Flow of Funds: Net Borrowing as a Percent of U.S. GDP at Annual Rates by Sector

Corrections displays the Fed Flow of Funds data.  The Flow of Funds data breaks the economy, for example, into seven sectors:  the household sector, nonfinancial corporate businesses, nonfinancial noncorporate businesses, state and local governments, federal government, rest of world, and financial sector.  Net lending in the world must add up to zero:  there are two sides to every loan.  The flow of funds breaks up the U.S. and the rest of the world, and then breaks up the U.S.  Nevertheless, the sum must still be zero.

Below, Corrections displays the flow of funds for each sector over time (click to enlarge).

The same graph zoomed into the recent period is depicted graphically below (click to enlarge).  Note that while the Federal government is borrowing much more than it used to, as a country we're receiving less than we used to from the rest of the world:  the Federal deficit is being made up by the financial sector. 

What is the financial sector?  The lending portion is made up of the Monetary authority, chartered banks, foreign banking offices in the U.S., credit unions, insurance companies, private and public pension funds, money market mutual funds, mutual funds, closed-end funds, exchange-traded funds, government sponsored enterprises, agency and GSE-backed mortgage pools, ABS issuers, finance companies, real estate investment trusts, brokers and dealers, holding companies, and funding corporations.  The borrowing portion is similar.  We organize these sources into four main sources:  1) private/stock market, such as mutual funds, exchange traded funds, and private pensions 2) government sans monetary authority, such as GSE-backed mortgage pools, government retirement funds 3)  foreign banking offices in the U.S. 4)  the monetary authority and funding companies (AIG and Bear Stearns, for instance).  We graph these four graphically below (click to enlarge):  they add up to the light blue line in the above graph.

From the second graph, we note that the Federal government is borrowing more and that this is financed by the financial sector.  We further note that within the financial sector, it is being financed primarily by domestic funds and the stock market, rather than the monetary authority.  

Indexed Employment by State

Below, Corrections displays indexed total employment by state over time.  The high outlier is North Dakota, the low outlier is Nevada.

The Impact of War on Economic Growth

Corrections took the dataset present in Growth Dynamics:  The Myth of Economic Recovery: Comment by Hannes Mueller and collapsed the dataset down to a single interesting table, giving the present period growth rates given whether a country was at war last year, this year, and next year.

Interestingly, lapsing back into war:  war last year, no war this year, but war next year, has the lowest growth rate, while failing to lapse back into war: war last year, no war this year, and no war next year, has the highest.

Bond Yields

Below, Corrections plots out bond yields in the recent period: corporate AAA, BBB, and CCC yields according to Merrill Lynch, and Treasury Constant-Maturity yields.  Note that during the housing crisis, corporate bond yields went up, while treasury bond yields went down:  

Women's Labor Force Decisions by Marital Status over Time

Below, Corrections depicts the proportion of women by marital status and labor force status over time from the Current Population Survey (click to enlarge).

What could cause this?  Why are married women working so much more?  It can't be anything special about being married--nonmarried and married women show the same pattern.  Women in general are working more.  

Even though we know it isn't anything to do with martial status, we then depict the proportion of married women by labor force status over time (click to enlarge).  Women used to work about 35% of time (other sources give 35% extending back further) to around 60% of the time, an increase of 35% (doubling the proportion of women working).  

Corrections then offers the proportion of women by martial status, educational status, and labor force status over time.  One trend dominates the pattern: women are becoming more educated, so both red lines decline while both blue increase (click to enlarge).

This is depicted more clearly below (click to enlarge):

Instead, we can normalize by the population of all women within an educational category (click to enlarge):

And we can normalize that to the proportion that was working in 1965, to see how much these proportions change over time (click to enlarge):

Households by Number of Earners

Below, Corrections depicts Households by number of earners from 1980-2011 (click to enlarge):

Making the same graph with proportions (click to enlarge):

State Unemployment

Below, Corrections displays the difference between state unemployment levels in a given month and its minimum from 2002 to that date (click to enlarge) and the difference between state unemployment levels in a given month and its maximum from 2002 to that date (click to enlarge).

U.S. Treasury Holdings by Foreign Entity

Who owns U.S. Treasuries?  Below, Corrections depicts holdings of U.S. Treasuries by Foreign Entity (non-US), from the Treasury International Capital System (click to enlarge).

