Archive for the ‘Economy’ Category

The Recession is over (as of June 2009)

Monday, September 20th, 2010

The NBER has finally declared that the recession of 2007 to 2009 ended in June 2009.  As we noted here last year the economic data that the NBER use to track the recession turned up indicating that the recession ended in the summer of 2009. There are two brief points I would like to review: (1) understanding what the end of the recession means and (2) why there was some uncertainty about when it ended.

The NBER has a particular view of the nature of economic dynamics; they believe that the economy experiences a series of expansions and contractions whose starts and ends they can clearly identify.  The reason that their findings are announced so far after the actual event (identifying the trough or cyclical bottom that occurred in June 2009 over a year later) is that they now believe that a new cyclical upturn started.  Any significant drop in economic activity later this year (a so-called double dip) would now be part of a different recession.  At some point in the future there will be clear evidence that a peak in economic activity has occurred and that the various indicators used by the NBER are declining.  The expansion that began in June 2009 will have ended, and the NBER will then identify the start of the next recession.

My uncertainty about the dating of the recession and the reason that the NBER found that the recession ended a few months before my post indicated was that the NBER used some new data that they had not previously used. The new NBER data, monthly GDP estimates not publicly available, hit bottom sooner than the other data, thus they declared the end of the recession at the end of June rather than later in the summer when the rest of their data hit bottom.

[Note to readers] The economic upturn has had many consequences, including a refocusing of my energies on other projects besides this site.  Future posting is highly uncertain and may occur as Milton Friedman said of monetary policy “with a long and variable lag”.  Readers wishing my views on specific issues can contact me at .

Tracking the 3.5 million jobs President Obama will save or create

Friday, January 8th, 2010

In an earlier essay I tried to explain President Obama’s notion of saving or creating jobs.  The stimulus plan bill was passed by both houses of Congress last night and the final plan was a bit smaller than the earlier version, so the President now asserts that the plan will save or create 3.5 million jobs.

This post will track the 3.5 million jobs.  There are a number of ways to measure jobs in the US.   Some people work several different jobs at a time while others change employers frequently, so measuring jobs is not as simple as it might seem.  There was a cartoon from the Clinton era showing the President speaking at a dinner that he had created 8 million jobs and an overworked waiter thinking that he had three of them.  Obama’s economic team define jobs as use the payroll data (see here for their original report).

Just before the stimulus bill passed the Department of Labor issued a report (see here).  The number of people working (see Table B1, about 2/3 of the way down, with the heading “Establishment Data”) was 134,580,000 (seasonally adjusted).  This is a preliminary measure and will be revised next month and probably revised again in a year.  Using the Obama team methodology, without the stimulus bill employment would be expected to fall by around 1,613,000 jobs during the next two years so that without the stimulus bill we would expect employment to be 132,967,000 in January 2011.

With the revised estimate of 3,500,000 jobs “saved or created”, employment should be 136,467,000, creating 1,887,000 in addition to the 1,613,000 jobs saved.

[Note I have revised the table format since the original post to make the tracking easier]

The table below will be updated with every new employment release to see how jobs have changed.  The first column is the actual number of payroll jobs starting with the month before the stimulus plan passed; the second column is the total change in employment since the month when the stimulus plan passed and the third column shows the gap remaining of jobs to be “created” in order to reach the target.

Date Number of Jobs Change in Jobs

After Stimulus

Number of Jobs needed

to reach target by Jan 2011

January 2009 134,333,000 1,887,000
February 2009 133,652,000 -681,000 2,568,000
March 2009 133,000,000 -1,333,000 3,220,000
April 2009 132,481,000 -1,852,000 3,739,000
May 2009 132,178,000 -2,155,000 4,042,000
June 2009 131,715,000 -2,618,000 4,505,000
July 2009 131,411,000 -2,922,000 4,809,000
August 2009 131,257,000 -3,076,000 4,963,000
Sept. 2009 131,118,000 -3,215,000 5,102,000
October 2009 130,991,000 -3,342,000 5.229,000
November 2009 130,995,000 -3,338,000 5.225,000
December 2009 130,910,000 -3,423,000 5.310,000

On March 6, 2009, the BLS released the revised January job numbers and the preliminary February numbers. I have revised the table to reflect these data. The loss of 651,000 jobs in February means that there will have to be an increase in jobs of a bit over 2.5 million over the next 23 months (111,000 per month) to reach the target number of jobs.

[update] On April 3, 2009, the BLS released preliminary March data (the change in February was unrevised). The loss in jobs means that there will have to be an increase in jobs of 3.2 million (148,000 per month) in the next 22 months to reach the target.

