I’ve heard a lot about the “four years ago” comparison. Four years ago, we were on the cusp of Don Luskin’s famous prediction (“… we’re on the brink not of recession, but of accelerating prosperity.”), and Phill Gramm had two months earlier decried the ongoing “mental recession”.  It seems to me the more appropriate marker is the last election, in 2008Q4. We can then assess what the data tells us about 2012Q2 vis a vis 2008Q4.
Aggregate Measures of Activity and Income
First, I look to several measures in addition to the standard real GDP indicator. Note that I divide by population, in order to express the variables in per capita terms (thereby controlling for population growth). In the graph below, when the series lie above the horizontal line at zero, then the indicator exceeds its level at 2008Q4. The numerical value can be interpreted as the percentage deviation from 2008Q4 levels, in log terms.
Figure 1: Log real per capita GDP (blue), GNP (red), Net National Product (green), Gross Domestic Income (purple), all SAAR in Ch.2005$, normalized to 2008Q4=0. Source: BEA, 2012Q2 second release.
In other words, aggregate output and income measures on a per capita basis exceed levels recorded at the end of the G.W. Bush Administration. (Given population growth, obviously absolute levels exceed by an even greater extent). If one looks to production by American-owned factors of production (as opposed to factors of production located within the geographical confines of the Nation), then output growth is even more pronounced.
If one refers the first derivative, then the case for improvement is reinforced. Per capita GDP was declining 10.3% (SAAR, log terms) in 2008Q4, and rose 1.0% in 2012Q2.
Consumption and Household Net Worth
Consumption is the object typically thought of as the measure entering into the utility function. We do not observe consumption, but we do observe consumption expenditures. Per capita consumption expenditures is above levels of 2008Q4. Consistent with that finding, net household worth is also above corresponding levels.
Figure 2: Log real per capita consumption (blue), SAAR in Ch.2005$, and net household worth (red), normalized to 2008Q4=0. Source: BEA, 2012Q2 second release, and Fed Flow of Funds.
Clearly, both measures exceed those recorded at the end of the Bush Administration; per capita consumption was 2.8% higher, while as of 2012Q1, net worth was 7.6% higher. I find both results remarkable, given the fact that these indicators are often depressed after large financial balance-sheet induced recessions.
The gradient is also relevant; per capita consumption declined 6.2% in 2008Q4, and rose 1.0% in 2012Q2 (SAAR, log terms).
The Private Sector
What about real compensation and productivity? BLS data indicate that both have risen.
Figure 3: Log real compensation (blue) and output per hour (red) in the nonfarm business sector, normalized to 2008Q4=0. Source: BLS via FRED, and author’s calculations.
The fact that real compensation is essentially at 2008Q4 levels is not encouraging. Nor is the fact that output per hour is far above compensation, as it indicates that a higher share of income is going to capital in the nonfarm business sector (this is confirmed by BLS data on the income share; it’s been decreasing since 2000 pretty much without stop). In this sense, the owners of capital have been doing particularly well. As an aside, on this Labor Day, it seems to me that measures that would allow labor to garner a larger share of national income are in order. Had the greater demand support measures been implemented, higher demand for labor would have induced greater upward pressure on wages. Unfortunately, additional measures to support aggregate demand have been stifled in the Congress.
As economists (actually, as sensible people), we should be interested in the difference relative to the counterfactual. The counterfactual (as well as conditioning) is a difficult concept for many people to understand, but I’ll forge ahead, using the CBO assessment of high-low range of impacts to determine the counterfactual level of GDP in the absence of the American Recovery and Reinvestment Act (ARRA). In my view, this range encompasses the range of multipliers associated with reasonable models of New Keynesian and Keynesian/neoclassical synthesis flavors.
Figure 4: Real GDP (thick black), WSJ mean forecast from August survey (thick blue), counterfactual GDP w/o American Recovery and Reinvestment Act, low impact (green), and high impact (red). NBER defined recession dates shaded gray. Vertical line at 2008Q4. Source: BEA, 2012Q2 second release, WSJ, CBO (Aug. 23) Table 1, NBER, and author’s calculations.
While by either measure, the positive impact of the ARRA has diminished, it should be remembered that in 2010, output was substantially higher (and unemployment lower, between 0.4 to 1.8 percentage points) than it otherwise would have been. Just because it’s in the past, doesn’t mean that it doesn’t matter. Furthermore, the midpoint of the CBO range would place economic output in 2012Q3 0.4 percent above what it would have been in the absence of the ARRA. (For discussion of why one might expect multipliers to be at the higher end, see here).
(As an aside, note that even when one uses the pessimistic measures of multipliers, output would have continued to decline after 2009Q2 in the absence of the ARRA. I do wonder how sensible people can continue to say “the stimulus didn’t work”. Once again, to me, this is the triumph of ideology over empirics.)
Or, what would Romney have done?
Finally, I think it useful to consider a regime that was completely laissez faire with respect to interventions should the economy encounter new challenges. We have one observation on Governor Romney’s policy perspective — namely what should have been done with Detroit in 2009. He recommended letting the companies go bankrupt.  While auto manufacturing does not represent all of the US economy, it remains important (e.g., ) and so this policy prescription should be taken as an important indicator of his approach.
Figure 5 depicts how motor vehicle production and new vehicle consumption has evolved since Governor Romney’s admonition to let Detroit go bankrupt.
Figure 5: US motor vehicle output (blue) and consumption of new vehicles (red), in bn. Ch.05$. Dashed line at 2008Q4 (November 2008). NBER defined recession dates shaded gray. Vertical dashed line at 2008Q4 (Romney’s NYT OpEd). Source: BEA, Table 7.2.6U. Real Motor Vehicle Output, August 29, 2012 release, and author’s calculations.
In other words, the real value of production of motor vehicles is now 40% above those recorded in 2008Q4 (in log terms). Once again, one can think of the counterfactual. After an uncontrolled bankruptcy, in which no financing was available due to the contemporaneous financial crisis, would the auto industry have risen like a phoenix, outshining what we have witnessed thus far? It’s possible. After all, I ran a 45 second 440 in high school. Plus or minus 25 seconds… 
Update, Tue, 9/4 8:30PM Pacific:
Reader tj adopts the “4 years ago” metric, which places things in the middle of the last year of the Bush Administration. Why. Beats me (well, we all know why). But here is a normalization based upon 2009M01; since the BLS data on employment are based on surveys mid-month, this is the latest observation before the beginning of the Obama Administration.
Figure 6: Log nonfarm payroll employment (blue), and Bloomberg consensus as of 9/4/2012 (light blue); log private nonfarm payroll (red), and consensus (pink); and civilian employment (green), all seasonally adjusted, normalized to 2009M01=0. Source: BLS, July release, Bloomberg, and author’s calculations.
Civilian employment and private nonfarm payroll employment exceed 2009M01 levels as of July. Nonfarm payroll employment as of July remained below 2009M01 levels, and Bloomberg consensus has it still below as of August. Since the difference between NFP and private employment is government employment, we can see that if you’re a government worker, or received services from government workers or learned from public school teachers, you are not necessarily better off.
Update, 9/6 6PM Pacific:The stock market price index now exceeds four years ago, or January 2009.
Figure 7: S&P 500, monthly averages of daily data (blue), and close on 9/6 (red square), and CPI (2010=1) deflated SP500 (green). NBER defined recession dates shaded gray. Long dashed vertical line 4 years ago at 2008M09, and short vertical dashed line at 2009M01. Source: FRED, NBER, author’s calculations.