While Chinese holdings of U.S. Treasuries have leveled off, the holdings of other countries have more than picked up the slack.  This can also be seen by normalizing each period to 100% and seeing the proportion of foreign-owned U.S. Treasuries owned by each actor (click to enlarge).

Measurement without Theory on the Empire State Manufacturing Survey

Corrections examines today's Empire State Manufacturing Survey:  the release contains 88 variables:, current and future estimates of New York manufacturers for: business conditions, orders, future shipments, delivery time, inventories, unfulfilled orders, prices paid, prices received  number of employees, and average employee workweek, along with future capital expenditures and future technology spending.

Corrections takes each series, normalizes it, HP filters it, multiplies it by -1 if it correlates negatively with the average of all series together (so variables that are high when conditions are bad flip sign), and finally sorts the series vertically based on the value in July 2009.  We then graph the series in a heat map:  red when the normalized, filtered, and signed variable is "high" and blue when it is "cold."

The result is blind economics (click to enlarge), without theory, seeing if these variables are telling us anything related to one another, suggesting they're measuring something.

Consumer Price Index Components and Transportation Breakout

Below, Corrections displays the components of the urban CPI (click to enlarge).  Weighted by expenditure, they make up the headline CPI.

Transportation looks interesting enough to break out further.  Below, we display the components of the transportation component (click to enlarge):

Ricardian Equivalence

From the Flow of Funds accounts, Corrections presents net private and net government savings, along with net total U.S. savings (click to enlarge).  Notice the mirror trends of net private and government savings as a percent of GDP.  

The graph certainly doesn't discount Ricardian Equivalence as a first-order effect:  households know that deficit spending means future taxes:  to smooth consumption, they save more  If true, this equivalence would suggest that financing government consumption through deficits or through taxation is equivalent, to a first order approximation:  households know net present value of taxation will rise and save for the event.

The Federal Reserve, "Printing Money" and Debt Monetization

From 2008 to the present, the Federal Reserve has created large quantities of reserves: has "printed" a lot of money: more than doubling the supply.  We know that in the long run, inflation moves 1:1 with money.  Why hasn't there been inflation?

Note that Federal Reserve liabilities have increased dramatically:  liabilities rose dramatically due to emergency lending measures.  These measures were temporary, but the reserves were reinvested into other assets.  Using the Federal Reserve's H.41 historical release plus discontinued series, Corrections displays the Federal Reserve's assets, in a graph inspired by James Hamilton  (click to enlarge).

In another inspired graph, we can clearly see emergency lending dying down (click to enlarge).

Rather than create money and then retire it, the Federal Reserve reinvested money created for loans into mortgage backed securities and treasury bonds.  This is using printed money to finance a deficit, no doubt.

So why no inflation?  One answer, losing credibility for the last five years and counting, is that prices just haven't had enough time to react.  A better answer is that unlike a money drop, the Federal Reserve printed money and put it into the economy, purchasing assets.  Those assets slowly mature, and take money back out of the economy, allowing the Fed to retire reserves.  

Imagine the Fed creates $1 and purchases a Mortgage-Backed Security that matures 10 years from now selling at a 50% discount.  For 10 years, the economy will have one more dollar, and one less MBS in circulation.  Come the end of those ten years, the MBS pays back $2, which the Fed takes to retire reserves, having originally pressed a button to create only $1.  Then the economy now has one less dollar in reserves than before all this, and the Federal Reserve has one more.  

There is no magic in all this:  the point is only to emphasize that the Fed isn't just printing money, creating a liability.  It creates liabilities and assets, generally at a one-to-more-than-one ratio in the long run.  This allows it to have low inflationary effect in the long run, because in the long run, no new reserves are created.

Corrections displays a very simple example of how money creation to purchase assets is inflation neutral (click to enlarge):  anyone can do what the Fed did, in theory:  go into debt to purchase assets, "freeing" money that other people can spend (the person who sold the asset now has a credit in their bank account equal to your debit).  Use the returns for those assets to retire your debt, and the economy is back to where it was: nothing new is created in the end.  

