About


The Atlanta Fed's macroblog provides commentary and analysis on economic topics including monetary policy, macroeconomic developments, inflation, labor economics, and financial issues.

Authors for macroblog are Dave Altig, John Robertson, and other Atlanta Fed economists and researchers.


January 11, 2019


Are Employers Focusing More on Staff Retention?

Many people are quitting their current job. According to data from the Job Openings, Layoffs, and Turnover Survey (JOLTS) from the U.S. Bureau of Labor Statistics, workers are voluntarily leaving their current workplace (for reasons other than retirement or internal company transfers) at rates not seen since the late 1990s.

A high rate of quits is consistent with a tightening labor market. In such a market, the promise of more pay elsewhere lures an increasing share of workers away from their current employer. The potential benefit of changing jobs is also evident in the Atlanta Fed's Wage Growth Tracker, which typically exhibits a positive differential between the median wage growth of people changing jobs versus those who don't.

However, this Wage Growth Tracker pattern has changed in recent months. Since mid-2018, the median wage growth of job stayers has risen sharply and is at the highest level since 2008. As the following chart shows, the gap between the wage growth of job switchers and stayers has narrowed.

Is it possible that this new pattern is just noise and will disappear in coming months? After all, the wage growth for job stayers has spiked before. But taken at face value, it suggests that the cost of losing staff (recruiting cost, cost of training new employees, productivity loss from having a less experienced workforce, etc.) might be forcing some employers to boost current employees' pay in an effort to retain staff. Interestingly, the pattern is most evident among prime-age workers in relatively high-skill professional occupations.

Firms are always focused on staff retention and employ various strategies to reduce turnover in key positions. One of those strategies—larger pay increases—might now be taking a more prominent role.

We will closely monitor the Tracker for further evidence on this emerging trend. However, that might be later rather than sooner because the source data is provided by the U.S. Census Bureau, which has cut most services because of the current lapse in congressional funding. But when information becomes available, we'll be ready to parse it—so stay tuned!

January 11, 2019 in Employment , Labor Markets , Wage Growth | Permalink | Comments ( 0)

December 04, 2018


Defining Job Switchers in the Wage Growth Tracker

Among the questions we receive about the Atlanta Fed's Wage Growth Tracker, one of the most frequent is about the construction of the job switcher and job stayer series. These series are derived from data in the Current Population Survey (CPS) and are intended to show how median wage growth differs for those who change their job from last year versus those who are in the same job. However, the monthly CPS does not actually ask if the person has the same job as a year ago.

So how to proceed? The CPS does contain information about the person's industry and occupation that we aggregate into consistent categories that can be compared to the person's industry/occupation reported a year earlier. If someone is in a different occupation or industry category, then we can reasonably infer that the person has changed jobs. To illustrate, for 2017, 18.8 percent of people in the Wage Growth Tracker data are in a different industry category than they were in 2016, and 28.6 percent are in a different occupation category. Of those who remain in the same industry, 23.9 percent changed occupation group, and of those in the same occupation group, 13.5 percent are in a different industry. However, this information doesn't allow us to identify all job switchers because being in the same industry and occupation group as a year earlier does not preclude having changed jobs.

Fortunately, the CPS also has questions based on who a person said their employer was in the prior month. It asks if that person still works there and if the employee's activities and job duties are the same as last month. If someone answers either of these questions in the negative, then it is likely that person is also in a different job than a year earlier. In 2017, 1.5 percent of people in the Wage Growth Tracker data said they have a different employer than in the prior month, and 0.9 percent report having different job responsibilities at the same employer.

Unfortunately, the dynamic structure of the CPS means that the responses to these "same employer/activity" questions can only potentially be matched with an individual's response in the prior two months, and not a year earlier. Moreover, some responses to those questions are blank, even for people whom we identify as being employed in the prior month. For those individuals, we simply don't know if they are in the same job as a month earlier. Of the non-null responses, the vast majority do not change job duties or employer from one month to the next. So if we assume the blank responses are randomly distributed among the employed population, it's reasonable to also treat the blanks as job stayers.

