Monday, July 27, 2009

The American Century? Hardly.

There is a great website that I discovered called . It has all kinds of data including economic data for lots of different countries. I was able to make the following plot of GDP per capita from 1820 to the present for a few different countries: US, Germany, France, Japan, Canada, China, India and Brazil.

The 20th century has often been called The American Century. The US did indeed do well over this century. They became the foremost economic and military power. However, when you look at GDP per capita (a measure of the wealth of a nation), you find that they don't exactly stand out. For example, France and Germany didn't grow their GDP must slower than the US over this period. In fact, they started out a few percent behind the US and ended up about 30% lower. Ok, but they were both nearly destroyed by world war. Germany lost both wars and suffered a period of hyperinflation in between. They both suffered from stagnant population growth unlike the US which grew population rapidly. In addition, the US had more abundant natural resources. Canada basically tracked the US. No one calls it the Canadian Century.

However, If GDP growth is your metric for naming centuries, you would have to choose Japan. Japan is basically equal to the US in terms of present day GDP per capita. But in 1820, GDP per capita was half that of the US. In 1900, it was 28% of the US. In 1950 (after losing WWII to the US) it was only 20% of the US. Now they are tied. So the GDP growth crown for the last century and especially the past 50 years has to go to Japan.

The darlings of today's investment crowd, China, India and Brazil have lagged miserably. In fact India has hardly grown real GDP per capita at all in almost 200 years. Brazil has done OK, since 1900 but was quite far behind.

Wednesday, July 22, 2009

Swine Flu

Lets make some numerical estimates about how severe swine flu could be for the US.

Start with the US population which is about 300 million people.

Lets discuss two scenarios. One is the mild scenario and the other is the severe one. Obviously, it could be anywhere in between and even outside of these two bounds. But I would say, to my knowledge, that these two scenarios contain about 80% of the possibilities.

Lets pick a time scale. Assume it lasts about two years with some tail before and beyond that.

Now we need to estimate the attack rate. What percentage of the population will get it? From what I have read, I think this is between 20% and 50% of the population. These are my mild and severe scenarios.

Then you can discuss things related to the severity of the illness. For example, what percent of people infected will have it severe enough to require hospitalization. What percentage of those hospitalized will require a ventilator? What percentage of those infected will die.

Lets start with the death rate since it is pretty well studied wit this current strain. Most think it is between 0.2% and 0.5%. Obviously this is dependent on the quality of health care received which will be affected by the number of people that get dangerously sick compared to the heath care resources available.

So right off, we can estimate the number of people that will die using the mild and severe scenarios. This is

MIld case: 300 million x 0.2 x 0.001 =60 thousand Americans or 30 thousand per year
Severe case: 750 thousand Americans or 375 thousand per year

Obviously that is a huge difference. Consider that this flu seems to be more severe for young people rather than the usual target for influenza which is infants and the elderly. See age distribution below from a study in Mexico.

Lets look at the three age groups, 5-14, 15-24 and 25-34 who will probably suffer about half the case of H1N1. Lets look at their morbidity statistics, that is the chance of dying in any one year. This is usually done in deaths per 100,000 people. There are some numbers here for these age groups. It has all causes and individual causes including Influenza (which is linked with Pneumonia, see below).

So lets translate our numbers above in morbidity statistics and add them to these numbers. Our death rate per 100,000 individual per year (for each of the two years) is 100,000 x 0.2 x 0.001 x 0.5 = 10 for the mild case and 100,000 x 0.5 x 0.005 x 0.5 = 125 for the severe case. Lets put these in the table with All causes and normal influenza.

Age GroupAll causesInfluenzaH1N1 mildH1N1 severe
5-1421.7 0.4 10 125
15-2489.6 0.6 10 125
25-34126.7 1.4 10 125

Even for the mild case, the number of people in these age groups that will die of influenza will be 7 to 25 times higher than usual. It will be the major cause of death next year for this age group. The mild case will increase the change of a 5-14 dying (of any cause) by 31%. The severe case would double the chance of dying for 25-34 year olds. For 5-14 year olds, the chance of dying has increases by a factor of 6.7.

People often say that the flu kills 36,000 people per year so this H1N1 is nothing to worry about. But this is wrong. First of all, as I mentioned Pneumonia is linked with Flu and "influenza-like illness". Flu by itself actually kills less than 1000 per year and less about 142 children under the age of 18. source .

Using these number above and assuming the average high school has about 1000 students, the mild case would imply that one in every five high schools will have a swine flu death (in two years). The severe case would say that every school would average 2.5.

