Below is a summary of my post-CPI tweets. You can follow me @inflation_guy. And, given where all of this seems to be going…you ought to.
- Core inflation only up 0.122%. But housing continues to accelerate! Apparel -0.5% this month.
- Core dips slightly to 1.734% from 1.766% y/y. At odds with our forecast, due to the continued weakness in core goods.
- Still think core ends 2013 over 2%, but depends on core commodities coming up some. Our housing forecast looks good.
- Primary Rents stays at 3%, OER at 2.2%.
- Medical Care 2.4% y/y from 2.3%. And that’s with “health insurance” falling to 2.5% from 2.9%. Obviously, that’s all pre-ACA.
- Accel Major groups: Medical Care (7.2%). Decel: Apparel, Recreation, Educ/COmm (16.3%). Everything else sideways.
- This really IS mostly about the apparel decline. Bad back-to-school adjustment probably.
- I think given apparel, what we know will happen in medical care, and the housing stuff…next month may be over 0.3% on core.
This has all the signs of one of those numbers (and we’re seeing a lot of them this month) that should be averaged with next month’s number because of data collection quirks. Actually, we probably ought to average September, October, and November data together to get a “before, during, and after” average around the government shutdown. The apparel decline hit women’s apparel, men’s apparel, and girls’ apparel, but boys’ apparel inflation accelerated. Medical care prices re-accelerated slightly, as I think is destined to happen because the current run-rate is significantly due to the effect of the sequester on Medicare reimbursements, but we can already see that the “insurance” category is going to be accelerating markedly in the next few months because of the large number of cancellations and re-policying that is going on around the implementation of Obamacare. While direct consumer purchase of insurance and/or medical care is just a small part of overall inflation, a big jump will still be felt in the overall data.
The key conundrum continues to be the softness in core goods, but as I’ve argued previously the biggest part of the effect is from the very low readings from medicinal drugs and medical equipment – both of which accelerated this month. If the apparel reading really is a quirk, then core inflation is going to start heading higher with alacrity now. All of the “interesting” parts of it already are.
Here is a summary of my tweets after the CPI release this morning. You can follow me @inflation_guy.
- CPI +0.1%/+0.1% core, y/y core to 1.8%. Core only slightly weaker than expected as it rounded down to 0.1% rather than up to 0.2%.
- Housing CPI was weak, second month in a row. Rents will eventually catch up w/ housing prices…but not yet.
- Apparel CPI was weak after a couple of strong up months. I’ll have the whole breakdown in a bit.
- Core was actually only 0.13%, suggesting last August’s 0.06% and this August’s number might merely be bad seasonals.
- Market was only looking for 0.17% or so, so it’s not a HUGE miss. Still disappointing to my forecasts as upturn in rents remains overdue.
- Core CPI now 1.766% y/y. More difficult comparison next month although still <0.2%.
- Accelerating major grps: Apparel, Medical Care, Educ/Comm, Other (20.9%); decel: Food/Bev, Housing(!), Transp (73.1%), unch: Recreation
- Housing deceleration actually isn’t worrisome. Primary rents were 3.0% y/y vs 2.8% last. OER was 2.23% vs 2.19% last.
- Housing subcomponent drag was from lodging away from home, household energy, other minor pieces. So housing inflation story still intact.
- Core services inflation unch at 2.4% y/y; core goods inflation up to 0% from -0.2%. Source of uptick: mean reversion in core goods.
- So OER still reaches a new cycle high at 2.23%…it’s just not accelerating yet as fast as I expect it to. Lags are hard!
The initial reading of this number, as the tweet timeline above shows, was negative. The figure was weaker-than-expected, and Housing CPI decelerated from 2.26% to 2.17%. This seemed to be a painful blow to my thesis, which is that rising home prices will pass through into housing inflation (expressed in rents) and push core inflation much higher than economists currently expect.
Housing CPI is one of eight major subgroups of CPI, the other seven being Food and Beverages, Medical Care, Transportation, Apparel, Recreation, Education and Communication, and Other. Housing receives the most weight, at 41% of the consumption basket and an even heavier weight in core inflation. So, a deceleration in Housing makes it very hard for core inflation to increase, and vice-versa. If you can get the direction of Housing CPI right, then you’ll have a leg up in your medium-term inflation forecast (although it isn’t very helpful in terms of projecting month-to-month numbers, which are mostly noise). Thus, the deceleration in Housing seemed discouraging.