[update] On May 8, 2009, the BLS released preliminary April data. The loss in jobs means that there will have to be an increase in jobs of 3.8 million (181,000 per month) in the next 21 months to reach the target.

[update] On June 5, 2009, the BLS released preliminary May data. The loss in jobs means that there will have to be an increase in jobs of 4.1 million (194,000 per month) in the next 20 months to reach the target.

[update] On July 3, 2009 the BLS released preliminary June data. The loss in jobs means that there will have to be in an increase in jobs of 4.5 million (225,000 per month) in the next 19 months.

[update] On August 7, 2009 the BLS released preliminary July data. The loss in jobs means that there will have to be an increase of jobs of 4.7 million over the next 18 months to reach the target.

[update] On September 4, 2009 the BLS released preliminary August  data. The loss in jobs means that there will have to be an increase of jobs of 5 million over the next 17 months to reach the target.

[update] On October 2, 2009 the BLS released preliminary September data.   The loss in jobs means that there will have to be an increase of jobs of 5.3 million over the next 16 months to reach the target.

[update] On January 8, 2010 the BLS released preliminary December data.  The loss in jobs (following a small revised increase in November) means that there will have to be an increase of 5.3 million jobs over the next 13 months to reach the target.

Technical note:  The Obama team issued their forecast in January before the release of the January employment data; expectations were for January employment to be down around 500,000.  One could infer that the Obama team expected such a job loss and that they actually now expect a loss of only 1.1 million jobs over the next 23 months, changing the number of jobs created relative to jobs saved.  But as I have not seen a revised version of the CEA memo reflecting the 3.5 million number (and in fact the White House still provides links to the 4 millions job memo), I will make the calculations in a way that provides a “best case” to the Obama team.  If there is a revised version of the memo to show the 3.5 million job estimate (with a revised forecast), please send me a note to ckendall at-sign umrkt dot com and I will update this page accordingly. [Update May 13, 2009: The President's staff is still using the 3.5 million number, see here in Chapter 2]

After the recession ends, when will unemployment fall?

Monday, January 4th, 2010

The unemployment rate typically continues to rise for a few months after recessions end, but in recent years the timing seems to have changed.  After the end of recessions from 1948 until the 1980s, rapid economic growth followed quickly with a sharp improvement in unemployment.  But after the last two recessions (recessions that ended in March 1991 and November 2001) there was a considerably longer period before the economy grew rapidly and unemployment fell.  This article will document this finding.

During recessions, the economy slows and the rate of unemployment rises. When a recession ends and the economy starts to grow, there is often a period of a few months (or more) until the unemployment rate falls.  I believe that the data indicate that the recession that started in December 2007 probably ended in July or August of 2009 (see here for the data), but the NBER, the agency that determines when recessions end, may not make an official decision until the evidence is much clearer, probably sometime in 2010 or 2011.

If the recession did end in the third quarter of 2009, when will the rate of unemployment fall?  As I write, in early January 2010, unemployment is around 10%, somewhat higher than it was when the recession ended.  There is a rough rule of thumb that the economy must grow around 3% to generate enough jobs to keep the rate of unemployment constant (see here for a more detailed analysis).  The economy grew at 2.2% in the third quarter of 2009, so it is not surprising that the unemployment rate continued to rise.

Here is a table that lists the end dates of the postWorld War II recessions, the lag from the end of the recession until the GDP grew above 4% and the change in the unemployment rate in the twelve months following the recession.  The 4% level was chosen as representing a rate of growth sufficient to substantially decrease the unemployment rate.

End of Recession Time to 4% Growth 12 month Change in Unemployment Rate
October 1949 1 Quarter -3.7%
May 1954 1 Quarter -1.6%
April 1958 1 Quarter -2.2%
February 1961 1 Quarter -1.4%
November 1970 1 Quarter +0.1%
March 1975 2 Quarters -1.0%
July 1980 1 Quarter -0.6%
November 1982 1 Quarter -2.3%
March 1991 4 Quarters +0.6%
November 2001 7 Quarters +0.4%
August 2009 (E) NA +0.3% (August to November)

Sources:  NBER, BEA and BLS (data obtained from FRED).

Following the recession that ended in October 1949 (during the fourth quarter), the economy grew at 17.2% in the first quarter of 1950, one quarter later.  Twelve months later, in October 1950, the unemployment rate was 4.2%, 3.7% lower than the 7.9% when the recession ended.

For the period from 1945 to the 1980s, the economy grew relatively rapidly (above 4%) shortly after the recession ended. In the year following all but one of these recessions, the rate of unemployment dropped, usually around 1 to 2%.