Now on to another issue: health care and its ability to handle the situation. I will only focus on one issue, the availability of ventilators. Severe cases of swine flu require the patient to put on a ventilator in order for them to breath, sometimes for weeks. The US has about 100,000 ventilators and 80,000 are in use in normal times. During flu season, most of the others are in use as well. Lets look at how swine flu could impact this. For the mild case, we assumed that 0.1% of people who get swine flu will die. Lets assume that 10 times this amount are put on ventilators so that 90% of people that are put on ventilators will live. So this is 1% of swine flu cases requiring ventilators. If we assume that each requires it for a week, this works out to be 3,000 people needing ventilators for swine flu if you evenly distribute it over two years. However it won;t be evenly distributed. It will probably peak about three times this average so that is 9,000 ventilators. This is beginning to stress the system but is probably manageable with some rationing. For the severe case, there are five times more cases which means 14,000 ventilators are needed (evenly distributed) and maybe 42,000 in peak periods. Now we are talking major panic. There will have to be major rationing. The system is completely overwhelmed. Consider that the ventilators will not likely be in the optimal places. What if New York has too many but San Francisco has too few. Imagine the chaos and ethical dilemmas that hospital and politicians will face. Canadian hospitals have been reporting that young people suffering from swine flu have required a special kind of ventilator call an oscillatory ventilator rather than the usual kind. How many of those do we have?

Other questions to consider. How many does of Tamiflu and Relenza do we have? If we run out of that, the death rate will surely rise. There is the issue of vaccines. Hopefully these will be finished in time to slow the spread. I have my doubts.

Finally, there is the effect on the economy. With chaos ensuing and fear rising, most people will not fell like shopping. Many will stay home from work to take care of sick family members. Many will be sick themselves and some of course will die. The effect on the economy means, less spending, less consumption, less production.


The Washington Post has a story on the stress on the health care system.

Consider the following statistics which they give

The swine flu virus, also known as H1N1, could infect up to half the U.S. population, making as many as 1.8 million sick enough to need hospitalization, including as many as 300,000 who might need intensive care, according to a presidential advisory council estimate.
... "There will be millions and millions of people seeking care in a relatively short period of time," said Eric Toner of the University of Pittsburgh's Center for Biosecurity, noting that the nation has only about 85,000 critical-care beds.

What he didn't say is that most critical care beds are being used at all times. Hospitals don't plan to have many more beds than are usually needed. That would be too costly. So we might assume, optimistically, that 70% on average are being used at any given time which leaves only 26,000 free beds for possibly 300,000 critically ill people. Now, it might not be that bad. But even if it is 100,000 critically ill people with 33,000 available beds, that will be a disaster. Two-thirds of the sick people can't get an ICU bed and may die because of it. Now consider that as much as half of the healthcare workers could be out of work due to infection of themselves or their dependents. Can they even operate at 100% capacity with half the workforce and three times as many people needing emergency ICU care?

----------------- Update --------------

Looks like there is more evidence on the death rate being much, much smaller than first thought. Looks like it is more like 0.1% or maybe even 0.05%. So this would be even milder than my mild scenario. Lets hope this is correct.

Sunday, July 19, 2009

US Private Consumption

Here are the numbers for private consumption, in $ Trillion, for the US and some other areas

US 10.1
Eurozone 7.6
Japan 2.8
UK 1.6
China 1.6
India 0.6
Korea 0.5

Source: CEIC

Many economists expect the US, Europe and Japan to slow their consumption after this great loss in wealth and also due to demographics. They also expect emerging markets like China, India etc to pick up some of the slack.

Those three combined are $20.5T. Lets says, they reduce consumption by 10%. That is a loss in demand of $2T. If China, India and Korea were to pick up all of this slack, they would have to increase their private consumption by 76%. They would have to do this while their export markets (their largest employer) are undergoing this historic bust. Does that make sense to anyone? Of course not. There is no other consumer capable of replacing the US and European consumers. The conclusion is unavoidable excess capacity.

Saturday, July 18, 2009

Real Estate Declines in Chicago

My wife and I visited some friends of ours in nearby Portage Park, a neighborhood in Chicago. Portage Park has long been an immigrant community, mostly Polish in the past but now hispanics are prevalent there as well. The Park itself is rather nice and the neighborhood may be considered up-and-coming if not exactly there yet.

Our friends have had a foreclosed home right next door that had been on and off the market a few timed over the past couple of years. It has recently sold. Our friends couldn't remember exactly the sale price but they thought it was around $130K. They also told us that it once sold for $450K near the peak of the housing bubble. I thought that sounded a little far-fetched so I went to to try to find the sales records. Yup. he was right. Here it is.

4933 W. Byron
Sale History
05/04/2009: $134,000
03/21/2007: $450,000
05/15/2006: $390,000

That is a 70% decline between actual sales in two years.

Here are some more examples in that neighborhood.

For example, here is 4136 N. Monitor Ave. .
Sale History:
05/06/2009: $134,000
01/17/2007: $480,000
12/17/1998: $189,000

That is a 72% decline in actual selling price in a little over two years!

5146 W. Dakin St.
Sale History
06/23/2009: $90,000
11/03/2004: $289,000

A 69% decline between sales and that is only 2004, two years before the bubble peaked.

5706 W. Wilson Ave.
Sale History
01/22/2009: $174,000
06/29/2006: $410,000
12/08/2005: $337,000

A 57% decline in 2.5 years.

