The Atlanta Fed's macroblog provides commentary and analysis on economic topics including monetary policy, macroeconomic developments, inflation, labor economics, and financial issues.
- BLS Handbook of Methods
- Bureau of Economic Analysis
- Bureau of Labor Statistics
- Congressional Budget Office
- Economic Data - FRED® II, St. Louis Fed
- Office of Management and Budget
- Statistics: Releases and Historical Data, Board of Governors
- U.S. Census Bureau Economic Programs
- White House Economic Statistics Briefing Room
March 05, 2015
Could Reduced Drilling Also Reduce GDP Growth?
Five or six times each month, the Atlanta Fed posts a "nowcast" of real gross domestic product (GDP) growth from the Atlanta Fed's GDPNow model. The most recent model nowcast for first-quarter real GDP growth is provided in table 1 below alongside alternative forecasts from the Philadelphia Fed's quarterly Survey of Professional Forecasters (SPF) and the CNBC/Moody's Analytics Rapid Update survey. The Atlanta Fed's nowcast of 1.2 percent growth is considerably lower than both the SPF forecast (2.7 percent) and the Rapid Update forecast (2.6 percent).
Why the discrepancy? The less frequently updated SPF forecast (now nearly a month old) has the advantage of including forecasts of major subcomponents of GDP. Comparing the subcomponent forecasts from the SPF with those from the GDPNow model reveals that no single factor explains the difference between the two GDP forecasts. The GDPNow model forecasts of the real growth rates of consumer spending, residential investment, and government spending are all somewhat weaker than the SPF forecasts. Together these subcomponents account for just under 1.0 percentage point of the 1.5 percentage point difference between the GDP growth forecasts.
Most of the remaining difference in the GDP forecasts is the result of the different forecasts for real business fixed investment (BFI) growth. The GDPNow model projects a sharp 13.5 percent falloff in nonresidential structures investment that largely offsets the reasonably strong increases in the other two subcomponents of BFI. Much of this decline is due to petroleum and natural gas well exploration; a component which accounts for almost 30 percent of nonresidential structures investment and looks like it will fall sharply this quarter. The remainder of this blog entry "drills" down into this portion of the nonresidential structures forecast (pun intended). (A related recent analysis using the GDPNow model has been done here).
A December macroblog post I coauthored with Atlanta Fed research director Dave Altig presented some statistical evidence that in the past, large declines in oil prices have had a pronounced negative effect on oil and mining investment. Chart 1 below shows that history appears to be repeating itself.
The Baker Hughes weekly series on active rotary rigs for oil and natural gas wells has plummeted from 1,929 for the week ending November 21 to 1,267 for the week ending February 27. The Baker Hughes data are the monthly source series for drilling oil and gas wells industrial production (IP) and one of the two quarterly source series for the U.S. Bureau of Economic Analysis's (BEA) estimate of drilling investment (for example, petroleum and natural gas exploration and wells). The other source series for drilling investment is footage drilled completions from the American Petroleum Institute, released about a week before the BEA publishes its initial estimate of GDP.
Chart 2 displays three of these indicators of drilling activity. The data are plotted in logarithms so that one-quarter changes approximate quarterly growth rates. The chart makes clear that the changes in each of the three series are highly correlated, suggesting that the Baker Hughes rig count can be used to forecast the other series. The Baker Hughes data end on February 27, and we can (perhaps conservatively) extrapolate it forward by assuming it remains at its last reading of 1,267 active rigs through the end of the quarter. We can then use a simple regression to forecast the February and March readings of drilling oil and gas wells IP. Another simple regression with the IP drilling series and its first-quarter forecast allows us to project first-quarter real drilling investment. The forecasts, shown as dashed lines in chart 2, imply real drilling investment will decline at an annual rate of 52 percent in the first quarter. This decline is steeper than the current GDPNow model forecast of a 36 percent decline as the latter does not account for the decline in active rotary rigs in February.
A 52 percent decline in real nonresidential investment in drilling would likely subtract about 0.5 percentage point off of first-quarter real GDP growth. However, it's important to keep in mind that a lot of first quarter source data for GDP are not yet available. In particular, almost none of the source data for the volatile net exports and inventory investment GDP subcomponents have been released. So considerable uncertainty still surrounds real GDP growth this quarter.
TrackBack URL for this entry:
Listed below are links to blogs that reference Could Reduced Drilling Also Reduce GDP Growth?:
January 09, 2015
Gauging Inflation Expectations with Surveys, Part 3: Do Firms Know What They Don’t Know?
