Repo Made Simple

The flair up in the repo market back in September caught a lot of people off guard—including the Federal Reserve. The issue area is arcane and complex, and I’m not the best expert, but I think I understand it enough to try and lay out what happened and why.

Subscribers Only

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

Validate Twitter

Not a subscriber yet?

Subscribe

Powered by

Validation unsuccessful.

Please subscribe to access the premium content.

Since there’s a non-zero chance this issue flairs up again around year-end, it’s important to have the understanding of the underlying issues—if for no other reason than to see more quickly thru the bullshit.

There are two main types of change post-GFC that led the desired amount of excess reserves to be much higher than anyone had anticipated.

One, volatility in reserves increased dramatically, for two reasons:

  1. For various reasons related to the GFC, the Fed took actions that allowed the Treasury to keep its all of its money at the Fed, whereas before the Treasury kept almost all of its money outside of the Fed Funds system, in banks. And since one dollar of Fed funds in the Treasury account reduces reserves available to others by one dollar, this, as the Treasury built up and drew down on its account, led to much greater volatility in the excess reserves available to banks.
  2. The Fed during the crisis removed the cap to and restrictions on foreign banks participating in the repo market. Oscillations in their demand for repo operations also added volatility to the amount of available reserves.

If the volatility in available reserves is higher, it becomes optimal for banks to keep a larger cushion of them.

The second type of structural change was behavioral: the preference for reserves over Treasuries (as an HQLA) shifted significantly.

  1. Reserves settle t+0, while Treasuries settle t+1. You can always meet a large unexpected demand for settlement with reserves, but not always with Treasuries.
  2. Reserves now pay interest, and do so at a rate of interest roughly equal to T-Bills, so there is no monetary opportunity cost to holding reserves instead of Treasuries.
  3. The stigma of having to go to the discount window increased sharply post-GFC. (If you get caught short of reserves and need to make same day payment, the discount window is your only option.)
  4. PTSD and the conveyed expectations of the bank supervisors leads strongly to erring on the side of caution.
  5. Capital charge on lending in repo is the same as for unsecured bank lending under bank leverage rules, which further inhibited the cautious banks from trying to take advantage when the repo rate spiked from 2 to 10 percent.

This was the reserve backdrop going into September. And then the repo market got hit with a two-sided shock.

On one side, the Treasury issued an unusually large slug of bonds, which primary dealers had to finance in the repo market. This increased the demand to borrow thru repo.

On the other side, corporations around the same date had to make large estimated tax payments, the funds for which they needed to withdrawal from the overnight market, a market which lends heavily via repos. So, we ended up with a positive demand shock and a negative supply shock. And this happened against a backdrop of cautious banks with a strong preference for reserves over treasuries.

The bottom line is there are plenty of excess reserves in the system for settlement purposes, but the banks have been extremely reluctant to part with them for the structural and behavioral reasons given above. This creates a sense of hoarding. The Fed now understands this and they are working with the banks to figure out ways to reduce the volatility in available reserves and loosen some of the de facto and de jure factors leading to reserve hoarding.

Bitcoin: FOMO, Patterns, and Where do we go from here?

Even for those not involved, it was hard to miss the dramatic reversal in bitcoin on Friday, ostensibly on news that Chinese president Xi Jinping expressed support for the development of blockchain technology. (China banned bitcoin and cryptocurrency exchanges in 2017).

Subscribers Only

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

Validate Twitter

Not a subscriber yet?

Subscribe

Powered by

Validation unsuccessful.

Please subscribe to access the premium content.

Sentiment had also gotten bearish on a fairly rapid fall from 10k to 7.5k, and bitcoin trading volume on the ten top exchanges had reportedly fallen from $4 billion per day a few months ago to under $200 million a day last week. And the chart pattern looked like bitcoin was poised to fall further, continuing the “echo bubble” dynamic.

On Friday the price jumped from 7500 to 7600 at about 6:45EST and then a few hours later spiked sharply from 7700 to 8500 in less than 10 minutes. Friday evening (again, EST) prices jumped one more time from 8500 up to 10400, before trading in a range between 8800 and 10000 since. It’s roughly 3pm EST right now and bitcoin is quoted at 9660 on Coinbase.

What do the charts now say to do, based on this price action?

