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
November 14, 2016
Is There a Gender Wage Growth Gap?
The existence of the "gender wage gap" is well documented. Although the gap in the average level of pay between men and women has narrowed over time, studies conducted in the past few years find that women still tend to make about 20 percent less than men. Researchers estimate that between one half and three quarters of the gap can be accounted for by observable differences between men and women in the workforce such as labor market experience, educational attainment, as well as job characteristics (see here , here, and here). This estimation leaves one quarter to one half of the gap that is the result of other factors. While some pin the remainder on discrimination or unfair hiring practices, others suggest the remaining gap may reflect subtle differences in work preferences, such as women choosing jobs with family-oriented benefit packages or flexible work arrangements.
A related question is whether there are differences between the average wage growth of men and women. Since 2010 the Atlanta Fed's Wage Growth Tracker has revealed a disparity between the pay raises of continuously employed men and women, as depicted in the following chart.
Between 1997 and 2010, wage growth of men and women was about equal. Since 2010 however, a gap has emerged. On average, men have been experiencing about 0.35 percentage points higher median wage growth than women. Can differences in characteristics such as experience and job choice explain this gap?
To answer this question, I aggregated individuals into groups based on their potential labor market experience (0–5 years, 5–9 years, 10–24 years, and 25–48 years) education (degree or no degree) family type (married, whether your spouse works, and whether you have kids); industry (goods versus services) occupation (low, middle, or high skill); sector (public versus private); and if the person switched jobs recently. I then computed the median wage growth for each unique group in each year. Using a statistical technique called a Oaxaca Decomposition, I separated out the difference between men and women's wage growth that can be pinned on differences in the way men and women are distributed among these groups (the "endowment" effect).
The following chart shows median wage growth after removing this endowment effect.
After removing the difference in wage growth that is the result of differences in gender-specific characteristics, wage growth of men and women is much more similar. In particular, these differences appear to almost entirely account for the gap that had emerged after 2009. What explains the gap in wage levels between men and women is still an open question, but this analysis suggests that much of the difference in wage growth through the years has to do with family/job choices and other individual characteristics.
November 07, 2016
The Price Isn't Right: On GDPNow's Third Quarter Miss
The U.S. Bureau of Economic Analysis's (BEA) first estimate of third quarter annualized real gross domestic product (GDP) growth released on October 28 was 2.9 percent. A number of nowcasts were quite close to this number, including the median forecast of 3.0 percent from the CNBC Rapid Update survey of roughly 10 economists. The Atlanta Fed's GDPNow model forecast of 2.1 percent? Not so close.
What accounted for GDPNow's miss? The table below shows the GDPNow forecasts and BEA estimates of the percentage point contributions to third quarter growth of six subcomponents that together make up real GDP. The largest forecast error, both in absolute terms and relative to the historical accuracy of the projections, was for the contribution of real net exports to growth. The published contribution of 0.83 percentage points was much higher than the model's estimate of 0.07 percentage points.
Why did GDPNow miss so badly on net exports? For both goods and services real net exports, the GDPNow forecast is a weighted average of two forecasts. The "bean counting" forecast uses the monthly source data on the nominal values and price deflators of exports and imports. The econometric model forecast uses published values of 13 subcomponents of real GDP for the last five quarters to predict real net exports for both goods and services. The statistically determined weights on the bean counting forecast increase as we get closer to the first GDP release and accumulate more monthly source data. (More details are provided here.)
For real net exports of services, 89 percent of the weight was given to the bean counting forecast. This weighting worked out well last quarter as the forecasts of the contribution of real services net exports to third quarter growth from both the bean counting and combined models were within 0.01 percentage points of the published value of −0.14 percentage points. But for real net exports of goods, the bean counting forecast received only 59 percent of the weight in the final GDPNow forecast. It projected that real net exports of goods would add 0.76 percentage points to growth—reasonably close to the BEA estimate. In contrast, the econometric model projected a subtraction of 0.58 percentage points from growth.
Since GDPNow had monthly price and nominal spending data through September on goods imports and exports, why didn't it place more weight on the source data? One of the important reasons is that it's difficult to match the quarterly inflation rate of the BEA's import price deflator for goods. The BEA constructs its price deflator with detailed price indices from the Bureau of Labor Statistics (BLS) producer price index and import/export price index programs as well as a few other sources. GDPNow uses the BLS's import price data at higher levels of aggregation than the BEA uses and also differs in the manner that it handles seasonality. The chart below plots the difference between the BEA's quarterly goods import price inflation rate and the GDPNow proxy. These inflation measures have differed by 5 percentage points or more on a number of occasions. Since goods imports are 12 percent of GDP, a miss of this magnitude on the price deflator would lead to a miss on the real net exports of goods contribution to growth of 0.5 percentage points or more, even if the other ingredients in the calculation were all correct.
