The Madness of Methodology

April 28th, 2013

A fight between economists is not as exciting as a dinosaur smackdown (Jurassic Park), but the controversy can be as damaging.  Politicians and pundits love to trot out those economic studies and theories which justify their actions or political point of view.  In 2009, two economists Carmen Reinhart and Kenneth Rogoff (now affectionately known as RR), published a study which showed that a country’s GDP growth becomes slightly negative when its debt grows above 90% of its GDP.  The study was cited by many politicians and pundits in Europe and the US, including VP candidate Paul Ryan, as they proposed various forms of austerity to curb the explosive growth of national debt.

Here’s what the debt to GDP ratio looked like 1940 – 1960

In the years 1947 – 1959, we had an annualized growth rate of 3.6% but a strong component of this growth was our strategic advantage in exports, being the manufacturing capital of the world after much of the production capacity of the developed world was destroyed in WW2.

Here’s what it looks like now; the same spike of debt.

But we have lost the advantage of being the leading manufacturer.

Given the assignment of replicating an existing economic study, Thomas Herndon, a PhD candidate at UMass, discovered some glaring spreadsheet errors in the original data set compiled by RR.  You can read an Alternet article summarizing the details here.

Some quick background.  There are two categories of economic policies.  Fiscal policy encompasses taxing and spending measures by a government.  Monetary policy is conducted by a country’s central bank and are targeted at the supply of money and interest rates.  Economists argue over which policy is more effective in a given circumstance.  Each of us goes about our daily lives under the influence of both fiscal and monetary policy. 

During the 1930s depression, the economist John Maynard Keynes proposed that governments borrow and spend money during recessions to make up for the lack of aggregate demand in the economy.  After the economy recovered, governments would then raise taxes to pay back the borrowed money.  Another leading economist, James Buchanan, predicted that nations who followed Keynes’ ideas would have permanent deficits.  While Keynes’ economic model was elegant, Buchanan argued that there was no incentive for a politician to raise taxes.

In 1963, with the publication of A Monetary History of the U.S., economists Milton Friedman and Anna Schwartz argued that the Depression had been largely a result of failed monetary policy by central banks.  During the 1970s, when government fiscal policies of increasing intervention in the economy failed to ingnite growth or curb inflation, Keynes’ policies fell into disfavor. 

The age old debate about the effectiveness of fiscal and monetary policy never dies. The recession that began in 2008 revived Keynes’ ideas.  In the late 1990s and early 2000s, economist Paul Krugman and Federal Reserve chairman Ben Bernanke were proponents of monetary solutions for Japan’s moribund economy.  As the world economy imploded in 2008, both men changed course and became advocates for fiscal policy as the most effective solution for the country’s economic woes.

In a recently published paper UMass professors Michael Ash and Robert Pollin (Herndon’s advisors), explained their methodology and took RR to task for their lack of follow up on incomplete data analysis after several years.  What they had missed was a follow up paper by RR in February 2011 and another published in the summer of 2012.  In these papers, RR modified their initial findings, saying that GDP growth slowed but did not necessarily turn negative.

In a WSJ blog post , RR answered the critique from the UMass Professors.  They admitted their spreadsheet error but reaffirmed their other assumptions in the study and their amended conclusions.

Paul Krugman weighed in (or waded in?), voicing his disappointment with RR’s methodology and their conclusion.  Krugman does make a point oft repeated in the social studies: correlation is not causation.  Does high debt cause slow GDP growth?  Or, does slow GDP growth cause high debt?  Or can we say that there is some indication that they accompany each other?

At Econbrowser, U. Cal professor James Hamilton, reviewed RR’s methodology and Ash and Pollin’s critique. (Link)  To which, Professors Ash and Pollin responded with some good points.