Will the Fed retire those reserves?  Skeptics will say that historically, governments don't do that.  They're wrong.  Throughout its history Britain would go to war, and cease convertibility to gold. It would "print" money (go into debt).  Then, in the years after the war, it would run surpluses and reduce monetary aggregates until the gold standard could be resumed at the same level it was previously.  For example, from 1797 to 1821 Britain departed the gold standard during the Napoleonic Wars, returning to where it left after 24 years.  This happened again in the U.S. after the Civil War, as the U.S. returned to prewar parity.  This happened yet in Britain after WWI in 1925, as they resumed at prewar parity.  

Perhaps the Fed will permanently keep an enormous amount of reserves in circulation.  But the Fed has stayed politically independent since the 1951 Accord, when William McChesney Martin replaced Thomas McCabe as Chairman of the Federal Reserve, betrayed Truman (who had just appointed him) and ended Fed support of low Treasury rates.  While the Fed had agreed to support the Treasury with $200 million worth of purchases over five years, they spent that amount in the first few days, and then refused any further help.  This kicked off a new age of Fed independence.  If this independence continues then debt monetization seems highly unlikely.

Real GDP per Capita as Percent of Previous Peak: 1947Q1-2012Q3

Using the first revision of GDP up to 2012Q3, Corrections displays GDP/capita as percent of previous peak (click to enlarge).

In 2007Q4, real GDP/capita in 2005 dollars peaked at $44,005.  It reached a trough in 2009Q2 at $41,389, (94.1% of its previous peak) and is now at $43,352 (98.5% of its previous peak), growing at an annualized rate of 1.4% per year from trough to now (0.3571% quarterly) 

If the trend between trough to now continues, we will reach the previous peak in the fourth quarter of 2013 (by the first month of the fourth quarter, if we broke things down monthly), having deviated from our previous peak for six years.  The previous longest postwar peak was 2.25 years, in the nine quarter between and including 1981Q2 and 1983Q2.

Bond Maturity Portfolios

Below, Corrections displays the returns over time of $1 invested in 1942 into a constant-maturity portfolio (e.g. purchase a bond of a given maturity, wait a short time, and sell it again, purchasing another bond of that given maturity with all of your proceeds (gains or losses). Data from CRSP.

The log version of the graph scales things appropriately, so that two increases mean the same thing.  Note the higher slopes in the mid to late 70's due to inflation expectations.

Some International Growth Rates

Below, using the IMF's International Financial Statistics Corrections depicts several measures of GDP and GDP/capita for Australia, Canada, France, the U.S. and the U.K. from 1959 to 2011.  

First, nominal GDP/capita and the appropriate constant growth rate (compounded annually) in the legend next to each country's name.  This is only mildly useful, given varying inflation rates but is interesting to view (click to enlarge).  Note that the volatility of non-U.S. countries is noise added by exchange rates.

More useful is perhaps indexed real GDP with compound growth rates (click to enlarge):

Of course, population increases, so indexed real GDP/capita with compound growth rates might be better (click to enlarge).

Finally, real GDP/capita as a fraction of U.S. GDP per capita.  The recent crisis, as well as Australia's dodge of the crisis via a natural-resource and Asia-driven boom (click to enlarge)

Tax Rates

Using NBER's TaxSim, a program that calculates expected income tax rates from a random sample of actual IRS returns, Corrections produces a graph of after-tax income against before-tax income for several states.  This includes only State and Federal Income Tax, along with FICA taxes.  It assumes a married earner with 2 children, a rent of $1500, and no other deductions or income.  Tax rates are for 2010.

This is a worst case scenario in terms of total income:  all is labor income, which is punished most severely.  Capital income, quite rightly, is generally taxed at lower rates.

Corrections also plots marginal tax rates over labor income: how much is taken of each extra dollar you earn.

Yield Curve

Below, Corrections depicts the daily yield curve from Jan1990-Aug2012 (click to enlarge).

Below, Corrections takes each day's yield curve and breaks it up into a constant, slope (by duration), and quadratic (by duration squared) term (click to enlarge):

Finally, we normalize each components to have mean zero and standard deviation one, and graph them together (click to enlarge).  Insofar as interest rates predict bad times, the three components point to lower growth (the low level, shallow slope, and less curvature).   

Joint Distribution of Coke and Pepsi Stock Returns

For fun, Corrections displays the joint distribution of daily Coke and Pepsi stock returns from 1990 to present, from two different angles (click to enlarge).