Are there any lessons here for improving GDPNow? Ideally, GDPNow would be able to closely map the monthly source data to real goods net exports so that most of the weight would go to the bean counting forecast once all of the data are in—much as it does with nonresidential structures and residential investment. The BEA's estimates of real petroleum imports are based on similar data in the monthly international trade data publication. Because petroleum imports account for so much of the volatility of inflation for goods imports, it may be better to use the monthly real petroleum imports data directly and only worry about replicating the price index for nonpetroleum goods.
That said, a previous macroblog post illustrated that the method GDPNow currently uses has a reasonable forecasting track record for net exports when compared with several consensus estimates from professional forecasters. Net exports may remain difficult to nowcast even with refinements to GDPNow's methodology.
GDPNow has established a commendable track record. But sometimes when it misses the mark, an analysis of the error can provide insight into how GDPNow works and the limitations of the model.
October 24, 2016
Is Wage Growth Accelerating?
The Atlanta Fed's Wage Growth Tracker came in at 3.6 percent in September, up from 3.3 percent in August and 3.4 percent in July, but the same as the 3.6 percent reading for June. By this measure, there are no obvious signs of an acceleration in wage growth for continuously employed workers during the last few months.
However, the headline wage growth tracker is a three month moving average of each month's median wage growth. Interestingly, for September, the median wage growth (using data that are not averaged, sometimes called "unsmoothed") was 4.2 percent, up from 3.6 percent in August, and the highest since late 2007. This pop in median wage growth can be seen in the following chart, which compares the median wage growth (smoothed using a three-month average) with the unsmoothed monthly median.
Even though this looks like a pretty large increase, the standard error on the difference between the unsmoothed August and September medians is also quite large at 0.5 percentage points. So the 0.6 percentage point difference in medians is not statistically significant—it could just as easily be sampling noise. But it is definitely something to keep an eye on going forward. As noted in a previous macroblog post, the correlation between the unemployment gap and the Wage Growth Tracker suggests that we should be seeing the wage growth tracker level off if the economy is stabilizing at full employment.
October 18, 2016
Unemployment Risk and Unions
A recent paper by the Economic Policy Institute (EPI) argues that increased unionization would have broad economic benefits and, in particular, could help improve the wage stagnation facing many lower-skilled workers. Yet union membership has been declining, down by about 3 million between 1983 and 2015, and membership is down 4.5 million in the private sector. (Union membership in the United States is discussed in this U.S. Bureau of Labor Statistics report and in this database, maintained by Barry Hirsch at Georgia State University.)
The overall membership decline in private-sector unions reflects a combination of lower employment in some traditionally unionized industries such as the steel and auto industries and lower unionization rates within industries. For example, the rate of unionization for goods-producing industries (largely manufacturing and construction) is down from 28 percent to 10 percent, and the rate in service-producing industries has declined from 11 percent to 6 percent. In contrast, union membership in the public sector has increased, mostly as a result of broad unionization among public safety, utility, and education occupations coupled with the fact that employment in these occupations has tended to grow over time.
For goods-producing industries in particular, unionized employment is down by about 4.2 million since 1983, and nonunionized employment is up by around 2.5 million. Many factors may have contributed to this shift away from union membership. A possibility I explore here is the role of wage rigidity. In particular, if union wage contracts prevent employers from adjusting wages in the face of an unexpected decline in output demand, then employers may adjust along the employment margin instead. The monopoly power of unions leads to higher wages for continuously employed union workers but also makes layoffs more frequent.
It is the case that unionized workers tend to earn more than their nonunion counterparts. For 1983 to 2015, I estimate that prime-age union workers in goods-producing industries earn an average of about 25 percent more (on a median hourly basis) than comparable nonunion workers (about 50 percent more in construction and about 10 percent more in manufacturing). In addition, the median wage growth of union workers is less cyclically sensitive. The following chart uses the Atlanta Fed's Wage Growth Tracker data, and it shows the annual median wage growth of continuously employed prime-age workers in goods-producing industries, by union status.
