A Dash of Insight

August Employment Report Preview

Each month we ask the question, "What change in payroll employment would be consistent with other economic data from the same time period (the middle of the prior month)?

This is not a forecast, per se, since we do not posit any causal relationship among these variables.  They are all concomitant indicators of economic activity. 

• We use the four-week moving average of initial unemployment claims, culminating in the week of the employment survey.  This is the best direct indicator of new lob losses.  The average was 571K for this month compared to 567K last month.  No improvement.

• We look at the University of Michigan sentiment survey, which we found more useful than the Conference Board's sentiment index.  Michigan uses a panel, where some families are carried over from month to month.  This is a good technique.    Sentiment is influenced by employment.  When people have lost jobs, or are worried about losing jobs, it shows up in sentiment.  It is a good concurrent indicator.  The Michigan index is now at 65.7, about the same as last month's 66.  Normally there is very high correlation between Michigan and the Conference Board.  In the last month the Conference Board estimate improved from 47.4 to 54.1.  We are surprised at the difference between the two surveys.  (There is a good technical reason for not using both in the model -- something called multi-colinearity.  We might consider combining the two surveys).  Meanwhile, there is some "upside risk" in our estimate.

• We us the ISM manufacturing index, which showed dramatic improvement from 48.9 to 52.9  While this is modest expansion in the manufacturing sector, it is quite bullish for the overall economy.  The ISM's research shows that this rating, if annualized, corresponds to a 3.7% increase in GDP.

Our long-term record has been pretty good, especially when compared to the final revised data.  This makes sense because our model was derived from the final data.  In recent months we have been too bearish.  We are still evaluating possible reasons for this.  One possibility is unusual seasonal factors in July.

This Month's Prediction

Our indicators suggest a net job loss of over 400,000, significantly worse than the Street estimates of about 225K.  Since the 90% confidence interval on the payroll survey (for sampling error alone) is +/- 100K jobs or so, it is unwise to place much emphasis on small differences, but we are outside this range.

While economic indicators have improved, some showing actual GDP growth, all of the results are far below economic potential.  In addition, businesses may be slow to re-hire, waiting for additional confidence of a recovery.  We are many months away from really good news on employment.

Other Forecasts

It is always interesting to compare the job forecasts from different sources.  We follow several because of the interesting and widely varying methods they use.  A wise interpretation would be to consider all of  these disparate sources of information.

ADP has proprietary data because of its payroll management business.  ADP sees losses of 298K.

TrimTabs also uses real time data.  Their estimates are based upon tax deposits for salaried employees.  They see a net job loss of 335K.

WANTED Technologies, a relatively new entrant in this field, has a model based upon online help-wanted advertising.  This is an innovative and different approach to real-time data.  They see a net job loss of only 190K

Conclusion

Each of the sources we cite is attempting to measure the actual net job change.  A wise stat prof once said, "Suppose God whispered into your ear and told you the TRUTH."

The BLS is attempting to do the same thing, with dramatically different methods.  The BLS result is not TRUTH.  It is a statistical estimate.  Actual TRUTH for a specific month will not be known for many months, when the state employment data are analyzed.  The BLS tries to count all of the jobs in one month, all of the jobs in the next month, and then report the difference.  They do this very well, but it is inherently difficult.  It does not focus direclty on the actual changes, as other methods attempt to do.

Meanwhile, the forecasters will all be graded by how well they predicted the BLS number -- the BLS estimate of TRUTH.

That is the wrong attitude.  The BLS number is just another estimate -- and one which will not be official until all of the revisions are in.  Despite this, the market will trade on the preliminary estimate revealed Friday morning.

Briefly put, everyone is trying to estimate monthly changes in a work force of over 130 million.  The error band is small.  The BLS -- and all of the other sources -- are doing a great job with various differing approaches.

Market participants would like to have more data, faster data, more accurate data.

Trading Conclusion

There is a very wide spread in this month's estimates.  The non-BLS sources were all too bearish last month.

As usual, the best trade is the short side.  If the number is poor, the market sells off.  Meanwhile, if that prediction is wrong, one can depend upon the Bearish Blogging Network to attack the BLS, seasonal adjustments, the Birth/Death adjustment, or anything else to save the bacon of hedge funds shorting before the number.  Since there will be no one to refute these claims (the BLS only responds to media inquiries), the short trade has good risk/reward.

