the pump price given that oil is priced in dollars, the above illustrates that as the oil price falls it has a weaker and weaker effect on inflation.

There is also a secondary effect of weak oil prices, because as prices fall, the percentage in the inflation basket will fall. Thus, energy price falls become less significant with regard to headline inflation.

The down draft in oil prices is coming closer and closer to an end, with its effect on inflation reducing the further it falls. Given other areas of scarcity, such as in the labour market, we would be surprised to see inflation stay this low over 2016. The downward impetus of falling oil prices on inflation will continue to become less meaningful as the bear market in oil progresses.

One of the first rules of economics is that the equilibrium market price is generated by relative supply and demand. Limited supply or excess demand should result in an increase in price. One of the questions that has arisen in the post financial crisis world is why have wages not increased despite unemployment heading towards historically low levels? Given the improvement in data such as headline employment, jobless claims, and the JOLTS (Job Openings and Labor Turnover Survey) data, the US economy should be experiencing higher wage pressures by now.

15.10.28 blog RW1

Thinking back to my tutorials from ex Bank of England guru Charlie Been, economists should not only take note of supply and demand when examining equilibrium rates. There was also lots of discussion of both nominal and real wages when examining the relationship between labour market equilibriums, and the cost of that labour.

Nominal wages in the US and UK are at historically low levels, implying that the relationship between supply and demand in the labour market is different this time. However, when we look at the relationship expressed as the real cost of labour it can be seen that the relationship between a tight labour market and rising wages is indeed present as expected.

15.10.28 blog RW3 15.10.28 blog RW2

Those that argue that the current low rate of wage growth in nominal terms is the true depiction of the labour market would have to reflect on the problem of the mid 1970s experience, when a weak labour market was accompanied by rapid nominal growth in wages of 7 to 8 percent as seen in the US chart above.  I argue that the labour market is tightening and wage growth is there in real terms. Looking at the data, it can be seen that the real cost of labour is more correlated with the unemployment rate (inverted in the charts above) than the nominal wage rate. The UK chart shows that nominal wage rates were pretty flat during the economic collapse and recovery in the UK, while real wages responded accordingly to the weakness and subsequent strength of the labour market.

We can therefore argue that the current data, and likely future data, imply the Phillips curve is still alive and well. The challenge the central banks face in hitting their inflation targets is that the price of other stuff (mainly oil) has (temporarily) collapsed, thus the strength in the labour market is currently masked. The economic rules of excess supply and demand are amply demonstrated in the depressed commodity markets, the low yields available in the QE-led European bond markets, and in the labour market.

The Fed and the Bank of England should recognise that labour data suggests rising inflationary pressures. Further attempts to strengthen growth via loose monetary policy will enable inflation targets to be hit sooner than would otherwise be the case, but once the deflation outside the labour market halts, central banks will be faced with the traditional problem of reducing (not increasing) inflation. This need for higher rates in the future implies that long term government bond yields near record lows do not need chasing.

The Bank of England’s Monetary Policy Committee (MPC) are due to meet on Thursday and most economists expect a dovish set of minutes to accompany the announcement of no change in the BoE base rate. Additionally, the minutes will likely emphasise the risks of a persistent undershoot in UK inflation given the continued fall in commodity prices and waning global demand. Despite these risks, the MPC are also likely to indicate that the first interest rate hike remains on the horizon. The decision to eventually hike interest rates will not be taken lightly, and the MPC will be having a long think about the potential labour market impact when a rate hike eventually occurs.

Monetary policy works with a lag; the BoE estimates this to be approximately two years in the UK. The relationship between rate hiking cycles and unemployment figures has therefore historically taken time to fully feed through the economy.

Since 1971, the BoE have undertaken six separate rate hiking cycles in the UK. After the initial rate increase in each of these, unemployment has continued to fall on average for the next 22 months, which is just shy of the BoE estimation of a 24 month lag. Given that the UK unemployment rate has been on a downward trajectory since mid-2012, history suggests that even if the MPC hikes rates, it will likely fall significantly further from the current level of 5.2%.