Ash and Pollin have made the original data available.  Some have accused RR of purposefully leaving five countries out of their data, saying that these five countries would have weakened or invalidated their findings.  The Excel file shows that this was a simple – but dumb – mistake, not some nefarious plan by RR.  The countries left out are on the last five worksheets which are arranged in alphabetical order.  What surprises many is that two prominent economists could publish a paper based on work that had so little verification before publication. 

What I question is RR’s decision to include many of the smaller countries at all in their analysis.  Finland and Ireland each have less than 2% of the GDP of the U.S. 

What I do hope is that this controversy will spur more analysis of the relationship between a nation’s debt load and its economic growth.  What I am afraid of is that this will discourage researchers from sharing their working data.  Reinhart and Rogoff are to be highly commended for doing so.

Readin’, Writin’ and Arithmetic

April 21, 2013

In any lively discussion of public education – its effectiveness, the spending and taxes required – some people bring out their swords, others their shields, and some are armed with both.  Armed only with a crayon, I will examine some of these trends.

Let’s look first at higher education spending.  The National Center for Education Statistics (NCES) at the U.S. Dept of Education reported that real – that is, inflation adjusted – spending per pupil had increased 233% in the past 31 years, an annual growth rate in real dollars of 2.8%.

NCES reports a slower spending growth in K-12 education – 185% in 28 years, or an annual growth rate of 2.2%.

But the annual growth rate during the past decade, 1999 – 2009,  has slowed to just under 2%.


When we zoom in on the spending growth during the 1960s and 1970s, we see a real growth rate of 3.6%

What we see in the per pupil data is a gradual slowing down of the real growth rate of spending.  Those who claim that there have been spending cuts in education have not looked at the data.  There have been no cuts in real spending, only reductions in the rate of growth. 

Some decry “austerity” policies recently undertaken in some European countries – the U.K. is an example – claiming that a country pursuing these policies has cut spending.  When we look at the spending data, we find that there have been no decreases in real spending, only in the growth of spending.  This misconception is common and results from a comparison of what we expect and what happens

If we have usually received a wage or salary increase of 3% each year, we come to expect a 3% increase.  If we get a 2% increase this year, it is 33% less than our expections and feels like a cut.  A retiree who has become accustomed to an annual 8% return on her investments, may feel that she has lost money if her investments only gain 5% this year.  It does no good to mention that she has really not lost anything.

Let’s get up in our hot air balloons and travel to California, where the size of its economy puts the state above many  countries.  California has often been the leading edge of trends that spread to the other states.  Ed-Data reports that per pupil spending has flattened since the recession started in 2008.  In real dollars, there has been NO GROWTH in per pupil spending in the past ten years.

Another complaint from teachers is that money is increasingly being spent on administrative costs, not teaching.  In California, teachers still command the lion’s share of spending  – more than 60%.

The proportion of teacher spending has remained relatively constant – above 60% – in the past ten years.

What has been growing?  On a per pupil basis, “Services and Other Operating Expenses” have grown 4% per year, or 1.8% real annual growth,  above the 2.2% annual growth in inflation.  Administration expenses have grown at the same rate of inflation so that real growth has been flat.  However, spending on teacher salaries has declined in real money at an annual rate of .7%.  However, their benefits expenses have grown 1.4% annually in real dollars.  Again, most people do not “feel” the cost of a benefit increase.  The bottom line to most of us is what we bring home.  It does not pay to tell a K-12 teacher that they are actually receiving a slight increase in real total compensation.

In California, as in many states, property taxes are a major component of revenues for K-12 education.  Over the past nine years, revenues from property taxes for education have declined 3% annually in real money.  For each student, there is $500 less money available from property taxes than it would have been if property tax revenues had kept up with inflation.  As a percent of total revenue for K-12 education, property taxes make up a little over 60%.

In 2011-2012, property tax revenues essentially paid teacher salaries.  Ten years ago, the percentage of revenues from property taxes was about 6% higher.

Other State revenues have had to make up for the shortfall in property taxes; the gap is about $1000 per student.  The problem would be even worse if it were not for the slight decline in students for the past 8 years.

