Improving Krugman – A Story About Brazil

Paul Krugman tells a story about the Brazilian economic crisis of 2015-2016. I am usually a fan of his large economic brain. You can disagree with his politics, and even quibble with his economics, but any objective reading of the outturn since the GFC makes it clear that he got the big things right when very few others did.

His story about Brazil, however, while not outright wrong, could be substantially improved. He argues that the trifecta of a commodity shock, large domestic debt burden, and application of expansionary austerity (contractionary fiscal and monetary policies) did them in.

First, Brazil is not a big commodity exporter as a share of GDP. The pervasiveness and persistence of this misconception goaded me into posting this a few years back.

The TL;DR is that Brazil’s exports are not only about 11 percent of GDP (very low by EM standards), but the ratio actually declined over the course of the boom years (from 15% of GDP to 11%). Hardly what you’d expect to see in an export-led boom—even if you account for currency appreciation on the export share of GDP.

This doesn’t mean that the terms-of-trade shock didn’t matter. It certainly had a negative effect on output, especially when you consider the negative signaling effect to foreign investors who deeply bought into that narrative, and had, by that point, already become a material part of Brazil’s domestic debt financing. But at 11% of GDP, this part of the story was not nearly as nasty as it was made out to be.

Second, he lays a lot of blame at the feet of the Brazilian policy response. And in a very strict sense, he’s right. Raising rates in an economy that just built up a lot of domestic debt (more on that in a sec) while contracting the fiscal position will hit an economy like Brazil’s hard. Government outlays there are a large share of GDP (IIRC just under 40%).

The important omission here though is “how risky would simulative policies have been”? In most developed economies—and certainly in the US–policy credibility allows for expansionary fiscal and accommodative monetary when there’s a downturn. This is not the case in emerging market countries. In fact, many would argue that this is the defining distinction between EM and DM countries.

To the point, the legacy of liquidity and solvency crises in emerging markets results in running the risk that expansionary countercyclical policies scare investors—rightly or wrongly—and induce more aggressive retrenchment by foreign portfolio investors and further capital flight by locals. The muscle memory in Latina America on this runs particularly deep.

As a rule of thumb, the greater the perception that a country is or could easily be faced with a liquidity-cum-solvency crisis, the greater the risk that expansionary countercyclical policies prove counterproductively tragic. At some point along the credibility spectrum countries decide, understandably, that it’s not a policy risk worth running.

The real culprit in the Brazilian crisis was the sudden stop. Not the classic sudden stop à la Calvo, but the sudden stop of domestic credit. And this is the part where Krugman gets the emphasis right, citing Atif Mian et al.

From 2003 to 2014, domestic credit to the private sector in Brazil went from 23% of GDP to 70%. This, not exports, is what fueled the economic boom (and in many other EM countries in the same period). It left Brazil with a credit overhang it had to digest and a lot of debt that needed to be rolled over at higher rates, as Brazil jacked up the SELIC (Brazil’s policy rate) from 7% to 14%.

Maybe this all seems like one big quibble, but to me there are two important take-aways: (1) It’s a myth that Brazil’s s a country driven by commodity exports, and (2) the role of policy credibly needs to be a larger part of the discussion of optimal policy responses to adverse shocks. This may be obvious to someone who has spent a career toggling between EM and DM sovereign analysis, but is regularly absent when DM economists–even great ones like Krugman–go tourist.  (Side note: This is particularly important in all the current chatter surrounding MMT, but that’s a post for another day.)

Charts and Thoughts

There’s a lot going on right now. Tons of cross-currents. Lots of risk, but lots of good risk/reward set ups as well.

Subscribers Only

Access all of BehavioralMacro's premium content and the private Twitter feed via Premo

Validate Twitter

Not a subscriber yet?


Powered by

Validation unsuccessful.

Please subscribe to access the premium content.

Here are the charts and thoughts that to me matter most.

First, and most importantly, we are very, very oversold by historical standards.

The above is the McCellan advance-decline summation index. If you line up all the other times since the GFC that the indicator has been at this level with the SPX, you will see that either contemporaneous to or within a few days of the local low, stocks started a meaningful bounce. The three times this did not hold true were in the July 2008-March 2009 window. There is also a pronounced divergence in the NYSE cumulative AD line and its RSI.