This proved out last month.  The positive number caused a modest pop in stocks.  (We covered our short hedges at a small loss.)  Then the market moved higher.  Two months ago there were rumors of a "fat finger" error, something that makes no sense to anyone who understands the reporting process.

The market continues to place too much emphasis on an estimate that is very difficult to do in real time.

Sell in September? Time for Reality!

Anyone paying attention to market news must know that September is the weakest month for stocks.  Mark Hulbert, one of our favorite writers, calls it The Cruelest Month.  He provides a table of returns, summarized as follows:

Notice from the table that in all but one of the last 11 decades, September was a below-average performer. In more than half the decades, in fact, the month's rank was dead last.

Why, given such an overwhelming record, would anyone question September's bad record? Because there is no good theory for why the month should be such an awful month for the stock market. And, without such an explanation, there's the distinct possibility that the statistical pattern is just a fluke.

As one can see, Hulbert is well aware of data mining.  He mentions the popular "butter production in Bangladesh" example.

Hulbert next considers a number of hypotheses and invites readers to share their own ideas.  His thought is that a good explanation would make the September story more convincing, although he shares the information that day traders do not wait for any hypothesis testing!  The sophisticated audience of "A Dash" may not find that very convincing.

Hulbert's Market Watch colleague, Irwin Kellner, has a number of reasons for September weakness, including the possibility of a self-fulfilling prophecy.

Why Both Articles are Wrong

Here at "A Dash" we are veteran debunkers of mythical market lore.  Most people (including our employers when we started in the business) just want to see the data --- all of it!  Traders all believe in patterns.  The more data the better.

The idea of random results is lost on most, like the "day traders" Hulbert mentions.  There are twelve months.  There will be a distribution.  Some will be good and others will be bad.  Always.

What if there is a reason?  A hypothesis does not really help.  When you already know the outcome, any smart person can invent a compelling reason.  New readers can revisit our discussion of this topic, where a group of very smart grad students were given a list of findings and asked to provide reasons.  They did very well.  Only after the class were they told that all of the relationships were reversed!

The scientific method works only when one begins with the hypothesis.

Let us try an experiment.  Instead of taking the currently constituted months of September, instead put all of the individual trading days in a basket.  (We know that this basket has a negative bias, but bear with us).  From this basket we create trading months.  From the days in the other months, we create comparison months.

This is an interesting approach to getting beyond the data mining issue.  It is (unfortunately) not our idea but that of Andrew Moe.  He writes as follows:

When comparing the months composing September to a random basket of days the results are random. Attempts to find seasons of non-randomness are frequently subject to data mining bias, as the same permutation test debunking the September drift is easily used to identify (falsely) statistically significant periods.

The study. Running a bootstrap permutation study on Dow data from 1960 to 2008 we estimate the empirical distribution of differences in monthly return between September and other months. We test the hypothesis that a random September is no more bearish than a composition of random days sampled with replacement. We find that the mean difference between populations is 0.0695%, yielding a p-value of 0.3612 – random.

Our Take

Here are four good reasons to ignore the September weakness articles.

1. This is a widely advertised theory.  Even if you do not believe in completely efficient markets, one would expect some anticipation.
2. We have already had a decline of nearly 2%, exceeding the expected monthly decline on the first day.  Should we now expect normal trading for the rest of the month?
3. The evidence shows that the September pattern is not a statistically significant deviation.
4. Other seasonal methods (Sell in May, Presidential Cycle) have not worked well in this time of turmoil

Quite obviously we do not know whether stock prices will mover higher or lower during the rest of September.  We do expect trading to reflect fundamental information about economic changes, as well as perceptions and trader lore.

The Perils of Data-Fitting

We are delighted to see the Felix Salmon piece (highlighted by Abnormal Returns) reviewing the supposed relationship between Japan in the 90's and the current US market.  Check out the article to see the "great fit."