Previously, we had treated the blank "same employer/activity" observations as job switchers, but that approach almost certainly misclassified some actual job stayers as job switchers. Instead, we now define a job switcher as someone in the Wage Growth Tracker data who is in a different occupation or industry group than a year earlier, or someone who says no to either of the "same employer/activity" questions in the current or prior two months. We label everyone else a job stayer.

Does the definition matter for median wage growth? The following chart shows the annual time series of the difference between the median wage growth of job switchers and job stayers based on both the old and new definitions.

As you can see, the results are not qualitatively different. Job switchers have higher median wage growth during strong labor market conditions and lower growth during bad times. Not surprisingly, the gap in median wage growth is generally lower (more negative) using the old definition.

The next chart shows the annual time series of the share of job switchers in the Wage Growth Tracker data based on the new and old definitions. A caveat: we have been unable to construct occupation and industry groupings for 2003 that are completely consistent with the groupings used in 2002. This results in an erroneous spike in measured job switching for 2003.

Notice that the share of job switchers under the new definition peaked prior to the last two recessions, declined during the recessions, and then recovered. That share is now at a cyclical high. A discrete jump in the number of blank responses recorded for the "same employer/activity" questions in the CPS starting in 2009 masked this cyclicality under the old definition.

In about a week from now, the next update of the Wage Growth Tracker data will implement the new and improved definition of job switchers—I hope you'll check it out, and I'll be writing about it here as well.

December 4, 2018 in Employment , Labor Markets , Wage Growth | Permalink | Comments ( 0)

November 29, 2018


Cryptocurrency and Central Bank E-Money

The Atlanta Fed recently hosted a workshop, "Financial Stability Implications of New Technology," which was cosponsored by the Center for the Economic Analysis of Risk at Georgia State University. This macroblog post discusses the workshop's panel on cryptocurrency and central bank e-money. A companion Notes from the Vault post provides some highlights from the rest of the workshop.

The panel began with Douglas Elliot, a partner at Oliver Wyman, discussing some of the public policy issues associated with cryptoassets. Drawing on a recent paper he cowrote, Elliot observed that there are "at least four substantial market segments" that provide long-term support for cryptoassets:

  • libertarians and techno-anarchists who, for ideological reasons, want a currency without a government;
  • people who deeply distrust their government's economic management;
  • seekers of anonymity, who don't want their names associated with transactions and investments; and
  • technical users who find cryptoassets useful for some blockchain applications.

Besides these groups are the speculators and investors who hope to benefit from price appreciation of these assets.

Given the strong interest of these four groups, Elliot argues that cryptoassets are here to stay, but he also asserts that these assets raise public policy issues that regulation should address. Some issues, such as anti–money laundering, are being addressed, but all would benefit from a coordinated global approach. However, he observes that of the four long-term support groups, only the technical users are likely to favor such regulations.

Another paper, by University of Chicago professor Gina C. Pieters, analyzed the extent to which the cryptocurrency market is global using purchases of cryptocurrency by state-issued currencies. She finds that more than 90 percent of all cryptocurrency transactions occur using one of three currencies: the U.S. dollar, the South Korean won, and the Japanese yen. She further finds that the dominance of these three currencies cannot be explained by economic size, financial openness, or internet access. Pieters also observed that transactions involving bitcoin, the largest cryptocurrency by market value, do not necessarily represent a country's cryptomarket share.

Warren Weber, former Minneapolis Fed economist and a visiting scholar at the Atlanta Fed, discussed so-called "stable coins," one type of cryptocurrency. The value of many cryptocurrencies has fluctuated widely in recent years, with the price of one bitcoin soaring from under $6,000 to more than $19,000 and then plunging to just over $6,000—all within the period from October 2017 to October 2018. This extreme price volatility creates a significant impediment to Elliot's technical users who would like some method of buying blockchain services with a currency controlled by a blockchain. In an attempt to meet this demand, a number of "stable coins" have been issued or are under development.

Drawing on a preliminary paper, Weber discussed three types of stable coins. One type backs all of the currency it issues with holdings of a state-issued currency, such as the U.S. dollar. A potential weakness of these coins is that they incur operational costs that require payment. Weber observed that interest earnings might cover part of these expenses if the stable coin issuer holds the dollars in an interest-bearing asset. Additionally, charging redemption fees might offset some or all of the expense.