In the previous two macroblog posts, we introduced you to the inflation expectations of firms and argued that the question you ask matters a lot. In this week's final post, we examine another important dimension of our data: inflation uncertainty, a topic of some deliberation at the last Federal Open Market Committee meeting (according to the recently released minutes).
Survey data typically measure only the inflation expectation of a respondent, not the certainty surrounding that prediction. As a result, survey-based measures often use the disagreement among respondents as a proxy for uncertainty, but as Rob Rich, Joe Tracy, and Matt Ploenzke at the New York Fed caution in this recent blog post, you probably shouldn't do this.
Because we derive business inflation expectations from the probabilities that each firm assigns to various unit cost outcomes, we can measure the inflation uncertainty of a respondent directly. And that allows us to investigate whether uncertainty plays a role in the accuracy of firm inflation predictions. We wanted to know: Do firms know what they don't know?The following table, adapted from our recent working paper, reports the accuracy of a business inflation forecast relative to the firm's inflation uncertainty at the time the forecast was made. We first compare the prediction accuracy of firms who have a larger-than-average degree of prediction uncertainty against those with less-than-average uncertainty. We also compare the most uncertain firms with the least uncertain firms.
On average, firms provide relatively accurate, unbiased assessments of their future unit cost changes. But the results also clearly support the conclusion that more uncertain respondents tend to be significantly less accurate inflation forecasters.Maybe this result doesn't strike you as mind-blowing. Wouldn't you expect firms with the greatest inflation uncertainty to make the least accurate inflation predictions? We would, too. But isn't it refreshing to know that business decision-makers know when they are making decisions under uncertainty? And we also think that monitoring how certain respondents are about their inflation expectation, in addition to whether the average expectation for the group has changed, should prove useful when evaluating how well inflation expectations are anchored. If you think so too, you can monitor both on our website's Inflation Project page.
TrackBack URL for this entry:
Listed below are links to blogs that reference Gauging Inflation Expectations with Surveys, Part 3: Do Firms Know What They Don’t Know?:
January 07, 2015
Gauging Inflation Expectations with Surveys, Part 2: The Question You Ask Matters—A Lot
In our previous macroblog post, we discussed the inflation expectations of firms and observed that—while on average these expectations look similar to that of professional forecasters—they reveal considerably more variation of opinion. Further, the inflation expectations of firms look very different from what we see in the household survey of inflation expectations.
The usual focal point when trying to explain measurement differences among surveys of inflation expectations is the respondent, or who is taking the survey. In the previous macroblog post, we noted that some researchers have indicated that not all households are equally informed about inflation trends and that their expectations are somehow biased by this ignorance. For example, Christopher Carroll over at Johns Hopkins suggests that households update their inflation expectations through the news, and some may only infrequently read the press. Another example comes from a group of researchers at the New York Fed and Carnegie Mellon They've suggested that less financially literate households tend to persistently have the highest inflation expectations.
But what these and related research assume is that whom you ask the question of is of primary significance. Could it be that it's the question being asked that accounts for such disagreement among the surveys?
We know, for example, that professional forecasters are asked to predict a particular inflation statistic, while households are simply asked about the behavior of "prices in general" and prices "on the average." To an economist, these amount to pretty much the same thing. But are they the same thing in the minds of non-economists?You may be surprised, but the answer is no (as a recent Atlanta Fed working paper discussed). When we asked our panel of firms to predict by how much "prices will change overall in the economy"—essentially the same question the University of Michigan asks households—business leaders make the same prediction we see in the survey of households: Their predictions seem high relative to the trend in the inflation data, and the range of opinion among businesses on where prices "overall in the economy" are headed is really, really wide (see the table).
But what if we ask businesses to predict a particular inflation statistic, as the Philly Fed asks professional forecasters to do? We did that, too. And you know what? Not only did a majority of our panelists (about two-thirds) say they were "familiar" with the inflation statistic, but their predictions looked remarkably similar to that of professional forecasters (see the table).
So when we ask firms to answer the same question asked of professional forecasters, we got back something that was very comparable to responses given by professional forecasters. But when you ask firms the same question typically asked of households, we got back responses that looked very much like what households report.
Moreover, we dug through the office file cabinets, remembering a related table adapted from a joint project between the Cleveland Fed and the Ohio State University that was highlighted in a 2001 Cleveland Fed Economic Commentary. In August 2001, a group of Ohio households were asked to provide their perception of how much the Consumer Price Index (CPI) had increased over the last 12 months, and we compared it with how much they thought "prices" had risen over the past 12 months.The households reported that the CPI had risen 3 percent—nearly identical to what the CPI actually rose over the period (2.7 percent). However, in responding to the vaguely worded notion of "prices," the average response was nearly 7 percent (see the table). So again, it seems that the loosely defined concept of "prices" is eliciting a response that looks nothing like what economists would call inflation.
