Wednesday, September 18, 2013

How I would propose to taper - my response to Harding in the FT on Sept 18th

Despite the fact that the FOMC decided not to "taper" today, I believe that it missed an opportunity to inject less funds into the market and begin the process of exiting from QE3.  Part of the reason that the Fed decided not to "taper" this month is that there are considerable uncertainties regarding US fiscal policy and whether the labor market improvements will be sustained at current levels through into next year.

In today's FT (see here), Robin Harding outlined how he thinks the FOMC will implement the taper of monetary policy which I've been discussing in my most recent blogs.  Although it will not occur now, I think it is instructive as a dynamic "exit strategy" from QE3.  So this is my (edited) response ( - posted as a response earlier today after Harding's article) in the comments section of the FT:

"The taper is largely misunderstood, but it clearly needs to happen sooner rather than later in my opinion, as I believe that inflation expectations have now bottommed, and are on the rise again. The taper though is a dynamic animal, and although Harding is right in assuming that the FOMC will rightly err on tapering Treasury purchases at a faster rate than MBSs, the static scenario he sets up likely will not be enough "forward guidance" on what the Fed has planned for the rollback. If they do not give some guidance on the taper going forward then every FOMC meeting from now on is going to be a financial media circus and lead to a lot of uncertainty.  And I make this statement knowing that Bernanke will likely want to put in place a dynamic policy that anchors expectations about Fed actions going forward, but that can be modified as needs dictate.

Harding asks, for example, "if a taper were 100% Treasuries, then does that communicate no reduction in MBS purchases until Treasuries are reduced to zero? Or just no reduction in MBS with the first tapering move – in which case how does the Fed make that clear?" 

I think that if the FOMC really deliberates on this point they will come to the conclusion that although monetary policy should be neutral by June 2014 - in other words, there is no "net" purchasing occurring in June, the only way to taper and allow for transparency is to make the rollback dynamic. That means that purchases need to be ramped back month by month, but perhaps so that Treasury purchases reverse before June, but MBSs reverse after June.  As far as I am aware, the Fed has been purchasing $40bn of MBS per month, and so must have been purchasing about $45bn of Treasuries.  That implies that if the taper is proportional roughly equal amounts of Treasuries and MBSs would not be bought.  Obviously that will have a much bigger impact on the housing market than the Fed would like at this point in the housing recovery. 

So given this, I would modify Hardings options to read:

Option 1: Bond purchases to fall by $9bn per month, and MBS purchases to fall by $1bn per month.  That means in terms of bond purchases, this will likely reverse purchases in about 5 months into sales, but MBS purchases would continue beyond the June 2014 "end" date and well into 2015. 

Option 2: Bond purchases to fall by about $7bn per month, and MBS purchases to fall by around $3bn per month.  This would put more onus on adjustment in the housing market, but would nevertheless would have the Fed selling MBSs by 2015. 

Option 3: Bond purchases to fall about $4.5bn per month and MBS purchases to fall by $4bn as well.  This would have the Fed neutralize policy by June of 2014, and then decide how much to sell going forward.

One thing I think is not known by anyone is how the Fed's buying or selling will change the market.  Why?  Because both markets are very different beasts.  The market for Treasuries is extremely international and dominated by foreigners because of the US's hegemonic place in the current economic world order.  MBSs on the other hand might be way more difficult to unload internationally, given the lessons of the finaniclal crisis, and so the buyers for these will more likely come from financial institutions based in the US.  That poses a problem for the Fed, as unloading these securities is going to require some judgement in terms of how much help the US housing market requires.  I am glad that I"m not on the FOMC making this decision today!!"


Tuesday, September 17, 2013

Taper talk and inflation expectations

Any student of economics knows from his or her money and banking course that there are two different effects that occur when you inject money into the economy.  The first is called the "liquidity effect" and it operates when the money supply is increased.  It operates in the short run when prices are sticky, so that no price adjustments take place.  Using the diagram below, you would just increase the supply of money, hence shifting the vertical M curve to the right in the diagram below.  That lowers interest rates.  Of course that should lower interest rates in normal circumstances, unless you hit extremely low interest rates in which case you could find yourself on the flat portion of the L or money demand curve.  In this case, as Keynes pointed out, you find yourself in a so-called "liquidity trap".  In a liquidity trap, increasing money supply will not lower interest rates further, so will not stimulate the economy. We used to teach this as an academic curiosity until it occurred in Japan ( - a zero bound on interest rates), but now most monetary economists realize that it is not just a curiosity - it can happen, and it did, even in the US!!

That is the reason why we have QE, or quantitative easing.  It is a way of stimulating the economy without relying on pushing official interest rates lower. In the longer run though, prices are flexible, and they adjust to changes in the money supply, according to the quantity theory of money.  The mechanism whereby this transition happens though is related to the so-called Fisher effect.  The Fischer effect basically says that higher inflation rates should be reflected one for one in higher nominal interest rates. So as inflation begins to rise after a monetary injection, at some point we should see interest rates rising. Obviously though the Fisher effect only works if you have a response in inflation.  At the moment, as the chart below shows, we really don't see too much response in inflation during 2013 ( - this includes the data release for August, released today, September 17th).
The key thing though is that it is really not actual inflation that matters as interest rates are a forward looking variable.  The interest rate is how much you charge or are charged for lending or borrowing from now into the future.  So it is really inflation expectations that are important here, as they are the equivalent forward looking variable, rather than the current level of inflation.

Luckily the Federal Reserve Bank of Cleveland has come up with some new methodology for teasing out inflation expectations from inflation swsps (a financial derivative in which investors swap a fixed payment for payments based on the CPI), which run the gamut from one to 30 years. The results of this academic work by Joseph G. Haubrich, George Pennacchi, and Peter Ritchken of the Cleveland Fed is updated every month on a special Cleveland Fed website which can be found here.