2015-09 blog RW

Indeed, if the BoE rate hiking cycle were to commence in October of this year, the graph above forecasts that the corresponding unemployment rate could potentially bottom out at 4.25% in August 2017. Building on this idea, if we extrapolate the current downward trajectory in unemployment and assume a later rate hiking cycle starting in April 2016, this predicts an unemployment rate trough as low as 3.61% by February 2018. Unless the BoE acts soon, the anticipated fall in UK unemployment from current levels could potentially return us to the multi-decade low unemployment rates last seen in the inflationary early 1970’s.

The trauma the BoE experienced during the recession was typified by Sir Mervyn King’s comments which I wrote about in 2012, when he said that we were not yet half way through the crisis. Sir Mervyn suggested that rates would stay at 0.5% until the end of 2015. Well, we are now approaching the end of that time frame window and the BoE has only fear itself as to when to take the economic action to slow a potential inflation problem. The current members of the MPC should take inspiration from the appropriately titled Lily Allen song “The Fear” which was the UK no.1 single the last time the committee moved rates in March 2009:

“And when do you think it will all become clear?
‘Cause I’m being taken over by the fear”

The graph below shows US unemployment alongside the Fed rate over a period of 45 years. From this you can observe the broad relationship between the two, specifically the time delays between Fed rate hikes and the upturn in employment which has historically followed. This time the Fed have delayed the rate hike for a number of reasons, but if history is anything to go by, we can perhaps use this data to make some predictions with regards to the potential timing of the first hike.

The inherent monetary policy lag

1) After the initial rate hike, the unemployment rate tends to bottom out in 2.25 years

If assessing the timing of an interest rate hike, we need to also consider the time lag inherent in monetary policy. We have therefore looked at each rate hiking period in isolation and determined how long it has taken for the unemployment rate to reach its trough after the first Fed rate hike in its tightening cycle. Although this has varied from 13 quarters in 1976 to 5 quarters in 1994, over the whole period it has taken an average of 2.25 years for unemployment to bottom out after the Fed’s initial rate hike. Therefore, if the Fed was to hike later today, based on previous experience, unemployment may bottom out in September 2017, although it could be between September 2016 and September 2018.

After the start of a rate hike cycle, it has taken 9 quarters (on average) for unemployment to reach its trough

2) Unemployment tends to fall 1.2% after a rate hike cycle
If the Fed hikes rates later today, where will unemployment trend from its end of Q1 reading of 5.6%? Again glancing backwards, when unemployment has bottomed out, it has tended to have fallen by 1.2% on average after the first hike, suggesting that consistent rate hikes commencing in June could result in unemployment eventually falling to 4.4% at the cyclical low.

After the start of a rate hike cycle, US unemployment has historically bottomed out after falling an average of 1.16%

3) Depending on the timing of the rate hike, unemployment could fall to 4.4-3.6%
Building on the observations of the graphs above, we can use these historical averages to predict a range that unemployment  could move in, should the Fed hike rates in any of its next four meetings. If the Fed was to delay hiking until March next year, unemployment could fall to 3.6% in June 2018, which would be the lowest US unemployment reading for at least 45 years.

Unemployment forecasts

The Fed has delayed its rate hike as wage pressure has been limited so far and the extent of the financial crisis and the lack of flexibility with regard to rate setting at the zero bound necessitated a more than a traditional easing cycle. However, as we have commented before, employment markets are healthy and full employment approaches. As we can see from the above analysis, extrapolating the current trend in employment growth, it is unlikely the Fed will delay its hike until next year and should signal a rate hike very soon given current economic trends and the inherent lag in monetary policy.