Not only is wage growth among union workers less variable over time as the chart shows, research has noted that union wages are less dispersed—even controlling for differences in worker characteristics. Joining a union leads to wages that tend to be higher, wages that vary less across workers, and wage growth that responds less to changes in economic conditions.
But what about unemployment risk? Do union workers get laid off at a greater rate than nonunion workers? Using matched data from the Current Population Survey, the following chart shows an estimate of the probability that a prime-age worker in a goods producing industry is unemployed 12 months later, by union status.
The probability of unemployment rises during economic downturns for both union and nonunion workers, but is higher for union workers. The union worker displacement rate reached 13 percent in 2009 versus 8 percent for nonunion workers.
However, recall provisions are often built into collective bargaining agreements, so perhaps looking at the total unemployment flow overstates the permanent job loss risk. To investigate, the following chart shows the likelihood of being on temporary layoff (expected to be recalled within six months) versus indefinite (permanent) layoff.
The likelihood of being recalled by your previous employer is much higher for union than nonunion workers, whereas the incidence of permanent layoff is about the same for both types of worker.
Admittedly, I'm not controlling for all the things about workers and employers that could influence employment and wage outcomes. But taken at face value, it appears that the likelihood of permanent job loss is no greater for union workers in goods-producing industries than for nonunion workers. At the same time, union workers are more likely to experience a spell of temporary unemployment. I view this as some evidence in support of my wage rigidity story, which holds that unionized firms use layoffs more intensively because wages are less flexible (I find that this same result holds if I look at the manufacturing and construction industries separately). However, this mechanism itself isn't able to account for much of the secular decline in union participation. The decline seems to be more about where the jobs are created than where they are lost.
October 14, 2016
Cumulative U.S. Trade Deficits Resulting in Net Profits for the U.S. (and Net Losses for China)
The United States has run trade deficits for decades (1976 is the last year with a recorded surplus). To illustrate this, chart 1 depicts the cumulative U.S. trade deficit since 1980, which now surpasses $10 trillion. As a result, a drastic deterioration in the U.S. net foreign asset position—the difference between the amount of foreign assets owned by U.S. residents and the amount of U.S. assets owned by foreigners—has occurred. That is, as Americans borrow from the rest of the world to finance the recurring trade deficits, the national net worth goes deeply into the red. Not long ago, many commentators predicted that as a result of this increasing U.S. foreign debt, the U.S. dollar was set to collapse, which would trigger a stampede away from U.S. assets. Of course, this has not happened.
Much of the rising U.S. deficit is the by-product of deficits with one country in particular: China. Chart 2 shows that U.S. bilateral trade deficits with China have been growing steadily during these years. In 2015, the total U.S. goods trade deficit was about $762 billion, and the goods deficit with China alone made up nearly half of that total ($367 billion). This situation is not unique to the United States, as many countries find themselves in similar trade positions with China. During the last few decades, China has been running protracted trade surpluses with the rest of world and has accumulated a positive and sizeable net foreign asset position.
Yet, despite accumulating a positive and sizeable net foreign asset position, China is facing increasing losses in net income on its foreign assets. Put differently, China has been accumulating negative returns on its increasingly large portfolio of foreign assets. Chart 3 shows this observation, made in a paper by Eswar Prasad of Cornell University at a recent conference cosponsored by the Atlanta Fed and the International Monetary Fund.
The net income on foreign assets measures the return a nation receives from the foreign assets it owns minus the return paid on domestic assets held by foreigners. In sharp contrast to China, however, the U.S. international net financial income has remained positive and has even increased. This increase comes despite the fact that the United States has consistently run trade deficits, and its net foreign asset position has deteriorated. Chart 4 shows the U.S. income from foreign assets and foreigners' income on U.S. assets, which is reported as a negative number for this series because it is regarded as a liability for the United States. The difference between these amounts is depicted by the middle line, which shows the net foreign income of the United States.
How is this possible? Ricardo Hausmann (Harvard University) and Federico Sturzenegger (currently, the chairman of Central Bank of Argentina) came up with an explanation more than ten years ago: the United States gets a far higher return on its foreign assets than the other way around. Indeed, U.S. foreign direct investments (FDI) often generate a relatively high rate of return. In part, U.S. FDI are benefiting from business expertise, brand recognition, and research and development in new product and service lines. Comparatively, foreigners tend to earn substantially less return on the American assets they own. Foreigners often desire to hold their dollar assets in the form of safe, liquid assets, which—following the "low-risk, low-return" principle—have relatively low returns.
