Search Results: 181
Pantheon Macroeconomics aims to be the premier provider of unbiased, independent macroeconomic intelligence to financial market professionals around the world.
Sorry, but our website is best viewed on a device with a screen width greater than 320px. You can contact us at: firstname.lastname@example.org.
181 matches for " payrolls":
We expect a 350K print for October payrolls today. The ADP report was stronger than we expected, suggesting that the post-hurricane rebound will recover more of the ground lost in September than we initially expected.
We aren't perturbed by the undershoot in December payrolls, relative both to the October and November numbers and all the leading indicators.
The downside surprise in April payrolls reflected weakness in just three components--retail, construction, and government--compared to their prior trends. Of these, we think only the construction numbers are likely to remain soft in May. Had it not been for the Verizon strike, then, we would have expected payrolls to rise by just over 200K in May, but the 35K strike hit means our forecast is 170K.
As a general rule, the best forecast of ADP payrolls in any given month is the official estimate for private payrolls in the previous month. This partly reflects the simple observation that payroll trends, once established, tend to persist, but it also is a consequence of ADP's methodology. The ADP number is generated from a model which combines both data collected from firms which use ADP for payroll processing, and lagged official data. The latter appear to be more important in determining in the month-to-month swings in the ADP number. ADP does not hide the incorporation of lagged official data in its model--you can read about it in the technical guide to the report--but neither does it shout it from the rooftops.
Whatever you think is the underlying tr end in payroll growth, you probably should expect a modest undershoot in today's report, thanks to the persistent tendency for the first estimate of September payrolls to undershoot and then be revised higher. The good news is that the initial September error tends not to be as big as in August--the median revision from the first estimate to the third over the past six years has been 49K, compared to 66K--and it has declined recently. Over the past three years, September revisions have ranged from only 18K to 27K. Still, we can't ignore six straight years of initial undershoots.
Today's December payroll number was a tricky call even before yesterday's remarkably strong ADP report, showing private payrolls soaring by 271K.
The ADP employment report suggests that the hit to payrolls from Hurricane Florence was smaller than we feared, so we're revising up our forecast for the official number tomorrow to 150K, from 100K.
All the signs are that ADP will today report a solid increase in February private payrolls; our forecast is 200K, but if you twist our arms we'd probably say the mild weather last month across most of the country points to a bit of upside risk.
We have had a modest rethink of our June payroll forecast and have nudged up our number to 150K, still below the 180K consensus. Our forecast has changed because we have re-estimated some of our models, not because of the 172K increase in the ADP measure of private payrolls. ADP is a model-based estimate, not a reliable survey indicator.
March payrolls were constrained by both the impact of colder and snowier weather than usual in the survey week, and a correction in the construction and retail components, which were unsustainably strong in February.
The dip in payroll growth in September was due to Hurricane Florence. We expect a clear rebound in payrolls in October; our tentative forecast is 250K.
The ADP employment report for September showed private payrolls rose by 135K, trivially better than we expected.
The 253K increase in May private payrolls reported by ADP yesterday was some a bit stronger than our 225K forecast. Plugging the difference between these numbers into our payroll model generates our 210K forecast for today's official number.
We're expecting a hefty increase in private payrolls in today's August ADP employment report. ADP's number is generated by a model which incorporates macroeconomic statistics and lagged official payroll data, as well as information collected from firms which use ADP's payroll processing services.
We're expecting a hefty increase in February payrolls today, but even a surprise weak number likely wouldn't prevent a rate hike next week. The trends in all the private sector employment surveys are strong and improving, and jobless claims have dropped to new lows too, though we think that's probably less important than it appears.
The 20K increase in February payrolls is not remotely indicative of the underlying trend, and we see no reason to expect similar numbers over the next few months.
Our payroll model, which incorporates survey data as well as the error term from our ADP forecast, points to a hefty 225K increase in November employment. We have tweaked the forecast to the upside because of the tendency in recent years for the fourth quarter numbers to be stronger than the prior trend, as our first chart shows.
Our below-consensus 125K forecast for today's February payroll number is predicated on two ideas.
We have few doubts that labor demand remained strong in January, but the chance of a repeat of December's 312K payroll gain is slim.
Private sector payroll growth has averaged 190K over the past year, but the six-month average has slowed to 150K. The downshift is consistent with the weakening in survey-based measures of hiring intentions, which began to soften at the turn of the year.
Everything but the weather points to a strong headline payroll number for March. Our composite leading payroll indicator has signalled robust job growth since last fall, and the message for March is very clear.
The huge rebound in September's ISM non- manufacturing survey, reported yesterday, strongly supports our view that the August drop was more noise than signal.
We chose last week to ignore the payroll warning signal from the ISM non-manufacturing employment index, which rolled over in January and February, because the danger seemed to have passed. The ISM is not always a reliable indicator--the drop in the index in early 2014 was not replicated in the official data, but the plunge in early 2015 was--and usually it operates with a very short lag, just a month or two.