If the recession that ended in 2009 were a typical 1940s/1980s recession, then we would expect the rate of unemployment to decline from the 9.7% level in August 2009 to around 8% in August 2010.  But if the recession is of the 1990s/2000 variety, then it would not be surprising to see 9-10% employment in late 2010.

Understanding the Recession that started in December 2007

Tuesday, December 29th, 2009

On December 1, 2008, the NBER announced that a recession in the US started in December 2007 (for more about how the NBER came to this conclusion, see here; for the NBER announcement, see here). [Note that after this article was initially posted, the data have been updated and the table includes updates; see below for details.]

The NBER described the five monthly series which they have followed, (1) payroll employment, (2) real personal income less transfer payments, (3) real manufacturing and wholesale-retail trade sales, (4) industrial production, and (5) employment estimates based on the household survey.  This last measure is not included in their prior research, but generally is similar to the payroll employment number.

Here are these data expressed as a percentage of December 2007 values, the date that the NBER denoted the peak of the previous business cycle.

Date Payroll Emp Real P I Ind Prod Sales HH Emp
September 2007 99.6 100.4 99.6 100.2 99.9
October 2007 99.8 100.4 99.1 100.6 99.7
November 2007 99.9 100.2 99.7 100.7 100.3
December 2007 (NBER PEAK)
100.0 100.0 100.0 100.0 100.0
January 2008 99.9 99.6 99.9 100.5 100.0
February 2008 99.8 99.3 99.6 98.8 99.9
March 2008 99.8 99.0 99.3 99.0 99.8
April 2008 99.6 98.8 98.8 100.0 100.0
May 2008 99.5 98.5 98.5 99.3 99.8
June 2008 99.4 98.0 98.2 99.3 99.6
July 2008 99.3 97.7 98,2 98.2 99.5
August 2008 99.2 98.0 97.2 96.7 99.3
September 2008 99.0 97.8 93.2 94.2 99.1
October 2008 98.7 97.9 94.5 93.4 98.9
November 2008 98.3 98.6 93.2 92.0 98.5
December 2008 97.8 98.6 91.1 91.4 98.0
January 2009 97.2 96.4 89.1 90.0 97.1
February 2009 96.7 95.0 88.3 90.2 96.9
March 2009 96.3 94.2 86.9 89.2 96.3
April 2009 95.9 94.5 86.5 88.6 96.4
May 2009 95.7 94.5 85.5 88.0 96.1
June 2009 95.3 94.0 85.2 87.6 95.8
July 2009 95.1 94.2 86.1 88.6 95.7
August 2009 95.0 94.1 87.3 88.3 95.5
September 2009 94.9 94.1 87.7 88.5 94.9
October 2009 94.8 94.1 87.7 88.9 94.5
November 2009 94.8 94.3 88.4 NA 94.7

Source: Department of Commerce, Bureau of Labor Statistics, Federal Reserve Board and author’s calculations; all data expressed as percentage of December 2007 values.


The NBER defines a recession as a significant decline in economic activity spread across the economy, lasting more than a few months, and these data indicate the economy is in recession.  All five indicators have been in decline for more than a few months, with their peaks coming between September 2007 (real Personal Income) and April 2008 (household employment).

Now that we know that a recession has started, the question is when the recession will end.  I will continue to update this table with current data; as of the most recent data (November 2008), the recession continues.

[Update Dec 25 2008] Just before Christmas, the BEA released November personal income data that were up sharply (largely due to the drop in prices) almost to the level at the start of the recession; October sales data were also up.  But personal income also remained stronger than the other economic measures during the 2001 recession and sales remain well below the peak level.

[Update Jan 11 2009] Employment data are still very weak; a benchmark revision to the household employment number is now presenting a similar picture to the payroll data.

[Update Feb 1 2009] Sales and Industrial production continue to fall; personal income is steady, not much below the level at the start of the recession, but this measure did not track other variables during the 2001 recession either.  One wonders why the NBER still include personal income in their analysis.

[Update Feb 12 2009] Employment continued to fall in January; further, a benchmark revision of the payroll employment data make the payroll job loss steeper than it had been.  The BLS does not revise the previous household numbers, but the sharp drop in the household number between December 2008 and January 2009 is in part due to the revision (see here for more).

[Update Feb 19 2009] Industrial production continues to decline.

[Update March 2 2009] PI was revised higher but the latest numbers are down; sales continue to decline.

[Update March 8 2009] Continued weak employment numbers with a slight downward revision for the payroll data.

[Update March 16 2009] Continuing weak industrial production with a slight downward revision.

[Update April 1 2009] Continuing weak personal income and sales data.