To see this, chart 5 shows the sources of the net financial income of the United States. The U.S. government net income is negative—mostly the by-product of interest payments in government debt held by foreigners. The U.S. gets most of its financial return from FDI. Although much has happened in the world economy during the last decade, the implications of Hausmann and Sturzenegger's analysis remain intact. In sum, the differential return from these foreign assets and liabilities appear to largely compensate for the trade deficits.
Eswar Prasad also showed that China is in a starkly different situation. Most of its foreign liabilities are in the form of FDI, while the vast majority of the foreign assets are reserve assets and foreign exchange reserves—not surprisingly, largely U.S. dollars and U.S. Treasury securities. The rate of return foreigners make on Chinese assets is around twice the rate of return China gets on its foreign assets.
This analysis suggests that focusing on a country's net foreign asset position conveys an incomplete picture of the profitability of foreign assets because it fails to account for the differences in rates of returns that countries earn on their foreign assets. Overall, the United States makes a sufficiently high return on foreign assets that it maintains positive net income on foreign assets. The situation is similar to role leverage in investing; debt can be profitable if you can devote it to purposes that earn a higher rate of return than your cost of borrowing it. Therefore, when viewed in terms of the net income earned on foreign assets the United States holds, the sizable U.S. trade deficits may not be as much of a concern as commonly thought.
October 05, 2016
The Slump in Undocumented Immigration to the United States
Immigration is a challenging and often controversial topic. We have written some on the economic benefits and costs associated with the inflows of low-skilled (possibly undocumented) immigrant workers into the United States here and here. In this macroblog post, we discuss some interesting trends in undocumented immigration.
There are no official records of undocumented immigration flows into the United States. However, one crude proxy for this flow is the number of apprehensions at the U.S. border. As pointed out in Hanson (2006), the number of individuals arrested when attempting to cross the U.S.-Mexico border, provided by the Department of Homeland Security (DHS), is likely to be positively correlated with the flows of attempted illegal border crossings (see chart 1).
The apprehensions series displays spikes that coincide with well-known episodes of increased illegal immigration into the United States, such as after the financial crisis in Mexico in 1995 or during the U.S. housing boom in the early 2000s. Importantly, the series also shows a sharp decline in the flows of illegal immigration at the U.S.-Mexico border during the last recession, and those flows have remained at historically low levels since then.
A better proxy for illegal immigration from Mexico would adjust the number of apprehensions for the intensity of U.S. border enforcement (for example, the number of border patrol officers). The intuition is straightforward: for the same level of attempted illegal crossings, greater enforcement is likely to result in more apprehensions. Chart 2 shows the border patrol staffing levels as an indicator of enforcement intensity.
As the chart shows, the sharp decrease in apprehensions after the Great Recession occurred despite a remarkable increase in border enforcement, indicating that the decline in migration flows in recent years may have been even more abrupt than implied by the (unadjusted) border apprehensions shown in chart 1.
The measure of inflows shown in chart 1 is largely consistent with estimates of the stock of undocumented immigrants in the United States, such as those provided by a new study by the Pew Research Center based on data from the U.S. Census Bureau. After having peaked at 12.2 million in 2007, the stock of unauthorized immigrants fell during the Great Recession and remained stable afterwards, most recently at 11.1 million in 2014. Also, the composition of this stock has shifted since the Great Recession. Although the population of undocumented Mexican immigrants fell by more than one million from its 6.9 million peak in 2007, the number of undocumented immigrants from Asia, Central America, and sub-Saharan Africa remained relatively steady as of 2014 and even increased in some cases. For example, the population of unauthorized immigrants from India rose by about 130,000 between 2009 and 2014. However, a lot of this type of unauthorized immigration is a result of overstayed visas rather than from people crossing the border without a visa.
What do these numbers suggest about the future? It is likely that the flows of undocumented immigrant labor between Mexico and the United States reflect differences in demographic patterns and economic opportunities between the two economies. In the United States, the baby boom came to an abrupt halt in the 1960s, causing a notable slowdown in the native-born labor supply two decades later. In contrast, Mexico's fertility rate remained high for much longer, hovering at 6.7 births per woman in 1970 versus 2.5 in the United States (see chart 3).