So it turns out that the question you ask matters—a lot—more so, evidently, than to whom you ask the question. What's the right question to ask? We think it's the question most relevant to the decisions facing the person you are asking. In the case of firms (and others, we suspect), what's most relevant are the costs they think they are likely to face in the coming year. What is unlikely to be top-of-mind for business decision makers is the future behavior of an official inflation statistic or their thoughts on some ambiguous concept of general prices.
In the next macroblog post, we'll dig even deeper into the data.
TrackBack URL for this entry:
Listed below are links to blogs that reference Gauging Inflation Expectations with Surveys, Part 2: The Question You Ask Matters—A Lot:
January 05, 2015
Gauging Inflation Expectations with Surveys, Part 1: The Perspective of Firms
Central bankers measure inflation expectations in more than a few ways, which is another way of saying no measure of inflation expectations is entirely persuasive.
Survey data on inflation expectations are especially hard to interpret. Surveys of professional economists, such as the Federal Reserve Bank of Philadelphia's Survey of Professional Forecasters, reveal inflation expectations that, over time, track fairly close to the trend in the officially reported inflation data. But the inflation predictions by professional forecasters are extraordinarily similar and call into question whether they represent the broader population.
The inflation surveys of households, however, reveal a remarkably wide range of opinion on future inflation compared to those of professional forecasters. Really, really wide. For example, in any particular month, 13 percent of the University of Michigan's survey of households predicts year-ahead inflation to be more than 10 percent, an annual inflation rate not seen since October 1981. Even in the aggregate, the inflation predictions of households persistently track much higher than the officially reported inflation data (see the chart). These and other curious patterns in the household survey data call into question whether these data really represent the inflation predictions on which households act.
Even if you're unfamiliar with the literature on this subject, the above observations may not strike you as particularly hard to believe. Economists are, presumably, expert on inflation, while households experience inflation from their own unique—some would suggest even uninformed—perspectives.
We have yet another survey of inflation expectations, one from the perspective of businesses leaders. We think this may be an especially useful perspective on future inflation since business leaders, after all, are the price setters. Our survey has been in the field for a little more than three years now—just long enough, we think, to step back and take stock of what business inflation expectations look like, especially in comparison to the other survey data.
Our initial impressions are reported in a recent Atlanta Fed working paper, and the next few macroblog posts will share some of our favorite observations from this research.
We have been asking firms to assign probabilities to possible changes in their unit costs over the year ahead. From these probabilities, we compute how much firms think their costs are going to change in the coming year and how certain they are of that change (see the table). What we find is that the inflation expectations of firms, on average, look something like the inflation predictions of professional forecasters, but not so much like the predictions of households.
But we also find that there is a significant range of opinion among firms, more so than the range of opinions that forecasting professionals express. Some of the variation among firms appears to be related to their particular industries and are broadly correlated with the uneven cost pressures shown in similar industrial breakdowns of the Producer Price Index from the U.S. Bureau of Labor Statistics (see the table).
So what we have now are three surveys of inflation expectations, each yielding very different inflation predictions. What accounts for the variation we see across the surveys? Our survey allows us to experiment a bit, which was one of the motivations for conducting it. We didn't just want to measure the inflation expectations of firms; we wanted to learn about those expectations. In the next few macroblog posts, we'll tell you a few of the things we've learned. And we think some of our initial findings will surprise you.
TrackBack URL for this entry:
Listed below are links to blogs that reference Gauging Inflation Expectations with Surveys, Part 1: The Perspective of Firms:
December 04, 2014
The Long and Short of Falling Energy Prices
Earlier this week, The Wall Street Journal asked the $1.36 trillion question: Lower Gas Prices: How Big A Boost for the Economy?
We will take that as a stand-in for the more general question of how much the U.S. economy stands to gain from a drop in energy prices more generally. (The "$1.36 trillion" refers to an estimate of energy spending by the U.S. population in 2012.)
It's nice to be contemplating a question that amounts to pondering just how good a good situation can get. But, as the Journal blog item suggests, the rising profile of the United States as an energy producer is making the answer to this question more complicated than usual.
The data shown in chart 1 got our attention:
As a fraction of total investment on nonresidential structures, spending on mining exploration, shafts, and wells has been running near its 50-year high over the course of the current recovery. As a fraction of total business investment in equipment and structures, the current contribution of the mining and oil sector is higher than any time since the early 1980s (and generally much higher than most periods during the last half century).