(NB: For an aggressive trading style, there is a big difference in sizing a position in anticipation of an immediate bounce, and sizing it for a bounce that you think is likely to happen after the next couple/few days.)

Second, shorter-term, many of the major and sector indices are running into natural resistance spots–either popular moving averages or overhead supply points. Most investors right now are some combination of skittish, wounded, and underperforming; Can’t afford to lose more, can afford even less to be left out of a year-end rally. This creates the exaggerated buy the rip, sell the dip dynamic that is often referred to as a ‘short gamma mindset’.

Here are the SPX, IWM, and QQQ:


You can also see overhead/natural resistance in the SMH and the XLF, among others. However, on a more constructive note, if you look closely at the above charts you will see that when they made lower lows in price, their RSIs (relative strength indices) made higher lows; in other words, positive divergences.

Third,  EMFX, metals, and some EM names bottomed in early September, and platinum, gold, and some gold miners have really promising patterns. It is true whatever safe-haven bid that may have crept into gold and gold minors would unwind in a meaningful equity bounce, but the fact that EMFX bottomed at the same time, and platinum is leading gold, is encouraging.

Here are: FXJPEMCS (JPM’s EMFX spot index), AU, ABX, GLD and BFR (one of the Argentine banks)

Fourth, and at odds with the last point, real and nominal yield patterns have shown no sign of stopping their rise.

Below are the 5-year real yield and the 10-year futures contract:

The bottom line on rates is that despite whispers of global slowdown, the data–in the US at least–are still running on the hot side and yield patterns still look higher. This could turn quickly, on even the smallest of comments from the Fed, but it is important to underscore that until it shows signs of doing so, it’s little more than wishful thinking.

Fifth, there are a lot of names that have asymmetric entry points. Stop outs are fairly close by, and if global markets do rebound, there would be a lot of room to run. Some, like Deutsche Bank, Argentine banks (BFR, BMA, GGAL), homebuilders (PHM, MTH, NVR, , are knife-catches–and in the cases of European banks and US homebuilders, they are bets that the US and global cycles haven’t ended. Others are strong secular growers–thinking about names like AMZN, Tencent, BIDU, SQ–that only give us entry points when we are least inclined to want to buy them.

It should also be pointed out that many of the charts that have bumped up into resistance of one type or another could easily pull back for a day or three and turn into a ‘W’ or an inverted Head-and-Shoulders–two common bottoming patterns–so the timing over the next couple few days could be tricky even if a meaningful bounce is upon us. But the signs, on balance, are quite constructive.

How you risk-manage (how to size, where to set stops, how much correlation across positions/risk concentration) is a function of your ‘mandate’, your personal style, your current level of ‘mental capital’, etc. But it is worth underscoring the importance of managing your risk in a way where you aren’t likely to get ‘chopped out’ because positions are too large, or the portfolio is too concentrated.

Good luck.


This is the time when we look for divergences. There are many ways to look at divergences and the approach should always be a weight-of-the-evidence kind of thing. It’s not binary. You simply want to see if there are enough indications that the odds have been tweaked in favor of a reversal.

Subscribers Only

Access all of BehavioralMacro's premium content and the private Twitter feed via Premo

Validate Twitter

Not a subscriber yet?


Powered by

Validation unsuccessful.

Please subscribe to access the premium content.

The classic divergence is when, in the re-test of a hard decline, price undercuts the previous low but other indicators don’t exceed the levels reached on the day of the first decline. If there are enough such divergences, the odds of a rebound improve significantly.

More often than not, though, the picture is somewhat muddied. And that’s the case here. A few indicators do show divergence, but one important one doesn’t.

Here they are:

First, I’m posting a chart of the SPY. I’m using SPY instead of futures so we can also get a clean read on volume.

You can see that today the October 11th low in price was undercut. But you can also see that it did so on lower volume. That’s one divergence.

Another common divergence go-to is the 52-week new low list. You want to see the new lows contract relative to the first low.

Here, you can clearly see that new lows expanded a fair amount today relative to October 11th, so no divergence.

Another common indicator is, of course, the VIX. You want to see it make a lower high.

You can see that here, the VIX did make a lower high, so we have another indicator suggesting the selling pressure today was less intense.

I also like to look at the number of declining names in the NYSE (common stocks only), which you also want to see contract.