The other two alternatives involve the creation of cryptofinancial entities or crypto "central banks." Both of these approaches seek to adjust the quantity of the cryptocurrency outstanding to stabilize its price in another currency. However, Weber observed that both of these approaches are subject to the problem that the cryptocurrency could take on many values depending upon people's expectations. If people come to expect that a coin will lose its value, neither of these approaches can prevent the coin from becoming worthless.

The question of whether existing central banks should issue e-money was the topic of a presentation by Francisco Rivadeneyra of the Bank of Canada. Summarizing the results of his paper, Rivadeneyra observed that central banks could provide e-money that looks like a token or a more traditional account. The potential for central banks to offer widely available account-based services has long existed. However, after considering the tradeoffs, central banks have elected not to provide these accounts, and recent technological developments have not changed this calculus. However, new technologies may have changed the tradeoff for token-based systems. Many issues will need to be addressed first, though.

November 29, 2018 in Capital and Investment , Technology | Permalink | Comments ( 1)

November 16, 2018


Polarization through the Prism of the Wage Growth Tracker (Take Two)

In a previous macroblog post, I thought I had discovered an interesting differential between the wage growth of middle-wage earners and that of low/high-paid workers. It turns out that what I actually discovered is that my programming skills could be improved upon. The following is an update to the post, written after correcting the coding error. Although there is no obvious wage growth polarization story, the wages of low-wage workers are currently rising at a faster median rate than for other workers.

Updated Post:

One of the most frequent questions we receive about the Atlanta Fed's Wage Growth Tracker (the median of year-over-year percent changes in individuals' hourly wage) is about the relationship between wage level and wage growth. For example, do high-wage earners also tend to experience greater wage growth?

When looking at wage growth by wage level, whether you use the prior or current wage level as the reference point matters—a lot. If we looked at wage growth categorized by the prior year's wages, we would find higher median wage growth for low-wage earners than for high-wage earners. This is because some workers who earned low wages last year earn middle or high wages this year, and some of last year's high-wage workers earn middle or low wages this year. If we instead categorized people based on current-year wages, we would see exactly the opposite: lower median wage growth for low-wage workers than for high-wage workers (see here for more discussion).

One way to lessen this wage-level base effect is to categorize an individual's wage growth according to their average wage across the two years. The following chart shows this categorization for the 2016–17 wage growth distribution of all workers in the Wage Growth Tracker data. (Note that since 1997, the annual salary for people whose earnings are only reported on a weekly basis is top-coded at $150,000 a year—these masked observations are excluded from the analysis). In the chart, the first quartile depicts the lowest-paid 25 percent of workers based on their average 2016–17 hourly wage, and so on. The center line of the box for each quartile is the median of that group's wage growth distribution, and the lower and upper boundaries of the box are the 25th and 75th percentiles, respectively. The outer lines are the thresholds for outlier observations (see here for the calculation.)

Macroblog - November 16, 2018 - chart 1: Distribution of Growth in Hourly Wage

The chart shows that the wage growth distribution across the average-wage quartiles does, in fact, differ. In particular, the median wage growth for the lowest-paid workers is higher than the median for other types of workers. The median wage growth from 2016 to 2017 for the lowest quartile is 3.8 percent, 3.0 percent for the second quartile, and 3.2 percent for the third and fourth quartiles.

However, the pattern of relatively higher median wage growth for low-wage workers is not uniform over time. This difference is apparent in the following chart, which plots median wage growth over time for each average-wage quartile.

Macroblog - November 16, 2018 - chart 2: Median Wage Growth by Average-Wage Quartile

As the chart shows, median wage growth of low-wage workers (the green line, representing the first quartile) currently exceeds that of higher-wage workers, but it was below the median for higher-paid workers in the wake of the Great Recession. This pattern is consistent with the both the severity of the recession and what we have been hearing more recently about emerging shortages of low-skilled workers. It also appears that the median wage growth of the highest-paid workers (the blue line, representing the fourth quartile) slows by a bit less than that of other workers during downturns but is otherwise not much different than for workers in the middle of the wage distribution.

So, relative to the incorrect charts I had in the previous version of this post, there is no obvious wage growth polarization story here. The wages of low-wage workers are currently rising at a faster median rate than for other workers, and these other workers are experiencing broadly similar median wage growth.