In a recent paper, economists Soren Andersen, Ryan Kellogg, and Stephen Salant explain why this matters:
We show that crude oil production from existing wells in Texas does not respond to current or expected future oil prices... In contrast, the drilling of new wells exhibits a strong price response...
In short, the investment piece really matters.
We've done our own statistical investigations, asking the following question: What is the estimated impact of energy price shocks in the second half of this year on investment, consumer spending, and gross domestic product (GDP)?
If you are interested, you can find the details of the statistical model here. But here is the bottom line: the estimated impact of energy price shocks is a very sizeable decline in investment in the mining and oil subsector relative to baseline and, more importantly, an extended period of flat to slightly negative growth in overall investment relative to baseline (see chart 2).
In our simulations, the "baseline" is the scenario without the ex-post energy price shocks occurring in the third and fourth quarters of 2014, while the "alternative" scenario incorporates the (estimated) actual energy price shocks that have occurred in the second half of this year. These shocks lead to a cumulative 8 percent drop in consumer energy prices and a 6 percent drop in producer energy prices by the fourth quarter of this year relative to baseline. By the fourth quarter of 2017, 2 percentage points of these respective energy price declines are reversed. In chart 2 above, each colored line represents the percentage point difference between the "alternative" scenario and the "baseline" scenario.
As for consumption and GDP? Like overall investment, there is a short-run drag before the longer-term boom, as chart 3 shows:
So is the recent decline in energy prices good news for the U.S. economy? Right now our answer is yes, probably—but we may have to be patient.
Note: We have updated this post since it was originally released, clarifying a sentence in the paragraph above chart 2 and providing the data for the charts. The original sentence stated: But here is the bottom line: the estimated impact of energy price shocks is a very sizeable decline in investment in the mining and oil subsector and, more importantly, an extended period of flat to slightly negative growth in overall investment (see chart 2).
TrackBack URL for this entry:
Listed below are links to blogs that reference The Long and Short of Falling Energy Prices:
November 04, 2014
Data Dependence and Liftoff in the Federal Funds Rate
When asked "at which upcoming meeting do you think the FOMC [Federal Open Market Committee] will FIRST HIKE its target for the federal funds rate," 46 percent of the October Blue Chip Financial Forecasts panelists predicted that "liftoff" would occur at the June 2015 meeting, and 83 percent chose liftoff at one of the four scheduled meetings in the second and third quarters of next year.
Of course, this result does not imply that there is an 83 percent chance of liftoff occurring in the middle two quarters of next year. Respondents to the New York Fed's most recent Primary Dealer Survey put this liftoff probability for the middle two quarters of 2015 at only 51 percent. This more relatively certain forecast horizon for mid-2015 is consistent with the "data-dependence principle" that Chair Yellen mentioned at her September 17 press conference. The idea of data dependence is captured in this excerpt from the statement following the October 28–29 FOMC meeting:
[I]f incoming information indicates faster progress toward the Committee's employment and inflation objectives than the Committee now expects, then increases in the target range for the federal funds rate are likely to occur sooner than currently anticipated. Conversely, if progress proves slower than expected, then increases in the target range are likely to occur later than currently anticipated.
If the timing of liftoff is indeed data dependent, a natural extension is to gauge the likely "liftoff reaction function." In the current zero-lower bound (ZLB) environment, researchers at the University of North Carolina and the St. Louis Fed have analyzed monetary policy using shadow fed funds rates, shown in figure 1 below, estimated by Wu and Xia (2014) and Leo Krippner.
Unlike the standard fed funds rate, a shadow rate can be negative at the ZLB. The researchers found that the shadow rates, particularly Krippner's, act as fairly good proxies for monetary policy in the post-2008 ZLB period. Krippner also produces an expected time to liftoff, estimated from his model, shown in figure 1 above. His model's liftoff of December 2015 is six months after the most likely liftoff month identified by the aforementioned Blue Chip survey.
I included Krippner's shadow rate (spliced with the standard fed funds rate prior to December 2008) in a monthly Bayesian vector autoregression alongside the six other variables shown in figure 2 below.
The model assumes that the Fed cannot see contemporaneous values of the variables when setting the spliced policy—that is, the fed funds/shadow rate. This assumption is plausible given the approximately one-month lag in economic release dates. The baseline path assumes (and mechanically generates) liftoff in June 2015 with outcomes for the other variables, shown by the black lines, that roughly coincide with professional forecasts.
