What should we base macro on?

[mpawder395]

For my first P123 post I have a couple of macro questions for the community.

The recent V-shaped turnaround since December has convinced me once and for all that the post-2008 market has truly come to believe, rightly or wrongly, that the Fed will rescue it if/whenever intervention is needed because of deteriorating economic conditions. And it will always seek to position itself beforehand in anticipation of such intervention (i.e. bidding stocks up whenever there is bad news in anticipation of Fed monetary easing and so forth). As such, all older paradigms about using backtested economic fundamentals (unemployment, yield curve, CPI, PMI, credit spreads, etc.) to predict market behavior prior to and/or during recessions need to be thrown out. In such a scenario what the markets of 1973-4, 1980-82, 2000-2, and 2007-9 did in response to the deterioration of their economic fundamentals becomes of much less importance to us as contemporary investors, as interesting as it might be for historical purposes. This is because the policy decisions of Fed officials, whatever they happen to be, become the most important determiner of the overall trajectory of the market. And Fed officials are regularly subject to whimsical political pressures. Their decision-making is highly capricious.

So today other, different macro paradigms need to take the place of older, fundamentals-based paradigms. Because of the difficulty of predicting Fed policy pronouncements, the new paradigms need to give much greater importance to the standard market sentiment indicators – breadth, advancing/declining volume, Hindenburg, 200 day avgs, consecutive down days, etc. Fundamentals factors need to be dramatically reduced in their importance in the new paradigms, although they probably ought not to be eliminated entirely. Either, (1) their percentage in our overall measurement cocktails should be reduced, or (2) they ought not to be activated at all in the new models unless/until the sentiment indicators ‘switch on’ and flash ‘bearish,’ thus opening a trap door to the gapingly unsupportive fundamentals. There are precision costs, to be sure, in reducing our prediction arsenal from larger, more sophisticated models that include an abundance of fundamentals factors to blunter, less predictive sentiment indicators. In our brave new world our macro predictions, which were always murky enterprises, will become much murkier. But any 1-yr, 2-yr, 5-yr, etc. macro predictions based on recession indicators, valuations, share supply/demand, or similar such metrics, may now need to be thrown out – or at the very least their importance may now need to be dramatically reduced.

In conclusion, I have two questions for the P123 community.

1. First, am I right in reaching this conclusion about the need now to greatly reduce the importance of backtested economic fundamentals in our macro models?
2. Second, what are the sentiment indicators that have you found to work best? Are you aware of any useful non-standard sentiment indicators that might be especially useful? I am interested particularly in non-price-based sentiment indicators, although price-based ones of course are welcome as well.

[yuvaltaylor]

The recent V-shaped turnaround since December has convinced me once and for all that the post-2008 market has truly come to believe, rightly or wrongly, that the Fed will rescue it if/whenever intervention is needed because of deteriorating economic conditions. And it will always seek to position itself beforehand in anticipation of such intervention (i.e. bidding stocks up whenever there is bad news in anticipation of Fed monetary easing and so forth).

I don’t know if what you state is the case or not, but the key words here are “rightly or wrongly.” Nobody thinks that there won’t be another recession. It’ll happen at some point. In 2019, 2020, 2021, 2030? I believe it’ll start this year, but I may be very wrong. When that recession actually happens, the belief system you posit will go up in smoke. The Fed is hardly all-powerful, and because interest rates are pretty low, government debt is extremely high, and Republicans are not very good at stimulus programs, any belief that the government will reverse the recession quickly will even more quickly evaporate, and the markets will tumble, as they always do at the outset of a recession.

As such, all older paradigms about using backtested economic fundamentals (unemployment, yield curve, CPI, PMI, credit spreads, etc.) to predict market behavior prior to and/or during recessions need to be thrown out.

Well, they certainly haven’t worked up to now. Nobody has been able to predict market behavior consistently and successfully, period. I reviewed the historical track record in this article: Market Timing, Tactical Asset Allocation, And Trading: A Dialogue | Seeking Alpha (with ads) and https://backland.typepad.com/investigations/2019/01/market-timing-tactical-asset-allocation-and-trading-a-dialogue.html (without ads). But I don’t think technical indicators have been any more successful.