{Update April 3 2009] Continuing weak jobs data

[Update April 16 2009] Annual revision of industrial production is mostly negative; IP is revised down for most of the last 18 months and now has a peak coincident with the recession.

[Update May 4 2009] Personal income continues to decline; real sales had a modest rise in February but remains at a low level.

[Update May 8 2009] Payroll employment continues to decline but household employment turns up slightly.

[Update May 15 2009] Continuing weakness in industrial production.

[Update June 7 2009] Little sign of any improvement; perhaps the rate of decrease is slowing, but no evidence of improvement in sales or employment.  Personal income is up a bit, but largely due to increased  payments by the government.

[Update June 16 2009] Continuing weakness in industrial production

[Update June 30 2009] Personal income remains weak; the increase in the headline number was caused by the increase in transfer payments which the NBER number subtract from the total.  Sales are also weak.

[Update July 2 2009] Another weak employment report, with both household and payroll numbers weaker.

[Update August 4 2009] Personal income numbers have been revised greatly changing the behavior of this series see here for details.  Revisions were largely negative.  IP is still weak; sales data will be revised in the next few days.

[Update August 14 2009] The sales data have been completely revised showing a modestly more serious decline during the recession period.

[Update August 28 2009] Both Industrial Production and Personal income are up in July (positive data will be indicated in bold); this is perhaps a preliminary indication of the end of the recession.  The NBER will wait to make a declaration that the recession is over until the evidence is much clearer.  Sales data were weaker in June.

[Update September 4 2009] Employment continued to decline in August at a slower rate.

[Update October 12 2009] BEA now calculates Personal Income minus Transfers and I am using their series that is very similar to the series I calculated; modestly down in August.  Household employment dropped sharply in September and is now just about where payroll employment is.  Industrial production was up again in August (indicative of the end of the recession).  Sales were up in August, another indicator that the recession may have ended this summer.

[Update December 29 2009] Personal Income, Industrial Production and Sales are all modestly up from their early summer minima.  The drop in payroll employment is close to zero, and household employment turned up in November.  All evidence seems to confirm that the recession ended in the summer of 2009, but the NBER is unlikely to declare an end until there has been a substantial increase in employment.

US Inflation After the Recession of 2007-2009

Tuesday, December 29th, 2009

To combat the economic slowdown and credit crisis of 2007-2009, the Fed took actions that were beyond anything attempted in the last 50 years; interest rates were cut to virtually zero and the Fed massively expanded its balance sheet by purchasing (in addition to the usual Treasury Bills) Treasury Bonds and Mortgage Backed Securities from Fannie Mae and Freddie Mac.  In another article I briefly described the policy actions the Fed has taken.  This article discusses (and tracks) the possible inflationary consequences of the Fed’s actions.

Many economists believe there is an approximate relation between changes in the money supply and changes in the price level i.e., inflation  (see, among many others here and here).  While the relation between money and inflation is not precise, few economists believe that large increases in the money supply will not eventually lead to large increases in prices.

Many economists are not worried about this relation in the short run, as they believe that while the economy is below full capacity (when output is below the maximum that the economy could produce or when unemployment is greater than normal) that monetary expansion will not result in inflation.  But as the economy recovers and output and employment recover, inflation is likely to accelerate.

So, when the economy recovers (assuming that the US will not face an extended crisis like the 1930s) the Fed will have to figure out how to undo the monetary expansion. Fed Vice Chairman Donald Kohn recently summarized the problem:

…the Federal Reserve’s actions to ease credit conditions have resulted in a tremendous increase in its assets and in bank reserves.  Some observers have expressed concern that these actions, if not reversed in a timely manner, are sowing the seeds of a sharp pickup in inflation down the road.  As I just noted, near-term prospects appear to be for a decline in inflation rather than an increase.  But my colleagues and I are acutely aware of the risk of higher inflation as the economic recovery gains speed.  We are firmly committed to acting in a way that preserves price stability, and we believe we have the tools to absorb reserves and raise interest rates when needed.  Moreover, we are working with the Treasury to introduce legislation that would enlarge our tool kit for moving away from the extraordinary degree of financial stimulus we have put in place when the time arrives.

This page will track some of the key variables associated with the possible inflationary surge. First, I will characterize the state of the economy, calculating how close production is to capacity.  Second, I will track several measures of the policy actions taken by the Fed and finally I will present several measures of inflation.