If you are fundamentally bullish bitcoin, I get that the FOMO here is irresistible. I don’t know how a bitcoin maximalist could remain on the sidelines or not buy more. But that doesn’t mean that you should.

Trading based on chart patterns doesn’t work like that. We use patterns precisely because they help us protect ourselves from emotions.  After a large selloff or downtrend there is typically a lot of technical damage that has to be repaired before you can get a safe entry point for a longer-term position. One large pop doesn’t do it.

There are two specific things to pay attention to: building a base and overhead resistance

An excellent example of “building a base” after a large decline is what we saw from December 2018 until March 2019. In fact, it was this base, with the requisite pattern of progressively higher lows and a quieting down/tightening up of the price action is what turned me technically bullish in late March.

I can’t get bullish on the chart until I see some kind base built and a sense that the technical/psychological damage has been repaired—even if I were fundamentally bullish the underlying asset.

The other problem is overhead resistance. From the charts above you can see that the support in the 9500-10000 area that was built from June through September has now become overhead resistance. The old saying is support that breaks becomes resistance, and resistance that breaks becomes support. This is a good example of that.

Because no base has been built and we are almost exactly at significant overhead resistance right now, the chart pattern is not giving you a buy signal. Is this a guarantee the price of bitcoin won’t continue going straight up from here? Of course not. It just tells you that the current chart pattern doesn’t put the odds in your favor. And playing the chart patterns has worked out unusually well in this asset for as long as I have been tracking it. This admittedly gets tougher and nosier now that volumes are considerably lighter and the bitcoin market is easier to push around in the near term, but typically this is noise that should influence your position sizing but not whether bitcoin ends up being a success or not.

If, however, you are a fundamental bitcoin bull and the FOMO is absolutely killing you (this happens to all of us), and you need to be involved, just be ultraconservative in your sizing. The general rules are 1) the closer you are to a natural stop on the charts, the larger the position you can afford to risk, and 2) the lower the volatility of the asset the larger position you can afford to risk. Given the magnitude of the pop, there are no natural stops for longs nearby right now, and volatility & gap risk are likely to remain quite high.

If the chart is saying bitcoin is not a good risk/reward buy here, does that mean it is a short? The basic answer there is also no.

The pattern of stair stepping down, consolidating, then breaking again to lower lows has ended. If you are bearish bitcoin and looking for an entry point to get structurally short, the odds aren’t great here either. The only thing working in your favor is that the overhead resistance offers a degree of protection.

If you are looking to trade it, you can also short price spikes up into resistance, but here too the sizing has to be more cautious because of the pick up in volatility, and you have to set a pretty tight stop and/or babysit the position closely.

Good luck.

My Performance Numbers

My career path has not been a common one. Not many policy economists take a wrong turn into finance, and those who do, do so as analysts, strategists, or pundits. Then they stay there. Very few cross over into front-line risk taker/portfolio manager—and those who try tend not to survive. I was repeatedly told economists are rarely good investors and could never be good traders.

Subscribers Only

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

Validate Twitter

Not a subscriber yet?

Subscribe

Powered by

Validation unsuccessful.

Please subscribe to access the premium content.

I saw this as a great challenge—like learning to shoot a basketball with your off hand (which, back when I was playing basketball, wasn’t done). Plus, putting your money where your mouth also seemed like fun—a lot more fun than writing reports, making recommendations, and being a portfolio manager’s scapegoat. So, when given the chance to transition into actually managing money, I jumped at it.

Over the years I went from mutual fund sovereign analyst, to mutual fund portfolio manager, to senior risk taker in a trading-oriented hedge fund. Along the way I figured out why economists tend to be bad risk takers, and, most importantly, I discovered the primacy of risk management. I also got to know myself, my risk tolerances, natural biases, and how similar my cognitive shortcomings were to everyone else’s.

For the past six-plus years I have been running money for myself and a handful of “friends and family”. This has allowed me to vastly improve the quality of life for my family and me, and, almost equally importantly, to run money the way I want to, suiting my abilities and personality.

More specifically, I had had in my head a trading approach that I thought improved upon what I had been doing as a hedge fund manager, and I wanted to test it.