While California faces challenges from declining property tax revenues, what about the rest of the country?  Let’s climb back in our data balloon and look at student enrollment throughout the country.  The NCES reports the same slight decline in K-12 enrollment.  However, they estimate a total 6% growth in K-12 enrollment in this decade.

As K-12 enrollment grew by a little more than 1 million in the 2000s, post secondary education enrollment grew by 6 million, or 37%, to over 21 million. (Source  The growth rate in older students, those aged 25+ is even faster, rising 42%. In this decade, “NCES projects a rise of 11 percent in enrollments of students under 25, and a rise of 20 percent in enrollments of students 25 and over.”

 The ratio of K-12 students to post-12 students was 28% in 2000; a decade later, it was 38%.  While K-12 enrollment is projected to increase for the rest of the decade, post-12 enrollment is estimated to be much faster.  How do these students pay for college?  The most recent data from NCES is at the start of the recession; I would guess that the need for aid has grown mightily since then. 

Put all of this in the blender: a declining work force (see my blog two weeks ago), a generational swelling of older people retiring, recovering but not robust state and local revenues, and more demand for K-12 AND post secondary education services.  How will politicians react in the midst of so many competing demands for money?

The increasing pressures for money from different segments of the population puts us in the precarious position that we can not afford to go into a recession, an impossible situation since the normal business cycle includes a recession every 7 – 10 years.  Europe is already in recession; China’s growth is still robust but slowing; on Friday, India announced a growth rate below 5%, the weakest in four years; in a hopeful sign, Brazil, the economic powerhouse of South America, is projecting GDP growth over 3%, rising up from an anemic 2.7% growth of the past 5 years.  (World Bank source)

Slackening demand around the world presents challenges for the U.S. economy, problems that a spastic Congress will only worsen. Y’all be careful out there…

Things That Spring

April 14th, 2013

Across the land, springtime wakens the trees and flowers, birds chirp and squirrels chatter.  From the buildings where the humans live comes the wailing and gnashing of teeth as many procrastinators spend this last weekend before the tax deadline in a spring ritual of angst.  The lost W-2 form is finally found beneath the Netflix DVD that has lain casually on the bookcase, waiting to be watched.  The 1099DIV form is found beneath a birthday card that was never sent.

Lay aside your problems; let’s climb inside the hot air balloon and look at the big picture.  A few weeks ago, economic growth for the fourth quarter of 2012 was revised marginally higher into positive territory, but dropping from the annualized growth rate of 3.1% in the 3rd quarter of 2012.  Let’s look at GDP from a per person basis since WW2.  Until the recession hit in late 2007, economic growth had consistently outpaced population growth.  Then POOF! went the economy and blew away a big gap in GDP.

Let’s zoom in on the past ten years to see the effect.  On a per person basis, the gap is $5,000 of spending that simply didn’t get spent.

Call it the GDP dust bowl of the 2000s, similar to the dust bowl of the 1930s when the wind blew the top soil from the prairie of the Oklahoma panhandle and forced many families from their farms.  In this case, the wind blew away a lot of jobs and chunks of home equity.

Policy makers in Washington want to close that $5000 per person gap.  If they could write a law forcing everyone to spend that $5000, they would.  Instead, they keep giving away money in unemployment benefits, food stamps, disability benefits, crop subsidies – all to keep people from not spending even less and making the problem worse.

Retail sales account for about 1/3rd of the total economy.  Including automobile sales and parts, consumers are still below twenty year averages.

This past Friday, the monthly report on retail sales showed little change from the past month.  When we look at per person real retail and food sales and take out automotive sales we get a feel for core sales, those that we make on a frequent basis.  Once again, we see the same gap that we saw in GDP.  Since mid-2009, this core consumer spending has grown 2.3% annually, above the 1.8% annual growth trend from 1992 through 2006, but it still down $2000 a year from what we would have spent if we had stayed on the same trend line before this past recession hit.