November 16, 2018 in Employment , Labor Markets , Wage Growth | Permalink | Comments ( 0)

November 14, 2018


Polarization through the Prism of the Wage Growth Tracker

One of the most frequent questions we receive about the Atlanta Fed's Wage Growth Tracker (the median of year-over-year percent changes in individuals' hourly wage) is about the relationship between wage level and wage growth. For example, do high-wage earners also tend to experience greater wage growth?

An earlier macroblog post explored this question. Unfortunately, answering it is not as easy as it might appear. When looking at wage growth by wage level, whether you use the prior or current wage level as the reference point matters—a lot. If we looked at wage growth categorized by the prior year's wages, we would find higher median wage growth for low-wage earners than for high-wage earners. This is because some workers who earned low wages last year earn middle or high wages this year, and some of last year's high-wage workers earn middle or low wages this year. If we instead categorized people based on current-year wages, we would see exactly the opposite: lower median wage growth for low-wage workers than for high-wage workers.

One way to lessen this wage-level base effect is to categorize an individual's wage growth according to their average wage across the two years. The following chart shows this categorization for the 2016 to 2017 wage growth distribution of all workers in the Wage Growth Tracker data. In the chart, the first quartile (labeled <$13.8) depicts the lowest-paid 25 percent of workers based on their average 2016–17 hourly wage, and so on. The center line of the box for each quartile is the median of that group's wage growth distribution, and the lower and upper boundaries of the box are the 25th and 75th percentiles, respectively. The outer lines are the thresholds for outlier observations (see here for the calculation.)

The chart shows that the wage growth distribution across the average-wage quartiles does, in fact, differ. For example, the median wage growth from 2016 to 2017 for the lowest quartile is 3.9 percent, 1.6 percent for the second quartile, 1.9 percent for the third quartile, and 3.2 percent for the top quartile.

The pattern of higher median wage growth in the lower and upper quartiles, compared with the middle part of the wage distribution, is reasonably uniform over time.  However, there is a cyclical difference between the median wage growth of high- and low-wage earners. This difference is apparent in the following chart, which plots median wage growth over time for each average-wage quartile.

As the chart shows, median wage growth of low-wage workers (the green line, first quartile) currently exceeds that of high-wage workers (the blue line, fourth quartile), but it was below the median for high-wage workers in the wake of the Great Recession. This pattern is consistent with the both the severity of the recession and what we have been hearing more recently about emerging shortages of low-skilled workers. In contrast, median wage growth for workers in the middle of the wage distribution (the orange and purple lines) remains lower than for either high- or low-wage workers. Overall, these findings reinforce the idea of polarization, where the demand for workers has generally grown more in the tails of the skill/wage distribution than in the middle.

November 14, 2018 in Employment , Labor Markets , Wage Growth | Permalink | Comments ( 0)

October 26, 2018


On Maximizing Employment, a Case for Caution

Over the past few months, I have been asked one question regularly: Why is the Fed removing monetary policy stimulus when there is little sign that inflation has run amok and threatens to undermine economic growth? This is a good question, and it speaks to a philosophy of how to maintain the stability of both economic performance and prices, which I view as important for the effective implementation of monetary policy.

In assessing the degree to which the Fed is achieving the congressional mandate of price stability, the Federal Open Market Committee (FOMC) identified 2 percent inflation in consumption prices as a benchmark—see here for more details. Based on currently available data, it seems that inflation is running close to this benchmark.

The Fed's other mandate from Congress is to foster maximum employment. A key metric for performance relative to that mandate is the official unemployment rate. So, when some people ask why the FOMC is reducing monetary policy stimulus in the absence of clear inflationary pressure, what they really might be thinking is, "Why doesn't the Fed just conduct monetary policy to help the unemployment rate go as low as physically possible? Isn't this by definition the representation of maximum employment?"

While this is indeed one definition of full employment, I think this is a somewhat short-sighted perspective that doesn't ultimately serve the economy and American workers well.  One important reason for being skeptical of this view is our nation's past experience with "high-pressure" economic periods. High-pressure periods are typically defined as periods in which the unemployment rate falls below the so-called natural rate—using an estimate of the natural rate, such as the one produced by the Congressional Budget Office (CBO).