We might also graph the joint distribution of what Coke's daily return was yesterday against what Pepsi's return is today, in the hopes that we can generate an actively-managed portfolio that buys and sells Pepsi based on Coke's conditional return (click to enlarge):

No such luck.  The two are slightly negatively correlated, with a slope of 0.01% return and an insignificant coefficient (even without correcting for serial correlation).  Note that while concurrent returns are moderately positively correlated, the two have no forecasting power for one another.

Entertaining. 

When the Crowd Isn't Wise

David Leonhardt writes a rather painful article in the New York Times "When the Crowd Isn't Wise" (July 7th, 2012).  He notes that Intrade thought Obamacare was more likely to be ruled unconstitutional than constitutional:

[Intrade markets] continued to show about a 75 percent chance that the law's so-called mandate would be ruled unconstitutional, right up until the morning it was ruled constitutional.

The market-the wisdom of crowds-turned out to be wrong. 

Corrections sees risk premia and self-insurance everywhere, but let's take the common view of Intrade markets, that the price/10 is equivalent to the probability the market puts on an event.

Are political writers like Leonhardt so ignorant of statistics to know what "wrong" means?  75% represents a distribution of outcomes!  If you looked at all Intrade markets that were at and around 75%, and then looked at the outcomes, if they were "right" (by Leonhardt's incorrect statistic) 100% of the time we should conclude these are poorly functioning markets.

No.  If you predict something should happen with 75%, then it should not happen 25% of the time, and happen 75% of the time.  Not happening 25% of the time means you are correct in your guess about the distribution.  75% represents a distribution of outcomes.  A 90% guess would be more wrong than 75% in general, when the outcome actually happens 75% of the time.  The relative entropy of the distributions is what helps us differentiate them statistically.

New York Times quality article.

Are Some Forecasters Really Better Than Others?

From the Journal of Money, Credit and Banking, (D'Agostino, McQuinn, and Whelan), the paper "Are Some Forecasters Really Better Than Others?" is entertaining.  (Sadly, not many easy-to-excerpt figures).  The abstract:

In any data set with individual forecasts of economic variables, some forecasters will perform better than others. However, it is possible that these ex post differences reflect sampling variation and thus overstate the ex ante differences between forecasters. In this paper, we present a simple test of the null hypothesis that all forecasters in the U.S. Survey of Professional Forecasters have equal ability. We construct a test statistic that reflects both the relative and absolute performance of the forecaster and use bootstrap techniques to compare the empirical results with the equivalents obtained under the null hypothesis of equal forecaster ability. Results suggest little support for the idea that the best forecasters are actually innately better than others, though there is evidence that a relatively small group of forecasters perform very poorly.

Corrections believes it.

10-year Eurozone Bond Yields: January 1993-April 2012

Below, Corrections displays 10-year Eurozone Bond Yields from January 1993-April 2012 (click to enlarge).  Did Ireland's "austerity" measures work like Iceland's did?  10-year bond yields have fallen dramatically.

The Religious Right Turns 33: What Have We Learned?

Jonathan Merritt writes an embarassingly wrong op-ed in The Atlantic:  The Religious Right Turns 33:  What Have We Learned? (June 8th, 2012).  In it, he attacks the Religious Right, arguing that the movement into politics of Evangelicals and the Religious Right has diminished interest in Christianity.

Economics has special ways of dealing with time-series theses like "The Christian Right got into politics, and their share of the population went down.  Therefore, it must have been because of the politics."  On its own, this has little empirical content:  post hoc ergo propter hoc.  But a good analysis can be convincing by showing parallel data.  We can examine other religions that didn't get into politics (or didn't change their relative immersion into politics), or look at factions of Christianity that went into politics more.

In other words, we can say "if that thesis is true, then it has testable implications."  Corrections offers two testable implications:

• Other branches of Christianity haven't gone into politics as much as Evangelicals:  therefore, Evangelicals should be suffering the most.
• Judaism, a religion strongly tied to the left for more than a century, has not changed its political position very much.  Therefore, it should be untouched by the last twenty years.

Obviously these aren't the only stories one can tell: Corrections is glad to entertain other testable hypotheses of Merritt's otherwise empty theory.  First, we use the Statistical Abstract of the United States to depict the proportions of different religions with a logarithmic scale (otherwise, Evangelicals, Muslims, and Jewish proportions are too small to distinguish) (click to enlarge).

