Sentiment has risen strongly over the last six months to a high level while earnings growth has accelerated, a perfect environment for the stock market which posted strong gains (up 8.9% over the last six months). Consensus estimates call for earnings growth to continue to accelerate, but sentiment already implies this so the odds of negative surprises and declining expectations has risen. Over the last month, the market has rotated to growth and defensive sectors, which implies that investors are beginning to realize that financial markets may be ahead of fundamentals.
Figure 1: Sentiment Readings Imply Moderate Caution: Growth drives returns, and growth is expected to improve to double digit rates in 2017. High expectations are not necessarily negative for returns if fundamentals deliver; however, expectations have priced in solid and improving growth. Negative surprises and declines in expectations are more likely from these levels.
Source: Spellman.
Level of sentiment is positively correlated (R2 = 0.63) with historical 1-year returns (figures 2-3, top charts below). Furthermore, even short-term changes in sentiment are correlated with short-term returns (figures 4-5, bottom charts). The stock market is a gauge for expectations of earnings and risk, so predicting sentiment is useful for predicting returns. Sentiment is likely to decline from these levels (sentiment declined in 53 of 72 months since February 2000 when sentiment was over 60%), and falling sentiment from these levels is associated with muted (slightly positive) returns.
Slower growth also reduces returns and EPS growth accelerated only 32 of 72 times from these sentiment readings. Still, consensus projections call for EPS growth to rise over the next few quarters, which is likely due to easy comparisons from 2016. Quarterly EPS growth bottomed at -6.7% in 1Q 2016 and has been accelerating each quarter and is expected to continue to improve to 13.8% in 1Q 2018; however, comparisons begin to toughen later in 2017 as we anniversary positive earnings growth of the prior year. 2018 has the potential to deliver solid growth if Congress does not delay or fail to pass tax reform, deregulation, and infrastructure investments. Last, it is probably unusual to experience higher growth in the face of more restrictive monetary policy, but so is fiscal expansion at this point of the cycle.
Figures 2-5: Level of and changes in sentiment are positively correlated with returns
Source: Spellman, FactSet.
Modeling Returns with Sentiment
Figures 5-6 illustrate how the level of and change in sentiment drive returns. Moderate or high sentiment levels that are increasing are associated with strong future returns; however, low and falling sentiment drives returns down. Moderate expectations have more room to rise, and as they increase, multiples may increase with underlying fundamentals (e.g., earnings). Low sentiment is not necessarily a buying opportunity if expectations decline and push down, albeit probably already low, multiples.
Sentiment can range from a theoretical low of 0% to a high of 100%. Sentiment made nearly a round trip since June 2014. In June 2014, sentiment was at its highest level (80%) since February 2000 (up from 5% from six months earlier), before dipping to 29% in February 2016 as the economy paused, and rising back to 69% today as restrictive US monetary policy was put on hold in 2016 and as world economies improved. As a historical comparison, the internet bubble / stock market peaked in March 2000 after sentiment rose 6% in the prior six months to 70%, just above today’s reading. The bottom of the S&P 500 during the financial crisis was made in February 2009 with sentiment at 19% (down 7% the prior six months), and the low in sentiment was 13% in November 2008 when sentiment was down 25% over the prior half year. However, the highs and lows of the stock market are not always consistent with highs and lows in sentiment. For instance, the S&P 500 peaked at 1549 (end of month October 2007) before the financial meltdown with a sentiment reading of 45%, which was down 5% over the past six months.
Figures 6-7: Moderate to high and improving sentiment leads to high returns
Source: Spellman, FactSet, February 2000-February 2016. The points on the graph represent sentiment readings six months before the date and the change in sentiment to the date.
Characteristics of Sentiment (figure 8)
Market prices measure expectations of underlying fundamentals; however, expectations can also impact fundamentals and vice versa. Financial markets are supposed to be forward looking (i.e., the value of a stock should be the discounted future value of cash flows), yet they are often coincident with underlying fundamentals. Still, they can be predictive by influencing future fundamentals. For instance, if the economy is strong, stock markets may rise and credit spreads may narrow. This, in turn, lowers the cost of capital and makes it easier for businesses to expand. Business expansion boosts sentiment, and cycle continues again in a positive feedback loop. The efficacy performance spiral may cause bubbles, and the unraveling of bubbles may result in a negative feedback loop that causes financial markets to overshoot intrinsic value on the way down.
While the “secret sauce” of the sentiment framework is proprietary, what follows is a review of the types of variables in the index. Given the importance of and interaction of financial markets and underlying fundamentals, the sentiment – or expectations – index includes indicators from both areas. 40% of the index is based fundamental variables, and 60% is based on financial market factors. To compute the composite, each variable is standardized over a three-year range. Thus, the index level and future change in sentiment measure growth in expectations. Comparing current levels against a three-year range allows for better assessment of historic levels for valuation and other variables that could otherwise – without this adjustment – look cheap or expensive if they are in secular trends or are influenced by other factors that could be in a secular decline or rise (e.g., generally, P/E multiples have risen as interest rates have experienced a secular decline). Furthermore, the standardization to a three-year range converts each variable to the same scale so the index can be a weighted-average sum of its components.