Date GDP Actual GDP Potential Output Gap
2007IV 100.0 100.0 0
2008I 99.8 100.8 -0.9%
2008II 100.2 101.5 -1.3%
2008III 99.5 102.3 -2.7%
2008IV 98.1 103.0 -4.8%
2009I 96.5 103.8 -7.0%
2009II 96.3 104.6 -7.9%
2009III 96.9 105.4 -8.1%

Source: Bureau of Economic Analysis and author’s calculations.  I assumed that the economy was at potential at 2007IV and that potential GDP grows at 0.75% per quarter (roughly 3% per year).

Date Fed Funds M Base M2 SA
September 2007 4.75% 98.8 98.7
October 2007 4.50% 99.0 99.1
November 2007 4.50% 99.6 99.5
December 2007 4.25% 100.0 100.0
January 2008 3.00% 99.3 100.7
February 2008 3.00% 99.1 101.7
March 2008 2.25% 99.5 102.6
April 2008 2.00% 99.2 102.8
May 2008 2.00% 99.7 103.1
June 2008 2.00% 100.3 103.3
July 2008 2.00% 101.2 103.9
August 2008 2.00% 101.3 103.5
September 2008 2.00% 108.5 105.0
October 2008 1.00% 135.7 106.6
November 2008 1.00% 172.9 107.3
December 2008 0-0.25% (see note) 199.9 109.6
January 2009 0-0.25% 205.8 110.7
February 2009 0-0.25% 187.7 111.0
March 2009 0-0.25% 197.9 111.9
April 2009 0-0.25% 210.7 111.2
May 2009 0-0.25% 213.3 112.1
June 2009 0-0.25% 202.4 112.6
July 2009 0-0.25% 201.0 112.3
August 2009 0-0.25% 204.9 111.6
September 2009 0-0.25% 216.4 112.0
October 2009 0-0.25% 232.8 112.4
November 2009 0-0.25% 243.2 112.8

Source: Federal Reserve Board; Monetary Base (SA) and M2 (SA) expressed as percentage of December 2007 levels. Fed Funds represent end of month values; on December 16, 2008 the FOMC announced that the target rate for Fed Funds was 0 to 1/4 percent.

Date CPI All CPI Core
December 2007 100.0 100.0
January 2008 100.4 100.3
February 2008 100.5 100.3
March 2008 100.9 100.5
April 2008 101.1 100.6
May 2008 101.6 100.8
June 2008 102.5 101.1
July 2008 103.2 101.4
August 2008 103.2 101.6
September 2008 103.3 101.7
October 2008 102.4 101.7
November 2008 100.7 101.7
December 2008 99.9 101.7
January 2009 100.2 101.9
February 2009 100.6 102.1
March 2009 100.5 102.3
April 2009 100.4 102.5
May 2009 100.5 102.7
June 2009 101.3 102.9
July 2009 101.3 103.0
August 2009 101.7 103.1
September 2009 101.9 103.2
October 2009 102.2 103.4
November 2009 102.6 103.5

Source: Bureau of Labor Statistics;

So what do we see in these data?  From the first panel, the economy as I write (May 2009) is 6% below full capacity.  From the second panel, starting in September 2008 (after the failure of Lehman) the Fed began a major balance sheet expansion, doubling the monetary base in roughly three months as the Fed purchased assets to try to ease the problems in credit markets.  The expansion of the base has only resulted in moderate expansion of broad money and in the third panel we can see that it has had little effect, so far, on inflation.  The Fed has a difficult job ahead, and I will track its progress by updating these tables in coming months.

[update June 18 2009 Monetary expansion continues in May, but there is still little evidence of inflation]

[update July 15 2009 the Monetary base declines significantly]

[Update August 14 2009 some evidence of modest monetary contraction in July with little evidence of inflation]

[Update September 26 2009 inflation up as gas prices rise but core fairly stable, monetary base up but M2 contracts]

[Update December 29 2009 The output gap continues to widen (growth in the third quarter below potential), with monetary expansion continuing and a bit more inflation]

Understanding when the Recession will End

Tuesday, June 23rd, 2009

The start and end of recessions are determined by the NBER, a “private, nonprofit, nonpartisan research organization” (see here for more).  While there has been much debate about why this group has the power to determine the dating of recessions, their proclamations are generally accepted by most economists.

The NBER determines what are the peaks and troughs of business cycles, that is, the date at which the economy started to decline and the date at which the economy hit bottom and started rising again.  Their announcement of the end of the 2001 recession provides a clear illustration of their thinking:

In determining that a trough occurred in November 2001, the committee did not conclude that economic conditions since that month have been favorable or that the economy has returned to operating at normal capacity. Rather, the committee determined only that the recession ended and a recovery began in that month. A recession is a period of falling economic activity spread across the economy, lasting more than a few months, normally visible in real GDP, real income, employment, industrial production, and wholesale-retail sales. The trough marks the end of the declining phase and the start of the rising phase of the business cycle. Economic activity is typically below normal in the early stages of an expansion, and it sometimes remains so well into the expansion.