Here at "A Dash" we have frequently exposed such  bogus relationships, even when the loud voices took the other side.  There is a simple process for generating a misleading relationship:

1. Examine a bunch of charts.  Look for variables that are all correlated to the economy in some way.
2. Find a couple of prominent features to the charts.  It is especially good if one of them (the leader) shows a move not yet reflected in the other.
3. Adjust the scales to make the prominent features line up.  Ignore any other activity.
4. Disseminate the predicted next move as widely as possible.

We exposed this process in the "Batman" chart allegedly showing a relationship between bank lending criteria and personal consumption expenditures.  This widely publicized chart was never defended by any of the original big-time proponents.  The collapse in PCE did not occur in the predicted time frame, as we showed here.  No one was keeping score.

It was an obvious case of data-fitting, although our analysis did not get the buzz that we see for Felix Salmon's piece.  Perhaps it will do better in the book version.

What You Do Not Understand

Let us assume that everyone now sees the peril of twisting old chart data to fit a pattern.  This is a lesson that can be applied to something you do not already know.

When someone reaches into history for an analogy -- -it is the same thing.  It is back-fitting of data, using the human mind, the most powerful computer.

When someone presents a trading system -- based upon many runs of old data, it is back-fitting.  The spectacular returns are not real.

Our Take

It is a shame that people are drawn to bad charts, bad analogies, and bad systems.

Too bad for them.  For the rest of us, it is a time of opportunity.

Negative Revisions in Government Data

Several observers have noted a pattern of revisions in government data -- the story always seems to get worse.  This has been particularly pronounced over the last few months.

Mike Panzner, writing at The Big Picture (one of our featured sites), does a nice job of explaining the significance of the issue.  He writes as follows:

Many market-watchers claim that U.S. economic statistics are increasingly being revised downward in subsequent periods, suggesting that the figures initially being reported by Washington are “puffed up,” so to speak, most likely for political purposes.

Panzner then does an initial check by looking at four different data series, summarized in this chart.

He draws a careful conclusion.  It summarizes both what we can see from the data and what we cannot.  Panzner writes:

Based on a quick read of a graph of the data (see below), it does seem as though the pattern of negative revisions has been trending higher lately, especially during the past year or so, suggesting that the cynics may be on to something.

That said, I am not a statistician, and the results may be nothing more than “noise.

This is an excellent introduction to a topic of interest.

In his influential weekly market letter John Mauldin generates new hypotheses.  He cites the Panzner work and also notices some other features from the chart.  His key insight pertains to the period of 2003-04, where initial estimates had upward revisions.  The Mauldin conclusion is that while there is no conspiracy, the government methods are poor at turning points and the birth/death methodology is "a wild-eyed guess based on past trends, which by definition we know will change at economic turning points."

Preliminary Analysis

First, this is an interesting and important topic.  There are some very naive people who believe that an entire government bureaucracy switches allegiance the day after an election.  There is a cottage industry aimed at convincing people to ignore data.  One of our missions at "A Dash" is to show that this is wrong.

Second, the evidence of the negative revisions is clear.  It is worth investigating, and we have been doing so for several weeks.

Third, we do not agree with the "turning point" argument.  We hope to convince John Mauldin of this in the weeks ahead.  The BLS  has several reports showing that the job creation modeling has done well with economic changes.  He is correct in his observations about the data, but we believe he is wrong about the cause.

So how does one explain the Mauldin observation about 2003-o4?

Preliminary Conclusions

In general, we like to have complete results before writing.  We are going to make an exception in this case to foster cooperation with others who are looking into the same topic.  We all seek the same objective.

We have been analyzing the payroll employment data, so that is the specific subject.  The results might apply to other series, but we do not yet know.

Our key finding is the following:  The negative revisions have little effect on the not seasonally adjusted data.  The revisions mostly relate to seasonally adjusted reports, by far the more publicized and more important.

We'll provide some data to support this in the near future.  We have a much more specific hypothesis for our research.  More later.

Significance

The most important implication of this result is, of course, that there is no conspiracy.  Anyone who understands government already knew that, but this is some additional supporting data.  If a month is revised downward because of a changed seasonal adjustment, some other month has to be revised upward.  For any twelve-month period, the unadjusted job changes and the seasonally adjusted job changes are (approximately) equal.