We would be quite surprised if today's official payroll number exceeded the 135K ADP reading; a clear undershoot is much more likely.
Barring some sort of out-of-the-blue shock, we are much more interested in the hourly earnings data today than the headline payroll number. The key question is the extent to which wages rebound after being depressed by a persistent calendar quirk in both February and March.
If the underlying trend in payroll growth is about 200K, then a weather-depressed 98K reading needs to be followed by a rebound of about 300K in order fully to reverse the hit. But the consensus for today's April number is only 190K, and our forecast is 225K.
We'd be quite surprised if the headline payroll number today turned out to be far from the consensus, 205K, or our forecast, 225K.
The terrible scenes from Texas will play out in the economic data over the next few weeks.
April payroll growth likely will be reported at close to 200K. Overall, the survey evidence points to a stronger performance, but they don't take account of weather effects, and April was a bit colder and snowier than usual. We're not expecting a big weather hit, but some impact seems a reasonable bet.
The 6.4-point rebound in the May ISM non-manufacturing employment index, to a very high 57.8, supports our view that summer payroll growth will be strong. On the face of it, the survey is consistent with job gains in excess of 300K, as our first chart shows, but that's very unlikely to happen.
Your correspondent is on the slopes this week, but the employment report deserves a preview nonetheless.
Don't bet the farm on today's October payroll numbers, which will be hopelessly--and unpredictably-- compromised by the impact of hurricanes Florence and Michael.
We argued yesterday that the August payroll number is unlikely to be a blockbuster, thanks to a combination of problems with the birth/death model and the strong tendency for this month's jobs number to be initially under-reported and then revised substantially higher. But these arguments don't apply to the unemployment rate, which is derived from the separate household survey.
The ADP measure of private employment hugely overstated the official measure of payrolls in September, in the wake of Hurricane Irma, but then slightly understated the October number.
It's always easy to find reasons to doubt single monthly observations of any economic time series, but our first chart makes it very clear that the labor market has strengthened markedly over the past few months. The underlying trend rate of growth in private payrolls is now above 300K for the first time in exactly 20 years, and we seen no reason to expect much change over the next few months.
We're expecting a 175K increase in December payrolls today. Our forecast has been nudged down from 190K in the wake of the ADP employment report, which was slightly weaker than we expected.
In the wake of yesterday's ADP report, which showed private payrolls rising by only 163K, we have pulled down our forecast for today's official number to 170K.
Our base forecast for today's February payroll number is an unspectacular 220K, though if you twist our arms we'd probably say that we'd be less surprised by a big overshoot to this estimate than an undershoot. The single biggest argument against a big print today is simply that February payrolls have initially been under-reported in each of the past five years and then revised higher.
A casual glance at our char t below, which shows the number of job openings from the JOLTS report, seems to fit our story that the slowdown in payrolls in April and May--perhaps triggered by the drop in stocks in January and February--will prove temporary. Job openings dipped, but have recovered and now stand very close to their cycle high.
We're guessing Fed Chair Yellen would have preferred to have another acceleration in hourly earnings and a dip in the unemployment rate along side the hefty 211K leap in November payrolls, but no matter. At its October meeting, the Fed wanted to see "some further improvement in the labor market", and by any reasonable standard a 509K total increase in payrolls in two months fits the bill.
The flow of data pointing to strength in the labor market has continued this week, on the heels of last week's report of a 250K jump in October payrolls.
ADP's report of a 235K increase in private payrolls in February is not definitive evidence of anything, but it is consistent with the idea that labor demand remains very strong.
The BLS offered no estimate of the impact on payrolls of the snowstorm which hit the Northeast during the March survey week, but it appears to have been substantial. All the leading indicators pointed to a solid 200K-plus reading, more than double the official initial estimate, 98K.
If our composite index of businesses' hiring plans could speak, it would say: "Told you payrolls were going to go nuts at the end of the year."
Neither the 33K drop in September payrolls nor the 0.5% jump in hourly earnings tells us anything about the underlying state of the labor market.
We're fully expecting to see a hit to September payrolls from Hurricane Florence, which struck during the employment survey week.
The November ADP employment report today likely will show private payrolls rose by about 180K. We have no reason to think that the trend in payroll growth has changed much in recent months, though the official data do appear to be biased to the upside in the fourth quarter, probably as a result of seasonal adjustment problems triggered by the crash of 2008. We can't detect any clear seasonal fourth quarter bias in the ADP numbers.
We look for a 210K increase in July payrolls. That would be consistent with the message from an array of private sector surveys, as well as the recent trend.
This is the final Monitor before we head out for our spring break, so we have added a page in order to make room to preview the employment report due next Friday, April 4. We expect a solid but unspectacular 175K increase in payrolls, slowing from February's unsustainable 242K, but still robust.