The committee waited to make the determination of the trough date until it was confident that any future downturn in the economy would be considered a new recession and not a continuation of the recession that began in March 2001.

The key point to note relative to the end of a recession is that it representts the end of the decline. So the NBER will declare that the recession has ended at the point when economic activity stops declining; the obvious problem is that you can only tell that the decline has stopped when economic activity has started to rise again.

Recessions are not like sporting events, where a winner is declared immediately after the match ends, because the end of the match is itself uncertain.  Economic commentators often describe recesssions as V or U shaped, depending on whether the rapidity of the rise in economic activity from the bottom,  A U shaped recovery (or even worse a W or L) means that activity remains low for some time before the recovery starts; in such situations (as in 2002-2003), the NBER will delay its determination of the end of a recession until the ensuing recovery is clear and the end of the recession was announced a year and a half later.

On another page on this site I am tracking some of the variables that the NBER uses to analyze the current recession.  As of the time I am writing (June 2009) most of the variables are low and still falling (personal income has been up a bit due to the increase in transfer payments associated with the government’s stimulus plan).  The NBER will most likely wait until the evidence is completely clear, that is when employment, sales, income and production are all rising and well above current levels.  Forecasters seem to expect that the economy will begin to improve in the second half of 2009, but the evidence that the recession has ended may not be clear until a year or so later.

Understanding Employment (or Obama and the 2.5 or 3 or 4 or [update] 3.5 million jobs)

Sunday, January 11th, 2009

President-elect Obama’s economic team says that he was developing a plan that “would save or create nearly four million jobs”.  Four million jobs is a large number; this essay will try to explain how large. (Note: earlier, Obama said his plan would “save or create at least two and a half million jobs”; later Obama used the number 3 million). [Note: As of February 12, with the final version of the stimulus bill, the number is now 3.5 million jobs; I will update the table when I find a revised report from Obama's economists.]

Depending on how you measure employment, as I write in January 2009, there are around 140 million jobs in the United States.  Calculating the number of jobs is not as simple as it seems; the labor force is constantly changing, with some people leaving school and entering the labor force while others go back to school or retire, people being hired and fired and some people holding multiple jobs.  The Bureau of Labor Statistics calculates two measures of employment, the household measure (derived from a survey of households) and the establishment data (derived from a survey of employers); see this essay for more information on how the data are assembled.

Obama’s economic team use the payroll data (see here for their report).  As of the end of December, the BLS estimated that businesses employed 135,489,000 workers (what the Obama advisers call “payroll employment”).  They estimate that with the Obama stimulus plan, employment will rise to 137,550,000, resulting in an increase of roughly 2 million more jobs; but without the stimulus, employment would only be 133,876,000, so the Obama plan results in almost 4 million more jobs than would otherwise be the case.

The Obama team presents these numbers as the result of forecasts of what will happen without the stimulus compared to what will happen with the stimulus.  Here are the job data for the last five recessions and the estimates for the current recession:


Recession Start First 12 months Next 12 months Next 12 months
November 1973 +350 -580 +2558
January 1980 +231 -474 -1576
July 1981 -2084 +915 +4348
July 1990 -1539 +476 +2247
March 2001 -2080 -524 +877
December 2007 no stimulus -2589 -807 [est] -806 [est]
December 2007 with stimulus -2589 +1030 [est] +1031 [est]

Change in Payroll employment measured in thousands of workers.

To review the history of the recessions of the last 40 years; during every recession there was some stimulus, usually a combination of Fed rate cuts and an increase in the budget deficits. The Congressional Budget Office is already forecasting that the budget deficit in 2009 will be 8.3%, the largest deficit in the last 60 years.

The Obama team is forecasting that the recession that started in 2007, without the Obama stimulus, will be the worst recession in the last 40 years in terms of job loss, even with the large budget deficit and the aggressive rate cuts that the Fed has already made. They forecast larger job losses during the period after the recession started than has occurred in the last five recessions.  But economic forecasts are not 100% accurate; for example, it was only a nine months ago that candidate Obama was in favor of raising taxes when he took office although he noted that “we don’t know what the economy’s going to look like at that point.” At this point his team seems much more certain about where the economy will be in a few years.

Is the US in Recession in 2008? YES

Monday, December 1st, 2008

[Updated as of November 26 data; on December 1, 2008, the NBER announced that in December 2007 a recession started.  I will no longer update this article with but I will write more about the recession here.]

