This statement is indeed accurate if one focuses on real GDP. I present the log of real GDP in Chained 2000$, normalized to 0 at the NBER-defined trough in 2001Q4. I also present for reference log GDP in the previous two expansions, normalized to 0 in the previous troughs in 1991Q1 and 1982Q4. (For those interested in output gaps, the mean WSJ forecast predicts output will be 2.7% below CBO's estimate of potential output by end-2008, in log terms.)

The graph clearly indicates that GDP is now substantially below (4%, in log terms) where it was at this time in the previous expansion. It is 12% below the corresponding level in the 1982-00 expansion.

I also plot the implied level of GDP assuming the mean forecast from the WSJ survey for GDP is realized. In that case, by 2009Q2, output will be 7% below the corresponding level in the previous expansion.

While the mean forecast is of interest, so too is the dispersion in forecasts. In the figure below, I plot the mean forecast, and the highest and lowest forecasts (using as the reference for highest and lowest the 2008 Q4/Q4 growth rate), from James F. Smith of Western Carolina University and Parsec Financial Management (2.82%), and Paul Ashworth of Capital Economics (0.24%), respectively.

The mean forecast projects a bump up in growth in 2009Q1. The high forecast by James Smith projects rapid growth all the way through the forecast period. Of course, Smith in the November 2007 survey provided the highest forecast of the surveyed economists for every quarter from 2007Q4-2008Q4, as well as the highest for 2008 Q4/Q4. (In March 2008, Smith also had the highest forecast for 2008Q1 growth, at 2.6%, and tied for highest Q4/Q4 growth, at 3.5%.) The lowest Q4/Q4 forecast comes from Paul Ashworth; he forecasts a 0.24% growth rate for 2008.

In Figure 3, I present a histogram of the 2008 Q4/Q4 growth rates, as implied by the q/q growth rates reported by the respondents to the survey. Smith is the single observation in the top bin. Ashworth is one of four respondents in the bottom bin. Highlighting the skewed nature of the distribution, Smith is more than 2 standard deviations from mean, while Ashworth is about 1.4 standard deviations below mean.

If we take the most optimistic forecaster's view, then by 2009Q2, GDP will be only 4.8% below the corresponding point in the previous expansion. Taking the pessimistic view, it will be 8.3% below.

These forecasts were conducted in the first week of July. Hence, they predate the surprise jump in CPI inflation. Higher than anticipated inflation could have a variety of offsetting impacts. First, it could signal greater demand pressures, so GDP forecasts would be revised upward. On the other hand, taking into account the Fed’s reaction function, one might think that monetary policy would therefore be tighter than otherwise, and hence output -- at least several quarters out -- would be lower than previously projected.

Today, additional information was released, for industrial production. Here the picture is mixed. Both industrial production and manufacturing production (recall industrial production includes output from utilities which can be influenced by weather factors, even when seasonally adjusted) rose slightly (I think "edged up" is the phrase).

Note that by the metric of the Nation’s factories, utilities and mines, the economy is not growing, over a 3 month moving period.

Finally, both Macroeconomic Advisers and e-Forecasting released estimates of monthly GDP today (May for MA, and June for e-Forecasting). These series are presented in Figure 4. At this point we have a substantial disjuncture between the implications arising from the two organizations' assessments. E-Forecasting’s figures indicate that in 2008Q2 GDP has declined 0.6% q/q on an annualized basis, while MA is predicting a 2.9% q/q growth rate, taking into account the April and May estimates and a forecasted 0.6% growth (annualized) in June.

Without knowing more about the track records of each of the estimates of GDP, I'm reluctant to say which one better tracks the state of the economy (or, less ambitiously, track the state of the advance GDP release…) . Any informed commentary or references from readers welcome.

What is true is that over the 1992-current period, an estimated error correction model assuming cointegrating between the two series (with long run unitary elasticity) indicates that the MA series reverts more strongly to conditional mean than does the e-Forecasting series. In a statistical sense, then, one might think the last observations of the e-Forecasting series are more likely to be closer to the underlying trend represented by both series (recalling that both series are calibrated to match the quarterly actual GDP series reported by BEA). In an informal sense, I think of this statistical result as meaning that the MA series bounces around more than the e-Forecasting series. This is not a conclusion with economic content, though, as it doesn't say which series better represents the economy's state. It might be the case that economic output does jump around a lot, month to month.

Technorati Tags: potential GDP, recession, output+gap, industrial+production, forecasts.

Several observations:

• As in recent years, the NIIP to GDP ratio continues to deteriorate.
• However, the NIIP to GDP ratio as of end-2007 is improved relative to the originally reported end-2006 NIIP/GDP ratio.
• In the last year, the dollar change in the NIIP deviates substantially (i.e., is more positive) than the corresponding current account reported on a NIPA basis. This repeats the pattern from the previous five years.
• Interestingly, 2005 stands out by far as an outlier, wherein the the NIIP improves while the CA is in substantial deficit.
• There appears to be a measurable correlation between dollar depreciation against other major currencies and the deviation between change in NIIP and CA.