The US unemployment report for April highlighted the continuation of the economic recovery. The market is now in the habit of viewing a job creation number of anything less than 200,000 as a weak result for the labour market and anything more than 300,000 as a strong result. Anything in between and the conclusion amongst economists is this: the Federal Open Market Committee (FOMC) is on hold, economic growth is reasonable, inflation is not a concern, and interest rate increases are not imminent.

To me this seems a bit complacent.

As the unemployment rate falls the US economy will get closer to the point where the low supply of labour will cause wages to increase. Inflation will get an upward push, and the FOMC will likely feel like it has to remove some of its ultra-loose monetary policy by raising interest rates. In the theoretical extreme, were we to get to full employment, then focusing on the absolute payroll jobs number does not make sense. By definition, at full employment very few jobs can be created, as there is no spare labour. An employment report of 100,000 or less when the US economy is operating at full capacity would indicate a vibrant economy with inflationary pressures.

In order to analyse the labour markets as they approach full capacity we have created the chart below. The chart shows the ratio of jobs created as a percentage of the spare labour available.  This is an attempt to move the focus away from a headline payroll number, to the true tightness of labour markets.

Blog_Richard_May2015-1y-avg

As can be seen, labour markets by this measure are historically very tight, and I therefore think wage pressures are stronger than the market consensus. Pre-emptive FOMC action is warranted sooner than the market is currently expecting (Eurodollar 90 day futures are pricing in a rate hike in December). Plotted alongside historic fed rates, we can see how far the normal FOMC interest rate response is to the labour market data in this cycle. Historically, when non-farm payrolls as a percentage of those unemployed has been this tight (around 2% level), the FOMC has begun a tightening cycle (in 1986, 1993, 1999).

In economics it is easy to focus on the absolute number, however one should always delve beneath this to analyse the relative. Given the strength in the US economy I would not be surprised to see higher wage growth numbers, rising inflationary pressures, and FOMC action despite the non-farm payroll number staying well behaved in absolute terms. At some point, strong job creation and low unemployment rates will result in higher compensation, and this would not be taken well by the bond markets.

It has been a while since we have discussed the economics of the single currency, but once again the issue of its suitability for all its members is at the forefront of economic concerns, as Greece faces some difficult decisions.

The financial crisis has taught us a number of lessons: fiscal policy works, monetary policy works, better regulation is beneficial for the financial sector, confidence is key, and importantly exchange rates matter.

Throughout the crisis, one of the economic mechanisms that aided the economies in most stress was the exchange rate. This can be seen with the collapse in sterling in 2007 on a trade weighted basis, the weakening of the dollar from 2009 until 2011, the yen in 2013 onwards, and the euro of late, see chart below. At the heart of these fx moves lie the currency vigilantes, as we discussed here in 2010.

15.04.16 blog RW

These external foreign exchange rate moves are a textbook play in terms of making labour cheaper and therefore helping economic recovery. However, as we know, in the Eurozone this mechanism does not exist, due to the creation of monetary union. I think exchange rates have become relatively more important in determining national economic outcomes and this is particularly relevant now to Greece.

The three main macro-economic levers are monetary policy, fiscal policy, and the exchange rate. Fiscal policy is still in the hands of politicians and therefore can be used to provide a strong impetus when required to differentiate national outcomes (although less so in Europe). Monetary policy has basically approached the zero bound in the main G7 economies, which means short rates have become hugely correlated. Without room to differentiate economic outcomes by cutting rates, national economic flexibility has been reduced, which means the exchange rate has to play a more important role than has historically been the case.

This is working between the main economic blocks. However, as the need for this most famous “invisible hand” has become greater, it has not been available within the Eurozone. This means that Greece has to somehow adjust with no fiscal room, no monetary room, and no exchange rate flexibility.

It would take at least a generation for Greece to solve its problems via structural reforms given the constraints it sits with. The short term solution is therefore for Greece to be bailed out via fiscal transfers directly, or in a quasi manner by allowing a Greek default. These are obviously hard to achieve given the political dilemmas many countries would face in providing this relief.