Many market analysts (e.g., Warren Buffett, Martin Feldstein and Alan Greenspan), have argued that the U.S. economy is in a recession. These pronouncements are relayed to the public without a clear definition of what a recession is and who actually determines whether the U.S. is in a recession. The common definition of two negative quarters of GDP growth is not correct, despite its widespread acceptance in the financial press.

This essay will explore the precise definition of a recession and provide a regularly updated view of the data that the committee that determines recessions will look at.

The NBER business cycle dating committee makes the official determination of when a recession starts (the peak of the previous economic cycle) and when one ends (the trough). On their site they write:

“The NBER does not define a recession in terms of two consecutive quarters of decline in real GDP. Rather, a recession is a significant decline in economic activity spread across the economy, lasting more than a few months, normally visible in real GDP, real income, employment, industrial production, and wholesale-retail sales.”

The committee track four specific indicators:

(1) personal income less transfer payments, in real terms and (2) employment. In addition, we refer to two indicators with coverage primarily of manufacturing and goods: (3) industrial production and (4) the volume of sales of the manufacturing and wholesale-retail sectors adjusted for price changes.

The committee also looks “at monthly estimates of real GDP such as those prepared by Macroeconomic Advisers” but notes that “there is no fixed rule about which other measures contribute information to the process.”

Here are the most recent data for the four key series that the NBER follows:

Date Real Personal Income Employment Ind Prod Sales
September 2007 8504.5 137837 112.3 973,028
October 2007 8507.5 137977 111.8 979,520
November 2007 8498.9 138037 112.3 973,785
December 2007 8495.0 138078 112.4 963,243
January 2008 8466.1 138002 112.6 968,367
February 2008 8464.9 137919 112.3 954,429
March 2008 8465.1 137831 112.0 954,909
April 2008 8435.2 137764 111.4 965,041
May 2008 8420.8 137717 111.2 962,718
June 2008 8384.1 137616 111.3 960,562
July 2008 8358.3 137550 111.4 952,280
August 2008 8370.6 137477 110.1 939.032
September 2008 8336.6 137138 105.9 921,459
October 2008 8424.7 136899 107.3 NA

Source: Department of Commerce, Bureau of Labor Statistics, Federal Reserve Board and author’s calculations; real personal income in billions of chained 2000 dollars, employment in thousands, and sales in billions of chained 2000 dollars.


Here are the same data expressed as a percentage of September 2007 values (some analysts have claimed that September 2007 represented a peak in the economy; in any event, values can be compared with January 2008 levels, claimed by others as a peak). The drop associated with the average of the last five recessions prior to 2001 is indicated at the bottom of the table in red.

Date Real Personal Income Employment Ind Prod Sales
September 2007 100.0 100.0 100.0 100.0
October 2007 100.0 100.1 99.6 100.7
November 2007 99.9 100.1 100.0 100.1
December 2007 99.9 100.2 100.1 99.0
January 2008 99.5 100.1 100.3 99.5
February 2008 99.5 100.1 100.0 98.1
March 2008 99.5 100.0 99.7 98.1
April 2008 99.2 99.9 99.2 99.2
May 2008 99.0 99.9 99.0 98.9
June 2008 98.6 99.8 99.1 98.7
July 2008 98.3 99.8 99.2 97.9
August 2008 98.4 99.7 98.0 96.5
September 2008 98.0 99.5 94.3 94.7
October 2008 99.1 99.3 95.5 NA
Recession Level NA 98.9 95.4 NA

Source: Department of Commerce, Bureau of Labor Statistics, Federal Reserve Board and author’s calculations; all data expressed as percentage of September 2007 values.


Then NBER has described the typical recession:

A recession involves a substantial decline in output and employment. In the past 6 recessions, industrial production fell by an average of 4.6 percent and employment by 1.1 percent. The Bureau waits until the data show whether or not a decline is large enough to qualify as a recession before declaring that a turning point in the economy is a true peak marking the onset of a recession.

As noted above, data as of October 2008 (industrial production and employment) are very close to recession levels and a recession is likely to be declared by the NBER in the next month or two.

Understanding Personal Consumption Expenditures (PCE)

Wednesday, August 20th, 2008

The U.S. is a consumption driven economy; personal consumption expenditures (PCE), that is consumption by households and nonprofits represent approximately 70% of GDP; in comparison, Japanese private consumption is a bit less than 60% of GDP (Japanese GDP data are available here) and German final consumer expenditure is around 57% (see here for German GDP data).