Most of the indicators we follow point to another very strong payroll report today; we look for a 260K gain, matching May's performance. The best single leading indicator, the NFIB small business hiring intentions number, points to a massive 320K leap in private payrolls, as our first chart shows.
Today's October ADP measure of private payrolls likely will overshoot Friday's official number.
We're expecting to see November payrolls up by about 200K this morning, but our forecast takes into account the likelihood that the initial reading will be revised up. In the five years through 2014, the first estimate of November payrolls was revised up by an average of 73K by the time o f the third estimate. Our forecast for today, therefore, is consistent with our view that the underlying trend in payrolls is 250K-plus. That's the message of the very low level of jobless claims, and the strength of all surveys of hiring, with the exception of the depressed ISM manufacturing employment index. Manufacturing accounts for only 9% of payrolls, though, so this just doesn't matter.
It is possible that the broad-based softness of September payrolls captures a knee-jerk reaction on the part of employers, choosing to wait-and-see what happens to demand in the wake of stock market correction. But that can't be the explanation for the mere 136K August gain, because the survey was conducted before the market rolled over. Even harder to explain is the hefty downward revision to August payrolls, after years of upward revisions. All is not yet lost for August--the last time the first revision to the month was downwards, -3K in 2010, the second revision was +56K--but we aren't wildly optimistic.
In the wake of yesterday's ADP report, which showed private payrolls up 250K in December, we have revised our forecast for today's official headline number up to 240K from 210K.
We're pretty sure that the unemployment rate didn't drop by 0.3 percentage points in November. We're pretty sure hourly earnings didn't fall by 0.1%. And we're pretty sure payrolls didn't rise by 178K. All the employment data are unreliable month-to-month, with the wages numbers particularly susceptible to technical quirks.
We had hoped that the statistical problems which have plagued the initial estimates of August payrolls in recent years had faded, but Friday's report suggests our judgement was premature.
Long-time readers will be aware of our disdain fro the ADP employment report, which usually tells us nothing more than that the monthly changes in payrolls tend to mean-revert. That's because the published ADP employment number for each month is generated by regressions which incorporate lagged official payroll data, as well as information from companies which use ADP for payroll processing.
The undershoot in April payrolls, relative to the consensus, is a story of a fluke number in just one sector. Retail payrolls reportedly shrank by 3K, after rising by an average of 52K over the previous six months. Our first chart shows clearly that the retail payrolls are quite volatile over short periods, with sudden and often inexplicable swings in both directions quite common.
We don't believe that payrolls rose only 138K in May. History strongly suggests that when the May payroll survey is conducted relatively early in the month, payroll growth falls short of the prior trend.
The models which generate the ADP measure of private payrolls will benefit in May from the strength of the headline industrial production, business sales and jobless claims numbers.
The trend rate of increase in private payrolls in the months before Hurricane Katrina in 2005 was about 240K per month.
The 351K net increase in payrolls reported Friday--a 261K October gain and a 90K total revision to August and September--puts the labor market back on track after the hurricanes temporarily hit the data.
We don't directly plug the ADP employment data into our model for the official payroll number. ADP's estimate is derived itself from a model which incorporates lagged official payroll data, because payrolls tend to mean-revert, as well as macroeconomic variables including oil prices, industrial production and jobless claims -- and actual employment data from firms which use ADP's payroll processing services.
The first thing to ask after a payroll number far from consensus is whether it is supported by other evidence. We are happy to argue that November's blockbuster report is indeed consistent with a range of other numbers, notwithstanding the unfortunate truth that there are no reliable indicators of payrolls on a month-to-month basis.
Payroll growth rebounded to 223K in May, after two sub-200K readings, and we're expecting today's June ADP report to signal that labor demand remains strong.
August could be the month to break the mold with regards to payroll figures, Ian Shepherdson, chief economist at Pantheon Macroeconomics, said.
You may have seen the chart below, which shows what appears to be an alarming divergence between the official jobless claims numbers and the Challenger survey's measure of job cut announcements. We should say at the outset that the chart makes the fundamental mistake of comparing the unadjusted Challenger data with the seasonally adjusted claims data.
It's hard to find anything to dislike in the February employment report.
To the extent that markets bother with the NFIB survey at all, most of the attention falls on the labor market numbers. But these data--hiring, compensation, jobs hard-to-fill--haven't changed much in recent months, and in any event most of them are released the week before the main survey, which appeared yesterday. The message from the labor data is unambiguous: Hiring remains very strong, employers are finding it very difficult to fill open positions, and compensation costs are accelerating.
Our forecast for a 0.3% increase in the September core PPI, slightly above the underlying trend, is even more tentative than usual.