5919 W. Warwick
Sale History
09/17/2008: $5,000
06/25/2004: $340,000
08/29/2002: $240,000

This one takes the cake, a 98.5% decline! Ok, that might have had a tax lien or something but still.

6028 W. School St.
Sale History
03/23/2009: $118,000
09/11/2006: $425,000

A 72% decline in 2.5 years.

Amazing stuff. This area must have been big into the subprime lending fiasco. Most of these houses must have been foreclosures that were auctioned off or sold by the bank.

Wednesday, July 15, 2009

Long Beach Ports Traffic in June

Here is a plot of the container traffic in June at Long Beach Ports which is America's second busiest seaport. Units are Twenty Foot Equivalent Units (TEU)s.

The number of TEU loaded in, a measure of imports, has declined 28% from last June and 39% from June 2007. This is a major contraction of imports and shows how dramatically the US consumers has put an end to their 20 year shopping spree.

Monday, July 6, 2009

Earnings and PE ratios in the Great Depression

Stock prices have risen about 40% from the low a few months ago. The main reason appears to be that people think the recession will end soon and that stocks are simply pretty cheap. Though I disagree about the recession ending, I would have to agree that stock look pretty cheap with the emphasis on the look.

For example the 30 stocks in the Dow Jones Industrial average have an average PE ratio of about 13. It is about the same whether you use trailing earning or forward analyst estimates for next year. Well 13 is indeed pretty cheap compared to the long term average of about 15.

But looks can be deceiving. The question is what the earnings will be in the future. Corporate profits as percent of GDP were recently at an extreme value. The long term trend for real earnings (as compiled by Robert Shiller) puts the S&P 500 earnings at about 50. That would be the trend value not the value one might expect in a very deep recession. Typically earnings fall to about half the trend during deep recessions. So earnings for the S&P 500 might be more like 30 over the next few years. With the S&P 500 at 900 that would put the PE ratio at about 30. Not so cheap looking anymore.

Let's look at the data for the Depression years 1928-1936 as an example.

This plot shows the aggregate earnings (black), stocks prices (blue) and the (trailing) PE ratio in red. The data is from Robert Shiller's data-sets. The stock prices have been divided by 17.

In late 1929, the bubble popped and stock prices fell as market participants realized that corporate earnings would fall and that stocks were overvalued. The PE ratio fell quickly from about 20 in 1929 to about 13 by the beginning of 1930. People naturally expected a quick economic recovery and decided that stocks were cheap. After-all, a PE of 13 (just like today) is pretty enticing. So began the rally of 1930. The PE ratio jumped up to about 18 before the rally fell apart. The market then fell steadily with occasional rallies until the bottom in late 1932. The amazing thing is that the PE ratios never strayed far from 17 during the rest of the bear market except for right at the end when it fell quickly to bottom at 10. The blue line is price over 17 so the fact that the blue line traces the black line demonstrates this. It is just that earnings fell steadily until they fell to about 7 from the peak at 26, a 73% decline. The story of the great bear market was really one of corporate earnings not valuation.

The moral of the story of course is that a PE ratio is pretty meaningless by itself. What matters is where earnings will be in the future. The best guide to that is simply mean reversion. Corporate profits tend to revert to about 6% of GDP. They recently peaked at almost twice that. If they bottom at half this average that would be a 75% fall just like in the Depression. If that happens we might expect the S&P to fall to at least a PE of 15. The S&P 500 earnings peaked about about 85 so a 75% decline would be 21 and a 15 PE would bring the index to a horrifying 315 a 65% decline from here.

Of course this might be overly pessimistic or it might not. But the point is that stocks will follow future earnings and that those will likely go down from here. Stocks are not nearly as cheap as they look.

Wednesday, July 1, 2009

The savings rate

The US personal savings rate might be the key indicator to tell us when the economy is at the bottom. Basically the savings rate needs to increase from its bottom of about zero in 2006 to its historical average of around 8%. In fact it may need to stay above 10% for a few years to rebuild much of the wealth that was squandered over the past decade.

So here is the data from the government's Bureau of Economic Analysis (BEA).

The good news is that we almost half way there. The bad news is that this is probably incorrect. Trim Tabs, an independent economic research firm says that the BEA data is flawed for a few different reasons. The main reason is that the BEA uses income data that is six months old. The real time data that Trim Tabs uses shows that incomes have been falling rapidly over the past few months. So the savings rate is actually more like 0.9% and not 6.9%. In other words, we haven't made any progress on saving. This is Irving Fisher's "paradox of thrift" in action. If everyone tries to save at once, incomes fall even faster and no one makes any progress on reducing their debt burden. That is probably an over-simplification but the point is that it does not appear that American's have made much progress on repairing their balance sheets. It does not appear that spending is reaching the bottom. Keep in mind that consumer spending accounts for roughly 70% of the US economy. Another 10% decline in consumer spending would take another 7% out of GDP. This doesn't bode well for an economic recovery any time soon. In fact we may only be half way through this recession.

trimtabs BLS