Greece has faced difficulties before, however we now have further pressure to find a solution as the economic policy options outlined above become more focused on the exchange rate conundrum, and the political environment in Greece points to a government more willing to take radical measures in the face of a great depression. The ability of Greece to provide for its citizens is damaged like the famous Venus de Milo statue. It could well be that politicians recognise the invisible hand of the exchange rate is still an important tool, and a free floating Drachma, the “Expulso” solution (see 2011 blog), though painful, might be the best shot at providing an economic solution given the extent of Greece’s problems.

The Fed has basically used three major themes in response to the financial crisis from a monetary point of view:

  1. Lower short term rates
  2. Quantitative easing
  3. Operation twist – an attempt to flatten the yield curve

The Fed has communicated that it now expects the first move in normalising rates as the economy recovers will be to increase short term rates. Personally I think there are other alternatives.

Firstly, looking back at 2014, we can see it was a year of further monetary easing. This was done primarily through QE, which ended in the autumn, and also through a bull flattening of the yield curve, as illustrated below.

The process the Fed is currently discussing is one of raising short rates first (reversing theme 1), and then looking at reversing the other monetary actions at a later date. This need to remove easy monetary policy is well documented and there is a significant focus on how the Fed will tighten post the financial crisis.

An early move with regard to tightening was engineered with the taper tantrum of May 2013. The below graph shows the yield curve ahead of the Fed’s signalling and where it was as at the end of 2013. This was an effective tightening of policy as discussed here.

This tightening has been unwound in 2014 as already mentioned. Surely one option open to the Fed is to normalise the curve as opposed to pushing up short rates ?

Firstly, I have always thought that monetary tightening should logically follow the reverse pattern of monetary easing. This would mean reversing operation twist, then QE, and only then pushing short rates higher. Removing any QE inspired asset bubbles seems logical to me.

Secondly, the Fed has acknowledged with operation twist in 2011 that the shape of the yield curve matters. When we look at the yield curve before operation twist and now, we can see the dramatic flattening of the curve, partly inspired by operation twist. Surely it makes as much sense to normalise the curve as a policy option as it does to put short rates up? Indeed p7 of this IMF paper on unconventional monetary policy suggests the term premium is currently distorted by around 100 basis points as a result of QE, although monetary policy and economic conditions outside of the US are surely also pushing the term premium lower. The Fed should consider selling long dated securities and buying shorter dated ones. This would also have the side effect of reducing the potential profit and loss impact on the Fed’s balance sheet, while making the eventual unwinding of QE more manageable as the shorter dated assets mature in an orderly manner.

Thirdly, by simply reversing QE, the Fed would further reduce the balance sheet risks as described above. This in turn would then make conventional policy tightening through raising short rates more practical and normal by reducing the risk of flattening (inverting?) the yield curve. Also if the Fed, like other central banks, wishes to look through the short term effect on inflation of a falling oil price, then surely leaving short rates where they are and pushing up longer rates by reversing QE makes sense economically. After all, the Fed has focused on forward guidance verbally and physically for a number of years, why should they stop now ?

When conducting pre-crisis monetary policy, the main central bank tool was short rates. Since the crisis other weapons have had to be used to achieve economic goals. In moving back to conventional monetary policy the Fed should look to remove unconventional measures from the market place ahead of or alongside conventional rate hikes, not after them.

Matt’s and James’s recent blogs outlined some of the issues markets face when rates go negative. This is obviously no longer just a theoretical debate, but has real investment implications. Why do investors accept sub-zero rates when they can hold cash ?

To recap using Swiss Francs for example, it makes sense for a saver from a purely economic view not to deposit a Swiss Franc note into a negative yielding bank account, as it will be worth less when it gets returned, due to negative rates.

However, the saver faces risks by holding physical cash as they don’t receive the security benefits from using a bank account (ie paying for an electronic safety box). The use of the old fashioned lock and key is not as convenient as a bank account, but would make increasing sense to use as deposit rates get more and more negative. This demand for owning physical as opposed to electronic cash is not confined to cash accounts. In theory, as the term structure of interest rates falls below zero, bond investors should sell their bonds and own “cash in a box”  instead. How efficient is it to do this ?