The alternative scenarios span the range of eight possible outcomes for low inflation/baseline inflation/high inflation and low growth/baseline growth/high growth in the figures above. For example, in figure 2 above, the high growth/low inflation scenario coincides with the green lines in the top three charts and the red lines in the bottom three charts. Forecasts for the spliced policy rate are conditional on the various growth/inflation scenarios, and "liftoff" in each scenario occurs when the spliced policy rate rises above the midpoint of the current target range for the funds rate (12.5 basis points).
The outcomes are shown in figure 3 below. At one extreme—high growth/high inflation—liftoff occurs in March 2015. At the other—low growth/low inflation—liftoff occurs beyond December 2015.
One should not interpret these projections too literally; the model uses a much narrower set of variables than the FOMC considers. Nonetheless, these scenarios illustrate that the model's forecasted liftoffs in the spliced policy rate are indeed consistent with the data-dependence principle.
By Pat Higgins, senior economist in the Atlanta Fed's research department
TrackBack URL for this entry:
Listed below are links to blogs that reference Data Dependence and Liftoff in the Federal Funds Rate:
September 29, 2014
On Bogs and Dots
Consider this scenario. You travel out of town to meet up with an old friend. Your hotel is walking distance to the appointed meeting place, across a large grassy field with which you are unfamiliar.
With good conditions, the walk is about 30 minutes but, to you, the quality of the terrain is not so certain. Though nobody seems to be able to tell you for sure, you believe that there is a 50-50 chance that the field is a bog, intermittently dotted with somewhat treacherous swampy traps. Though you believe you can reach your destination in about 30 minutes, the better part of wisdom is to go it slow. You accordingly allot double the time for traversing the field to your destination.
During your travels, of course, you will learn something about the nature of the field, and this discovery may alter your calculation about your arrival time. If you discover that you are indeed crossing a bog, you will correspondingly slow your gait and increase the estimated time to the other side. Or you may find that you are in fact on quite solid ground and consequently move up your estimated arrival time. Knowing all of this, you tell your friend to keep his cellphone on, as your final meeting time is going to be data dependent.
Which brings us to the infamous “dots,” ably described by several of our colleagues writing on the New York Fed’s Liberty Street Economics blog:
In January 2012, the FOMC began reporting participants’ FFR [federal funds rate] projections in the Summary of Economic Projections (SEP). Market participants colloquially refer to these projections as “the dots” (see the second chart on page 3 of the September 2014 SEP for an example). In particular, the dispersion of the dots represents disagreement among FOMC [Federal Open Market Committee] members about the future path of the policy rate.
The Liberty Street discussion focuses on why the policy rate paths differ among FOMC participants and across a central tendency of the SEPs and market participants. Quite correctly, in my view, the blog post’s authors draw attention to differences of opinion about the likely course of future economic conditions:
The most apparent reason is that each participant can have a different assessment of economic conditions that might call for different prescriptions for current and future monetary policy.
The Liberty Street post is a good piece, and I endorse every word of it. But there is another type of dispersion in the dots that seems to be the source of some confusion. This question, for example, is from Howard Schneider of Reuters, posed at the press conference held by Chair Yellen following the last FOMC meeting:
So if you would help us, I mean, square the circle a little bit—because having kept the guidance the same, having referred to significant underutilization of labor, having actually pushed GDP projections down a little bit, yet the rate path gets steeper and seems to be consolidating higher—so if it’s data dependent, what accounts for the faster projections on rate increases if the data aren’t moving in that direction?
The Chair’s response emphasized the modest nature of the changes, and how they might reflect modest improvements in certain aspects of the data. That response is certainly correct, but there is another point worth emphasizing: It is completely possible, and completely coherent, for the same individual to submit a “dot” with an earlier (or later) liftoff date of the policy rate, or a steeper (or flatter) path of the rate after liftoff, even though their submitted forecasts for GDP growth, inflation, and the unemployment rate have not changed at all.
This claim goes beyond the mere possibility that GDP, inflation, and unemployment (as officially defined) may not be sufficiently complete summaries of the economic conditions a policymaker might be concerned with.
The explanation lies in the metaphor of the bog. The estimated time of arrival to a destination—policy liftoff, for example—depends critically on the certainty with which the policymaker can assess the economic landscape. An adjustment to policy can, and should, proceed more quickly if the ground underfoot feels relatively solid. But if the terrain remains unfamiliar, and the possibility of falling into the swamp can’t be ruled out with any degree of confidence...well, a wise person moves just a bit more slowly.
Of course, as noted, once you begin to travel across the field and gain confidence that you are actually on terra firma, you can pick up the pace and adjust the estimated time of arrival accordingly.