But on closer inspection, the main portions of Housing CPI are doing about what I expected them to do. Primary Rents (aka “Rent of primary residence”) is now above 3%, in sharp contrast to the expectations of those economists and observers who thought that active investor interest in buying vacant homes would drive up the price of housing but drive down the price of rents. Though I never thought that was likely…the substitution effect is very strong…it was a plausible enough story that it was worth considering and watching out for. But in the event, primary rents are clearly rising, and accelerating, and Owners’ Equivalent Rent is also rising although less-obviously accelerating (see Chart, source BLS).
So, it is much less clear upon further review that this is a terribly encouraging CPI figure. It is running behind my expectations for the pace of the acceleration, but it is clearly meeting my expectations for what should be driving inflation higher. As I say above, econometric lags are hard – they are tendencies only, and in this case the lags have been slightly longer, or the acceleration somewhat muted, from what would typically have been expected from the behavior of home prices. Some of that may be from the “investors producing too many rental units” effect, or it might simply be chance. In any event, the ultimate picture hasn’t changed. Core inflation will continue to rise for some time, and will be well above 2% and probably 3% before the Fed’s actions have any meaningful effect on slowing the increase.
Here are my post-Employment tweets. You can follow me @inflation_guy.
- Pretty weak NFP number since the payrolls figure (169k) plus revisions (-74k) is way worse than forecast. Decline in rate irrelevant.
- Actually think Fed spent so much time talking about starting taper that they may do it anyway, but have an excuse now to delay.
- Nothing like a weak NFP number to help the beleaguered bond market. Bounce may temporary but in Sep you don’t wanna fade rallies.
- I don’t watch it much, but avg hrly earns at 2.2% is highest since brief pop to 2.3% in mid-2011. Y do people hate TIPS here?
So 10-year note yields broke above 3% overnight, the highest level since 2011. More importantly, 10-year real yields had been approaching 1% (reaching 0.93% overnight) as fear-of-taper has investors quite reasonably fleeing fixed-income.
I said above that I don’t look much at average hourly earnings. This is because the evidence is that wages follow prices, rather than prices following wages in a mythical “wage-push” inflation. Moreover, we can intuit that this is the case because if wages led inflation, we would really like inflation since we would tend to see our wages increase before inflation did…we would be doing better all the time, rather than worse. In fact, we know intuitively that is wrong.
With that giant caveat, it is worth pointing out that average hourly earnings are above median CPI (which right now is a better measure of the central tendency of inflation because of the large one-off effects in medical care) by the most they have been since 2011 (see chart below, source Bloomberg).
The unemployment rate declined, but only because the Participation Rate plumbed a new post-Carter low at 63.2%. You have to go back to July 1978 to find participation rates this low, and back then there were a lot fewer women in the workforce.
All in all, this is a pretty ugly employment report, but the FOMC has carefully lined up its doves and even gotten a few hawks to say that tapering ought to begin this month. I suspect it is still likely that they start down that path, but probably the first steps are fairly small. Still, given how far rates have risen and the possibility that this will lead to some “taper: off” talk, and given the strong seasonal tendency for rates to decline in September and early October, I would not want to fade a bond market rally.
Here are my post-CPI tweets from this morning. You can follow me @inflation_guy:
#inflation +0.2% core. But here’s the thing: that’s with housing showing unexpected softness. And housing markets are bubbling.
- Unrounded core inflation 1.698%. That’s the last we’ll see of 1.6% handles for years.
- Core inflation actually barely rounded up, at +0.155% m/m. But, again, that’s with housing inexplicably weak.
- Core services 2.4%. Core goods still plodding along at -0.2%, and holding overall core inflation down. That won’t persist.
- CPI major groups accelerating: Food/bev, Housing, Apparel, Transp, Rec, Educ/Comm (89.5%). Decelerating: Medical and Other (10.5%).
- …but housing only accelerated b/c household energy. OER was unch at 2.2% and primary rents 2.8% from 2.9%. That’s a quirk.
- certainly nothing in today’s inflation data to scare the Fed into a faster taper.
- bonds are breaking lower; although the convexity overhang has been worked off, we never got the expected bounce! Not sure why they’re weak.