The great problem with using paper money as a saving instrument is that its inherent best in class liquidity also makes it vulnerable to fire and theft. From an individual’s perspective, the use of a secure fire proof safe deposit box in a bank or a secure location away from home is the best starting point. The optimum solution however relies on economies of scale. Is this easily achieved?

As an investor, diversity makes sense. Therefore, an individual should spread their cash over a wide selection of safety deposit boxes in a wide variety of very secure locations. An improvement, but currently not that practical. But there could be a way to achieve the above goals relatively efficiently and cheaply.

In a negative interest rate environment  there are likely to  be enough investors who want to own bearer cash for a network of highly secure safe deposit boxes to be developed by a bank or institution. This means there would be a high degree of security and diversification regarding the location for the cash. In order to make the cash available to the investor easily, certificates of deposit could be issued physically or preferably electronically. This would allow the investor to transfer money easily, as they would only need to go to their nearest depository to deposit or access the cash, or their nearest bank if a bank agrees to physically deposit or withdraw cash for them.  Basically this ends up being a bank account where the cash does not get lent, but has a custodian holding charge. In theory, in the extreme, you could even develop markets in exchange traded derivatives issues that are linked to cash held in a depository, to allow individuals and large institutions to manage cash as a saving instrument with no negative yield. A new efficient savings industry could be developed in a negative yield environment, so limiting the downside to the sub-zero bound for short and long term interest rates.

One side effect would be that all these savings would have to be held in real cash, which will mean an increase in the demand for physical notes. If cash is held in custody and is not lent on then the supply of money in the economy for normal transactions will fall. This begins to sound deflationary, and runs counter to why sub-zero policy is being pursued.

As long as cash exists in a physical bearer form it is hard to see how you can have significant negative rates of interest in an economy where government debt and cash are the obligations of the same entity, as they are truly fungible. At its worst, monetary policy of sub-zero rates could encourage a deflationary spiral. Maybe the only policy left to create inflation is real and not conservative QE (see my last blog).

It has been a while since we talked about QE, but we covered this substantially in the past (see for example ‘Sub Zero?’,  ‘QE – quite extraordinary‘ and ‘Quantitative easing – walking on custard‘). It now appears, at least for the time being, to be a part of monetary history in the UK, and more recently the US. However, it is being reapplied in Japan and about to do a grand tour of Europe. Our earlier blogs were an attempt to analyse a new experiment. What do we think now we have had the practical implication of the theory?

Let’s go back to basics first. Monetary policy reaches the zero bound, so short term interest rates can no longer be cut. Therefore it is time to print money. Being prudent, the central bank needs to be able to tighten policy again at some point by destroying the printed money. So it favours purchasing large, liquid, risk free debt and buys government bonds in huge quantities.

This drives longer term interest rates down to the zero bound, and therefore should encourage long term borrowing, discourage long term saving, increase asset prices that are a function of long term rates (property and equity), and therefore via the wealth effect, stimulate growth.

The above effects and especially the wealth effect are seen as proof by its supporters that QE has worked via higher asset prices and this chart is often used as evidence of the correlation between the two.

UK QE and asset prices

Asset prices have risen and growth has indeed returned, but where is the inflation?

Inflation has indeed been temporarily induced in countries where the exchange rate has collapsed (for example the UK and Japan). However, this has proved to be a blip in the UK, and will likely be the same in Japan once the yen, which is down more than 50 percent versus the dollar over the past three years, finds a new stable equilibrium.

There are principally two reasons inflation has not returned. Firstly, inflation is not just a monetary creation but is a function of other factors, ranging from the oil price, to productivity, technology, inflationary expectations and free markets. The first of these has been exceptionally volatile, creating cyclical inflation and disinflation spurts, whilst the last four have been a constant driver of structurally low inflation for many years.