As the CBO defines it, the natural rate is "the unemployment rate that arises from all sources other than fluctuations in demand associated with business cycles." These "other sources" include frictions like the time it takes people to find a job or frictions that result from a mismatch between the set of skills workers currently possess and the set of skills employers want to find.

When the actual unemployment rate declines substantially below the natural rate—highlighted as the red areas in the following chart—the economy has moved into a "high-pressure period."

For the purposes of this discussion, the important thing about high-pressure economies is that, virtually without exception, they are followed by a recession. Why? Well, as I described in a recent speech:

"One view is that it is because monetary policy tends to take on a much more 'muscular' stance—some might say too muscular—at the end of these high-pressure periods to combat rising nominal pressures.

"The other alternative is that the economy destabilizes when it pushes beyond its natural potential. These high-pressure periods lead to a buildup of competitive excesses, misdirected investment, and an inefficient allocation of societal resources. A recession naturally results and is needed to undo all the inefficiencies that have built up during the high-pressure period.

"Yet, some people suggest that deliberately running these high-pressure periods can improve outcomes for workers in communities who have been less attached to the labor market, such as minorities, those with lower incomes, and those living in rural communities. These workers have long had higher unemployment rates than other workers, and they are often the last to benefit from periods of extended economic growth.

"For example, the gap between the unemployment rates of minority and white workers narrows as recoveries endure. So, the argument goes, allowing the economy to run further and longer into these red areas on the chart provides a net benefit to these under-attached communities.

"But the key question isn't whether the high-pressure economy brings new people from disadvantaged groups into the labor market. Rather, the right question is whether these benefits are durable in the face of the recession that appears to inevitably follow.

"This question was explored in a research paper by Atlanta Fed economist Julie Hotchkiss and her research colleague Robert Moore. Unfortunately, they found that while workers in these aforementioned communities tend to experience greater benefits from these high-pressure periods, the pain and dislocation associated with the aftermath of the subsequent recession is just as significant, if not more so.

"Importantly, this research tells me we ought to guard against letting the economy slip too far into these high-pressure periods that ultimately impose heavy costs on many people across the economy. Facilitating a prolonged period of low—and sustainable—unemployment rates is a far more beneficial approach."

In short, I conclude that the pain inflicted from shifting from a high-pressure to a low-pressure economy is too great, and this tells me that it is important for the Fed to beware the potential for the economy overheating.

Formulating monetary policy would all be a lot easier, of course, if we were certain about the actual natural rate of unemployment. But we are not. The CBO has an estimate—currently 4.5 percent. The FOMC produces projections, and other forecasters produce estimates of what it thinks the unemployment rate would be over the longer run.

For my part, I estimate that the natural rate is closer to 4 percent, and given the current absence of accelerating inflationary pressures, we can't completely dismiss the possibility that the natural rate is even lower. Nonetheless, with the unemployment rate currently at 3.7 percent, it seems likely that we're at least at our full employment mandate.

So, what is this policymaker to do? Back to my speech:

"My thinking will be informed by the evolution of the incoming data and from what I'm able to glean from my business contacts. And while I wrestle with that choice, one thing seems clear: there is little reason to keep our foot on the gas pedal."

October 26, 2018 in Economic conditions , Federal Reserve and Monetary Policy , Monetary Policy | Permalink | Comments ( 2)

October 01, 2018


Demographically Adjusting the Wage Growth Tracker

In a recent report, the Council of Economic Advisers (CEA) referred to the Atlanta Fed's Wage Growth Tracker, noting its usefulness as a people-constant measure of wage growth because it looks at the over-the-year changes in the wages for a given set of individual workers. The CEA's preferred version of the Wage Growth Tracker is the one created by my colleague Ellie Terry and described in this macroblog post. It weights the sample of individual wage growth observations so that the worker characteristics resemble the population of wage and salary earners in every month. However, the CEA report also noted that this measure does not adjust for the fact that the characteristics of wage and salary earners have changed over time.

The following table, which shows the percent of workers in different age groups for three years (in three different decades), illustrates this point. The statistics are shown for the unweighted Wage Growth Tracker sample (the green columns), and for the population of wage and salary earners (the blue columns).