The first two observations are illustrated in Figure 1.

The second two observations are illustrated by Figure 2.

The last observation is illustrated by the scatterplot in Figure 3, applying to data over the 1992-07 period.

While the regression coefficient is not significant, allowing for a dummy for the anomalous 2005 observation (Brad Setser's decomposition attributes this effect to mostly market valuation changes, if I read his graph correctly), one finds that each 10% depreciation in the dollar against a basket of major currencies induces a 2.5% increase in the gap between the NIIP change and the current account (all expressed as a ratio of GDP). Specifically:

dev = 0.01 + 0.25dep + 0.097dum05

Adj-R² = 0.68, SER = 0.017, N=16, all coeffs sig at 5% msl.

Should the dollar appreciate in 2008 against other major currencies, then one should expect the valuation effect to be reversed. A cessation of dollar depreciation will mean on average the gap between the change in NIIP and the CA (as a share of GDP) will be only 1%. (Another point I've highlighted is that the cumulative current account and NIIP converge during recessions -- see "Can Gravity Be Defied?".

In the title of this post, I mentioned "What's Missing". One interesting point highlighted in Frank Warnock, et al.'s paper, presented at the UW-Madison conference on current account sustainability a couple months ago is that it is difficult to rely upon some dark matter explanation for the differential rates of return on US owned assets abroad versus foreign owned assets in the US. Rather, previous studies have mismeasured the stocks of assets, and mismatched flows to stocks, imparting substantial measurement error to rates of return.

In an extensive forensic analysis indicates Warnock et al. conclude:

In this paper we provided a brief summary of some of the theories of U.S. current account sustainability and viewed them through the lens of the relative reliability of various items in the international accounts. From the perspective of relative reliability, the dark matter view fails, as it rests on an assumption that income streams are the most accurate items in the entire set of international accounts. Given that the bulk of income streams are themselves estimates based on other items in the accounts, this assumption is false. The exorbitant privilege view also fails. In its original form it rested on the assumptions that the current vintage of revised positions and flows form a consistent dataset and that all "other adjustments" are best thought of as valuation adjustments. In this paper we show that this is not true, in part by calculating “other adjustments” by asset class and filling some known holes in the international accounts. The set of accounts we produce by doing so are entirely consistent with a small cross-border returns differential, suggesting that there is no evidence that the U.S. can earn its way to current account sustainability.

One particularly interesting point Frank identifies, which was news to me, was that real estate purchases are not included in the balance of payments data, and hence not in the international investment position estimates.

In principle, cross-border transactions and holdings of residential real estate should be included as part of direct investment, as is currently the case for commercial real estate. In practice, individual homeowners are not surveyed and hence these data are omitted from the recorded DI figures. To the extent that foreign activity in the U.S. residential real estate market is of the same magnitude as the level of activity of U.S. residents in the foreign real estate markets, there is no net impact on the international transactions accounts. However, recent surveys conducted by the National Association of Realtors (NAR) suggest that this may not be the case.

Additional discussion of the IIP release from Brad Setser.

Figure 1 depicts the time series for year-on-year nonresidential investment growth, and residential investment growth lagged four quarters. There's an obvious correlation, but clearly it's not a particularly strong one. There are periods where business fixed investment levitates above residential growth, such as the latter part of the 1980s (due to the dollar's depreciation), and during the 1990s, as well as the most recent few quarters. The relationship can be summarized using a OLS regression of 4-quarter growth on 4-quarter growth rates.

d(nresinv)_t = 0.035 + 0.311 d(resinv)_t-1 + u_t

Adj. R² 0.37, SER = 0.054, smpl 1967q1-08q1, bold face denotes statistical significance at 1% MSL.

Using quarter-on-quarter growth rates yields a similar result (i.e., the cumulative coefficients on four lags of changes in residential investment sum to 0.33, as opposed to the 0.311 coefficient above).

What does this predict for year-on-year business fixed investment growth in 2009Q1? That's shown in Figure 2.

The regression indicates a 3.8% 4 quarter decline (in log terms) in 2009q1. It's unclear to me how this fits in with other forecasts. The q4/q4 prediction I obtain of -2.9% is somewhat less than the 0.3% (non-log terms) decline forecasted by the participants of the Chicago Fed Automotive Outlook symposium [1].

An alternative approach would be to try to explain investment using variables that theory says might be important. As discussed in this post on the investment disconnect, corporate profits is one candidate. In Figure 3, I plot 4 quarter changes in nonresidential investment and in real corporate profits.

This relationship,

d(nresinv)_t = 0.034 + 0.317 d(cprofits)_t-1 + u_t

Adj. R² 0.29, SER = 0.058, smpl 1967q1-08q1, bold face denotes statistical significance at 1% MSL.

predicts that the 4 quarter change in nonresidential investment will be about 0.3 ppts. in 2009q1.

In either case -- using a lead-lag relationship [2], or one based upon a version of a cash-flow model -- investment growth will be slowing down over the next year.