To make it a bit clearer, let’s look again at that chart and compare the 15 year annual growth rate from 1992 to the longer 21 year growth rate.  It has fallen from 1.8% to 1.1% annual growth.

GDP measures spending; let’s look at Gross Domestic Income, or GDI.  A fundamental principles of economics is that it takes money to spend money.  A six year old asks a parent “Why can’t we just go out and get more money?” to which the parent replies “Whaddya think money grows on trees?!”  End of Chapter One in the Parent’s Guide to Economics.

When we compare the country’s income to spending, we find that a dip in income below production precedes recessions.

After the 2008 – 2009 crash and recovery in national income and spending, both are limping along.

A few weeks ago came the monthly New Orders, an indication of business confidence.  As regular readers know, I have been watching this declining trend since September of last year, when the percent change in New Orders was negative.  The recent rise has been a welcome sign of growing confidence but new orders fell 2.7% in February and now hover around the zero growth line. 

On a quarterly basis, the year over year (y-o-y) percent change is still firmly in negative territory, meaning that businesses are not putting up more money to invest in new equipment.  Why?  Because they are still not sure about consumer spending. The six month run up in the SP500 stock index might lead a casual observer to think that the economy and companies are gearing up.  New Orders indicates that there is much more caution out there than the stock index would indicate.

This past Friday, business’ caution to commit to new investment was only reinforced when the latest Consumer Sentiment index was released.  After climbing the past few months, confidence is sinking again.  Maybe it’s the extra 2% coming out of paychecks since January 1st.  Whatever it is, it doesn’t inspire many business owners to put a lot of money into expanding their production.

When the stock market is trading on hope, it looks six months ahead.  The recent run up is hoping for double digit profit growth in the second half of this year.  When the market trades on fear, it looks ahead about 2 seconds, faster than the normal investor can or should react.  Let me get out my broken record for another spin, cue the needle and play that same old song “Diversify.”

P.S. For those of you who are more active investors, check the latest post from Economic Pic in my blog link list on the right.  It shows the past 40 year returns for a strategy of selling the SP500 index in May and buying the long term government / credit index.  The iShares ETF that tracks this index is ITLB.  A comparable ETF from Vanguard is BLV.

Labor Participation Rate

April 6th, 2013

First I will look at a rather disappointing March Labor Report, released this past Friday.  Then I will zoom up and look at the big picture and some disturbing trends.  The net job gains this past month were 88,000, about half of the 169,000 average gains of the past year.  Remember that it takes about 150,000 job gains each month just to keep up with population growth.  Although the headline unemployment rate dropped .1% to 7.6%, it was because almost half a million people dropped out of the work force, meaning that they had stopped looking for a job in the past month.

Mitigating the meager job gains were revisions to previous months gains as more survey data was returned by employers. January’s job gains were revised from +119,000 to +148,000, and February’s gains were bumped upward from +236,000 to +268,000.  The two revisions added up to an additional 61,000 jobs; adding that to March’s gain of 88,000 gets close to the minimum gains needed of 150,000. The initial reaction of the market was a swift loss at Friday’s market opening of almost 200 points on the Dow.  By the end of the day, the market had regained much of the ground it lost, ending down about 40 points.

The average hours worked increased again to 34.6, a hopeful sign, but earnings saw no change.

Construction continued to show gains; the media’s attention to this area of employment probably gives the casual reader the impression that contruction jobs are a larger part of the work force than they actually are.

Compare that to Professional and Business Services, which has showed consistently strong gains and low unemployment.

Employment in the health care field continues to grow.  As a percent of total employment, health care continues to reach new heights, although its growth has moderated.  Taking care of the sick may be a sign of a compassionate society, but it consumes resources, prompting the question: what is the upper limit?  One in nine workers now work in health care.  Twenty years ago, the ratio was one in twelve. 