- higher rates->higher velocity->more inflation pressure, ironically. in this case, higher rates won’t affect money supply as offset to that
Of all of the places I expected to see a downside surprise, housing was not it. Of course, econometric lags aren’t the same as destiny, so the fact that the leading series all turned higher at the “right time” to cause a rise in Owners’ Equivalent Rent right about now is helpful information for investing, but not necessarily a timing tool!
At 2.2%, OER is still well above core inflation and primary rents at 2.8% are as well. But core goods continue to drag on the overall core inflation number (and to hold core inflation well below median inflation, which comes out later this morning).
I feel I should nudge lower my forecast for 2013 core inflation again, to a range of 2.4%-2.7% from 2.5%-2.8%. I am doing this for two practical reasons related to housing. One is that every month that passes without the expected acceleration is one less month over which inflation can accelerate to reach my year-end target. The other is that every month that passes without the expected acceleration increases the odds that I’m simply wrong, and something is holding down rents even though home prices are launching higher. I don’t think that’s true, but I want to be cognizant of overconfidence bias! However, at this point my nudging of the forecast is more about the former point: my 2014 forecast range remains 3.0%-3.6% for core.
Writing from the Netherlands after visiting future clients this week; here is a summary of my post-CPI tweets (Follow me @inflation_guy) :
- very surprising core inflation, barely rounded up to 0.1% month/month. Waiting for breakdown, but Shelter was still +0.1% so something odd.
- Core CPI ex-shelter only 1.39%, not terribly far from the 2010 lows.
- …part of the answer is that core commodities decelerated further, -0.1% y/y. But core services, most of which is housing, ALSO decel.
- Major groups; Accel: Food/Bev (15.3%); Decel: Apparel, Transp, Medical Care, Educ/Comm (34.3%). Balance unch on y/y rate rounded to tenths.
- Housing actually accelerated slightly. So decline in core was apparel, Medical Care, Education, and non-fuel transportation. Hmmm.
- …If you believe core is going to keep falling, you DON’T want to bet on it being led lower by Medical Care and Education.
- We expect this to be the low print on core. Our forecast remains 2.6%-3.0% for 2013, but only 0.6% has been realized so far
- It wd probably be prudent to lower our 4cast range; we will if there is another miss lower in May. But not by much: housing still the driver
I think the sixth bullet is the key point: core inflation is drooping because of Medical Care and Education & Communication decelerating. This is terrific news, but there’s about forty years of history that should lead one to be skeptical that these are the categories that will lead inflation lower.
Our forecast for 2.6%-3.0% is based on an expected acceleration in rents, based on the recent rise in home prices. We’re not changing that forecast yet because our model didn’t expect the acceleration to happen yet. However, it should begin to happen in the next 1-3 months.
If primary and Owners’ Equivalent rents don’t begin to accelerate in the next month or two, we will lower our 2013 forecast simply because it will be difficult to see a sufficient acceleration to reach our goal with only a half year or so to go. But the reason we don’t lower our forecast much is that the primary driver here is still rents, and there is no question which way rent inflation is headed. Only if we conclude that for some strange reason there is going to be a permanent shift in the capitalization rate of owner-occupied housing (that is, if there is a permanent shift in the ratio of rents to prices from what it has historically been) would we reconsider the direction of our forecast, and then only if home prices stopped launching higher.
Meanwhile, weak growth numbers, soft inflation numbers, and the seeming success of the Abe program in Japan as growth there has abruptly surprised higher (although it cannot be attributable to the BOJ monetization, since that program hasn’t been around long enough to affect the real economy even if there is money illusion at work) ought to cause any silly talk about the “taper” of the Fed’s buying program. That was always due for enormous skepticism, but with all of the arrows pointing the wrong way there is almost no chance that the FOMC will elect to taper purchases in the next few months. Indeed, I would expect the “hints” of such action to cease in short order. The only reason to talk about it is to (a) convince the world that Fed policy is credible, but a ruling on that credibility won’t be made until the episode is over, based on results, not at this time and based on what they say; or (b) because there is little cost of doing so, since the markets won’t panic if there’s no chance of near-term implementation.
- upward surprise + upward revision in #Payrolls – not too shocking, as I pointed out in last article. Weak hours though…
- Here is part of what’s happening in #payrolls: more jobs, fewer hours = employers cutting back hours to avoid Obamacare coverage
- Question is, which is better for confidence? More jobs, lower earnings & wages, or fewer, but better, jobs? Probably the former.
