Wednesday, May 29, 2013

QE and Stock Markets: Boom or Crash?

Here is a follow up to my previous post about what to expect when Quantitative Easing is reversed. The point I made earlier is that the more likely scenario for tightening of monetary policy is that this will happen when growth gets stronger. Krugman makes the same point in his latest post. But I can also see plenty of articles talking a much more pessimistic view and describing all the bad things that will happen when central banks change their current monetary policy stance from expansionary to neutral, including a large downward adjustment to stock markets. I am surprised that there are very few historical references in those articles, so let me produce two to refresh our historical memory.

What has happened to the stock market in previous cycles after the central bank decided to abandon its policy of low interest rates and start raising those rates? I will use US data and I will focus on the last two episodes:

1. After the 1990 recession, interest rates remain low until February 1994 when the Federal Reserve started increasing rates from a level of 3% to 6.6% in May of 1995.
2. After the 2001 recession, interest rates were lowered to 1% and then started to increase in June 2004 from that level to 5.25% in June 2006.

The image below plots the evolution of the stock market in the months that followed both episodes. The Dow Jones Index is rebased to be equal to 100 the month when interest rates started going up -- I also provide the previous 12 months to get a perspective on what happened during the previous year. [Note: I stop the series when I see interest rates declining again as growth slows down in 1997 or as we enter the recession of 2007.]

In both cases we see a similar patter. Initially the stock market moves sideways. During the first year there are gains of about 1%. But the years that follow, and once the increases in interest rates have stopped we see large increases in the index -- interest rates stopped increasing in the first episode 12 months after they started going up; in the second episode it happened about 24 months later. This is more or less when the stock market starts its climb. 

As always it is hard to know what the market was thinking during those months. We clearly do not see a fall in the stock market in any of the two episodes, we seem to see a period of wait and see while interest rates are increasing but once interest rates stabilize we see significant booms in stock prices driven by the positive economic dynamics (high growth) that drove interest rates up in the first place.

Antonio Fatás

Sunday, May 26, 2013

Looking forward to the end of QE

Good piece in the FT online today by Gavyn Davies on what wil happen to markets when QE (quantitative easing) ends. As we have witnessed last week, comments on the potential end to QE can generate significant volatility in financial markets. But there is something in this debate that still confuses me and it has to do with the interpretation that some make of the end of QE.

QE will end one day, this must be the assumption of anyone who understands monetary policy. The day the recovery is strong enough, when central banks feel that the economy is close enough to full employment we will see a normalization of monetary policy conditions. From an expansionary stance we will move towards a more neutral stance. This has always been the case in the past although we used to think only in terms of interest rates and not QE. One way to see how this will happen is the chart below (which I am borrowing from a recent Brad DeLong blog post).


This is a plot of long-term (10 years) and short-term (3 months) interest rates for the US. The pattern that we see after each of the last recessions is that short-term rates deviate from long-term rates as a sign that monetary policy is hitting the accelerator (the yield curve becomes steep). A few years after the recession is over, short-term rates are raised and they reach a level similar to long-term rates. After that we see the two continuing at similar levels until the next recession comes. [Note: the chart above plots nominal interest rates so we need to be careful in our interpretation, differences in inflation over different horizons can be relevant to explain the shape; I am ignoring inflation in my analysis.]

We expect a similar patter this time, the red line will catch up with the blue line at some point in the future and before that happens we will see a statement from the US Federal Reserve that QE is about to end (or to slow down at first). What is clear from the chart is that it is taking longer than before for short-term rates to go back up to meet long-term rates, a sign that the current recovery is much weaker than previous recoveries.

What will we learn the day Ben Bernanke announces that we are starting that path towards normalization? It might be that we simply learn that he is becoming optimistic about growth in the US. This will be good news. It might not be a surprise to some who expected that type of growth going forward, but it could be a positive surprise to others that thought growth would never come back. In this scenario, it is difficult to think about such an announcement as bad news. We know that QE will end one day, we know that short-term rates will have to increase, so if the announcement was to be a surprise in the sense that it is coming too early, it would mean that there is a positive surprise in terms of growth happening early than expected -- and this has to be good news.

There is a second and more pessimistic scenario: the day Ben Bernanke announces that QE is ending we learn that the economy is not doing much better but that the FOMC has simply changed their mind. That they do not care about low growth, that they want to be tough and that they are ready to stop QE to signal a change in policy. This would be bad news because it represents a change in policy and not a change in our expectations about growth.