Technorati Tags: nonresidential investment, residential investment, corporate profits.

The easiest thing to do is to pull down the CBO's measure (interpolated to quarterly frequency). This yields the following picture (in logs):

Two observations: (i) recessions do not necessarily coincide with negative output gaps (although they do seem to coincide with the beginning of periods of negative output gaps); and (ii) recoveries do not always coincide with positive output gaps.

This is obvious when one thinks about it, given the NBER BCDC definition of a recession as the following:

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. A recession begins just after the economy reaches a peak of activity and ends as the economy reaches its trough. Between trough and peak, the economy is in an expansion. Expansion is the normal state of the economy; most recessions are brief and they have been rare in recent decades.

That is a recession is the description of the first derivative of output taking on a negative value, while an output gap is a description of output relative to the output level consistent with the "normal" utilization of factors of production (also "full-employment"). For details of how CBO calculates potential GDP, see this document.

Looking at the last ten years provides another insight.

The US economy only barely made it close to full-employment in the expansion of 2001-2007/8/?, and is now declining again. I think I had realized this (especially in the course of various past debates over the surge in tax receipts, which I think has now been resolved -- see [1], [2] for a recollection.)

As I noted in the beginning of this post, I was finding it difficult to discern the most appropriate measure, so I will share the other measures I have investigated as indicators of potential (or "trend") GDP (which can then be used to calculate corresponding measures of the output gap): (1) the Hodrick-Prescott filter; (2) the Band Pass filter; and the quadratic time trend.

The Hodrick-Prescott filter is a ubiquitous two-sided filter used in time series macroeconometrics and elsewhere (I've even seen it used in the analysis of temperature data! [3]). Essentially, the HP filter calculates a trend that minimizes the weighted sum of squared deviations from trend, and squared changes in the the growth rate of the trend. This weighting is controlled by a parameter which is usually set at 1600 for quarterly data. As a public service, I'll note there are serious hazards associated with this filter, especially when trying to correlate various macro series that have been put through the same filter [4].

There is an additional problem (which is often ignored), namely that the HP filter is two-sided, so that running the filter up to the end point of data will tend to result in the trend being too close to the last data point (in our case, the output gap will be pulled to zero).

There are several band pass filters; the one most commonly used in the macroeconometric literature is the Baxter-King version.

Band pass filters are called this because (in the frequency domain) they pass through any cyclical components within a particular frequency band, and elimates the others. In the time domain, this means fluctuations that are shorter or longer than a specific length are ignored. In the business cycle area, then, in order to use this filter, one would have to have a prior on how long a "typical" business cycle is. More on both the HP and BP filters can be found in Tim Cogley's entry for the New Palgrave Dictionary of Economics [5].

The final means of identifying the output gap is to the simplest (at least in implementation) -- take the deviation from an estimated quadratic trend in time. This is not an uncommon procedure, but if you are pretty confident there is a unit root in log GDP, you might feel a little queasy about doing this [6]. But for completeness' sake, I'll show what happens when you do this as well.

Note that in order to circumvent the two-sided filter aspect of the HP filter, I have done a standard fix, which is to use an ARIMA(1,1,1) on log GDP over the 1980-08q1 period to dynamically forecast out 12 quarters, and then apply the HP filter to this "extended" series. I could have done a similar procedure for the BP filter, but opted to use the Christiano and Fitzgerald (2003) one-sided asymmetric version of the band pass filter to estimate the trend series.

First, the good news. Using the band pass filter, one finds that the output gap is still positive, at less than one percentage point of GDP in 2008q1. The HP filter indicates an essentially zero output gap. The quadratic trend indicates something similar to what the CBO indicates -- something close to a 2 percentage point negative output gap in 2008q1, versus 1.5 percentage points for CBO.

I don't want to say that there is one best version. Variations in the sample period, and the parameters used in each filter, will change the results. And of course, data revisions will mean the real-time output gaps will differ from the final revised output gaps we estimate today using either mechanical or judgmental approaches. For a pessimistic view regarding real time use, see Orphanides and van Norden (2004) [pdf].

Nonetheless, the distinction between the rate of change in economic output, and where output levels will gravitate to (and at what pace) is a useful one. And I think it will become of greater usefulness as the slowdown ends, if and when growth resumes. In particular, I'm thinking of whether the pattern of smaller -- but more persistent -- ups and downs continues (the half life of a deviation from potential has risen from 8 quarters in the 1970-90q1 period to 11.6 quarters in the 1990q2-2008q1 period).

Technorati Tags: potential GDP, recession, output+gap, full employment output, trend.

One point to keep in mind, when comparing against previous downturns: for the last few months, the indicators are either preliminary or once/twice revised, while viewing back in time, one will be looking at final, revised, data. For the issue of vintage data and revisions, see these posts (see Creative Destruction, as well as these posts [3], [4], [5]). To access vintage data, see the St. Louis Fed's ALFRED system.

With that caveat in mind, it looks to me like we've passed at least a local maximum, and indicators are trending down.