And here we have another divergence. Today’s 2,574 is lower than last week’s 3,209.

Lastly, even though I don’t place too much emphasis on relative strength indices, the RSI on the SPY did make a higher low today, which constitutes another divergence.

Bottom line: We do have a number of divergences. The odds of a rebound are decent. It would be a lot better if new 52-week lows also contracted, but still, decent.

Argentine Banks

Catching falling knives is a dangerous game. It might be a little bit easier in an emerging market economy that has just gone through a crisis, but figuring out when there’s the right amount of blood on the streets, when others are sufficiently fearful, is always a dicey task.

But it is also, potentially, a very lucrative one. There are no guarantees, but there are a couple things in situations like this you can do to tweak the odds in your favor.

Subscribers Only

Access all of BehavioralMacro's premium content and the private Twitter feed via Premo

Validate Twitter

Not a subscriber yet?


Powered by

Validation unsuccessful.

Please subscribe to access the premium content.

First, the currency leads. EM blowups are synonymous with big, uncontrolled currency depreciations. The local monetary authorities are usually forced to raise policy rates aggressively to short-circuit capital flight. They also often avail themselves of an IMF program to bolster reserves and confidence. Eventually, these policies start to work and when they do we tend to see it first in the currency.

This appears to be happening now in the Argentine peso.

The break of 37 last week signaled that currency stabilization/appreciation is, with a high probability, on. It helps when short term market rates are just over 70 percent.

The second tweaker is pattern analysis. Ideally, one would want to see a day or two of high volume bottom selling, followed by a period of solid bounce on high volume as well. Then you would want to see a drift down from that bounce to a higher low on light volume, and then a move out off of that higher low on a pick up in volume. This would constitute a classic double bottom retest. High quality pattern.

This is now what we are seeing in the Argentine banks. The four charts posted below all more or less show it.






Of these four, the first two show the best patterns. And the first three are probably “systemically important”. The last one, $SUPV, probably has the the diciest book of business (most consumer credit, least government credit) of the four. Of course, qualitative distinctions might not matter (at least for a while) if this indeed is the beginning of a recovery.

But in investing there are many ways to be wrong. There are plenty of headwinds. One, we’re in the midst of a bout of global risk aversion. Two, EM sentiment is terrible and investors have shown little interest so far in returning. Three, the peso might not be done depreciating. It’s not clear to me that the peso has overshot fair value as dramatically as is often the case in EM crises. There were many years in which high inflation far exceeded peso depreciation. If the peso is only ‘cheap’ and not ‘ridiculously cheap’, this would limit the scope for appreciation and probably dampen the amplitude of any rebound in the banks (I doubt with 70% rates we’ll see significantly more depreciation, but this too is possible). Four, the banks are black boxes at this stage. It’s hard to know how much damage there will be. In these kinds of plays if you wait for there to be more visibility onto the banks’ books you will have missed the opportunity. Some leap of faith in these circumstances is required. But not knowing exactly what’s on the books and how it will interact with high rates and a big currency move is obviously a big risk.

To mitigate the consequences of these risks, there are two things you can do. One, set stops below the recent higher low. How far below those lows will depend on your confidence in the thesis and how you size the overall position, but there are many points that will give you highly asymmetric payoffs. Two, you can set a stop based on USDARS staying below a level. 37 would probably be too aggressive. 38 would probably be more reasonable. If USDARS doesn’t stay down–and ideally continue lower–this investment thesis is much less likely to play out.

Good luck.

Currency swap basis vs Currency swap rate

I’ve run into some confusion lately on FinTwit with respect to ‘currency swap basis’ and ‘currency swap rate’. The discussion was about changes in currency hedging costs for euro and yen-denominated investors buying US fixed income assets. Admittedly, it’s easy for people who aren’t deep global fixed income practitioners to confuse them. I thought this simple description might help.

A currency swap is supposed to reflect the interest rate differentials between two countries. But it never works out to exactly that. There’s usually a small difference. And that difference is called the ‘basis’. ‘Deviation from theoretical’ is how I think about it. So, the basis is not the cost of the hedge, but it is one component of the overall swap rate, which is the cost of the hedge.

The original market point I was making is that the costs of hedging out the currency exposure in dollar-denominated assets has increased more than the yields on those assets have, and this has been an underappreciated force that has been weighing on the US bond market.