If you had only the NFIB survey of small businesses as your guide to the state of the business sector, you'd be blissfully unaware that the economic commentariat right now is obsessed with the potential hit from the trade tariffs, actual and threatened.
The key labor market numbers from the monthly NFIB survey of small businesses are released ahead of the main report, due today.
We argued in the Monitor yesterday that the plunge in capital spending on equipment in the oil sector could cost about 300K jobs over the course of this year. Adding in the potential hit from falling spending on structures, which likely will occur over a longer period, given the lead times in the construction process, the payroll hit this year could easily be 500K, or just over 40K per month.
A rate hike from the Fed this week would be a gigantic surprise, and Yellen Fed has not, so far, been in the surprise business. It would be more accurate to describe the Fed's modus operandi as one of extreme caution, and raising rates when the fed funds future puts the odds of action at close to zero just does not fit the bill.
The gratifyingly strong 222K headline June payroll gain, if repeated through the second half of the year, will put unemployment below 4% by December.
Yesterday's FOMC statement was a bit more upbeat on growth than we expected, with Janet Yellen's final missive describing everything -- economic growth, employment, household spending, and business investment -- as "solid".
The pace of layoffs might be picking up. Our first chart looks pretty convincing, but it's much too soon to know for sure. The claims data from mid-December through late January are subject to serious seasonal adjustment problems, partly because Christmas falls on a different day of the week each year and partly because the exact timing of post-holiday layoffs varies from year-to-year.
We've had some correspondence questioning our view on the "weakness" of February hourly earnings, which we firmly believe were depressed by a persistent calendar quirk. Almost nine times in 10 over the past decade, when the 15th of the month has fallen after the week of the 12th--the payroll survey week--the monthly gain in wages has undershot the prior trend.
In theory, the headline labour market data in France should be a source of comfort and support for the new government.
As we reach our Sunday afternoon deadline, Hurricane Irma is pounding Florida's west coast with an intensity not seen since Andrew, in 1992.
Yesterday's French industrial production data were worse than we expected. Output slipped 1.1% month-to-month in September, pushing the year-over-year rate down to -1.1% from a revised +0.4% in August. Mean-reversion was a big driver of the poor headline, given the upwardly-revised 2.4% jump in August.
The easiest way to track the impact of the rising dollar on real economic activity is via the export orders component of the ISM manufacturing survey. We have been profoundly skeptical of the value of the ISM headline index, because it suffers from substantial seasonal adjustment problems, but the export orders index seems not to be similarly afflicted.
The weekly jobless claims numbers tend to be choppy around the turn of the year, and our take on the seasonal adjustments points to a clear increase in today's report, for the week ended January 11, even without the impact of the government shutdown.
Back in the dim and increasingly distant past the semi-annual Monetary Policy Testimony--previously known as the Humphrey-Hawkins--used to be something of an event. Today's Testimony, however, is most unlikely to change anyone's opinion of the likely pace and timing of Fed action.
For some time, the Fed has been locked in a loop of endless inaction. Every time the economic data improve and the Fed signals it is preparing to raise rates, either markets--both domestic and global-- react badly, and/or a patch of less good data appear. The nervous, cautious Yellen Fed responds by dialling back the talk of tightening, and markets relax again, until the next time.
At the headline level, much of the recent U.S. macro dataflow has been disappointing. January retail sales, industrial production, housing starts, and both ISM surveys--manufacturing and non-manufacturing-- undershot consensus, following a sharp and unexpected drop in December durable goods orders.
The chance of a zero GDP print for the first quarter diminished--but did not die--last week when the president signed a bill granting full back pay to about 300K government workers currently furloughed.
Back on May 14, we argued--see here--that the stars were aligned to generate very strong second quarter GDP growth, perhaps even reaching 5%.
The core economic narrative in U.S. markets right now seems to run something like this: The pace of growth slowed in Q1, depressing the rate of payroll growth in the spring. As a result, the headline plunge in the unemployment rate is unlikely to persist and, even if it does, the wage pressures aren't a threat to the inflation outlook.
This week's GDP figures showed that firms invested only sparingly in 2016, but their financial fortunes have been bolstered by a recovery in profits. The gross operating surplus of all firms rose by 4.5% quarter-on-quarter in Q4, the biggest increase for 11 quarters. This pushed the share of GDP absorbed by profits up to 21.3%, just above its 60-year average of 21.2%.
The key market risk in the August employment report is the hourly earnings number. The consensus forecast is for a 0.2% month-to-month increase, in line with the underlying trend, but the balance of risks is firmly to the downside.
After a slew of media reports in recent days, we have to expect that the president will today announce that Fed governor Jerome Powell is his pick to replace Janet Yellen as Chair.
Last week, the MBA's measure of the volume of applications for new mortgages to finance house purchase rose 1.7%.