Mexico's labor force expanded rapidly during the 1980s, which, juxtaposed with the Mexican economic slump of the early 1980s, unleashed a wave of Mexican migration to the United States (Hanson and McIntosh, 2010). Also encouraging this flow was the steady U.S. economic growth during the "Great Moderation" period from the mid-1980s up through 2007 (Bernanke, 2004). More recently, however, Mexico's fertility rate has fallen (as in some Central American economies), and economic growth there has mostly outpaced that of the United States. Therefore, it is perhaps not too surprising that demographic trends—along with greater enforcement—have caused the inflows of undocumented migration at the U.S.-Mexico border to slow in recent years. Shifts in demographic and economic factors across countries are likely to continue to influence undocumented immigration in the United States.
Note: The views expressed here are those of the authors and do not necessarily reflect the views of the Federal Reserve Banks of Atlanta or Boston.
September 30, 2016
A Quick Pay Check: Wage Growth of Full-Time and Part-Time Workers
In the last macroblog post we introduced the new version of the nominal Wage Growth Tracker, which allows a look back as far as 1983. We have also produced various cuts of these data comparable to the ones on the Wage Growth Tracker web page to look at the wage dynamics of various types of workers. One of the data cuts compares the median wage growth of people working full-time and part-time jobs. As we have highlighted previously, the median wage growth of part-time workers slowed by significantly more than full-time workers in the wake of the Great Recession. The extended time series allows us to look back farther to see if this phenomenon was truly unique.
The following chart shows the extended full-time/part-time median wage growth time series at an annual frequency.
The chart shows that the median wage increase for part-time workers is generally lower than for full-time workers, with the average gap about 1 percentage point. The reason for the presence of a gap is a bit puzzling. Could it be that part-time workers have lower average productivity growth than full-time workers? It is true that a part-time worker in our data set is more likely to lack a college degree than a full-time worker, and the median wage level for part-time workers is lower than for full-time workers. But interestingly, a reasonably systematic wage growth gap still exists after controlling for differences in the education and age of workers. So even highly educated prime-age, part-time workers tend to have lower median wage growth than their full-time counterparts. If it's a productivity story, its subtext is not easily captured by observed differences in education and experience.
Changes in economic conditions might also be playing a role. The wage growth gap exceeded 2 percentage points in the early 1980s and again between 2011 and 2013, both periods of considerable excess slack in the labor market, as we recently discussed here. In fact, in each of 2011, 2012, and 2013, half of the part-time workers in our dataset experienced no increase in their rate of pay at all.
To explore this possibility further, it's useful to separate part-time workers into those who work part-time because of economic conditions (for example, because of slack work conditions at their employer or their inability to find full-time work) from those who work part-time for noneconomic reasons (for example, because they have family responsibilities or because they are also in school). The following chart shows the median wage growth for full-time, voluntary part-time, and involuntary part-time workers.
Admittedly, there are not that many observations on involuntary part-time workers in our data set. But it does appear that their median wage growth has tended to slow by more after economic downturns than those working part-time for a noneconomic reason—at least prior to the Great Recession. After the last recession, however, the wage growth gap was just about as large for both types of part-time workers. In that sense, the impact of the last recession on the median wage growth of regular part-time workers was quite unusual.
Since 2013, median wage growth for part-time workers has been rising, which is good news for those workers and consistent with the labor market becoming tighter. With the unemployment rate reasonably low, employers might have to worry a bit more about retaining and attracting part-time staff than they did a few years ago.
September 27, 2016
Back to the '80s, Courtesy of the Wage Growth Tracker
Things have been a wee bit quiet in macroblog land the last few weeks, chiefly because our time has been devoted to two exciting new projects. The first is a refresh of our labor force dynamics website, which will feature a nifty tool for looking at the main reasons behind changes in labor force participation for different age groups. More on that later.
The other project has been adding more history to our Wage Growth Tracker. The tracker's current time series starts in 1997. The chart below shows an extended version of the tracker that starts in 1983.
Recall that the Wage Growth Tracker depicts the median of the distribution of 12-month changes of matched nominal hourly earnings. In the extended time series, you'll notice two gaps, which resulted from the U.S. Census Bureau scrambling the identifiers in its Current Population Survey. For those two periods, you'll have to use your imagination and make some inferences.