Understanding what will happen to markets when QE ends requires to decide about which of the two scenarios above is more likely. I personally see the first scenario more likely than the second one. I do not see a policy reversal in the near future but I do see the end of QE as good news accumulate. But this is my view, what matters is how the stock market will read the communications of the central bank. The words chosen to communicate their actions at that point as well as their credibility will make a great difference.

One final point to notice in the chart above: while the pattern of deviations between the red and blue line are clear, there is also a clear downward trend in long-term rates (in nominal terms in the chart but we know that there is a similar downward trends in real rates). As Gavyn Davies points our in his article, there is some uncertainty about this trend. Is this trend real or is it the outcome of a bubble? This is a second potential source of uncertainty. When the red line starts climbing, what will happen to the blue line? Will it stay around where it is and the two will meet at a level that will be low by historical standards? Or will we see a significant increase in long-term rates as expectations of monetary policy are challenged? I have argued before that there are fundamental reasons why long-term interest rates are low that are independent of what monetary policy is doing. But others argue that long-term rates are distorted and ready to go up when QE ends. If this is the case, then stopping QE will suddenly increase all rates up and it will certainly be be bad news for financial markets.

Antonio Fatás

Sunday, May 19, 2013

Search for yield

More on the issue of bubbles in the stock market. Here is a chart from Google Trends on the worldwide trends for the search of the term "stock market rally". An upward trend but still below the levels in October 2010 or 2011 and much lower than at the height of the financial crisis (2009-2009).


Thursday, May 16, 2013

The myth of liquidity and bubbles in financial markets

Gillian Tett at the Financial Times wonders today about how long the stock market rally will continue. Her argument is that we see data that are at odds with historical norms: the stock market goes up when unemployment is still rising in Europe, earnings come short of expectations and the rally continues. uncertainty is still high but spreads are low. How can this be? Her explanation:

"It is easy to identify the reason for this: as this data has gone haywire in the past couple of years, western central banks have been unleashing an estimated $7tn worth of quantitative easing. This has lowered interest rates on government bonds, forcing investors to search elsewhere for yield."

So it is the Central Banks and quantitative easing that is supporting the behavior of financial markets and we need to worry about how sustainable it will be. Gillian Tett refers to the 2003-2007 experience when we witnessed something similar and we know how it ended.

This is now a common argument and a concern, that the behavior of central banks and the enormous amount of liquidity that they are introducing in financial markets is simply creating mispricing and bubbles in financial markets.

While it is possible that periods of low interest rate create behavior in financial markets that is associated to bubbles and potential instability, the argument that puts all the burden in central banks and liquidity is simply wrong. In fact, articles like this one reminds me of how disconnected the analysis of prices in financial markets is from the analysis of fundamentals and it is this disconnect and what you read in the article that makes me worry about instability in asset prices. Let me explain.

The first source of confusion is to that rates and returns are low just because central bank actions. As I have argued before, interest rates are low because of a trend that started in the mid 2000s of increased saving in some emerging markets and the effects of the great recession that increased saving in advanced economies and made investment collapse.  When no one wants to invest or consume, interest rates are low. And they are unusually low this time because the patterns of investment and saving are driven by a  crisis that is very large compared to historical patterns. As a reminder, interest rates are low everywhere not just in countries where quantitative easing is taking place:

The chart above (from the most recent IMF Asia Pacific Regional Outlook) shows that real rates are negative in all countries in the Asia-Pacific region. And, sure, you can still blame Ben Bernanke for all this by arguing that the liquidity that he is creating is crossing oceans and reaching all these countries but that is just nor right. The right explanation is that in a world with fairly well integrated capital markets, returns are similar everywhere.

And this basic fact: that it is the balance between real flows that is mostly behind the abnormal behavior in financial markets is wildly misunderstood. Here is a quote from the Financial Times piece:

"(Quantitative easing) has lowered interest rates on government bonds, forcing investors to search elsewhere for yield. But what has intensified the crunch is that, while central banks have been gobbling up bonds, the supply of assets has declined."

Why is the supply of assets declining? There must be a reason why companies or individuals do not borrow (or borrow less). Equilibrium is about demand and supply. Why is it that companies are not borrowing in what seems to be extremely favorable financial conditions? When asset prices are too high and the cost of capital is below its fundamental value we should observe an excessive amount of (real) investment. But we do not see it. Isn't it more natural to think about an equilibrium where relative high saving and low investment are resulting in low returns and yields across all assets? It is, but it is probably less interesting to write an article about it.