 

Wage Growth Tracker Sample

Wage and Salary Earner Population

 

16-24

25-54

55+

16-24

25-54

55+

1997

10.0

77.8

12.2

15.5

73.3

11.2

2007

8.5

71.7

19.8

14.1

69.2

16.7

2017

7.5

65.8

26.7

12.8

65.1

22.1

Source: Current Population Survey, author's calculations

The table shows that the Wage Growth Tracker sample in each year has fewer young workers (and more old workers) than does the population of all wage and salary earners, a fact for which the weighted version of the Wage Growth Tracker adjusts. However, the weighted version doesn't adjust for the fact that the workforce has also become older over time—the share of workers over 54 years old has risen nearly 11 percentage points since 1997.

Shifts in the distribution of demographic and other characteristics over time could matter for measures of wage growth because, for example, wage growth tends to be much higher for young workers. Young workers switch jobs more often, whereas workers aged 55 and older tend to have the lowest rates of job switching. Other changes in the composition of the workforce could also be important, such as changes the mix of education, the types of jobs, etc.

To investigate the impact of changes in workforce characteristics over time, we developed another version of the Wage Growth Tracker. This one weights the sample for each month so that it is more representative of the wage and salary earner population that existed in 1997. So, for instance, it always has about 15.5 percent aged 16-24, 73.3 percent aged 25-54, and 11.2 percent over 54 (the blue columns in the 1997 row of the table above).

As the following chart shows, the shifting composition of the workforce has put some additional downward pressure on median wage growth in recent years. That is, median wage growth would be even stronger if the sample each month looked more like it came from the population of wage and salary earners in 1997.

All three versions of the Wage Growth Tracker—unweighted, weighted to each month's workforce characteristics, and weighted to 1997 workforce characteristics—are available in the data download section of the Wage Growth Tracker web page. Which one you prefer depends on the question you are trying to answer. The monthly weighted version makes the Wage Growth Tracker more representative of the characteristics of the employed in each month, and in doing so gives young workers more influence, but it does not control for the fact that today's workforce has a smaller share of young workers than in the past. The 1997-weighted version fixes the workforce characteristics at their 1997 levels. It says that the median growth in individual wages would be higher than it is today if the composition of the workforce had not changed (other things equal). Nonetheless, any version of the Tracker you consult in the previous chart tells a pretty similar overall story: median wage growth is significantly higher than it was five or six years ago, but it hasn't shown much acceleration over the last couple of years.

October 1, 2018 in Employment , Labor Markets , Wage Growth | Permalink | Comments ( 1)

August 23, 2018


What Does the Current Slope of the Yield Curve Tell Us?

As I make the rounds throughout the Sixth District, one of the most common questions I get these days is how Federal Open Market Committee (FOMC) participants interpret the flattening of the yield curve. I, of course, do not speak for the FOMC, but as the minutes from recent meetings indicate, the Committee has indeed spent some time discussing various views on this topic. In this blog post, I'll share some of my thoughts on the framework I use for interpreting the yield curve and what I'll be watching. Of course, these are my views alone and do not reflect the views of any other Federal Reserve official.

Many observers see a downward-sloping, or "inverted," yield curve as a reliable predictor for a recession. Chart 1 shows the yield curve's slope—specifically, the difference between the interest rates paid on 10-year and 2-year Treasury securities—is currently around 20 basis points. This is lowest spread since the last recession.

The case for worrying about yield-curve flattening is apparent in the chart. The shaded bars represent recessionary periods. Both of the last two recessions were preceded by a flat (and, for a time, inverted) 10-year/2-year spread.

As we all know, however, correlation does not imply causality. This is a particularly important point to keep in mind when discussing the yield curve. As a set of market-determined interest rates, the yield curve not only reflects market participants' views about the evolution of the economy but also their views about the FOMC's likely reaction to that evolution and uncertainty around these and other relevant factors. In other words, the yield curve represents not one signal, but several. The big question is, can we pull these signals apart to help appropriately inform the calibration of policy?

We can begin to make sense of this question by noting that Treasury yields of any given maturity can be thought of as the sum of two fundamental components:

  • An expected policy rate path over that maturity: the market's best guess about the FOMC's rate path over time and in response to the evolution of the economy.
  • A term premium: an adjustment (relative to the path of the policy rate) that reflects additional compensation investors receive for bearing risk related to holding longer-term bonds.