The latest survey evidence strongly supports our view that momentum is building in the industrial economy, but the official production data continue to lag. Yesterday's March Philly Fed survey was remarkably strong, with the correction in the headline sentiment index -- inevitable, after February's 33-year high -- masking increases in all the subindexes.
We can't remember the last time a single economic report was as surprising as the December retail sales numbers, released yesterday.
In the wake of last week's rate increase, the fed funds future puts the chance of another rise in September at just 16%. After hikes in December, March and June, we think the Fed is trying to tell us something about their intention to keep going; this is not 2015 or 2016, when the Fed happily accepted any excuse not to do what it had said it would do.
The FOMC statement did enough to keep alive the idea that rates could rise in March, but the ball is now mostly in Congress' court. If a clear plan for substantial fiscal easing has emerged by the time of the meeting on March 15, policymakers can incorporate its potential impact on growth, unemployment and inflation into their forecasts, then a rate hike will be much more likely.
We can't recall a time when we have disagreed so strongly with the consensus narrative, in both the media and the markets, about the state of the U.S. economy. We think both investors and the commentariat are too bearish on growth and too complacent about inflation risks, and as a result, insufficiently worried about the speed with which interest rates will rise over the next couple of years.
The case for expecting a robust January jobs number is strong, but it is not without risks.
The partial government shutdown is now the longest on record, with little chance of a near-term resolution.
We expected a consensus-beating ADP employment number for February, but the 298K leap was much better than our forecast, 210K. The error now becomes an input into our payroll model, shifting our estimate for tomorrow's official number to 250K; our initial forecast was 210K.
Mean-reversion is a wonderful thing; it's what gives the ADP employment report the wholly unjustified appearance of being a useful leading indicator of payroll growth. Over time, the best single forecast of payroll gains or losses in any particular month is whatever happened last month.
The shock of the weak May payroll report means that the June numbers this week will come under even greater scrutiny than usual. We are not optimistic that a substantial rebound is coming immediately. The headline number will be better than in May, because the 35K May drag from the Verizon strike will reverse.
The odds favor a robust January payroll report today. The key leading indicator--the NIFB hiring intentions index from five months ago--points to a 275K increase, while the coincident NFIB actual employment change index suggests 260K.
Where to start with the January employment report, where all the key numbers were off-kilter in one way or another?
The least-bad way to forecast the ADP employment number is to look at the official private payroll number for the previous month. ADP's methodology generates employment numbers from a model incorporating lagged data from the Bureau of Labor Statistics as well as information from companies which use ADP for payroll processing.
The March employment report didn't tell us what we really want to know. The underlying trend in wage growth remains obscured by the calendar quirk which depresses reported hourly earnings when the 15th of the month--pay day for people paid semi-monthly -- falls after the payroll survey week.
The ADP report yesterday has not changed our view that tomorrow's payroll number will be about 180K, well below our estimate of the underlying trend, which is about 250K. ADP's numbers are heavily influenced by the BLS data for the prior month, and tell us little or nothing about the next official report.
The 90-day truce in the trade wars between the U.S. and China, brokered on Saturday at the G20 meeting in Argentina, is a big deal for financial markets in the euro area, at least in the near term.
The Fed surprised no-one yesterday, leaving rates on hold, saying nothing new about the balance sheet, and making no substantive changes to its view on the economy. The statement was tweaked slightly, making it clear that policymakers are skeptical of the reported slowdown in GDP growth to just 0.7% in Q1: "The Committee views the slowing in growth during the first quarter as likely to be transitory".
We didn't believe the first estimate of Q1 GDP growth, 0.7%, and we won't believe today's second estimate, either. The data are riddled with distortions, most notably the long-standing problem of residual seasonality, which depressed the number by about one percentage point.
Today brings the first glimpse of the post-hurricane employment picture, in the form of the September ADP report.
Markets and the commentariat seemed not to like the April ADP employment report yesterday but we are completely indifferent. We set out in detail in yesterday's Monitor the case for expecting a below consensus ADP reading--in short, the model used to generate the number includes lagging official data, some of which were hugely depressed by the early Easter--so it does not change our 200K forecast for tomorrow's official number.
We very much doubt that Fed Chair Powell dramatically changed his position last week because President Trump repeatedly, and publicly, berated him and the idea of further increases in interest rates.
The headline 250K October payroll number looked great.
In the wake of the ADP report released Wednesday, we moved up our payroll forecast to 150K from 100K, but we've now taken a closer look at the post-Florence path of jobless claims.
The official payroll numbers seem not to be consistently affected by seasonal adjustment problems when Easter falls in March, probably because the earliest possible date for the holiday, the 23rd, comes long after the payroll data are captured. The BLS data cover the week of the 12th.
Behind all the talk of slowdowns and Fed pauses, we see no sign that the labor market is loosening beyond a very modest uptick in jobless claims, and even that looks suspicious.