The final confusion is about how low interest rates (or equilibrium returns) should be reflected in asset prices. If markets suddenly learn that the fundamentals of the world economy are pushing the real rate on safe assets down, we expect asset prices to increase to adjust to this new equilibrium. What we should NOT expect is that stock market returns are higher going forward. In fact, stock prices should go up so that equilibrium returns are lower across all asset classes. And this is the biggest source of confusion in the commentary we often read about the stock market rally. The question is always: "How long will the rally last?" as opposed to "Are current stock prices supported by fundamentals?" Bubbles occur when markets think in terms of arrows and not in terms of levels. Some of the rally in the stock market now, as well as in 2003-2007, is associated with the fact that the market is learning that interest rates and equilibrium returns are going to be lower for longer than what they though (this is what some call the search for yield). Asset prices increase to reflect this new equilibrium. But this increase, this "rally", leads to a reading in terms of the high short-term returns that it has generated. And then investors extrapolate and assume that they have finally found a place where yields are high! This is a bubble. When asset prices are not consistent with fundamentals.

The current value of the stock market is consistent with a world where returns are likely to remain low (by historical standards) across all assets, and this is why stock prices are high. But this story is inconsistent with a world where stock prices keep going up and generate an excessive return in the future. If we keep looking at the stock market in terms of arrows we will end up once again in a very unstable world. And it will not be the liquidity of central banks who caused it but the lack of understanding about some basic principles of asset pricing.

Antonio Fatás

Monday, May 13, 2013

Time travel in Euroland

Unfortunately, this is not news by now, but the president of the Euro group, Jeroen Dijsselbloem in an interview with CNBC yesterday dismissed the role that fiscal policy and monetary policy can have to address the economic crisis (emphasis is mine):

"Monetary policy can really not help us out of the crisis. It can take away the pressure, it can accommodate new growth, but what we really need in all countries is structural reforms in the first place. I'd just like to stress the point that in the policy mix of fiscal policy, monetary policy and structural reforms — I'd like the order to be exactly the other way around. Structural reforms in the first place, fiscal policy and viable targets in the mid-term for all regions in second place — and monetary policy can only accommodate domestic economic problems in the short-term."

It is not exactly clear what to make out of his statement but it seems that long-term solutions should come first before we implement those that will help us in the short term. It is surprising that even today there is such a great confusion about long-term versus cyclical problems.

This confusion comes from a basic belief that some hold that there is nothing inherently different in the dynamics of an economy when one looks at the short run and the long run. This is part of a never-ending academic debate but when it comes to policy makers and politicians it seems to be more a matter of beliefs.

What it is not always understood is that we are dealing with two separate problems and therefore we need two different set of tools or solutions to deal with them.

It is possible that irresponsible behavior, excessive spending and accumulation of debt (private or public) are the cause of the Great Recession. And if this is true, it will require future adjustments to spending plans, deleveraging, and fiscal discipline to avoid a repetition of this event in the future.

But once the crisis started we are dealing with a second problem: a recession that moves us away from full employment. This is a cyclical phenomenon that is well described in macroeconomic textbooks and to deal with it we use monetary and fiscal policy. The fact that potentially debt and excessive spending were the cause of this cyclical event does not mean that we need to deal with these imbalances now to get out of the crisis. We are dealing with two separate phenomena that are only related because one possibly led to the second one, but the dynamics associated with each of them are very different and the recipe to get out of them can be, in some cases, the opposite.

This is what we write in all macroeconomics textbooks: what works in the short run might not work in the long run. As an example, we emphasize the importance of saving in the long run to drive investment and growth. But when we talk about the short run we emphasize the importance of spending to understand fluctuations in economic activity. Excessive spending hurts growth in the long run but it is spending and demand what drives growth in the short run.

There will be a day when we will have to debate about whether the cyclical phenomenon has already been addressed because we are back to full employment and therefore all our focus should be on the long term, but it is very hard to argue that this is where Europe is today. My point is not to deny that there are many deep structural issues to be addressed among Euro countries, but to recognize that we are dealing with two set of dynamics that require different solutions and until we invent time traveling the short term still comes before the long term.

Antonio Fatás

Tuesday, May 7, 2013

Let's get real about the stock market.

As reported by the financial press, the stock market continues to hit fresh record-high levels in many advanced economies. The Dow Jones passed the 15,000 mark, the Nikkei just went over 14,000, and the DAX just went above its previous record. It seems to be the time to talk about bubbles in asset prices - an important issue given how these bubbles have dominated the last business cycles in these economies.

Except that we are looking at the wrong numbers. It is remarkable that the discussion on the value that these indices are reaching ignores two fundamental issues:
1. These are nominal values and as we were told in the first economics lesson, we need to look at real variables and not nominal ones.
2. Asset prices are not supposed to stay constant (in real terms). In many cases its appreciation will reflect real growth in the economy, earnings and/or the expected return that these assets should provide in equilibrium.