The second question relates to understanding seasonal adjustments.  The BLS uses the Census Bureau ARIMA X12 method, an approach with many decades of development and support.  We are doing some tests with this model, available free to any researcher, and expect to report more soon.  We wonder how anyone can have a strong opinion without some personal experience with seasonal adjustments via this model.

And finally, we wonder if the revisions will turn upward at some point.

This is a very difficult subject, and we are trying to cooperate with others who are taking a look.  It would be a great topic for thesis research.  As John Maudlin wisely suggests, it might not have much immediate payoff for those of us managing investments.

A Fresh Look at the Payroll Employment Report

Each month the market reacts and over-reacts to the payroll employment report.  The problem is that the BLS has been asked to measure the monthly change in employment.  Why?  Because that is what we want to know.

The problem?  It is an impossible job.  The BLS attempts to solve the problem in a very professional fashion, by estimating the total employment in each of two months and subtracting one from the other.  They use excellent statistical methods in their process, but even a small error in the total number of jobs leads to an apparently large error in the change. 

At some point we will have something approaching an actual count using the state unemployment data.  That is not available for six months or so and by then, no one cares any more.

The Fresh Look

A wise stat prof once explained problems like this by telling us to imagine that God whispered into our ear and told us the Correct Answer!  A statistical process is only an estimate of that answer.  The methodology, if it includes a survey, involves both sampling and non-sampling error.

The BLS attempts to learn the correct answer and announces their result on the first Friday of each month.  (See our Payroll Employment Game and play yourself to get some idea of the range of results even when one knows the Correct Answer.)

ADP has plenty of data about private firms and uses this actual data to measure monthly changes.

The NewArc Investments staff has developed a model that asks what change in jobs corresponds to other economic indicators.  Please note that the job change is not a "true" dependent variable.  Each indicator in the analysis is a measure of some other cause -- change in the economy.

The Twin Danger for Modelers

There are many problems in analyzing payroll employment, but two stand out:

1. The BLS, even though it is only doing an estimate of the Correct Answer, is deemed to have the official result.  Put another way, anyone making a prediction about non-farm payrolls is not trying to predict the truth, but rather to predict the BLS estimate of the truth.  This estimate has a confidence interval of +/- 100,000 from sampling error alone.
2. Most of us base models on the final, official data -- after both the monthly and the benchmark revisions.  This makes sense because we are attempting to represent what is actually happening, something that is not known until later.  It also makes sense because there is no predictable pattern to BLS revisions.  They come from a delayed response from some companies and eventual benchmarking.  We have seen no good method of predicting revisions, despite frequent claims by some pundits.  Claims that the BLS does revisions to improve the outlook for political reasons are just silly.

The Modelers May be Right

At "A Dash" our methods are based upon the final revisions.  We use various economic indicators and ask what job change would be consistent with things like Michigan sentiment, the ISM report, and the four-week moving average of weekly jobless claims.  Our result for this month suggests a job loss of about 200,000, a very disturbing result.

There is a problem.  One of our indicators, the initial claims series, may be distorted.  We described this today on RealMoney as follows:

Why not accept the result at face value? There was a big spike in initial claims a few weeks ago. It was attributed to the emergency extension of benefits and state efforts to communicate with those who might be eligible. The result was that many people who had been eligible for some time all showed up for first-time claims. I would expect this effect to last only for a few weeks, but claims have remained high. My guess is that the unusual circumstances caused the spike, but there has also been a build-up in claims. As a result, the four-week average is a bit higher than it otherwise would be.

To conclude, this many-faceted report is even more difficult to preview than usual. Interested readers can get more background on the methods I use in my article from last year on little known facts about the payroll employment report.

This situation causes us to have even less confidence than usual in a prediction.

ADP also works from final BLS data.  In a typically excellent article, David Altig analyzes the ADP estimate of a job loss of 33,000.  Readers should see his entire article for a complete analysis and charts, but here is his conclusion on ADP:

In words, when the ADP stat on employment growth has exceeded the initial BLS number, there has been a tendency for ultimate revisions to the official BLS job growth number that are in the upward direction.

Our Conclusion

The employment data is showing, so far, job losses consistent with slightly positive economic growth from the payroll jobs perspective.  If one uses the employment rate or hours worked (noted by Menzie Chinn), the data are at mild recession levels -- so far.