I have reproduced the current chart of Inflation expectations from the Cleveland Fed's methodology in the chart on the left.  What is striking is that if we use the ten year swaps we appear to be at a turning point in terms of expectations.  Inflation expectations now appear to be potentially moving up again. And that means that if 10 year bond rates are yielding just over 2.8%, that given that inflation expectations are roughly 2%, that the real interest rates, in other words the real gain lenders get from loaning their money out is around 0.8%.

The real interest rate is important in an economy because it signals the rewards from lending.  For very short term loans these are now negative - in other words it is not worth lending short term for most banks.  We can see this if we calculate the short term real interest rate - which is given as say a 2 year bond yield minus the expected inflation rate over a 2 year horizon.

Short term real rates are about -3%.  This means that the Fed has really pushed short term interest rates down to an incredibly low level - well we know this already from my previous blog which you can read here.

But in terms of policy implications, and what needs to happen this week at the Fed's monetary policy meeting, is that these short term lending rates need to rise to turn the real interest rate positive again.  That means that in fact the Fed should, if anything, extract much more short term credit from the market when it tapers than long term credit so as to allow short term nominal interest rates to run to more normal levels again and make it profitable to lend short term.  At the moment, in one sense, the Fed's critics are right - the Fed's monetary policy is distorting the yield curve, and the sooner the Fed extricates itself from this the better.

Sunday, September 1, 2013

Growth Cycles, Interest Rates and the Yield Curve

Watching the news this week we saw the recent ramp up in bond yields pull back with terrible US housing data.  Although the news was somewhat overshadowed by the breakdown at the NASDAQ ( - more on that some other time), it did cause stocks to resume their climb after a few days of heavy selling. Most economists think of this as "overshooting" - in other words, interest rates may have gone up too much too fast, which has caused a stall in the housing pickup leading to the Fed stating that it would be flexible if the economic sputters rather than slowly accelerates, which in turn leads bond yields to fall again.

US real GDP growth - quarterly at an annual rate
Source: BEA
But in my opinion for whatever reason what we are seeing here is the growth cycle at work. Growth does not occur in a straight line as most economists assume in their models - growth appears to occur in small cycles, and this is pretty clear from the graph of real GDP shown above.  So you might ask why we are seeing a slow down in the housing market and other data, and yet the 2nd quarter growth rate was revised upwards.  Well part of the reason is that the real GDP release is a "rear view" measure - as it only captures growth up until the end of June, and the spike in interest rates largely occurred in July.  So likely we are in for a lower real GDP growth figure for the 3rd quarter as interest rates increased in the 3rd quarter.

Some commentators are now saying that this might give the Fed pause in terms of starting it's "tapering" program to scale back purchases or bonds and other financial instruments.  I really don't think so as US growth is still positive and at a 2.5% rate, hardly sluggish now.  And remember that the "taper" only slows down the purchases, leaving the Fed the room to vary the rate of decline in these purchases if it wishes.

So "tapering" is really going to be happening sooner rather than later in my view.  But we know that expectations usually act ahead of policy actions, and indeed some of the run up in bond yields has already occurred, but I thought it would be good to review exactly what might happen to the yield curve over the next year or so given where we currently are in the business cycle.  The yield curve is a plot of the bond yields of different maturity government bonds at any given point in time.  So the exercise here is to try and deduce where bond yields should be at this stage of the business cycle without the monetary stimulus that the Fed has injected into the monetary side of the economy.

The big problem is what to compare the current level of bond yields to, so I decided to take two dates, so as to make an informed judgement about how far current yields are below where they should be at this stage of the business cycle. Given that we are now about 4-5 years since the end of the recession in June 2009, we are now 4 years from the end of the recession. Given this, I thought it would be good to compare the yield curves for roughly 4 years after the previous 2 recessions to where we are now.  Here are the results.

Yield curve comparisons
This comparison in one sense is not the best comparison as the peak of the business cycle after the 2001 recession occurred in 2006, and this can be seen as the flattened yield curve in 2006, but it does give us an idea about where interest rates should be under normal circumstances.  As can be seen, 5% appears to be roughly where bond yields should be - but current bond yields are far below this level, with only the really long term bonds now approaching roughly 1% below the 5% level.  This is due to "operation twist" ending, and the surge in long term bond yields that has occurred this year.

Here, we compare current yields to where the yield curve was about 5 years after the 1991-2 recession, and in 1997 the yield curve did have it's typical positive slope, as a recession was still a few years off in 2001, but in this instance yields were between 5 and 7%, way higher than where they are now.  Nevertheless we have the same general result - that short term rates are way lower than where they usually were at this stage of the business cycle.

The two charts really beg another question though - how will the yield curve move through time?  Obviously the "taper" implies that bonds will still be bought by the Fed, which will depress the short end of the curve for some time.  Some are betting that it's the 5 and 10 year bond yields that will move upwards most quickly now, but either way it is clear that the short end is where the biggest moves will occur once the Fed's exit strategy is fully realized.

Much will depend on economic developments over the next couple of years and how the new Fed Chairman decides to execute the Fed's exit strategy.

[Investment tips: If you want to ride the wave to higher bond yields at the shorter end of the yield curve, two ETF investment vehicles that would correspond to this movement are DTUS and TBZ, both inverse short term maturity bond ETFs ( - TBZ is a 2x leveraged ETF); if you are interested in the middle of the yield curve, TBX and TYNS are the inverse non-leveraged ETFs and TYO is the 3x leveraged equivalent]

Featured Post

Free Trade on Trial - What are the Lessons for Economists?

This election season in the US there has been an extraordinary and disturbing trend at work: vilifying free trade as a "job kille...

Popular Posts

Search This Blog