As we have emphasized previously, the Wage Growth Tracker is not a direct measure of the typical change in overall wage costs because it only looks at (more or less) continuously employed workers. But it should reflect the amount of excess slack in the labor market. This point is illustrated in the following chart, which compares the Wage Growth Tracker with the unemployment gap computed from the Congressional Budget Office's (CBO) estimate of the long-run natural rate of unemployment.
As the chart shows, our measure of nominal wage growth has historically tracked the cyclical movement in the unemployment rate gap estimate fairly well, at least since the mid-1980s. We think this feature is potentially important, because the true unemployment rate gap is very hard to know in real time and hence is subject to potentially large revision. For example, in real time, the unemployment rate was estimated to have fallen below the natural rate in the fourth quarter of 1994, but it is now thought to have not breached the natural rate until the first quarter of 1997—more than two years later. The Wage Growth Tracker is not subject to revision (although it is subject to a small amount of sampling uncertainty) and hence could be useful in evaluating the reliability of the unemployment rate gap estimate in real time.
This also is important from a monetary policy perspective if we are worried about the risk of the economy overheating. For example, President Rosengren of the Boston Fed described why he dissented at the most recent Federal Open Market Committee meeting in favor of a quarter-point increase in the target range for the federal funds rate. His dissent, he said, arose partly from his concern that the economy may overheat and drive unemployment below a level he believes is sustainable.
Currently, the CBO estimate of the unemployment rate gap looks like it is plateauing at close to zero. The fact that the Wage Growth Tracker for the third quarter slowed a bit is consistent with that. But it's only one quarter of data, and so we'll closely monitor the Wage Growth Tracker in the coming months to see what it suggests about the actual unemployment rate gap. We'll discuss what observations we make here.
September 08, 2016
Introducing the Atlanta Fed's Taylor Rule Utility
Simplicity isn't always a virtue, but when it comes to complex decision-making processes—for example, a central bank setting a policy rate—having simple benchmarks is often helpful. As students and observers of monetary policy well know, the common currency in the central banking world is the so-called "Taylor rule."
The Taylor rule is an equation introduced by John Taylor in a seminal 1993 paper that prescribes a value for the federal funds rate—the interest rate targeted by the Federal Open Market Committee (FOMC)—based on readings of inflation and the output gap. The output gap measures the percentage point difference between real gross domestic product (GDP) and an estimate of its trend or potential.
Since 1993, academics and policymakers have introduced and used many alternative versions of the rule. The alternative forms of the rule can supply policy prescriptions that differ significantly from Taylor's original rule, as the following chart illustrates.
The green line shows the policy prescription from a rule identical to the one in Taylor's paper, apart from some minor changes in the inflation and output gap measures. The red line uses an alternative and commonly used rule that gives the output gap twice the weight used for the Taylor (1993) rule, derived from a 1999 paper by John Taylor. The red line also replaces the 2 percent value used in Taylor's 1993 paper with an estimate of the natural real interest rate, called r*, from a paper by Thomas Laubach, the Federal Reserve Board's director of monetary affairs, and John Williams, president of the San Francisco Fed. Federal Reserve Chair Janet Yellen also considered this alternative estimate of r* in a 2015 speech.
Both rules use real-time data. The Taylor (1993) rule prescribed liftoff for the federal funds rate materially above the FOMC's 0 to 0.25 percent target range from December 2008 to December 2015 as early as 2012. The alternative rule did not prescribe a positive fed funds rate since the end of the 2007–09 recession until this quarter. The third-quarter prescriptions incorporate nowcasts constructed as described here. Neither the nowcasts nor the Taylor rule prescriptions themselves necessarily reflect the outlook or views of the Federal Reserve Bank of Atlanta or its president.
Additional variables that get plugged into this simple policy rule can influence the rate prescription. To help you sort through the most common variations, we at the Atlanta Fed have created a Taylor Rule Utility. Our Taylor Rule Utility gives you a number of choices for the inflation measure, inflation target, the natural real interest rate, and the resource gap. Besides the Congressional Budget Office–based output gap, alternative resource gap choices include those based on a U-6 labor underutilization gap and the ZPOP ratio. The latter ratio, which Atlanta Fed President Dennis Lockhart mentioned in a November 2015 speech while addressing the Taylor rule, gauges underemployment by measuring the share of the civilian population working their desired number of hours.