And finally, none of this has much to do with the much-maligned birth/death adjustment, which has consistently improved BLS forecasts when checked against actual data.  We expect some downward revisions from the upcoming benchmarking, but not nearly as great as those forecast by the leading critics of the BLS.

Developing and Evaluating Trading Systems

Improved technology, more power.  We would expect this to be good.

In fact, more power can enable us to do exactly the wrong thing.

This happens all of the time with the world's most powerful computer, the human mind.  A year ago we reviewed analysts who thought the market looked like a replay of the 1987 crash.  This type of analysis crops up all of the time, often using old charts as evidence.  With the power to search among thousands of choices, picking the time frame, and adjusting the scales, the human computer can "prove" nearly anything.

Those developing computer-based trading systems face the same problem.  The modern software makes it easy to include many variables --- too many!

Some Helpful Illustrations

Bill Rempel missed the Kentucky Derby by a few days, but his story highlighting horse race handicappers is excellent.  A group of handicappers were tested, using gradually increasing amounts of information.  The extra data increased their confidence, but not their performance!  (Read the entire discussion.)

Bill discusses Occam's Razor and points out the importance of reducing the number of independent variables:

I use this paring down or pruning technique at work as well as when examining trading strategies or opportunities. My first question, when faced with complex models, has for a long time been “I wonder how many of those variables actually do most of the work?”

This is pretty convincing to us, since Bill sounds just like our own Vince Castelli.  It is easy to develop a model using all of the available data and lots of variables.  You will generate a perfect "post-diction" but not anything useful for prediction.

The result:  Over-fitting and over-confidence, a dangerous brew!

Unfortunately, consumers of system strategies, including a few big-time "gatekeepers" we have met, have become accustomed to seeing eye-popping (and unrealistic) results.  They apply an automatic discount, regardless of the methodology employed.

The TCA Model Applied to the S&P 500

For the purposes of comparison, the chart below shows our TCA Model (Trend, Cycle, Anticipation) as applied to the S&P 500.  Without giving away the store, we can say that the model uses a relatively small number of variables -- some designed to choose between trend and cycle, and others representing indicators for each.  Much of the power comes from advanced techniques for filtering and smoothing data, thereby improving signal to noise.  The chart below is not a back-test, but the signals actually used in trading during the last year.

The overall performance shows a gain of about 6% during a time when the S&P declined by a few percent.  It accomplishes this while reducing risk by staying out of the market for significant periods.

A key point is that the model gets the investor into the market to enjoy the big moves.  The cost?  There are some losses at times of rapid changes or churning.

Finding the big moves is very important.  Some traders have trouble joining in when the market has already made a move.  They are reluctant to "chase."  It is difficult to show gains when missing the big rallies.

Anyone interested in trading systems should join us as regular readers of The Rempel Report, where he updates and reports on several interesting trading systems.  One of these is similar to our own sector rotation approach.

TCA-ETF Update

Each Thursday (a day late this week) we share with the investment community a recent report from our ETF ratings.  We have been doing this in real time for eight months.  Our purpose is partly to gain visibility for the approach (free report available on request), but also as information for other ETF traders, and most importantly to provide a laboratory for others trying to develop trading systems.  We discuss the issues surrounding system development in many of the articles in this series.

As we noted last week, we have expanded the ETF universe, and we seek more additions.  Adding more targets is helpful, as long as they can be shown to have characteristics suitable for one's model.

The current ratings show some dramatic changes from recent weeks, and include one of the new ETF's, KOL.

This looks good, too Good

Readers of "A Dash" do not need us to tell them the old common-sense idea, "If it sounds too good to be true...."

Or so we would think.

Meanwhile, those watching financial television are bombarded with self-serving ads from brokerage firms suggesting that the intelligent viewer can design his/her own trading system with their free software.

Other ads offer a pre-packaged approach, like Robert Taylor and the Xyber9 system.  We can't find the specific claims on the website, but the TV ads talk about accuracy in the 80's.  It is a great marketing program, with references to a Nobel Prize nomination and the Harvard Business Review.

So we wonder....

Analyzing Trading Systems

A standard method for analyzing a system includes some back testing.  Everyone does it, so the question is not "whether?" but "how?".