To put all of this a bit more formally, an individual FOMC participant’s “reaction function”—the implicit rule that connects policy decisions to economic conditions—may not depend on just the numbers that that individual writes down for inflation, unemployment, or whatever. It might well—and in the case of our thinking here at the Atlanta Fed, it does—depend on the confidence with which those numbers are held.
For us, anyway, that confidence is growing. Don’t take that from me. Take it from Atlanta Fed President Lockhart, who said in a recent speech:
I'll close with this thought: there are always risks around a projection of any path forward. There is always considerable uncertainty. Given what I see today, I'm pretty confident in a medium-term outlook of continued moderate growth around 3 percent per annum accompanied by a substantial closing of the employment and inflation gaps. In general, I'm more confident today than a year ago.
Viewed in this light, the puzzle of moving dots without moving point estimates for economic conditions really shouldn’t be much of a puzzle at all.
By Dave Altig, executive vice president and research director of the Atlanta Fed
TrackBack URL for this entry:
Listed below are links to blogs that reference On Bogs and Dots:
July 21, 2014
GDP Growth: Will We Find a Higher Gear?
We are still more than a week away from receiving the advance report for U.S. gross domestic product (GDP) from April through June. Based on what we know to date, second-quarter growth will be a large improvement over the dismal performance seen during the first three months of this year. As of today, our GDPNow model is reading an annualized second-quarter growth rate at 2.7 percent. Given that the economy declined by 2.9 percent in the first quarter, the prospects for the anticipated near-3 percent growth for 2014 as a whole look pretty dim.
The first-quarter performance was dominated, of course, by unusual circumstances that we don't expect to repeat: bad weather, a large inventory adjustment, a decline in real exports, and (especially) an unexpected decline in health services expenditures. Though those factors may mean a disappointing growth performance for the year as a whole, we will likely be willing to write the first quarter off as just one of those things if we can maintain the hoped-for 3 percent pace for the balance of the year.
Do the data support a case for optimism? We have been tracking the six-month trends in four key series that we believe to be especially important for assessing the underlying momentum in the economy: consumer spending (real personal consumption expenditures, or real PCE) excluding medical services, payroll employment, manufacturing production, and real nondefense capital goods shipments excluding aircraft.
The following charts give some sense of how things are stacking up. We will save the details for those who are interested, but the idea is to place the recent performance of each series, given its average growth rate and variability since 1990, in the context of GDP growth and its variability over that same period.
What do we learn from the foregoing charts? Three out of four of these series appear to be consistent with an underlying growth rate in the range of 3 percent. Payroll employment growth, in fact, is beginning to send signals of an even stronger pace.
Unfortunately, the series that looks the weakest relates to consumer spending. If we put any stock in some pretty basic economic theory, spending by households is likely the most forward-looking of the four measures charted above. That, to us, means a cautious attitude is the still the appropriate one. Or, to quote from a higher Atlanta Fed power:
... it will likely be hard to confirm a shift to a persistent above-trend pace of GDP growth even if the second-quarter numbers look relatively good.
This experience suggests to me that we can misread the vital signs of the economy in real time. Notwithstanding the mostly positive and encouraging character of recent data, we policymakers need to be circumspect when tempted to drop the gavel and declare the case closed. In the current situation, I feel it's advisable to accrue evidence and gain perspective. It will take some time to validate an outlook that assumes above-trend growth and associated solid gains in employment and price stability.
By Dave Altig, executive vice president and research director, and
Pat Higgins, a senior economist, both in the Atlanta Fed's research department
TrackBack URL for this entry:
Listed below are links to blogs that reference GDP Growth: Will We Find a Higher Gear?:
July 10, 2014
Introducing the Atlanta Fed's GDPNow Forecasting Model
The June 18 statement from the Federal Open Market Committee opened with this (emphasis mine):
Information received since the Federal Open Market Committee met in April indicates that growth in economic activity has rebounded in recent months.... Household spending appears to be rising moderately and business fixed investment resumed its advance, while the recovery in the housing sector remained slow. Fiscal policy is restraining economic growth, although the extent of restraint is diminishing.
I highlighted the business fixed investment (BFI) part of that passage because it contracted at an annual rate of 1.2 percent in the first quarter of 2014. Any substantial turnaround in growth in gross domestic product (GDP) from its dismal first-quarter pace would seem to require that BFI did in fact resume its advance through the second quarter.
We won't get an official read on BFI—or on real GDP growth and all of its other components—until July 30, when the U.S. Bureau of Economic Analysis (BEA) releases its advance (or first) GDP estimates for the second quarter of 2014. But that doesn't mean we are completely in the dark on what is happening in real time. We have enough data in hand to make an informed statistical guess on what that July 30 number might tell us.