I now have six full years of running the Trading strategy. It’s levered, hyper-liquid, and macro and, on average, short term. When not in cash waiting for opportunities, it tends to be very aggressive. It relies a lot on pattern recognition, risk management and interpreting the pendular swings of market narratives.

This last January I had Ursa Fund Solutions compile a Schedule of Investment Performance from 2013-2019. I’m pleased to say that the results are, on the whole, good. Despite one disastrous year (actually, one disastrous quarter), the average annualized return was 24.3%. The cumulative return was 269%. I have attached that schedule here. I know six years is not a huge sample size, but, at a minimum, it’s an encouraging start.

There are some salient take-aways.

  1. The style is uncorrelated to the S&P. I haven’t run the numbers, but even a cursory glance at them (and me living them) suggest it’s true.
  2. It can also be uncorrelated to my own ideas. Anyone following me closely in 2018 knows how many things I got wrong—themes both big and small. Yet the fund returned 42 percent.
  3. Well over half of the returns were generated from only three assets: S&P futures, gold (precious metals), and the euro. (For those who are interested, bitcoin was the third largest contributor in 2018; and the outsized return in April 2013 was the collapse in gold and precious metals).

I didn’t have these numbers verified because I plan to launch a proper fund or raise money. I did it to put an official stamp on the process. And I wanted the stamp for three reasons: to show the subs on @behavioralmacro that the process works, and that hopefully can be learned from; to have hard numbers in my back pocket to shut up the occasional insistent naysayer who insists pattern recognition and TA can’t work, and, most of all, to document for myself that I had learned to shoot with my off hand.

Bonds, Pendular Swings, and Currencies

I look at sentiment in Fed analysis like a pendulum, one that swings from an excessively hawkish interpretation of the Fed’s stance at one end to an excessively dovish interpretation of the Fed’s stance at the other. Back and forth. Over and over again. The amplitude of each swing varies, as can its length. But the back and forth is always there, oscillating around the unobservable notion of a moving center of gravity. If you get good at gauging where we are in this process, you have a real leg up in bond and currency trading.

Subscribers Only

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

Validate Twitter

Not a subscriber yet?

Subscribe

Powered by

Validation unsuccessful.

Please subscribe to access the premium content.

It’s obviously not easy. But it can pay off well if you keep your losses small all the times you’re wrong, and really push your advantage when right.

Right now, I think the pendulum started swinging a few weeks ago away from an excessively dovish take on the Fed and too much fear of worsening slowdown. It had reached a local extreme and started swinging back.

The question at this point is how much farther should we expect the pendulum to swing back?

As a proxy, I’m using the 10yr UST futures. On the 5yr chart, you can see that if we continue the pattern of the last few years, this local lower high would lead to a lower low. My trading conclusion is I wouldn’t want to bet on the 10yr UST future falling quite that far, but I would expect it to fall further–maybe a lot–from here in the context of a pendular swing. After all, these pendular swings typically last many months, and this one only started a few weeks ago.


However, in the short term there is significant support right around these levels, as you can see in the 1yr chart of the 10yr UST futures.

It may turn out that the news is so bullish and/or the positioning so wrong-footed that we cut right thru the red area and proceed lower as the pendulum continues to swing back. Or maybe not.

Either way, the broad direction for precious metals and currencies in the near term is likely to depend on how this pendular swing in bonds plays out.

Awakening the Bear

If you’ve been following my tweets—especially on @behavioralmacro—you will have noticed a shift in my long-standing bullishness over the past week. Here’s an edited DM to a friend who is more bearish bonds than I am that kinds of lays out the bare bones of why. I welcome questions.

Remember, predictions of a bear market or recession are mostly bullshit. These are complex behavioral phenomena that defy timing. But, like a good doctor, even if you may not be able to forecast when or if a patient will have a heart attack, you can identify vulnerabilities to one, look out for early warning signs, and prepare measures in the hope that if it happens you will recognize it early and come out strong on the other end.

I’ll be honest: I’m less bullish on virtually all time horizons than I was even a couple of weeks ago. Base case still a longer cycle in the US, China muddles and the EU doesn’t implode. But odds of this have dropped and risks are more skewed to the downside.