When we were developing and analyzing trading systems in the late 80's, there was an interesting phenomenon.

Every system we saw "predicted" the Crash of '87.

We should say "post-dicted."  Most of us now understand why, since we have all read Fooled by Randomness ( on our recommended reading list if you missed it).   If a system developer did not forecast the Crash, that system would never have seen the light of day.  It certainly was not presented to trading firms like ours.

The founder of  our group, who assembled a team of top options traders at the Chicago Board Options Exchange (CBOE), was Ralph Katz.  In the days before computer-generated option prices, Ralph could glance at a complicated page of quotes and tell you which specific prices were out of line.  Ralph understood immediately why the system development process might be wrong.  He would insist that the system developer go to a different time period, (e.g., Did the system capture the Gulf War Decline?),  one not contemplated by the original developer.

Ralph's approach is now used by the best system testers.

Analyzing Systems

When you get a system that seems "too good", it is time to take a close look.  CXO Advisory Group, one of our featured sites, does a great job of determining "guru grades."   Taylor's idea of using gravitational pull to predict stock performance was intellectually unsatisfying for us, so we asked CXO for an opinion.  They identified the strange definition of "trend" used by Taylor and concluded as follows:

In summary, Robert Taylor's accuracy rate probably derives not from forecasting   ability but from defining targets that are very hard to miss. The accuracy rate   seems high only if one ignores the peculiar way he defines trends.

Consumers Do Not Understand

The key to building a successful investment management business is marketing, and that seems to include making big claims.  The consumer is looking for a home run.  This usually means looking at what worked last year, and chasing that performance.

In a very helpful and honest article, Barry Ritholtz described the process of presenting his methods on a road show.  (Thanks to Barry for taking the time to send the pointer for an article we remembered.)  Barry wrote as follows:

Money raisers and some GPs have long ago figured this out. You have a few choices: you can answer the investors' questions honestly -- or to quote Ray Davies, you can give the people what they want (or think they want):

"We expect gains of 35-45%, with minimal risk or leverage. Our black box algorithms  have been backtested, and generate better numbers than that, but we would rather under-promise and outperform."

Read the entire article for the full flavor of this process.  Based upon our own experience, it is a very realistic account.  This method of choosing managers has led many to illusory low-risk returns based upon very high leverage and minimal true advantage (alpha).

If the big fund of fund managers are using this approach, imagine the challenge for the individual investor.

A Dash of Realism

The best investment managers, those that we admire greatly like Warren Buffett and Bill Miller, have periods where their style does not seem to be working.  That says more about the market than it does about the manager.

Any investor who wants to trade equities with little turnover (meaning low transaction costs) needs to understand what results qualify as strong.  Beating the market by several points a year is excellent, and it probably comes with some volatility.

TCA-ETF Update

Each Thursday we have been providing the rankings of the top ETF's in our universe along with a (slightly delayed) account of our actual trades.  We define a "cycle" as beginning with the last day where we were completely out of the market.  The current cycle, the third we have reported, shows results that lag the S&P (we were a little slow getting back in -- the TCA does not follow the "Gong Model" which was faster) and a little better than the Nasdaq.  There are many attractive sectors right now, including foreign ETF's and energy holdings.

 

The Use and Abuse of Data: Two Excellent Commentaries

In the current issue of Barron's Mike Santoli has some special insight for those trying to interpret the many historical analogies making the rounds.  He writes:

• Every time the U.S. unemployment rate has risen by at least 0.3% in a month, as it did in December, a recession has occurred. So a recession is a sure thing. 

• Prior to the onset of recession, there's typically at least a 25% one-year rise in weekly unemployment claims. The increase in claims in December was less than 7%. So a recession is not imminent. 

• But, then: The average decline in the S&P 500 from a pre-recession peak to a trough since 1945 has been 25%, just a few percent more than the index had lost from its 2007 peak to its intraday low Wednesday of last week. So, maybe the market has mostly discounted a typical recession scenario? 

• Hold on, because a bearishly inclined forecaster suggests that in recessions, 70% of the prior bull market's upside is undone, implying another 20% or so downside risk from today's levels. Scared yet?

• Not so fast, because every one of the 23 times since 1987 that the ratio of bears to bulls in the weekly American Association of Individual Investors poll has exceeded two (as now), the market was up 12 months later, by an average of 21%.