One can see that Christianity and Judaism have declined while Athiests, Muhammadans, and Evangelicals have seen an increase in their proportions.  It would be easier to compare them all to their 1990 proportion, to see the change (click to enlarge):

This figure tells our whole story:  if being political has hurt Christians, then why has Judaism, which hasn't changed its political orientation seen a larger fall, while the subset of Evangelicals in Christianity seen the largest rise?  

As a note, it is true one can begin to tell stories (ex:  Evangelicals rose by draining other Christians while the rest left, Judaism has its own thing going on, etc.) to make sure Merritt's claim is devoid of testable hypotheses.  Such a tack would ironically and safely bring one's own politics into a religious (non-testable) sphere.

Balance Sheet of U.S. Households and Nonprofit Organizations 1949-2011

Below, Corrections displays the Balance Sheet of Households and Nonprofit Institutions in the United States from 1949-2011, in real 2011 U.S. dollars (click to enlarge).

The Labor Wedge

The labor wedge is a difference between the marginal rate of substitution (MRS) between consumption and leisure, and the marginal product of labor (MPL).  That is, how willing you are to trade off leisure for consumption, and the degree to which you are able to do it.  If an individual may do so perfectly, then the labor wedge is zero.  It is given its name because all real taxes have distortionary effects, and most, if not all, have effects on labor.  If we look at measures of MRS and MPL, we can say "what tax rate explains this gap?"  This is what the labor wedge is:  essentially a structural "this is what taxes seem to be, given distortions in the economy."

Below, Corrections offers the labor wedge as offered in Rob Shimer's book, with data from Cociuba, Prescott, and Ueberfeldt (Simona Cociuba's website).  We graph two possible labor wedges:  one with a low Frisch elasticity of 0.5, and one with a Frisch elasticity of 4 (used for most macro settings).   The Labor Wedge is depicted graphically below (click to enlarge).

U.S. Employment and Hours as Percent of Previous Peak

Below, Corrections plots U.S. Employment (nonfarm and total private) and Hours (aggregate hours of nonsupervisory and production employees), along with dated NBER recessions (click to enlarge).

Business Employment Dynamics: Where Jobs Losses and Gains Come From

Below, Corrections graphically depicts transformed Business Employment Dynamics data.  The two data series are the proportion of gross job losses generated by closing establishments, rather than contracting establishments (click to enlarge).  Similarly for gross job gains generated by opening establishments, rather than expanding establishments.

Three things seem to jump out of the figure:

• Generally, around 20% of gross job gains and losses come from opening and closing establishments.
• Compared to the proportion of gross job losses that come from closings, generally a higher proportion of gross job gains come from openings.
• There has been a secular downward trend in the impact of closings and openings on employment.

The last point is probably bad news for the U.S. economy.

Business Employment Dynamics: 1992:Q3-2011:Q2

Below, Corrections shows Business Employment Dynamics data from 1992:Q3-2011:Q2.  We index to 1992:Q3=1, from data originally in levels.

• Gross job gains are the total people hired in a quarter (not subtracting losses).  U.S. generally has around 7.6 million total gains in a given quarter.  
• Expansions are businesses that reported more jobs than last quarter.  U.S. generally has around 6.1 million firm expansions in a given quarter.
• Openings are businesses that did not exist in the previous quarter.  U.S. generally has around 1.6 million firm openings in a given quarter.
• Gross job losses are the total separations in a quarter (not adding gains).  U.S. generally has around 7.4 million total losses in a given quarter.
• Contractions are businesses that reported fewer jobs than last quarter.  U.S. generally has around 6.0 million firm contractions in a given quarter.
• Closings are businesses that reported last quarter but are no longer active.  U.S. generally has around 1.5 million closings in a given quarter.

We generally think of having both gross job gains and gross job losses high as creative destruction:  while not much is moving, there's a lot of churn in the economy, generally very good.  We generally think of having both gross job gains and gross job losses low as stagnation or sclerosis:  not much is flowing in the economy.

The 1990's and the Great Recession both show prominently in the figure of BED data, depicted graphically below (click to enlarge).

US GDP, Log GDP, and Percent Deviations from Trend

GDP from 1947-2011, log GDP for the same period, and deviations from that log trend (which can be interpreted as percent deviations).