Secondly, let’s look at what QE actually does from a monetary perspective. The central bank simply exchanges cash for near cash. Holders of government bonds now own cash having sold them, while the central bank now owns government bonds. Interest rates are lower along the yield curve, but there is no actual new money circulating in the economy. Cash that has been created has been exchanged for another form of cash – government bonds.

Central banks have printed money in a very conservative way, so its growth and inflation effects are limited to wealth effects, and a reduction in long term rates.

The interest rate effect of driving the whole yield curve towards zero will reach its own zero bound, and cease to be effective, like short rates at the zero bound. The wealth effect will diminish as it will reach the bounds of investors’ rational market expectations (like inflation expectations) and asset prices will cease to rise as strongly. The assets that do rise are held by individuals who will reduce their marginal consumption as their wealth increases, or those that can’t access them as they, for example, are in a pension fund. Therefore QE itself, in its current form, reaches a zero bound.

When we first discussed QE, the great fear was that it would result in an inflationary spiral as money is printed prolifically. However, QE has been done in a responsible fashion so far. If it were to return to its philosophical roots, as outlined by ‘helicopter Ben’ in his 2002 speech before the National Economists Club, then you would get inflation. Printing money with nothing in exchange for it is inflationary. Printing money and swapping it for near money (ie government debt) is not quite the same.

Fortunately, monetary and fiscal policy has been effective in restoring growth, though inflation remains low. Will the new member of the QE fan club that is the ECB generate any meaningful long term inflation with its traditional QE programme? I doubt it. No one else has.

I blogged last year about the state of the US labour market and given the recent release of September’s initial jobless claims data, this seems like a good time to revisit these ideas.

US Initial Jobless Claims is an unemployment indicator which tracks the number of people who have filed jobless claims for the first time, representing the flow of people receiving unemployment benefits.  The September headline figure of 288,000 is strikingly low and is the lowest reported month end figure since January 2006. Still, this understates the current strength of the labour market as when adjusting to take into account the working population, jobless claims as a percentage of the US labour force is now at multi-decade historical lows.

Initial jobless claims as % of working age population

Initial Jobless claims data has been taken on a monthly basis whereas the OECD US working age population data is annual. Therefore what is particularly important to note is that the latter figure for 2014 is not yet available and so the 2013 data has been extrapolated and kept stale since year end. As a result, the graph is more conservative than reality since 2014 population growth to date has not been incorporated. If it were, the fall in this indicator would be even more pronounced.

Traditionally monetary policy has worked with the Fed tightening as the economy picks up steam and jobless claims fall. What is remarkable today is that the Fed hasn’t even begun to tighten interest rates. In the past, the Fed would have already ended the tightening cycle by the time jobless claims fell to the levels we see today.

Unemployment and Fed Policy

Time and again the Fed has stressed that rate decisions will be data dependent and on Tuesday and Wednesday next week the FOMC are due to decide on whether or not it ends its QE program. Given the above, it seems that the US economy is continuing its healthy response to the stimulus provided and momentum in the US labour force is gathering pace in a positive direction. With more people working and fewer claiming unemployment benefits, the downward trajectory of this indicator – as well as other labour market indicators – surely helps to paint a positive macro picture. However, the risk-off wobble in markets last week has left many questioning whether the volatility experienced will have any bearing on the QE decision. Considering however that the trigger was the underperformance of US Retails Sales data, it can perhaps be argued that this is a typically volatile number in itself and the data release therefore triggered an overreaction in bond markets (exacerbated by capitulations, technical trading level breaches etc).

If the FOMC is in agreement that the market response was overdone, one should expect their decision to be based on the fundamentals and the picture of an overall improving economy. If true to their rhetoric, markets should expect to see an end to asset purchases, right on schedule.

Author: Richard Woolnough

Get Bond Vigilantes updates straight to your inbox

Sign up to the BV Mailing List