Many of the indicator choices use real-time data. The utility also allows you to establish your own weight for the resource gap and whether you want the prescription to put any weight on the previous quarter's federal funds rate. The default choices of the Taylor Rule Utility coincide with the Taylor (1993) rule shown in the above chart. Other organizations have their own versions of the Taylor Rule Utility (one of the nicer ones is available on the Cleveland Fed's Simple Monetary Policy Rules web page). You can find more information about the Cleveland Fed's web page on the Frequently Asked Questions page.
Although the Taylor rule and its alternative versions are only simple benchmarks, they can be useful tools for evaluating the importance of particular indicators. For example, we see that the difference in the prescriptions of the two rules plotted above has narrowed in recent years as slack has diminished. Even if the output gap were completely closed, however, the current prescriptions of the rules would differ by nearly 2 percentage points because of the use of different measures of r*. We hope you find the Taylor Rule Utility a useful tool to provide insight into issues like these. We plan on adding further enhancements to the utility in the near future and welcome any comments or suggestions for improvements.
August 15, 2016
Payroll Employment Growth: Strong Enough?
The U.S. Bureau of Labor Statistics' estimate of nonfarm payroll employment is the most closely watched indicator of overall employment growth in the U.S. economy. By this measure, employment increased by 255,000 in July, well above the three-month average of 190,000. Yet despite this outsized gain, the unemployment rate barely budged. What gives?
Well, for a start, there is no formal connection between the payroll employment data and the unemployment rate data. The employment data used to construct the unemployment rate come from the Current Population Survey (CPS) and the payroll employment data come from a different survey. However, it is possible to relate changes in the unemployment rate to the gap between the CPS and payroll measures of employment, as well as changes in the labor force participation (LFP) rate, and the growth of payroll employment relative to the population.
The following chart shows the contribution of each of these three factors to the monthly change in the unemployment rate during the last year.
A note about the chart: The CPS employment and population measures have been smoothed to account for annual population control adjustments. The smoothed employment data are available here. The method used to compute the contributions is available here.
The black line is the monthly change in the unemployment rate (unrounded). Each green segment of a bar is the change in the unemployment rate coming from the gap between population growth and payroll employment growth. Because payroll employment has generally been growing faster than the population, it has helped make the unemployment rate lower than it otherwise would have been.
But as the chart makes clear, the other two factors can also exert a significant influence on the direction of the unemployment rate. The labor force participation rate contribution (the red segments of the bars) and the contribution from the gap between the CPS and payroll employment measures (blue segments) can vary a lot from month to month, and these factors can swamp the payroll employment growth contribution.
So any assumption that strong payroll employment gains in any particular month will automatically lead to a decline in the unemployment rate could, in fact, be wrong. But over longer periods, the mapping is a bit clearer because it is effectively smoothing the month-to-month variation in the three factors. For example, the following chart shows the contribution of the three factors to 12-month changes in the unemployment rate from July 2012 to July 2013, from July 2013 to July 2014, and so on.
Gains in payroll employment relative to the population have helped pull the unemployment rate lower. Moreover, prior to the most recent 12 months, declines in the LFP rate put further downward pressure on the unemployment rate. Offsetting this pressure to varying degrees has been the fact that the CPS measure of employment has tended to increase more slowly than the payroll measure, making the decline in the unemployment rate smaller than it would have been otherwise. During the last 12 months, the LFP rate turned positive on balance, meaning that the magnitude of the unemployment rate decline has been considerably less than implied by the relative strength of payroll employment growth.
Going forward, another strong payroll employment reading for August is certainly no guarantee of a corresponding decline in the unemployment rate. But as shown by my colleagues David Altig and Patrick Higgins in an earlier macroblog post, under a reasonable range of assumptions for the trend path of population growth, the LFP rate, and the gap between the CPS and payroll survey measures of employment, payroll growth averaging above 150,000 a month should be enough to cause the unemployment rate to continue declining.
- Using Judgment in Forecasting: Does It Matter?
- Does Lower Pay Mean Smaller Raises?
- Outside Looking In: Why Has Labor Force Participation Increased?
- Wages Climb Higher, Faster
- Is There a Gender Wage Growth Gap?
- The Price Isn't Right: On GDPNow's Third Quarter Miss
- Is Wage Growth Accelerating?
- Unemployment Risk and Unions
- Cumulative U.S. Trade Deficits Resulting in Net Profits for the U.S. (and Net Losses for China)
- The Slump in Undocumented Immigration to the United States
- December 2016
- November 2016
- October 2016
- September 2016
- August 2016
- July 2016
- June 2016
- May 2016
- April 2016
- March 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