The BEA's data-construction machinery for estimating GDP is laid out in considerable detail in its NIPA Handbook. Roughly 70 percent of the advance GDP release is based on source data from government agencies and other data providers that are available prior to the BEA official release. This information provides the basis for what have become known as "nowcasts" of GDP and its major subcomponents—essentially, real-time forecasts of the official numbers the BEA is likely to deliver.
Many nowcast variants are available to the public: the Wall Street Journal Economic Forecasting Survey, the Philadelphia Fed Survey of Professional Forecasters, and the CNBC Rapid Update, for example. In addition, a variety of proprietary nowcasts are available to subscribers, including Aspen Publishers' Blue Chip Publications, Macroeconomic Advisers GDP Tracking, and Moody's Analytics high-frequency model.
With this macroblog post, we introduce the Federal Reserve Bank of Atlanta's own nowcasting model, which we call GDPNow.
GDPNow will provide nowcasts of GDP and its subcomponents on a regularly updated basis. These nowcasts will be available on the pages of the Atlanta Fed's Center for Quantitative Economic Research (CQER).
A few important notes about GDPNow:
- The GDPNow model forecasts are nonjudgmental, meaning that the forecasts are taken directly from the underlying statistical model. (These are not official forecasts of either the Atlanta Fed or its president, Dennis Lockhart.)
- Because nowcasts are often based on both modeling and judgment, there is no reason to expect that GDPNow will agree with alternative forecasts. And we do not intend to present GDPNow as superior to those alternatives. Different approaches have their pluses and minuses. An advantage of our approach is that, because it is nonjudgmental, our methodology is easily replicable. But it is always wise to avoid reliance on a single model or source of information.
- GDPNow forecasts are subject to error, sometimes substantial. Internally, we've regularly produced nowcasts from the GDPNow model since introducing an earlier version of it in an October 2011 macroblog post. A real-time track record for the model nowcasts just before the BEA's advance GDP release is available on the CQER GDPNow webpage, and will be updated on a regular basis to help users make informed decisions about the use of this tool.
So, with that in hand, does it appear that BFI in fact "resumed its advance" last quarter? The table below shows the current GDPNow forecasts:
We will update the nowcast five to six times each month following the releases of certain key economic indicators listed in the frequently asked questions. Look for the next GDPNow update on July 15, with the release of the retail trade and business inventory reports.
If you want to dig deeper, the GDPNow page includes downloadable charts and tables as well as numerical details including the model's nowcasts for GDP, its subcomponents, and how the subcomponent nowcasts are built up from both the underlying source data and the model parameters. This working paper supplies the model's technical documentation. We hope economy watchers find GDPNow to be a useful addition to their information sets.
By Pat Higgins, a senior economist in the Atlanta Fed's research department
TrackBack URL for this entry:
Listed below are links to blogs that reference Introducing the Atlanta Fed's GDPNow Forecasting Model:
April 28, 2014
New Data Sources: A Conversation with Google's Hal Varian
New Data Sources: A Conversation with Google's Hal Varian
In recent years, there has been an explosion of new data coming from places like Google, Facebook, and Twitter. Economists and central bankers have begun to realize that these data may provide valuable insights into the economy that inform and improve the decisions made by policy makers.
As chief economist at Google and emeritus professor at UC Berkeley, Hal Varian is uniquely qualified to discuss the issues surrounding these new data sources. Last week he was kind enough to take some time out of his schedule to answer a few questions about these data, the benefits of using them, and their limitations.
Mark Curtis: You've argued that new data sources from Google can improve our ability to "nowcast." Can you describe what this means and how the exorbitant amount of data that Google collects can be used to better understand the present?
Hal Varian: The simplest definition of "nowcasting" is "contemporaneous forecasting," though I do agree with David Hendry that this definition is probably too simple. Over the past decade or so, firms have spent billions of dollars to set up real-time data warehouses that track business metrics on a daily level. These metrics could include retail sales (like Wal-Mart and Target), package delivery (UPS and FedEx), credit card expenditure (MasterCard's SpendingPulse), employment (Intuit's small business employment index), and many other economically relevant measures. We have worked primarily with Google data, because it's what we have available, but there are lots of other sources.
Curtis: The ability to "nowcast" is also crucially important to the Fed. In his December press conference, former Fed Chairman Ben Bernanke stated that the Fed may have been slow to acknowledge the crisis in part due to deficient real-time information. Do you believe that new data sources such as Google search data might be able to improve the Fed's understanding of where the economy is and where it is going?