My bullishness in the risk cycle has been underpinned by three things: 1) the old Rudiger Dornbusch saying that “you can’t commit suicide jumping out of the basement window” and 2) deep, pervasive PTSD from the GFC, and 3) various asset shortages. I think the PTSD and asset shortage parts are still broadly operative, but the US economy is no longer near the ground floor. We’re not on the top floor either, but the labor market and corporate positioning have been pumped up enough by natural cycle progression goosed by recent deficit spending that a meaningful shock from home or abroad could lead to a fall, even if not a deadly one.

Spot growth does look decent right here, right now, but we are definitely more vulnerable to negative shocks than we were before the fiscal impulse, even if the financial sector is clean and a proper recession isn’t that plausible to me. But our top end growth is clearly limited, and politics are about to get a lot more noisy. I’d be wrong if 2019 growth turned out anywhere close to 3%, and unless we pick back up closer to that, I don’t think rates are going to get away from us, and a market dislocation on every growth scare seems likely. I’ve locked in a good return already in my investment account this year. I like the idea of moving to cash and taking an option on being able to reinvest into a dislocation. I’ll take the risk that the market gets away from me to the upside. I don’t think–even with the artificial beginning-year marker–that the SPX is going to put up a 30% year. In my Trading book, I’ll be ready to be active either way, depending on more tactical opportunities.

Update On Bitcoin


I admit I haven’t been tracking bitcoin very closely for a while. The energy behind it seems dead and the long-promised institutional investors keep not materializing. On top of that, the futures that I had used to short bitcoin has been discontinued.

However, a pattern is a pattern. And bitcoin has historically been the most pattern-perfect asset I can remember seeing–both on the way up and the way down.

One of my subs asked me about the current set up, so I took a look.

Guess what? It’s bullish. It says nothing about the fundamentals. It could still be a dying asset. It could still be a bubble unraveling. But the pattern has gone from bearish to bullish. Higher-to-flat highs and higher lows is a bullish wedge (at least that’s what I call it). It is the opposite of the bearish wedge we saw in bitcoin from June to October last year. Bitcoin fell 50% on that drop.

The difference between a bullish and bearish wedge couldn’t be any cleaner than in this chart. Totally textbook.

Here’s the chart:

Feel free to follow up with questions on @behavioralmacro.

Good luck.

Argentine Banks, the Sequel

Back in October I posted this about Argentine banks, arguing it was a good risk/reward point of entry on the long side, and that appreciation and/or stabilization of the currency was the likely catalyst.

Subscribers Only

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

Validate Twitter

Not a subscriber yet?

Subscribe

Powered by

Validation unsuccessful.

Please subscribe to access the premium content.

Today, the chart patterns suggest we are at a good risk/reward point to add to these positions if you don’t already have a full one, or to put a position on if you missed the last window.

Argentine banks are not for the faint of heart. The stocks are fairly illiquid in any meaningful size, and the macro backdrop there, while improved–and supported by an IMF stabilization program–is still fluid and fraught with risks. Moverover, there are general elections (for the presidency, congress, and the governorships of many provinces) coming up in October, and there will for sure be populist, anti-IMF posing posing in the run up to it.

But using chart patterns to identify good risk/reward points and manage downside risk brings with it courage, and here the asymmetries are good-to-excellent. Argentina poorly managed its monetary policy and currency and had a very hard fall. You can see this very clearly in the long-term chart of the four banks included in the October post (BMA, BFR, GGAL, SUPV).

Historically, Argentina finds a way to crawl back from its blowups. Without getting into the weeds, the pattern has been that the country acts less responsibly when times are flush, and more responsibly when its back’s up against the wall. It could be that this time things turn out differently, but that wouldn’t be the high probability bet.

Technically, the patterns, in the main, show higher highs and higher lows, which also improve the odds on the long side.

So, what about the downside?

For the one year charts below, you can see the banks have had a meaningful downdraft/pullback that hasn’t violated that overall pattern of higher highs and higher lows.

This simple approach here would be to put a stop on any buys here (or perhaps on the entire position–depending on how you need to manage your risk) somewhere at ot below these recent local lows. IMHO, given the illiquidity, stops should be done on a closing basis, and probably based on more than one close below the level chosen. Obviously, the “slower” the closing trigger, the greater the scope for exit slippage, and this needs to be reflected in position sizing.