 

Such evidence of history's malleability is almost enough to make us all post-modern, and deny that any objective reality exists. More to the point, it's a little different every time, and the key is to figure out that difference.

These observations are excellent and the thoughtful followers of "A Dash" should read the entire column.  He reaches a conclusion that will be familiar to our readers:

But stocks have already been repriced to reflect expectations that published estimates are too high. The current S&P 500 forecast could be chopped by 10% and still market valuations wouldn't appear extreme at current prices. But the market's real-time response to muted corporate outlooks will deliver the true verdict on whether the market has built in enough fear.

The pricing of stocks compared to expected earnings and bond yields is our definition of a negativity bubble.

Barry Ritholtz, citing the Santoli column, offers some great advice for interpreting data.  He provides a list of questions--really good ones-- that anyone should ask.  Once again, readers should check out the entire article.

Applying the Ritholtz Advice

As we read Barry's advice, we immediately thought of many current "recession indicators".  We tried to illuminate this  problem with an earlier article on the problems inherent in forecasting unlikely events.  Let us try to apply the advice in this context, using several questions from the list.

• Do we have enough historical examples?  For recessions, the answer is clearly "no."  There are not that many of them, especially in recent years.  It is easy to tinker with a system to achieve a perfect retrospective model.
• Causation or correlation?  For recessions, many of the economic variables are concurrent or lagging indicators.  Most analysts look at directional movement versus absolute levels.  For example, unemployment claims are not close to prior recession levels.  They have moved higher, from some extremely low levels.
• Coincidence.  Are the factors really related?  The so-called Leading Economic Indicators come to mind.  Our own analysis finds little relationship between these indicators and payroll employment growth.  Other items, like the inverted yield curve, reflect unusual circumstances in the demand for US Treasuries, the "conundrum."
• Differentiating elements in other time periods.  This one is obvious.  Many of the recession forecasts reach deep into history.  Do we really think that markets and economic management from the Taft administration or the Great Depression are similar to current conditions?  What about corporate improvements in inventory control?  Is it possible that using computers makes a difference?
• Compare interest rates, inflation, dividend yield, P/E contraction or expansion, sentiment, overall market trend, business cycles -- across different eras. Might that account for potentially different outcomes?  Yes!  Circumstances in the early 70's, with very high interest rates, were different from now.  Another example: The Fed interest rate actions in the bubble era were much slower than they are now.
• Consider things in terms of probabilities, not outcomes.  Excellent!  We wonder how anyone can state with 100% certainty that we are "now in a recession."  Or the opposite.  If a recession occurs, something that no one knows yet, the actual dating of the beginning by the NBER will be the prior peak.  It might include the current period.  If there is a bounce in the economy, it will not.  Many of those so certain about a recession  have been (mis-)predicting the economy for years.  There is no statute of limitations.  The media points out that these people saw the problem early, as if it were praiseworthy to miss major market gains.
  

Conclusion

These two excellent articles are good resources.  Any time one sees a simple historical analogy, it would be wise to check out Barry's list of questions. 

Using Gambling to Learn about Investing

At "A Dash" we have tried to draw a clear distinction between investing and gambling.  We believe that the analysis of gambling situations frequently provides more data, allowing a long-run analysis.  The techniques used are the same that investors should use in evaluating stocks and sectors.

The Danger

Those buying stocks, whether their time frame is short-term or long-term, frequently focus on the return rather than the risk.  This can take two distinct forms:

1. Believing in the apparent consistency of returns.  This has been costly for investors in mortgage securities, either directly or through hedge funds that leveraged these instruments.
2. Swinging for the fences.  The investor looks at the potential gain rather than the risk.

The Gambling Lesson

Seeing the risk/reward error is extremely difficult when one already has an investment position.  The  gambling comparison can help one take a completely different perspective -- one where there is no  psychic stake in the outcome.

Accrued Interest  provides several lessons on this theme in an excellent post today.  The points are difficult to summarize, so we recommend reading the entire article.

Tom Murcko at InvestorGuide.com also provides a thoughtful comparison  of investing and gambling,  including a good intellectual framework.

Our own comparison is also helpful.