In such a scenario what the markets of 1973-4, 1980-82, 2000-2, and 2007-9 did in response to the deterioration of their economic fundamentals becomes of much less importance to us as contemporary investors, as interesting as it might be for historical purposes. This is because the policy decisions of Fed officials, whatever they happen to be, become the most important determiner of the overall trajectory of the market. And Fed officials are regularly subject to whimsical political pressures. Their decision-making is highly capricious.

So today other, different macro paradigms need to take the place of older, fundamentals-based paradigms. Because of the difficulty of predicting Fed policy pronouncements, the new paradigms need to give much greater importance to the standard market sentiment indicators – breadth, advancing/declining volume, Hindenburg, 200 day avgs, consecutive down days, etc.

Well, since economic fundamentals don’t apply, maybe we should look to astrology instead. The phases of the moon and the courses of the planets should guide us in our investing behavior. Or maybe we should consult the I Ching, but using yarrow sticks rather than coins. Tarot cards should also be considered.

OK, I’m overstating the case. People use technical analysis because it supposedly reflects investor sentiment, which supposedly doesn’t turn on a dime. But why would measures of recent investor behavior/sentiment prove any more predictive than lagging indicators of economic factors? They haven’t been predictive so far.

Fundamentals factors need to be dramatically reduced in their importance in the new paradigms, although they probably ought not to be eliminated entirely. Either, (1) their percentage in our overall measurement cocktails should be reduced, or (2) they ought not to be activated at all in the new models unless/until the sentiment indicators ‘switch on’ and flash ‘bearish,’ thus opening a trap door to the gapingly unsupportive fundamentals. There are precision costs, to be sure, in reducing our prediction arsenal from larger, more sophisticated models that include an abundance of fundamentals factors to blunter, less predictive sentiment indicators. In our brave new world our macro predictions, which were always murky enterprises, will become much murkier. But any 1-yr, 2-yr, 5-yr, etc. macro predictions based on recession indicators, valuations, share supply/demand, or similar such metrics, may now need to be thrown out – or at the very least their importance may now need to be dramatically reduced.

Frankly, I don’t see how they could get any murkier than they are right now.

In conclusion, I have two questions for the P123 community.

1. First, am I right in reaching this conclusion about the need now to greatly reduce the importance of backtested economic fundamentals in our macro models?
2. Second, what are the sentiment indicators that have you found to work best? Are you aware of any useful non-standard sentiment indicators that might be especially useful? I am interested particularly in non-price-based sentiment indicators, although price-based ones of course are welcome as well.

For extremely oversold conditions, try the Chaikin MONTHLY money flow index set to 15 months. When that’s well below 0, buy like hell. But make sure it’s the monthly indicator, not the daily or weekly.

The best way to deal with changing market conditions, though, is to STAY THE COURSE. Create a low-beta low-volatility stock-picking system that consistently outperforms and stick with it through thick and thin.

[SpacemanJones]

1. First, am I right in reaching this conclusion about the need now to greatly reduce the importance of backtested economic fundamentals in our macro models?

I’m new to macro, so take this with huge grain of salt, but I find value in the macro fundamentals.

I did not think the US fundamental data was looking recessionary during the Q4 drop. Data seemed more like a 2015/2016-grade deterioration with a few cautionary blips, but not lighting up red like in previous recession cases. Housing data and corporate spreads were looking scary there for a while, but other series looked OK, and as December was happening I was surprised at the severity of the market decline. I was equally - perhaps moreso - surprised by the market rise since then though; erasing the December collapse. I think the fundamental data is making the case, for the US at least, that trends weren’t as bad as initially feared. Now, things like unemployment uptick and retail sales trend are firing cautions, but present housing and corporate spread trends are less worrisome than they initially looked a month or two back.