Among other things, this premium may be related to two factors: (1) uncertainty about how the economy will evolve over that maturity and how the FOMC might respond to events as they unfold and (2) the influence of supply and demand factors for U.S. Treasuries in a global market.

Let's apply this framework to the current yield curve. As several of my colleagues (including Fed governor Lael Brainard) have noted, the term premium is currently quite low. All else equal, this would result in lower long-term rates and a flatter yield curve. The term premium bears watching, but it is unclear that movements in the premium reflect particular concerns about the course of the economy.

I tend to focus on the other component: the expected path of policy. When we ask whether a flattening yield curve is a cause for concern, what we are really asking is: does the market expect an economic slowdown that will require the FOMC to reverse course and lower rates in the near future?

The eurodollar futures market shows us one measure of the market's expectation for the policy rate path. These derivative contracts are quoted in terms of a three-month rate that closely follows the FOMC's policy rate, which makes them well-suited for this kind of analysis. (Some technical details regarding this market can be found in a 2016 issue of the Atlanta Fed's "Notes from the Vault.")

Chart 2 illustrates the current estimate of the market's expected policy rate path. Read simply, the market appears to be forecasting continuing policy rate increases through 2020, and there is no evidence of a market forecast that the FOMC will need to reverse course in the medium term. However, the level of the policy rate is lower than the median of the FOMC's June Summary of Economic Projections (SEP) for 2019 and 2020.

Once we get past 2020, the market's expected policy path flattens. I read this as evidence that market participants overall expect a very gradual pace of tightening as the most likely outcome over the next two years. Interestingly, the market appears to expect a slower pace of tightening than the pace that at least some members of the FOMC currently view as "appropriate" as represented in their SEP submissions.

For this measure, I find the short-term perspective most informative. As one looks further into the future, the range of possible outcomes widens, as many the factors that influence the economy can evolve and interact widely. Thus, the precision of any signal the market is providing about policy expectations—if indeed there is any signal at all—is likely to be quite low.

With this information in mind, I do not interpret that the yield curve indicates that the market believes the evolution of the economy will cause the FOMC to lower rates in the foreseeable future. This interpretation is consistent with my own economic forecast, gleaned from macroeconomic data and a robust set of conversations with businesses both large and small. My modal outlook is for expansion to continue at an above-trend pace for the next several quarters, and I see the risks to that projection as balanced. Yes, there are downside risks, chief among them the effects of (and uncertainty about) trade policy. But those risks are countered by the potential for recent fiscal stimulus to have a much more transformative impact on the economy than I've marked into my baseline outlook.

I believe the yield curve gives us important and useful information about market participants' forecasts. But it is only one signal among many that we use for the complex task of forecasting growth in the U.S. economy. As the economy evolves, I will be assessing the response of the yield curve to incoming data and policy decisions along the lines I've laid out here, incorporating market signals along with a constellation of other information to achieve the FOMC's dual objectives of price stability and maximum employment.

August 23, 2018 in Economic conditions , Federal Reserve and Monetary Policy , Monetary Policy | Permalink | Comments ( 2)

August 15, 2018


Does Loyalty Pay Off?

A newspaper article last week posed the question: Why do bosses pay new hires better than loyal staffers? The article looked at the Atlanta Fed's Wage Growth Tracker data on job stayers versus job switchers and noted that job switchers are getting a bigger percentage gain in their pay than job stayers.

Does that mean that people who switch jobs are paid better than those who stay with their employer? Well, it's useful to keep in mind that job switchers and job stayers differ along a number of dimensions, and perhaps the most important is that job switchers tend to earn less than job stayers. For example, using the data that go into constructing the Wage Growth Tracker we see that the median job switcher's pay in 2017 was around 9 percent lower than the median pay of those who stayed in their job. So even though the 2017 median wage growth for job switchers was 3.9 percent versus 3.0 percent for job stayers, those who change jobs are typically paid less than those who don't.

Why is the median pay higher for people who remain in their jobs? For one thing, job stayers in Wage Growth Tracker data are relatively older, with commensurately more work experience. In addition, job stayers tend to be more educated and hence more likely to be in jobs that require specialized skills. Economic theory also suggests that holding a higher-paying job reduces the likelihood of quitting. The argument goes that as a worker's wage increases, other employers will make fewer offers that exceed the person's minimally acceptable wage (their reservation wage). As a result, as an individual moves into better paying jobs, on-the-job search efforts and expected wage growth decline.