Shortest Market Take Ever

I thought this email captured my views pretty succinctly, so I decided I’d share it. Please ask questions if you have them in the @BehavioralMacro stream and I will answer them as best I can during jury breaks. Good luck.

“Hey X. Sorry again about the slow response. Still terrible with email and on top of that have been (and still am) on jury duty. I posted this on my blog back in April.

I still think, amid all the tape bombs and geopolitical noise, that this is the operating environment until roughly end of summer. If we get into a full blown trade war, it will of course get uglier, but whether we do is still too path dependent for us to know. I have been in ‘keeping P&L mode’ more than ‘making P&L mode’ since I posted that piece. I had a ridiculously good Q1 in my Trading style and don’t want to mess it up. In my Investment style I have been waiting for us to finish the consolidation pattern to deploy more cash. Not too worried about the Fed, about overheating, or about a recession for the foreseeable future. I think EM continues to underperform until the market begins to think it has some clarity on where the Fed’s terminal rate will be and when we are likely to get there. At least this is my working hypothesis. Hope all is well out there in the Hamptons.




Currency View. Right here. Right now.

Surfacing from pup training to share two words on currencies. I was prompted by a DM this morning from a sub who follows EM ccys very closely, and by the price action this morning/late last week, which to me has evolved in favor of seeing some USD softness soon. (Just look at today’s intra-day reversal in silver.)

I have posted below his question (pretty sure my assumption that the sub is a ‘he’ is correct) and my brief answer. Please feel free to follow up with questions in the stream or in DMs.


“Hi Mark, what is your best estimate of when a short USD / Long EM position is tradable? I take that this is to some extend data dependant and we probably need to see a bit more softer US data to conclude that the top is in for rates and the USD, but given the sharp corrections we have already seen in certain EMs, I am starting to getting increasingly tempted. Do you have any thoughts in terms of the timing here? Cheers!”


Hi. So, first one has to specify if we’re talking about a trade or a position. And if you are looking to position EM ccys, you have to start with what your current weighting relative to your benchmark/mandate/risk tolerance is. I would want to be modestly overweight here from a strategic POV. So, if you are underweight EM ccy, it makes sense to add here IMO. Need to leave room in case the EM unwind goes further (easily could), but I think the levels and the dynamic warrant adding. If you are flat, you can still start here to get overweight, but, again, make sure you leave room to add. On the tactical side, makes sense IMO to take a shot here at being more aggressively short USD, but with stops in mind/in place. Yes, for a short USD position to really run the data would have to cooperate somewhat, but the positioning and psychology is at least now somewhat favorable to try short USDEM here.”

Trump WIll Come for the Fed

Monday morning, Trump tweeted the following, ostensibly about currency manipulation:

But it wasn’t about currency manipulation. It was about the Fed. He was laying down a marker for later blame while at the same time preparing the ground for stepping up political pressure when and if the interest rate hikes start to bite. It also serves as a subtle first trial balloon, to see what kind of pushback/reaction he gets—with the ambiguity in phrasing serving as plausible deniability if not well received.

This is obviously conjecture, and it’s hard to get into Trump’s head. But when it comes to tweeting, there is usually a propagandistic method to his madness.

So, let take a look at what we do know.

We know Russia is not a major economy by any metric. And what exports it does have (commodities) are not exchange rate sensitive. And Russia doesn’t have a history of competitive FX manipulation. So, it wasn’t about Russia, which was likely thrown in to create the impression of a broader issue.

We know the Chinese exchange rate had been a real issue, but not so much any more. It is one Trump repeatedly mischaracterized/misunderstood during the campaign, but the Big De-peg of 2005 and the massive ULC-adjusted CNY appreciation thru 2013 pretty much took that issue off the front burner.

USDCNY, last 15 years

It’s true that Trump often tries to put lesser or bogus issues on the table in the hopes of getting something for negotiating them away later, and the US is in some sense in trade negotiations with China, but it’s a weak play, the issue wasn’t top-of-mind, and there was no need to reference Russia or the Fed if trumping up a negotiating point with China was the objective.