One final point: the currency. The Argentine peso didn’t have the scope for appreciation in its starting point compared to typical macro blow ups, so don’t be too concerned if it doesn’t strengthen significantly. It does however, at a minimum, need to broadly stable. Even slow depreciation is fine. But a surge in local demand for dollars–above and beyond and seasonal and transitory factors (and seasonal factors are large in Argentina)–means they are losing control of the macro framework. And this would show up quite quickly in the bank stocks.

Good luck.

The Mexican Peso

Bonds look well bid, the narrative has been swinging from Fed hawkish to Fed dovish, the US is settling back into its potential growth rate (fiscal stimulus is rolling off, and no sign of the supply-side afterburners), and the trend in the 5yr real yield has clearly broken lower.

Against this backdrop, EM currencies should do well. The Mexican peso got out ahead by rallying hard in December, making it tough to jump on here for those who haven’t been involved. But things could be lining up for a big move, if we can get this narrative to persist and a couple/few chart patterns break our way.

Correlated assets can be great leading indicators. In the case of the Mexican peso, there are 3 Mexico-specific assets I look at–beyond, of course, the peso itself.

Here’s USDMXN itself:

This is the five year chart. Betting on a continued move after the December move is tricky, because there are no natural near-by stops above. However, it’s clear that if this structure does break down (say, thru 19.00), there’d be some serious scope for some open field running below.

I also look at the yield on the five year TIIE swap. Here it is, zoomed into the one year chart:

As you can see, and as EM hands know, it correlates strongly to the bigger impulses in the peso itself–even if the relative volatilities are very different. You can also see it looks to be breaking down already.

I also look at the five year CDS on Pemex, the state oil producer. Here that is on a one year chart:

This one has yet to break but is on the verge. It too tends to correlate strongly with general Mexico risk.

Lastly, I look at spread between the yield on the US 10 year and the 10 year Mexican TIIE swap. It too is on the verge.

I watch all four of these, and when one or two break it is usually a sign that the others will follow–as long as the broader narrative remains intact.

The way I am playing this for the Trading style is by putting on a half position now, and adding the rest once I feel “it’s on”. As always, sizing and setting stops is key. Good luck.

The Big Dollar Swoon?

Subscribers Only

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

Validate Twitter

Not a subscriber yet?

Subscribe

Powered by

Validation unsuccessful.

Please subscribe to access the premium content.

There were so many false starts last year for those of us who were looking for the big dollar to roll over that it’s natural to be a bit gun-shy here. But the charts are unambiguous: It’s happening right now, and the odds that it sticks this time are as good as they’ve been in a long while.

Obviously, buckets matter. The funding currency bucket will behave differently from the risk/EM currency bucket. And the antipodean bucket will behave a little differently from the other two. But all of them look good against the dollar right now.

In poker, there’s a term called “pot odds”. It means that the even if you’re not confident that your cards are strong enough to win, you play out the hand because the return on your marginal bet to stay in is potentially so large. The analogy here is that there are certain setups that–whether you want to or not, whether you’re feeling it or not–your discipline says you just have to go for. The patterns on the charts posted below tell us that if you’ve been looking for a short dollar run, you have to be in now. At a minimum, you certainly can’t be long dollars.

This does not mean you go crazy with size or that you abandon your stops. In fact, when you have been faked out a number of time and are gun-shy, it is best to undersize the positions until you build up more mental capital (and pray the market gives you another decent entry point after you have). But you have to be in.

The EM charts look the best. USDTRY, USDBRL, USDZAR look great. And USDTRY and USDBRL look even better on the five year charts. The BBDXY has rolled over too, and the five year real yield have finally completed its drop, weak-ass bounce, then drop again pattern. (USDMXN looks good too, but it has already had a big run and positioning is not as favorable.)

Obviously, for the EM currency bucket to work, risk appetite will almost for sure need to comply. This is less true for the funding currencies and precious metals, which have looked good for a couple of weeks now.

Roll tape:

USDBRL on the one year
USDBRL on the five year
USDTRY on the one year
USDTRY on the five year
USDZAR on the one year
BBDXY on the one year
BBDXY on the five year
US five year TIPS real yield on the five year

Good luck.