Our Chief Eurozone Economist, Claus Vistesen, is covering the Italian situation in detail in his daily Monitor but it's worth summarizing the key points for U.S. investors here.
The rebound in the ISM non-manufacturing index in February was in line with our forecast, but behind the strong headline, the employment index dropped to an eight-month low.
The budget sequestration process, which cut discretionary government spending by a total of $114B in fiscal 2013 and fiscal 2014, was one of the dumbest things Congress has done in recent years.
The 62K jump in jobless claims for the week ended September 2 is a hint of what's to come. Claims usually don't surge until the second week after major hurricanes, because people have better things to do in the immediate aftermath, so we are braced for a further big increase next week.
At their March meeting FOMC members' range of forecasts for the unemployment rate in the fourth quarter of this year ranged from 4.4% to 4.7%, with a median of 4.5%. But Friday's report showed that the unemployment rate hit the bottom of the forecast range in April.
In the wake of the February employment report, the implied probability of a June rate hike, measured by the fed funds future, jumped to 89% from 71%. The market now shows the chance of a funds rate at 75bp by the end of the year at just over 60%. That still looks low to us, but it is a big change and we very much doubt it represents the end of the shift in expectations.
The headline May ISM non-manufacturing index today likely will mirror, at least in part, the increase in the manufacturing survey, reported Friday.
The FOMC flagged recent market developments as a source of risk to the U.S. economy yesterday, unsurprisingly, but didn't go overboard: "The Committee is closely monitoring global economic and financial developments and is assessing their implications for the labor market and inflation, and for the balance of risks to the outlook."
We expect August's GDP figures, released on Wednesday, to show that month-to-month growth slowed to 0.1%, from 0.3% in July.
Fed Chair Yellen's speech in Cleveland yesterday elaborated on the key themes from last week's FOMC meeting.
The next nine weeks bring three jobs reports, which will determine whether the Fed hikes again in September, as we expect, and will also help shape market expectations for December and beyond.
Chair Yellen's final FOMC meeting today will be something of a non-event in economic terms.
We have argued frequently that the ADP employment report is not a reliable advance payroll indicator--see our Monitor of May 4, for example-- so for now we'll just note that it is generated by a regression model which includes a host of nonpayroll data and the official jobs numbers from the previous month. It is not based solely on reports from employers who use ADP for payroll processing, despite ADP's best efforts to insinuate that it is.
With the FOMC decision now just seven days away, the forcefulness of recent Fed speakers has led many analysts to argue that only a spectacularly bad payroll report, or an external shock, can prevent a rate hike next week. External shocks are unpredictable, by definition, and we think the chance of a startlingly terrible employment report is low, though substantial sampling error does occasionally throw the numbers off-track.
Today's wave of economic reports are all likely to be strong. The most important single number is the increase in real consumers' spending in July, the first month of the third quarter.
The Fed left rates on hold yesterday, as expected, repeating its long-held core view that inflation will rise to 2% in the medium-term, requiring gradual increases in the fed funds rate.
If you need more evidence that the U.S. economy is bifurcating, look at the spread between the ISM non- manufacturing and manufacturing indexes, which has risen to 3.5 points, the widest gap since September 2016.
The Fed likely will do nothing today, both in terms of interest rates and substantive changes to the statement. We'd be very surprised to hear anything new on the Fed's plans for its balance sheet.
In a note to clients ahead of the report, Ian Shepherdson at Pantheon Macro said that while ADP isn't all that reliable of an indicator for the government's payroll release, set for Friday morning.
Chief U.S. Economist Ian Shepherdson on U.S. Employment
For analysts with a broadly positive view of the U.S. economy, it is tempting to argue that the slowdown in payroll growth this year reflects supply constraints, as the pool of qualified labor dries up.
The softening in payroll growth in November appears mostly to be a story about short-term noise, rather than a sign that tariffs are hurting or that the broader economy is slowing.
In the wake of the soft-looking ADP employment report released on Wednesday, the true consensus for today's official payroll number likely is lower than the 230K reported in the Bloomberg survey. As we argued in the Monitor yesterday, though, we view ADP as a lagging indicator and we don't use it is as a forecasting tool.
One of the possible explanations for the slowdown in payroll in growth in recent months is that the pool of labor has shrunk to the point where employers can't find the people they want to hire. That's certainly one interpretation of our first chart, which shows that the NFIB survey's measure of jobs-hard-to-fill has risen to near-record levels even as payroll growth has slowed.
We're expecting a 180K increase in today's May headline payroll number, a bit below the underlying trend--200K or so--for the second straight month.
We're expecting a strong-looking 225K increase in the May ADP measure of private sector payroll growth, due today. The consensus forecast is 180K.
We were pretty sure that the underlying trend in jobless claims had bottomed, in the high 230s, before the hurricanes began to distort the data in early September.