And we know that the timing makes more sense with respect to the Fed. Rates have been hiked now a half-dozen times, the growth optimism felt back in January has cooled/is cooling. The prospect of a slowing—or even moderating—economy has to terrify Trump, since he has pinned his political hopes on delivering outsized growth. It’s the only play he has. He both needs the Fed to blame if growth falls short and for the Fed to “cooperate” as much as possible. Moreover, his administration announced two new Fed Board governors the day of the tweet, so this issue was top-of-mind. And he has shown zero compunction in trampling on the integrity of any person or any institution that gets in the way of what he wants, whatever the systemic implications.

If you believe, as I do, that growth will soon settle back in with a two-handle and not a three-handle, the temptation for Trump to increase political pressure on the Fed will be too great to resist—unless the GOP and his base revolt in consequence, a bet that recent history suggests is not a good one.

Markets in the Medium Term

This week was unusual for me in that I had three institutional investor friends come a-calling here in my sleepy little hamlet. Always great to see old friends who speak the same language. Lots of good discussion.

The upshot is that it forced me to distill my thoughts for the next six months or so. You will have already picked up on them in bits and pieces if you’ve been tracking @Behavioralmacro tweets closely over the past month or so. And I insert here the usual caveat that what I’m about to say could turn out to be totally and utterly wrong. But, all of us need a base case, so here’s what I boiled mine down to:

I call it the ascending triangle scenario.

The shock and pain of the January-February selloff forced everyone to take a deep breath and reassess the backdrop. We now need to wring out that extreme reading of euphoria, digest the rate hikes that we had been shrugging off, and recognize that growth is much more likely to have a run rate with a two and than a three handle. In short, market and growth expectations have to continue the process of coming off the January boil

In a sense, the shock created a mini disaster myopia dynamic. This process often plays out in an ascending triangle type pattern (if you can conceptualize one for growth expectations as well). And it resolves to the upside once the oscillations decline in magnitude, volatility dies down, and the January shock starts fading in our memories.

This is what an ascending triangle looks like:


We are entering the “bad” season for risk taking, many are licking their wounds from Q1’s volatility, and fundamental and technical consolidation referenced above needs time to play out.

The implication is while we can expect to return to the highs, it’s hard to imagine us getting anywhere meaningfully beyond them until this rough triangle plays out and risk appetite is replenished, which I’m guessing would likely be sometime this Fall.

To set up for this period, I added to discounted fixed income products to increase carry, taken the overall beta of the Investment book down, and increased in selected sectors like US home builders, which, in my view got hit too hard out of overdone fears of higher rates and higher wage inflation. I also like names that have a secular growth story–to the extent one can find them and get good entry points.

I have not changed my views on EM, but I do expect it to under-perform over much of this upcoming period. In part because its relative strength is no longer a surprise to anyone, and in part because any decent USD rally would shake out EM investors who came to the party late.

Hope this is helpful. Good luck.

The Three Books

I get asked often about what books I would recommend for getting a better handle on behavioral finance, markets, macro, pattern recognition, etc.

Back in 2012 I posted The Book List that covered a lot of ground–especially on the behavioral side. But I got asked again today, and I boiled it down to three books, one on behavior, one on central banking, and one on patterns and markets. This is where I recommend someone start.

Subscribers Only

Access all of BehavioralMacro's premium content and the private Twitter feed via Premo

Validate Twitter

Not a subscriber yet?


Powered by

Validation unsuccessful.

Please subscribe to access the premium content.

Here are the books:

On the behavioral side, Thinking Fast and Slow–help you realize the extent to which our judgements are thoroughly damaged by emotions/cognitive biases. If you haven’t been exposed to Daniel Kahneman, it will blow you away. You’ll understand how there is no way markets can be rational or efficient–not even in the aggregate.

On central banking, Angel Ubide’s Paradox of Risk really gets you into the heads of the major CBs in the crisis and past-crisis period. You don’t have to agree with all of his prescriptions for this book to help you get past cynical market myths and aphorisms to see how CBs really think about equities, inflation, moral hazard. It also has excellent color on the debates the major CBs had at critical points in time. Angel has spent enough time in both policy circles and markets to see the whole picture in way few do. (Sorry, Jim Grant.)

On patterns/markets, Stan Weinstein’s Secrets For Profiting in Bull and Bear Markets. Ignore the old and cheesy cover, it’s a great introduction to patterns and stock cycles. Yes, the world has changed since then, but basic building blocks of patterns and cyclical tendencies have not.

Happy reading!