Over the past twenty years, the employment market has shifted markedly away from producing goods.  As a share of total employment, about 1 in 7 workers produces goods.  Just ten years ago, the ratio was 1 in 6.

What jobs did those workers find?  Serving food and drink to the ever growing share of people in Professional and Business Services.

The core work force, those aged 25 – 54, shows no growth over the past year.  I use the words “work force” to include only the employed.  “Labor force” includes both the employed and unemployed.  More on that in a bit.

I have written before about the year over year (y-o-y) percent change in the headline unemployment rate, or U-3 rate, and that past recessions usually follow when this change goes above 0.  The unemployment rate has benefitted remarkably from the number of people who continue to drop out and are no longer counted as unemployed.  Because of the drop outs the percent change in the unemployment rate is still in good territory.

A secondary indicator may be the y-o-y percent gain in the employed.  The long term average is 1.5%.  When the percent gain falls below that, recession soon follows.  The percent gain just fell below the long term 1.5% average.

Let’s zoom out to the past forty years to see how this percent gain in employment has preceded past recessions.  The exception was in 1973-74 when the Arab oil embargo created a sudden and deep recession in the country.

There was a decline in the number of people who dropped out but had been searching for work (but not in the past month) and were available to work.

The long term trend of those not in the labor force continues to reach new heights.  As a percent of the population, it  keeps climbing at an alarming rate.

Older workers are retiring, either voluntarily or involuntarily, at the rate of 800,000 a year.

Which brings us to several sometimes confusing concepts, the Civilian Labor Force (CLF), the Participation Rate, and other metrics.  The Civilian Labor Force is those people aged 16 and over who are either employed or unemployed.  To be counted as unemployed, a person is not working but has searched for work in the past month.  The unemployment rate is simply the percentage of unemployed in the Civilian Labor Force, which now totals about 155 million.  An unemployment rate of 7.6% means that about 12 million people are counted as unemployed. 

Then there is the Civilian Labor Force Participation Rate, or simply the Participation Rate, which is the percentage of the Labor Force to what the BLS calls the Civilian Non-Institutional Population (CNP).  Don’t go to sleep on me.  The CNP is those people who are aged 16 or more and who are not in prison or the military. 

So, the Participation Rate (PR) is the number of people who ARE working as a percent of people who CAN legally work; i.e. who are over 16 and not in some institutional setting that prevents them from working or finding work.

Let me give you some numbers and a pie chart.

The total population of the U.S. is estimated at 313 million; the CNP is estimated at 245 million.  The difference between those two figures are mostly children under 16 and people in prison and the military.  Here’s how the Labor Force compares with those not in the labor force and children under 16.

Why does the the Participation Rate (PR) matter?  As it declines, workers have to support more of those who are not working.  Many seniors feel that they “paid into the system” but the “system” – yes, your elected representatives in Congress – spent the additional money paid into the system over the past thirty years.  Social Security is a “Pay as you Go” system meaning that existing workers must somehow pay back into the system to pay benefits for those who retire.  Pay back = higher taxes. As the percentage of the population who works declines, taxes must rise or benefits decrease regardless of who paid into the system. 

This past month the BLS estimated a further decline of .2% to a level of 63.3%.  For comparison, Canada has a PR of 66.6%. 

Part of the decline is a natural demographic change as the population ages.   So how much has the aging of the population contributed to the decline in the PR?  What is the PR for those of working age 16 – 64?  Oddly enough, the time series figures are not easy to come by.  But before we get to that, let’s get to the surprises.

Since 2010, the older labor force, those aged 65+, has grown by 1.2 million. 

In 20 years, the participation rate among seniors has risen 50%, from 11.8% in 1990 to 17.4% in 2010.  The BLS projects that it will rise to 22.6% by 2020, a doubling in thirty years.  Seniors will continue to compete for jobs with the working age population.