Varian: Yes, I think that this is definitely a possibility. The real-time data sources mentioned above are a good starting point. Google data seems to be helpful in getting real-time estimates of initial claims for unemployment benefits, housing sales, and loan modification, among other things.
Curtis: Janet Yellen stated in her first press conference as Fed Chair that the Fed should use other labor market indicators beyond the unemployment rate when measuring the health of labor markets. (The Atlanta Fed publishes a labor market spider chart incorporating a variety of indicators.) Are there particular indicators that Google produces that could be useful in this regard?
Varian: Absolutely. Queries related to job search seem to be indicative of labor market activity. Interestingly, queries having to do with killing time also seem to be correlated with unemployment measures!
Curtis: What are the downsides or potential pitfalls of using these types of new data sources?
Varian: First, the real measures—like credit card spending—are probably more indicative of actual outcomes than search data. Search is about intention, and spending is about transactions. Second, there can be feedback from news media and the like that may distort the intention measures. A headline story about a jump in unemployment can stimulate a lot of "unemployment rate" searches, so you have to be careful about how you interpret the data. Third, we've only had one recession since Google has been available, and it was pretty clearly a financially driven recession. But there are other kinds of recessions having to do with supply shocks, like energy prices, or monetary policy, as in the early 1980s. So we need to be careful about generalizing too broadly from this one example.
Curtis: Given the predominance of new data coming from Google, Twitter, and Facebook, do you think that this will limit, or even make obsolete, the role of traditional government statistical agencies such as Census Bureau and the Bureau of Labor Statistics in the future? If not, do you believe there is the potential for collaboration between these agencies and companies such as Google?
Varian: The government statistical agencies are the gold standard for data collection. It is likely that real-time data can be helpful in providing leading indicators for the standard metrics, and supplementing them in various ways, but I think it is highly unlikely that they will replace them. I hope that the private and public sector can work together in fruitful ways to exploit new sources of real-time data in ways that are mutually beneficial.
Curtis: A few years ago, former Fed Chairman Bernanke challenged researchers when he said, "Do we need new measures of expectations or new surveys? Information on the price expectations of businesses—who are, after all, the price setters in the first instance—as well as information on nominal wage expectations is particularly scarce." Do data from Google have the potential to fill this need?
Varian: We have a new product called Google Consumer Surveys that can be used to survey a broad audience of consumers. We don't have ways to go after specific audiences such as business managers or workers looking for jobs. But I wouldn't rule that out in the future.
Curtis: MIT recently introduced a big-data measure of inflation called the Billion Prices Project. Can you see a big future in big data as a measure of inflation?
Varian: Yes, I think so. I know there are also projects looking at supermarket scanner data and the like. One difficulty with online data is that it leaves out gasoline, electricity, housing, large consumer durables, and other categories of consumption. On the other hand, it is quite good for discretionary consumer spending. So I think that online price surveys will enable inexpensive ways to gather certain sorts of price data, but it certainly won't replace existing methods.
By Mark Curtis, a visiting scholar in the Atlanta Fed's research department
TrackBack URL for this entry:
Listed below are links to blogs that reference New Data Sources: A Conversation with Google's Hal Varian:
- The Fed’s Inflation Goal: What Does the Public Know?
- Going to School on Labor Force Participation
- Bad Debt Is Bad for Your Health
- Working for Yourself, Some of the Time
- Gauging Firm Optimism in a Time of Transition
- Can Tight Labor Markets Inhibit Investment Growth?
- More Ways to Watch Wages
- Unemployment versus Underemployment: Assessing Labor Market Slack
- Does a High-Pressure Labor Market Bring Long-Term Benefits?
- Net Exports Continue to Bedevil GDPNow
- April 2017
- March 2017
- February 2017
- January 2017
- December 2016
- November 2016
- October 2016
- September 2016
- August 2016
- July 2016
- Business Cycles
- Business Inflation Expectations
- Capital and Investment
- Capital Markets
- Data Releases
- Economic conditions
- Economic Growth and Development
- Exchange Rates and the Dollar
- Fed Funds Futures
- Federal Debt and Deficits
- Federal Reserve and Monetary Policy
- Financial System
- Fiscal Policy
- Health Care
- Inflation Expectations
- Interest Rates
- Labor Markets
- Latin America/South America
- Monetary Policy
- Money Markets
- Real Estate
- Saving, Capital, and Investment
- Small Business
- Social Security
- This, That, and the Other
- Trade Deficit
- Wage Growth