- average weekly hours have stagnated since 2011, even as Unemployment has fallen.
Today’s Employment report was pretty straightforward: an upward surprise to payrolls and upward revisions; a decline in the Unemployment Rate, and declines in hours worked. The upward revisions to Payrolls is not really a surprise, although seeing the Unemployment Rate continue to decline when Consumer Confidence “Jobs Hard to Get” is increasing is unusual.
Two years ago, the “Average Hours Worked” was 34.4 hours and the Unemployment Rate was 9.0%. Today, average hours worked is still 34.4 hours and the Unemployment Rate is 7.5%.
What I said about Obamacare coverage should be expanded a bit. There have been anecdotal reports (see, e.g., here and here) that many employers are cutting back hours for some employees, because they are required to offer health insurance (at steep premium increases) to part-time employees working at least 30 hours per week. The incentives are large, especially for employers who are near the 50 employee cutoff, to cut back employee hours. The way this would show up in the data, if the behavior was widespread, would be (a) a decline in average hours, as more people work shorter shifts, and (b) potentially (but not automatically) an increase in the number employed, since an employer who cuts 100 hours of work from existing employees is now 10 hours short of the labor input needed. I suspect this is only partly the case – if you cut 100 hours, maybe you add three 25-hour part-timers (it still costs money to hire, after all) – but it may help explain why the payrolls number keeps rising and the jobless number keeps falling although the average hours worked is pretty stagnant.
It would also help resolve the conundrum between the “Jobs Hard to Get” survey result and the Unemployment Rate, although it is a small divergence at present. If respondents are answering the survey as if the question is whether good or full-time jobs are hard to get, it may well be the case that those jobs are getting more difficult to find while there are more part-time positions being offered.
This is mere speculation, and storytelling, but I think it’s plausible that this is happening and may be affecting the data.
Below is a summary of my post-CPI tweets. You can follow me @inflation_guy.
- Sigh.Not this month!Core #inflation +0.1%, waiting 4 data to see rounding. Big fall in apparel, & housing uptick not happening yet.
- Apparel decline most since April 2001. That will reverse.
- Core with rounding looks like 0.106%, 1.89% y/y. definitely weak. Will look at breakdown to see where.
- This is the most that y/y has been below our forecast track since January 2009!
- Core services fell from +2.6% y/y to +2.5%, but core goods fell to flat from +0.3%. Our thesis is that these will converge up, not down.
- Core goods is inherently harder to forecast. Core services is largely housing.
- Accelerating subgroups: Educ/Comm (6.8% of basket). Decel: Apparel, Transp, Food & Bev, Recreation (41.7%). Unch: Housing, Med Care, Other
- CPI drooping appears to be seasonal maladjustment. The Apparel burp itself is worth 0.04%.
- Housing – we’re still waiting for the upturn. This month primary rents rose to 2.81% from 2.74% y/y, Lodging Away from Home rose too.
- Owners’ Equivalent Rent (24% of CPI) fell to 2.083% vs 2.144%. Since Primaries are rising, probably just a lag issue. 1 or 2 more months.
- Apparel saw broad-based huge declines. Again, likely seasonal. Core decelerated 0.11%, and 0.06% of that was Apparel.
- Our forecast for full-year CPI doesn’t change as a result of this number. We’re still at 2.6%-3.0% on core for 2013.
I don’t want to overplay the Apparel card here. March is a big month for seasonal adjustment for Apparel, but it is possible that this marks the end of the two-year spike in Apparel prices. If it does, then it resolves one speculation I’ve had: that the rise in Apparel, after twenty years of flat-to-declining prices, indicated that in some areas the globalization dividend on inflation may have ended. It is far too early to say that speculation is wrong, especially with big seasonal adjustments in March. Prior to 1992, there were certainly setbacks in Apparel on a regular basis due to the difficulty in seasonal adjustment. So it’s too early to say this is wrong, but not too early to say it’s surprising. And, after all, there’s always the possibility that the screwy numbers were the last twenty-three of them, and not this one. But these are still the highest seasonally-adjusted Apparel prices we’ve had in March since 2001 (see chart, source BLS, below).
The small blip down in Housing is much less of a concern. Primary rents are still rising, and OER didn’t exactly decline aggressively. We’ve been waiting out a “flat part” in the lag structure – this just means we have to wait another month or two. The chart below is updated (multiple sources) through this morning.