The underlying trend in payroll growth probably has not changed significantly from the 228K average monthly gains recorded last year. But the average hides wide variations, some triggered by seasonal adjustment problems and others by one-time weather effects or unavoidable and random sampling error. January's below-trend 151K increase was likely a victim of seasonal problems, because payroll gains in recent years have tended to be soft at the start of the year after outsized fourth quarter increases.
This is the final Monitor before we hit the beach for two weeks, so want to highlight some of the key data and event risks while we're out. First, we're expecting little more from the FOMC statement than an acknowledgment that the labor market data improved in June. After the May jobs report, the FOMC remarked that "...the pace of improvement in the labor market has slowed".
Leading indicators all point to a solid August payroll number. Survey-based measures of the pace of hiring signal a 200K-plus increase, and jobless claims--a proxy for the pace of gross layoffs--are at a record low as a share of the workforce.
Over the past six months, payroll growth has averaged exactly 150K. Over the previous six months, the average increase was 230K. And in the six months to August 2015--a fairer comparison, because the fourth quarter numbers enjoy very favorable seasonals, flattering the data--payroll growth averaged 197K.
Since its October 2012 revamp, the ADP measure of private employment--the November survey will be released this morning--has tended to be little more than a lagging indicator of the official number.That's because ADP incorporates official data, lagged by one month, into the regression which generates its employment measure.
In the absence of market-moving data today, we want to take a closer look at the labor market, and, specifically, the idea that payroll growth is slowing because firms cannot find staff they consider suitably qualified for the jobs available. Every indicator of labor demand, with the sole exception of manufacturing-specific surveys, is consistent with very rapid payroll growth, well in excess of 200K per month.
We have no reason to think the underlying trend in payroll growth has changed--the 235K average for the past three months is as good a guide as any--but the balance of risks points clearly to a rather lower print for August. Two specific factors, neither of which have any bearing on the trend, are likely to have a significant influence on the numbers, and both will work to push the number below the 217K consensus.
Labor demand appears to have remained strong through August, so we expect to see a robust ADP report today.
We are expecting a hefty increase in the August ADP employment number today--our forecast is 225K, above the 175K consensus --but we do not anticipate a similar official payroll number on Friday. Remember, the ADP number is based on a model which incorporates lagged official employment data, the Philly Fed's ADS Business Conditions Index, and data from firms which use ADP for payroll processing.
Our payroll model relies heavily on lagged indicators of the pace of hiring, most of which have improved in recent months after a sustained, though modest, softening which began last spring. That's why we expected an above-consensus reading from ADP on Wednesday and from the BLS today.
The underlying trend in payroll growth ought to be running at 250K-plus, based on an array of indicators of the pace of both hiring and firing. The past few months' numbers have fallen far short of this pace, though, for reasons which are not yet clear. We are inclined to blame a shortage of suitably qualified staff, not least because that appears to be the message from the NFIB survey, which shows that the proportion of small businesses with unfilled positions is now close to the highs seen in previous cycles. If we're right, payroll growth won't return to the 254K average recorded in 2014 until the next cyclical upturn, but quite what to expect instead is anyone's guess.
The combination of astounding fourth quarter payroll numbers and weak GDP growth has prompted a good deal of bemused head-scratching among investors and the commentariat. The contrast is startling, with Q4 private payrolls averaging 276K, a 2.4% annualized rate of increase, while the initial estimate for growth seems likely close to 1%. Even on a year-over-year basis, stepping back from the quarterly noise, Q4 growth is likely to be only 2% or so.
Sooner or later, the surge in consumers' spending power triggered by the drop in gas prices and the acceleration in payrolls will appear in the retail sales data.
Markets are still discounting Banxico rate increases in the near term, despite the fact that Mexico's inflation is under control. Unless the MXN goes significantly above 18.7 per USD in the near term, or activity accelerates, we see little scope for rate increases until after the Fed hikes. After May's soft U.S. payrolls, and in light of the economic and financial risk posed by the U.K. referendum, we think a hike this week is unlikely.
The strength in payrolls in recent months is real. The three-month moving average increase in private payrolls now stands at 280K, despite adverse seasonal adjustments totalling 91K in the fourth quarter, compared to the same period last year.
The details of the substantial pay raises being offered to some 18K JP Morgan employees over the next three years are much less important than the signal sent by the company's response to the tightening labor market. In an economy with 144M people on payrolls, hefty raises for JP Morgan employees won't move the needle in the hourly earnings data.
You might want a strong coffee before you read today's Monitor, because we're going deep inside the relationship between the ADP employment report and the official numbers. We have long argued that ADP's headline reading is not a useful indicator of payrolls, because the numbers are heavily influenced by the official payroll data for the previous month, which are inputs to the ADP model. To be clear, ADP's employment number is not generated solely from hard data from companies which use the company for payroll processing; that information is just one input to their model.