Conclusion

Looking for understanding from different, but analogous situations, is a strong method for improving investment performance.

Many approaches are working well, but investors need to choose with a realistic assessment of risks.

ETF Update: Are you missing the rally?

How has your portfolio been doing?   If you are not happy with the answer, it might be time to ask why.

Best Use of ETF's

We know that many investors and fund managers have been caught in the negativity bubble.  They focus on a macro-economic debate, mostly between economists and non-economists.  Regular readers know that we have tried to expose this.

Meanwhile, the market is conveying a very different message, as noted in Dr. Brett Steenbarger's must-read article, "Why Aren't Stocks Forecasting Recession?"

Those following a disciplined investment method, adjusting for market cap size and style, have captured the move.  My friend and RealMoney colleague, Steve Birenberg of Northlake Capital Management, graciously provides occasional updates for his system.  On September 5th, he wrote on RealMoney (subscription required and worth it) as follows:

For the third consecutive month there were no changes to my Market Cap and Style models. The signals remain large cap and growth. Client portfolios own the S&P 500 Spyder (SPY) and the iShares Russell 1000 Growth (IWF) to take advantage of the current signals. The large cap signal from the Market Cap model remains one of the strongest readings in the monthly data I have going back to 1980. All ten factors covering a breadth of economic, interest rate, and stock market technical indicators are flashing a large cap signal. On their own, each factor has in the past shown predictive ability for anticipating relative performance of large caps vs. small caps as measured by the S&P 500 and Russell 2000. With the weight of the evidence from a broad array of previously accurate indicators lined up so solidly in favor of large caps, I feel very good about the prospects for large cap outperformance to continue for at least a few more months.

Our Own Method

Our own TCA-ETF model has captured another dimension, the decline of the dollar.  The effect is described in a good column from MarketWatch.  Since we use an iShares universe that includes foreign ETF's, we capture all of these effects through computer-based technical analysis.  This is the complete set of rankings, showing our actual holdings, entry dates, and returns through yesterday.   (you can click to enlarge)

This is a complete set of sector ratings.  The IEZ position was closed and IXC purchased because of the relative rankings.  The sectors marked with an asterisk are in what we call "The Penalty Box," stopped out for violating specific criteria.

The Sector Message

The most important conclusion from the rankings is that almost forty sectors qualify as a "buy" based upon our analysis.  The positive expectancy is broad-based.

The next message is that certain themes -- especially foreign ETF's, energy, and basic materials, show the greatest strength.

Why Not All Top Sectors

Our investment discipline attends to risk considerations and diversification, so we have limits on sector concentration.  Overweighting some groups is an element in beating the market, but risk control is important.

Stops and Exits

Part of risk control is the use of "stops."  This is a tricky topic, worthy of a separate analysis, so it will be part of a future weekly update.  All good things have an end, and exits are important.  The perceptive Adam Warner highlights the risk in one of our key holdings.

Fundamentals

Some investors focused on fundamentals have enjoyed similar gains.  Unlike our partnerships, our individual accounts include no foreign ETF's but there is considerable overlap in holdings.  Many US corporations have significant earnings from abroad and strong earnings prospects.

At "A Dash" we view fundamentals in terms of forward earnings prospects versus alternative investments.  The market is recognizing that bonds and real estate have weak prospects relative to stocks.  This is another way of saying "Fed Model," a valuation topic familiar to regular readers.

Implications for Investors

There are many traps for the investor.  Spending a lot of time reading pundits who start with the conclusion and then find the evidence is the single biggest trap.  Amazingly, some of these sites are the most popular and highest-rated on the Internet.  Go figure.

Another trap comes from the financial media.  The stories seem always to focus on the worries.  That seems to get reader attention.  Few provide good information on the multi-year growth of earnings -- far outstripping stock performance.

Meanwhile, the online brokerage commercials are a huge trap.  They lead the investor to think that his/her own market "feel" can beat the experts.  Some even encourage one to develop and backtest a trading system.  The main purpose of our TCA-ETF series is to describe the challenges in system development and testing, including dealing with the inevitable losing streaks.  Most people trying this at home will make all of the common mistakes and get terrible results, as described here.

A thoughtful reader can learn to do better.