None of this will help the commodities guys, nor the TIPS guys, in the short run. But it doesn’t change the big picture for inflation. It’s coming. We just have to wait another month or two for the evidence!
Here’s a summary of my post-CPI tweets:
- #CPI +0.7%/+0.2% ex food & energy. Y/Y core back up over 2% at 2.004%. And going higher.
- Gasoline prices +9.1%, so there you go for everyone who thought Jan’s number was low.
- Economists had been looking for a “soft” 0.2% on core core, and got it at +0.17% rounding up to +0.2%.
- y/y rise in Owner’s Equivalent Rent rose to 2.14%, a new post-2008 high. That’s on the way to 3% or 4% over the next 1-2 years, at least.
- Accelerating Major groups: whoa…EVERY major group accelerated y/y. All 8 of them. I’ve never seen that.
- Food & Bev: 1.625% from 1.558%. Housing 1.929% from 1.805%. Apparel 2.426% from 2.115%. Transp 2.361% from 0.712%.
- Recreation 0.890% vs 0.554%. Educ/Comm 1.740% from 1.622%. Other 1.803% from 1.574%. Only close one: Medical 3.107% from 3.095%
- #CPI didn’t surprise on the upside, but there is just nothing weak about this number. Anywhere.
- Well, core goods I guess. +0.3% from +0.4%. But core services went to 2.6%, mainly on housing strength. If core goods ever recover…
The mainstream media and many economists will yawn at this number and miss the big picture. There is just nothing important here that is weak. In particular, every major group accelerated on a year-on-year basis. That’s amazing. It’s not unprecedented, I am sure, but I don’t remember seeing it happen before. Usually some are accelerating, some are decelerating, even when inflation overall is accelerating or decelerating.
In particular, housing continues to quietly accelerate, as we’ve been predicting. It is going significantly higher.
Core inflation is going to accelerate further, although the next several months have solid year-earlier comparisons of +0.22%, +0.22%, +0.20%, and +0.21%. But we think we’ll see core inflation nevertheless accelerate over that time frame, and then there are six easy comparisons in a row with nothing above 0.17%. By year-end, we still think we will see core of 2.6%-3.0%.
You can follow me @inflation_guy!
There will be many more days ahead for the Fed, and many of them will have plenty of good news. It is a mistake to trya and read too much into one day’s economic releases. With that said, here is my attempt to do exactly that.
I tweeted the following real-time reactions (@inflation_guy) following the CPI release this morning:
- Ready for an exciting day…CPI, Claims, Philly Fed, a 30-year TIPS auction, wild commodity swings, 3 Fed Presidents…buckle up!
- Hello! Core inflation +0.3%, higher-than-expected. Look out above.
- Apparel +0.8%. Some will pooh-pooh the number on that basis, but Apparel has been trending higher for more than a year.
- To be fair, core inflation BARELY rounded up to +0.3%. But the market was looking for +0.16% or +0.17%.
- Core Services remains at +2.5% y/y, but core goods ticks up to +0.4%. The recovery of core goods has been something we’re looking for.
- Somewhat surprisingly, the +0.251% rise in core inflation did so without having a rise in Owners’ Equiv Rent. Went from 2.1% y/y to 2.08%
- Accel Inflation: Housing, Apparel, Educ/Commun, Other (54.7% of basket); Decel: Food/Bev, Transp, Med Care, Rec (45.3% of basket)
- In Transp, the drag was almost all fuel. New/used Cars, maintenance, insurance, airline fares, inter- and intracity transp all up.
- What’s amazing in the CPI today is how much it did with how little from the main driver of housing. That uptick is yet to come.
- …and, next month, headline will get upward pressure from the steep rise in gasoline, which also dampens discretionary spending.
The primary takeaway from the CPI release is this: yes, core inflation surprised a little bit on the high side. But it did so without the support of the main factor that I think will push core inflation almost certainly higher going forward: housing. Rents (both primary and OER) neither accelerated nor decelerated this month from the prior year-on-year pace. And yet, there is really no temporary factor that pushed inflation higher this month. It was fairly broad-based. Apparel stood out on the month-to-month change perspective, but here is the chart (source Bloomberg) on Apparel:
This month doesn’t appear to me as too much of a true outlier. The underlying dynamic there has simply changed.