Meanwhile, the participation rate of the core work force, those aged 25 – 54, is on a steady decline.

Now comes the biggest surprise, the decline in the working age Participation Rate.  To get the time series, I had to add a number of series together and take some population estimates by the Census Bureau.  Changing demographic shifts and 2010 census revisions make the series not entirely accurate but does give a good representation of the approximate 6% decline.

Let’s look at the last five years for the overall Participation Rate, which has declined about 5%. 

The aging of the population is contributing maybe 20% to the overall decline.  The bulk of the decline is a deterioration of the working age labor force.  Some are going back to school, some have given up looking for a job recently.  Many younger workers are finding it difficult to find a job. The Consumer Credit report released Friday shows another surge in student loans.  The FinAid student debt clock shows that student loans now exceed a trillion dollars. I have the sinking feeling that this will end badly. The participation rate for those aged 20 – 24 has declined about 7% and is now slightly less than the rate for all working ages. 

Payments under the Social Security Disability program, or SSDI, took about 10% of Social Security taxes in 1984.  They now consume 20% of SS taxes and are becoming an increasing burden on the Social Security program even as the boomers begin to retire.  The ranks of the disabled have grown more than 10% in the past three years.

A declining percentage of the population working to pay for an increasing number of benefits – this economic tension is sure to produce social and political conflict.  Many of us probably hold the vague hope that it will all work out somehow.  Some think that politicians in Washington will figure it out despite the fact that the solutions that Congress comes up with to most problems only exacerbate the problem or shift the problem to another area.

On the other hand, the baseball season is still young and anything is possible, right? 

Home Sweet Home

March 31st, 2013

From its catatonic state the housing market continues to make headlines.  On Tuesday came a somewhat disappointing report on new home sales for February; at 411,000 it was a bit below expectations of 425,000.   A real estate saleswoman told me this week that it’s now a seller’s market in Denver.  I presume that means that buyers are now having to offer the asking price or above when submitting a sales contract to a seller.

For a long term perspective, let’s zoom out fifty years.  Home sales are at past recession bottoms BUT they are better than last year and the year before and the housing and labor markets are hoping.

Will the patient stir, starting to rise, only to fall back on the bed?  PUH-LEEZ DON’T!

Housing Starts, which include multi-family dwellings, are on an upswing but are also coming from a deep trough.

What is more telling for the labor market is the ratio of home sales to housing starts, which continues to decline as more and more multi-unit apartment buildings and condos are being built.

Construction of multi-unit dwellings takes less labor per family unit and the type of construction is often skewed to a different kind of labor force than the construction of single family homes.  There is more steel, concrete and masonry work in multi-unit construction, employing trade skills unfamiliar to some in single family residential construction.  This shifting emphasis of skills in the work force may damper growth in the construction labor market.

Let’s go up in our hot air balloons and take a gander at home valuation for the past 130 years.  The Case-Shiller Home Price index surveys home prices throughout the nation and adjusts for inflation.  The homes of today offer more than the homes of 100 years ago, both in convenience, comfort and safety.  However, the index is approaching an upper range that may be less attractive to potential buyers.

Let’s look at housing evaluations from an affordability perspective.  The National Association of Realtors offers an affordability index based on a composite of mortgages.  I prefer a different measure, one that is based on disposable income – income after taxes.  For many of us, buying a house is the biggest purchase of our lives.  Before we make such a big commitment, we need to have some savings (except during the housing boom) to make a down payment, and we need to feel some certainty about our future income.  Mortgage payments will probably take the largest bite out of our income.  

When we look at a long term history of the growth of the home price index (purchases only) and the growth of inflation adjusted disposable income, they track each other closely – until the housing boom really took off in 2000.  Below is a graph of the past 20+ years, showing the relationship between the two.


The upturn in home prices is still above the trend line growth of disposable income and until personal income can resume or surpass a 3% growth rate, any rise in home prices will be constrained.