So I guess I’m not throwing out fundamentals as it still seems to provide good clues. As you mention, FED action is a wildcard though. Many macro bears seem upset that the FED didn’t keep tightening as they were counting on the FED to keep raising to drive us into recession. The story seemed to be the FED’s going to tighten too much because that’s what the FED always does (mostly a FED incompetence view that I don’t share). Now that the FED has changed the playbook many macro bears are upset, because the FED’s not doing what they “should” to bring about the end of the cycle. I think FED over-tightening was a key component of many macro bear theses - so it was off script when the FED went dovish pretty quickly when market puked following the 12-19 meeting, and has additionally indicated a substantial scaling back of QT. To me the change made sense within a mandate for full employment and controlled inflation, and I thought they might not even raise in the December meeting - but I was wrong on that. Inflation’s just not much of an issue presently.

In any event, I’ll still be watching the fundamentals with interest as they are reported. Many of the data series are at crucial levels and I expect the next several months of data could be telling. Delayed federal data at this point in time is problematic. For example: we still don’t have official December housing numbers. Those are critical. I’m guessing January is going to be far behind schedule for many series also.

As to sentiment measures:

• I include both the U of Mich and Conference Board indicators. I guess I consider them part of the fundamental model. It’s unclear to me if sentiment is more coincident rather than leading, but I do use.
• I’ve recently started looking at price based sentiment like sector strength models - cyclicals vs. defensive using the XLI, XLF, XLU etc family of ETF historicals. Work isn’t complete on that, but it’s something Stanley Druckenmiller said he uses, so I figured I ought to get up to speed there. The initial look at the data isn’t as clear cut as I was expecting - so things to suss out there.
• I wish I had indicators for what economists call global liquidity. I need to study on this. For example: when economists point out Chinese stimulus is behind some of the EM rally in the past couple of months, I would like to have better tracking an understanding of that. Despite deteriorating PMIs, I could see several EM markets turning positive via the price charts several months ago (EM bottomed in late October). I wonder if some global liquidity measures would help understand that turn better.

Again, I reiterate I’m only learning macro, so take all with a grain of salt. Just my $0.02.

[mpawder395]

Yuval,

1. Believe it or not, I actually hope you’re right about the Fed Put going up in smoke during the next recession. I agree that the Fed’s power is overestimated these days (…a late-Bull symptom), and as such I think the Fed only can dominate markets as long as the investors believe that it dominates markets. The explosion of that belief in a cleansing recession will once more facilitate meaningful price discovery.

2. I am in agreement with you about the worthlessness of most sentiment indicator readings. However, and I think you might agree with this because of your Chaikin reference, in my view a large number of mainstream sentiment indicators become valuable at their extremes. Breadth most of the time is meaningless, but when a large number of breadth indicators are printing extremes - and a bunch of them really, really did in September 2018 - then they become valuable. The downturn in the last months of 2018 was without a doubt predictable on the basis of August-September’s extreme sentiment indicators. The same goes for RSI, advance/decline, advancing volume/declining volume, etc. Usually worthless, but incredibly valuable whenever they are registering extremes.

3. If you do want to continue to rely on economic fundamentals, I suggest recessionalert.com. The site was down today, but Dwaine Van Vuuren is a trustworthy researcher whose WLEI has been more reliable in the past than unemployment, the yield curve and a bunch of other single indicators in predicting recession. But I’ve lost a lot of confidence post-2008 that fundamentals will matter going forward in the way they did during past recessions.

[primus]

1. First, am I right in reaching this conclusion about the need now to greatly reduce the importance of backtested economic fundamentals in our macro models?

If it as as you say, and the Fed put becomes caput, the price action from the return to fundamentals will greatly outweigh any short-term action from sentiment.

Moreover, I sense your market views are influenced by an anti-Keynesian political philosophy. Philosophy can be an important lens for truth seeking, but any lens also distorts what is beyond. It is a pernicious worldview to hope for a collapse just to vindicate one’s worldviews imho.

2. Second, what are the sentiment indicators that have you found to work best? Are you aware of any useful non-standard sentiment indicators that might be especially useful? I am interested particularly in non-price-based sentiment indicators, although price-based ones of course are welcome as well.

Yes. I am not going to give away the specifics. However, I have found that normalized and stationary time-series signals constructed from the deltas of actual versus expected market values are universally useful and unambiguous.

Best of luck.