No need to look for data that provides a better benchmark than just nominal indices. Robert Shiller provides all the necessary data in his web site. Adjusting for inflation is easy and below is a chart with the real price of the S&P500 index where the CPI has been used to convert nominal into real variables.
















After adjusting for inflation we can see that the index is far from its peak in 2000 and it is even below the peak in 2007. No sign of a record level yet.

Adjusting for inflation is not enough as the fundamentals (earnings) also grow because of real growth. The price-to-earnings ratio takes care of this adjustment (it also takes care of inflation because earnings are measures in nominal terms). Making this adjustment is more difficult but I will reply on Shiller's numbers again (his book and writings provide a lot of detailed analysis on his data).
















Once we adjust for nominal as well as real growth, the current levels look even less impressive. Very low compared to the bubble built in the 90s and significantly lower than the ratio observed in most of the years during the 2002-2007 expansion. We are very far from record-high levels if we use this indicator.

Of course, to do a proper analysis we need to bring a lot of other factors: expected earnings growth, interest rates, risk appetite,.... And there will be room there to debate whether the current valuation of the stock market is reasonable, too high or too low. But starting the analysis with a statement of record-high levels when measured in nominal terms and ignoring real growth in earnings is clearly the wrong place to start the debate.

Antonio Fatás

Thursday, May 2, 2013

Me not working hard?

Reading a footnote from a blog post written by Noah Smith comparing GDP per capita across countries led me to look back at statistics on how labor markets and effort matters for cross-country comparisons of GDP per capita. The footnote was about one of the most successful Asian economies, Singapore, with GDP per capita clearly above that of the US when adjusted for PPP.

Using data provided by the U.S. Bureau of Labor Statistics, in 2011 Singapore has a GDP per capita which is about 25% higher than that of the US. But a comparison of GDP per hour reveals a very different picture, Singapore has a GDP per hour which is 32% lower than that of the US. Other Asian countries display a similar pattern.

GDP
Per capita
GDP
Per worker
GDP
Per hour
Korea
63
58
45
Japan
71
64
66
Singapore
126
93
68
United States
100
100
100
Levels relative to the US. Year 2011. Source: Bureau of Labor Statistics.

In the table above I compare Korea, Japan and Singapore to the US and as we move from the first column to the second and third columns the relative position of these countries worsens relative to the US. Singapore is the most extreme example with GDP per capita of 125% of the US level but GDP per hour as low as 68% of the US level. How can we explain this difference? How hard do the Singaporeans work? There are two things that matter:

1. The ratio of employment to population which is affected by both demographics and labor force participation. This explains the change from the first to the second column.

2. The average number of hours worked. This explains the difference between columns 2 and 3.

Here is the labor market data for these three Asian countries in comparison to the US:

Employment to Population
Average Annual Hours
Korea
49%
2289
Japan
50%
1726
Singapore
61%
2409
United States
45%
1758
Year 2011. Source: Bureau of Labor Statistics.


In Korea and Singapore both ratios point in the same direction: employment to population as well as hours worked are higher than in the US. In the case of Japan number of hours worked is similar but the employment to population ratio is also higher. The differences are very large, more so in Singapore, and they explain the high levels of GDP per capita relative to GDP per hour worked.

Some European countries are on the other side of this comparison, with lower effort than the US as measured by employment or number of hours. The table below shows some of these countries as well as Canada and Australia.

GDP
Per capita
GDP
Per worker
GDP
Per hour
Germany
81
73
91
Italy
66
74
73
Australia
86
75
77
Canada
84
75
77
Spain
67
76
79
Sweden
86
80
85
France
73
80
95
United States
100
100
100
Levels relative to the US. Year 2011. Source: Bureau of Labor Statistics.

The most visible case is France, which has a much lower GDP per capita than the US but a very similar level of GDP per hour (the French are very productive...when they work). Another interesting comparison is Spain and Canada where we can see a slightly higher number for GDP per hour in Spain even if there is a significant difference in favor of Canada when it comes to GDP per capita. Northern European countries (like Sweden) look very close to the US when it comes to labor markets so when you move from one column to another you see very little change in their relative position.

All of these numbers make clear that looking at GDP per capita to assess growth and convergence can be misleading in the presence of significant differences in labor markets.

A final caveat: GDP per hour is not a perfect measure of productivity either. It is ignoring the productivity of other factors and it might give a distorted picture of productivity when there are large variations in the sectoral composition of GDP -- a sector-by-sector comparison would be a much better way to assess true differences in technology.

Antonio Fatás