The U.S. is in the process of changing the way it presents PCE data; in this article I will briefly describe the new method to be implemented in 2009, described by the BEA in an article published in May 2008 (see here for the pdf).  The main change are that the BEA is using the international System of National Accounts (SNA) classifications and making the PCE data more compatible with other data that the government compiles (such as the consumer price index, CPI).

Roughly 98% of PCE are household consumption expenditures (the rest are the consumption expenditures of nonprofit institutions).  Health and housing are the largest categories, each roughly 18% of PCE, followed by transportation (11%) and then recreation, food purchased for off-premises consumption (largely food consumed at home), financial services, food services and accomodations (food consumed in restaurants and hotel expenditures) and other goods, services and household furnishing and clothing, that comprise most of the rest of the category (between 4 and 9% each).

One way that BEA classifies PCE is by the type of good purchased; goods are divided into nondurable goods (23%), durable goods (13%) and services including purchased meals and beverages (64%).  The services category is greatly expanded by including the purchased meals and beverages category (formerly included in nondurable goods).  While the behavior of the US consumer did not change, this reclassification emphasizes the size of the US services sector.

Unlike GDP numbers, PCE is available both monthly and quarterly.  Normally the monthly PCE number comes out a day or two after the GDP number.  For example, the first revision of second quarter 2008 GDP (so-called preliminary GDP) will be released on August 28 and then July PCE numbers will be released on August 29.  This schedule creates a slightly odd result: at the end of August, we get a revision of 2008:II (April to June GDP numbers) and July PCE; at the end of September we get a final revision of 2008:II GDP and August PCE; by now we know 2/3 of the PCE data for the third quarter.  At the end of October the advance estimate of 2008:III GDP is released; this release includes the quarterly PCE.  Yet we do not learn the September monthly PCE until a few days later (there is some residual uncertainty about the PCE in the third month, even when the quarterly PCE is already known, as there are usually revisions to the earlier monthly numbers).

Understanding Data Announcements

Thursday, August 14th, 2008

If you look at a data calendar (see here for a calendar and here for my essay) you can find out when certain data are released (for example, the Employment Report is typically released on the first Friday of the month at 8:30 am Eastern Time).  The Employment Report essay (here) explains how the Bureau of Labor Statistics calculates the numbers that are included in the report; in this essay, we explore the announcement process, that is how analysts forecast the data, how the consensus forecast is obtained, what are “whisper” numbers, the lock-up and what happens at 8:30 am on the first Friday of the month. For the purposes of this essay I will use the monthly announcement of the Employment Report but the same analysis could be applied to other announcements.

Economic analysts regularly forecast upcoming economic numbers.  These forecasts range from sophisticated statistical models (see here for a pdf of a recent academic paper that examines 30 variables that could be used to forecast the change in employment) to “back of the envelope” calculations (like the one described here).  Companies that provide data services to traders (such as Bloomberg and Reuters) routinely ask economists for their forecasts and use them to find the average or median forecast.

There is a lengthy literature behind the idea (sometimes called the Delphi method) that the average or median forecasts typically outperforms any individual forecaster (see here for an online book with an extensive bibliography).  But economic analysts work for profit making enterprises and do not wish to give away their best forecasts for free.  So they play a game something like this: early in the process they give away to Bloomberg a free forecast, which may (or may not) reflect their actual thinking.  The consensus forecast is formed on the basis of these early forecasts.  But as more information becomes available (closer to the data of the announcement) the forecaster may only distribute his forecast to preferred clients; this forecast is referred to as a “whisper” number, that is not printed and distributed (there are a number of services that claim to provide whisper numbers for company earnings: see here, here and here but none that I know of that provide whisper numbers for economic data; please send me an email at ckendall ##at## if you know of any).

At 8:30 am the data are released and broadcast on the news wires and television; but before then there are approximately 20 journalists who have seen the data and written reports at the “lock-up” (see here for an excellent essay describing the lock-up).  Accredited journalists from major news organizations receive the data release a half-hour early and prepare their articles that can be released at exactly 8:30 (they are “locked up” in a room where they are not allowed to contact the outside world before the official release).  The data are usually available to senior government officials the night before the release so that the appropriate cabinet secretary can be prepared to speak just after the release.

Then at 8:30 the data are released to the public; typically the data are announced on television (CNBC and elsewhere), via paid data services to their clients, on news wires and on the web sites of the agencies that release the data.  Traders often react to the initial announcements and analysis by the reporters in the lock-up before; later analysts will release comments on what the data really mean through to their traders, through wire services and in letters to clients (a service performed by the author of this article).  But an hour later there is another announcement and a month later a revision and five years after that a benchmark revision and the data will look very different.

Bad Behavior has blocked 33 access attempts in the last 7 days.

FireStats icon Powered by FireStats