We expect the Fed to leave rates on hold today, but the FOMC's new forecasts likely will continue to show policymakers expect two hikes this year, unchanged from the March projections. We remain of the view that September is the more likely date for the next hike, because we think sluggish June payrolls will prevent action in July.
The underlying trend in payroll growth is running at about 225K-to-250K, perhaps more, and the leading indicators we follow suggest that's a reasonable starting point for our December forecast. The trend in jobless claims is extraordinarily low and stable--the week-to-week volatility is eye-catching, especially over the holidays, but unimportant--and indicators of hiring remain robust. The unusually warm weather in the eastern half of the country between the November and December survey weeks also likely will give payrolls a small nudge upwards, with construction likely the key beneficiary, as in November.
In the wake of the payroll report on Friday, several readers sent us a version of the chart reproduced below, showing the rates of growth of S&P earnings and private sector payrolls. The message from the chart appears to be that the current trend in payroll growth, a bit over 200K per month cannot be sustained.
Our hopes of a hefty rebound in payrolls in October, following the hurricane-hit September number, have been dashed by the imminent landfall of Hurricane Michael in Florida panhandle.
Normal service appears to have resumed in August, with payrolls rising by 201K, very close to the 196K average over the previous year.
We expect to see a 180K increase in November payrolls
We were happy to see the 255K gain in July payrolls, but we remain nervous about the sustainability of such strong numbers. The jump in employment was very large relative to some of the key survey-based indicators of the pace of hiring, even after allowing for the 29K favorable swing in the birth/ death model, compared to a year ago, and the 27K jump in state and local government education jobs, likely due to seasonal adjustment problems
We aren't revising our payroll forecast in the wake of the ADP report, which showed private payrolls rising by 241K in December. We expected a bigger increase because ADP tends to lag the official data for the previous month, and the BLS reported a 314K jump in private employment in November, but the "shortfall" is too small to matter.
The winter hit to payrolls is now ancient history. Private employment rose by an average of 273K per month in the second half of last year, so May's 262K has restored normal service, more or less. History strongly suggests the number will be revised up, so we are happy to argue that the data convincingly support our view that the weakness in late winter and early spring was temporary, substantially due to the severe weather.
Most of the evidence points to a robust December employment report today, though we doubt the headline number will match the heights seen in November, when the initial estimate showed payrolls up 321K. We look for 275K.
We are taking our spring break starting tomorrow so it makes sense to preview the employment report today. To forecast payrolls, we start with the underlying trend -- mean reversion is the most powerful force in payrolls, most of the time -- and then look for reasons why this month's number might deviate from it.
ADP's reported 158K increase in private payrolls was very close to our model-based estimate, so it doesn't change our 220K forecast for todays official payroll number, well above the 177K consensus.
Today's ADP employment report for December ought to show private payrolls continue to rise at a very solid pace
Looking back at the numbers over the past few weeks, it is pretty clear that the gap between the strong payroll reports and the activity data widened to a chasm in the first quarter. We now expect GDP growth of about zero--the latest Atlanta Fed estimate is +0.3% and the New York Fed's new model points to 0.8%--but payrolls rose at an annualized 1.9% rate.
Any model of payrolls based on the usual indicators--jobless claims, ISM hiring, NFIB hiring, and other sundry surveys--right now points to payroll growth at 250K or better. Indeed, the ISM non-manufacturing report on Wednesday is consistent with payroll growth closer to 400K, and the lagged NFIB hiring intentions number points to 300K. Yet the consensus forecast for today's October report is just 182K. Why so timid?
All the fundamentals point to a very strong payroll number for May. The NFIB hiring in tentions index, the best single leading indicator of payrolls five months ahead, signalled back in December that May employment would rise by about 300K. The NFIB actual net hiring number, released yesterday, is a bit less bullish, implying 250K, but the extraordinarily low level of jobless claims, shown in our first chart, points to 300K. Finally, the ISM non-manufacturing employment index suggests we should be looking for payrolls to rise by about 260K. Our estimate is 280K.
First, a deep breath: June payrolls, with a margin of error of +/-107K, missed the consensus by 10K. Adding in the -60K revisions and the miss is still statistically insignificant. The story, therefore, is that there is no story. Even relative to our more bullish forecast, the miss was just 37K. Nothing bad happened in June. But we hav e to acknowledge that payroll growth has now undershot the pace implied by the NFIB's hiring intentions number--lagged by five months--in each of the past four months. In June, the survey pointed to a 320K jump in private employment, overshooting the actual print by nearly 100K.
Ian Shepherdsonon why the ADP report is simply not a reliable indicator
pantheon macroeconomics, pantheon, macroeconomic, macroeconomics, independent analysis, independent macroeconomic research, independent, analysis, research, economic intelligence, economy, economic, economics, economists, , Ian Shepherdson, financial market, macro research, independent macro research