Misunderstanding Liquidity, Misunderstanding QT

Liquidity. It’s one of the most frequently used words in finance. It gets invoked to explain virtually everything and anything. But it’s often clear that those invoking it are just parroting things they learned somewhere along the way and don’t truly grasp the mechanics of it. Most don’t even make the basic distinctions among its various forms.

Here’s a rough TL;DR of what you need to know.

There are three basic types of liquidity: Systemic, Credit, and Transactional.

Systemic liquidity can be loosely thought of as the unencumbered resources in the banking system that can be used to settle intra-bank payments. Think Fed funds. And if Fed funds breaks down, payroll doesn’t get made and ATMs run dry. This is what we were on the cusp of in 2008.

But, importantly, Fed funds is a closed system. A bank can draw on its reserves to meet payments to other banks in the system, or, when necessary, get physical cash, but it can’t ‘lend them out’ to clients. Nor can it flood the equity or currency markets with them–contrary to the popular trope. They are not fungible in that way. Only the Federal Reserve can add or withdrawal from the system (with that small exception of physical cash). So, while the composition of reserves across banks can change, the aggregate level in the system cannot unless the Fed wants it to. This type of liquidity is exogenous; it’s all about the Fed.

Credit liquidity is the ability of borrowers to access credit–either to increase debt or roll over existing liabilities. Bank loans, bond issuance, trade finance, whatever. Credit availability is a function of risk appetite, not bank reserves.

It is really hard to disabuse people of the belief in the loanable funds model of credit availability we were all taught in school. This will surprise a lot of people, but the level of Fed fund reserves and credit extension are–even over the long run–uncorrelated.

Don’t believe me? Consider this:

From 1981 to 2006 total credit assets held by US financial institutions grew by $32.3 trillion (744%). How much do you think bank reserves at the Federal Reserve grew by over that same period? They fell by $6.5 billion.

Think about what that means. Bank reserves declined over the 25 year span of a generational credit boom of massive proportions. There’s no way this could happen if banks couldn’t, on their own, without regard to reserves, create money ‘out of thin air’.

Skeptical? Go over to FRED and verify these numbers for yourself.

Yes, in theory banks have capitalization ratios that at some point could constrain lending, but, as we’ve seen time and time again, banks find ways to get around regulations when their risk appetite runs hot. Moreover, cap ratios are about sufficient ‘asset coverage’; reserve levels are about sufficient ‘liability coverage’ and have nothing to do with lending.

Think about it this way: If I give my brother an IOU for $100, and he accepts it, we have created credit out of thin air. No cash needed, no reserves liquidated, no assets pledged. He can then sell it to my sister, if he so decides and she trusts my creditworthiness. She then has the claim on me, and we have just created money. If my reputation in her town is sufficiently creditworthy, she could then sell the claim to others, and so forth and so on. No one has to even think about systemic liquidity or the Federal Reserve’s balance sheet, much less be constrained by it. It all comes down to risk appetite, in this case specifically others’ perception of my creditworthiness and their perceived vulnerability should I not make good on it. This is what is called endogenous credit creation.

Transactional, or market-making liquidity is the ease with which market participants can buy and sell financial assets. This is often proxied by bid-ask spreads, volatilities, and market depth. Old hands know how pro-cyclical this type of liquidity is. It is driven by risk appetite, regulatory environment and market structure. This is where things like the Volcker Rule bite. It has virtually nothing to do with the size of the Fed’s balance sheet, either.

The bottom line: Only one of the three fundamental types of liquidity are directly in the hands of the Fed. The other two are pretty much entirely up to our risk appetite.

This is an extremely un-nuanced explainer of the basic types of liquidity and their drivers. Importantly, it abstracts completely from the psychological dimension of what people think the actual drivers are–something that genuinely matters, if only in a transitory way. And it also abstracts from the Fed’s signaling and other indirect effects, which can be significant. But, as a first cut at looking at the mechanistic links between the size of the Fed balance sheet and ‘liquidity’, this should be a good place to start. So, next time when someone comes on TV conflating different types of liquidity, you’ll know what time it is.

P.S. If you want to understand the mechanism through which banks lend, I recommend the Money and Banking chapter of L. Randall Wray’s “Why Minsky Matters“. Some of the concepts are counterintuitive, so keep reading it until it makes sense.