So what should you make of the higher median wage growth enjoyed by job switchers in the Wage Growth Tracker data? I view it as an indication that the demand for labor is strong and provides plentiful opportunities for less experienced and less educated workers to improve their circumstances by changing jobs. A job has an option value, and the possibility of getting a better-paying job offer is high when the worker's reservation value is low and the frequency of offers is high.

August 15, 2018 in Employment , Labor Markets , Wage Growth | Permalink | Comments ( 1)

August 08, 2018


Immigration and Hispanics' Educational Attainment

In a previous macroblog post, Whitney Mancuso and I wrote about the improved labor market outcomes for workers with the least amount of formal education. We attributed this improvement mostly to a combination of a secular decline in the supply of these workers over time and a shift in the composition of the low-skilled workforce toward Hispanic immigrants—a group that has an especially high rate of workforce attachment.

In a related article by colleagues at the St. Louis Fed, Alexander Monge-Naranjo and Juan Ignacio Vizcaino explore how the employment characteristics of the Hispanic population have grown increasingly concentrated in low-skilled occupations over time, and they relate this to the relatively smaller gains in the average educational attainment of the Hispanic population.

The authors ask why the education level of Hispanics has lagged behind other groups and suggest that it could be a consequence of intergenerational persistence; it takes a while for the children of poorly educated immigrants to catch up with the rest of the population. This explanation is likely to play a role, especially when considering why a relatively smaller share of U.S.-born Hispanics go to college. The study also notes differences across gender, showing that Hispanic men are less likely than Hispanic women to continue their education after high school, and although the college rate has been rising for all Hispanics, it is growing faster for women.

I also want to note that a large share of the Hispanic population in the United States are foreign born, and these immigrants have a much lower average level of educational attainment than do U.S.-born Hispanics. This observation is evident in table 1, which is based on data on individuals aged 25-54 (prime age) from the Current Population Survey. For instance, in 2017, 57 percent of the U.S. prime-age Hispanic population was foreign born, and 21 percent of these prime-age foreign born Hispanics had a college degree (associate degree or higher). In contrast, 36 percent of U.S.-born prime-age Hispanics had a degree.

Table 1: Selected Characteristics of the U.S. Prime-age Population (percent)

 

Foreign born

Completed a college/associate degree

 

Hispanic

Non-Hispanic

Hispanic

Non-Hispanic

 

 

 

Foreign born

U.S. born

Foreign born

U.S. born

1997

62

9

13

22

48

37

2007

64

12

15

29

56

43

2017

57

14

21

36

64

51

Source: Current Population Survey, author's calculations

As the St. Louis Fed study concludes, a primary factor distinguishing the Hispanic workforce in the United States is their lower average level of educational attainment. Further distinguishing between foreign and U.S.-born Hispanics shows the role that immigration has played in holding down the average education level since a large fraction of Hispanic immigrants have less education.

The Hispanic/non-Hispanic college completion gap remains large and has not closed over time. However, there has been relative improvement in high school completion, as table 2 shows.

Table 2: Selected Characteristics of the U.S. Prime-age Population (percent)

 

Foreign born

Completed 12th grade

 

Hispanic

Non-Hispanic

Hispanic

Non-Hispanic

 

 

 

Foreign born

U.S. born

Foreign born

U.S. born

1997

62

9

50

81

92

92

2007

64

12

55

87

93

94

2017

57

14

65

92

95

96

Source: Current Population Survey, author's calculations

Since 1997, the share of the prime-age foreign born Hispanic population who have finished 12th grade has increased by 15 percentage points. At the same time, the share of prime-age U.S.-born Hispanics completing high school has increased by 11 percentage points and is now not much lower than for non-Hispanics. While relatively low college attendance remains a major obstacle, greater high school completion is encouraging for Hispanics' future role in the workforce.

August 8, 2018 in Education , Immigration , Labor Markets | Permalink | Comments ( 0)

Google Search



Recent Posts


Archives


Categories


Powered by TypePad