So this month core inflation stayed at 1.9%; next month it is very likely to return to 2.0% as we are dropping off the weak February change from last year. And all of that, before the housing inflation hits the data.
Speaking of housing inflation, there is no sign yet of that abating. In today’s Existing Home Sales report, the year-on-year change in Median Existing Home Sales Prices rose to 12.61%, another post-2005 record, and the highest real price increase ever, outside of 2005. This is happening because the inventory of new homes has dropped to almost a record low – really! Sure, the chart below (source Bloomberg) ignores “shadow inventory,” but it is starting to look more like the inventory of new homes now.
Some of that is seasonal, but there’s no doubt that lower inventories are now helping the home pricing dynamic. And, as I’ve shown previously, the inventory of existing homes actually has a nice relationship with shelter inflation 1-2 years later (Source: Enduring Investments):
The current level of inventories translates into a 3.6% expected rise in CPI-Shelter over the course of 2014. So you see, we’re not only firing inflationary rounds but we’re also continuing to feed more ammunition into the gun for next year. Our model of housing inflation projects Owners’ Equivalent Rent no lower than 3% by year-end 2013. And if that happens, there is no way that overall core inflation is going to be at 2%.
Now, in addition to the bad news on prices and the news on home prices that are probably seen at the Fed as a guarded positive (after all, it means the mortgage crisis is essentially over as more borrowers will be ‘above water’ again every month hereafter), there was also a mild surprise on the high side from Initial Claims (362k versus 355k) and a bad miss on the Philly Fed index for February. This latter was expected at +1.0 after -5.8 last month; instead it dropped to -12.5. Philadelphia-area manufacturers have reported softening business conditions in three of the last four months, suggesting that December’s pop to +4.6 was the outlier. Now, there were similar one-month dips in August of 2011 and June of 2012, so we’ll have to see if it is sustained…but it is consistent with the report out of Wal-Mart and the worsening of business conditions in Europe.
Higher prices (and more coming, on the headline side, as retail gasoline prices have now risen in 35 consecutive days) and lower business activity. This is exactly the opposite of what the Fed wants. It has been a bad day at the Fed.
However, it is exactly what traditional monetarism expects: accommodative monetary policy leads to higher prices (check), and has no effect on real activity in the absence of money illusion (check). So score one point for Friedman today.
And so, what else would you expect after such a day? Bond yields are declining, inflation breakevens are narrowing, and industrial commodities (metals and energy) are sliding. As with so much else these days, that makes no sense, unless you just don’t know what’s going on. When we encounter these bouts with irrationality (or, more fairly, thick-headedness), the market can be frustrating for a long time – and the ultimate denouement can sometimes be jarring. As I said earlier in this post: buckle up!
Here are my post-Employment tweets, summarized and expanded on:
- Employment Report: Unemp Rate up to 7.923% from 7.849% – the consumer confidence report was the hint that’d happen.
- Always hard to look at the January Payrolls report especially. In addition to rockin seasonals, you get prior yr revisions.
- Avg Hourly Earnings +2.1% y/y, faster than any time in 2012. Wages lag inflation so this will be higher than this at year-end.
- …the bad news is that aggregate inflation will be higher as well.
- So after the benchmark revisions, avg 2011 payrolls gain was 153k. For 2012: 156k. Today: 157k. Nice stimulus.
- Nearly 5 years after the recession started, “Not in Labor Force, Want a Job Now” still near the highs.
Here is the BLS chart on the “Not in Labor Force, Want a Job Now” series.
And here is a chart (Source: BLS, Bloomberg) showing the trailing 12-months deficit (represented positive) versus Payrolls. You can see why there’s some reason to think the massive spending (blue line) curtailed further job losses in the recession, although it’s important to remember that we don’t know the counterfactual…that is, what would have happened in the absence of spending.
The graph does not imply that if the government had not run a huge deficit that we would have had continuing job losses, even though that is the tale our elected representatives would like you to believe. Indeed, look at the next chart, which shows the level of the deficit versus the 12-month acceleration/deceleration in job growth, lagged 12 months.
If there seems to be a correlation between big deficits and job market acceleration, it comes mainly from the big swings associated with the teeth of the crisis when the causality may have been going either direction. Take out that big “S” and you have similar jobs growth with huge government and with small government (and you can see that same fact on the prior chart).