Figure 8: The sentiment index consists of fundamental and financial market based variables
Source: Spellman
Fundamental variables are categorized into business and consumer indicators.
- Business conditions are tracked by corporate surveys and real underlying fundamentals. The surveys monitor the attitudes of the manufacturing and small business areas. Fixed investment spending measures corporate managers’ confidence in business expansions, a gauge of optimism about their firms’ prospects. Commercial loan and lease growth evaluates both bank health and bank managers’ views of the credit-worthiness of customers, and businesses’ willingness to increase risk by levering up.
- Multiple consumer confidence surveys are included in the sentiment index, plus the American Association of Individual Investor’s bull-bear survey. Furthermore, stock-bond flows and asset allocation to stocks-bonds assess whether consumers are acting on their moods (as quantified in the surveys).
Financial market sentiment includes variables from the major asset classes – bond markets, stock markets, and alternatives.
- Credit indicators measure monetary policy sentiment (real level of government interest rates), market implied inflation expectations, the yield curve, and credit spreads. Higher real rates, higher inflation expectations, a steeper yield curve, and lower credit spreads imply higher sentiment.
- Equity variables include valuation, earnings revisions, stock advances less declines, and price trends. Positive changes in valuation multiples imply improving buy-side sentiment, and positive revisions imply rising sell-side sentiment. Increasing advance-decline differential and high prices relative to a moving average also imply growing sentiment.
- Optimism in the alternatives space is determined by the dollar, commodities, volatility, and derivatives markets. Since the US is considered a safe-haven, a weak dollar is a sign of positive sentiment. Strong commodity prices (except for gold) imply that investors are optimistic about the economy (high sentiment), low volatility implies complacency (high sentiment), and aggressiveness in derivatives markets (put/call ratio, margin balances) indicates enthusiasm.
Sentiment Leads Changes in Real GDP and Earnings Growth
Figures 9-10: Sentiment Leads Growth
Source: Spellman, FactSet.
Sentiment leads real GDP and EPS growth by about six months (figures 9-10), which infers that people can project fundamentals, at least in the short-term. The R2 of sentiment and real GDP is 47% and 49% for sentiment and S&P 500 EPS growth. The S&P 500 also discounts future earnings, and leads GDP growth by about three months and EPS growth by around six months. Sentiment is currently high; this implies that people expect – or hope – that EPS and GDP growth will accelerate, and earnings must deliver to avoid disappointment.
Thus, there is a relationship between sentiment, returns, and growth. Higher business and consumer sentiment may lead to a pick-up in spending that shows up in GDP and earnings growth. Improving growth may be the catalyst to drive up financial assets (financial market sentiment), which lowers the cost of capital. Lower cost of capital adds modestly to earnings, and makes it easier to finance growth initiatives.
Given that sentiment predicts growth and sentiment is correlated with returns, adding earnings growth should improve the previously described predictive returns model (figures 6-7) which only included level of and changes in sentiment. The new model with growth added is illustrated in figures 11 and 12.
- Generally, accelerating growth results in modestly better returns than slowing growth.
- Level of sentiment is most important when sentiment is falling (high sentiment that is getting worse results in better returns than low sentiment that is deteriorating further).
- Change in sentiment has a substantial impact on returns. Changes in expectations are more likely than not in the same direction as changes in growth rates. Due to behavioral biases, people are poor at predicting turning points, so a changing growth rate is a surprise that drives up or down sentiment.
- The absolute highest returns occur when sentiment is moderate and rising while EPS or real GDP growth is accelerating (this was exactly the environment over the last six months). The level of sentiment at the start of the period can be considered one’s mood, if this mood is modest and improves while underlying growth rises, then the market has room to run.
Figures 11 and 12: Moderate to high and improving sentiment plus improving EPS and real GDP growth leads to high returns
Source: Spellman, FactSet, February 2000-February 2016.
Going forward, one can expect sentiment to fall from 69%, or at best hold steady given the steep rise of 22% over the past six months. Sentiment rose over the next six months in only 19 of the 72 periods when sentiment was above 60%. Returns were positive in all EPS growth and level of sentiment cases when sentiment rose. When sentiment fell, it was only positive when overall sentiment was above 60%, as it is now. Given that earnings growth comparisons from 2016 are relatively easy (they were negative in 2016) during the first six months of 2017, it seems reasonable to expect EPS growth to accelerate. An environment of high but falling sentiment with improving EPS growth has produced modest positive returns in the past. Although, if sentiment is high, like today, the market has also generated positive returns (barely) when sentiment declined as EPS growth slowed.
Changes in Sentiment Over Past Year Led by Financial Markets and Business
The rise in sentiment over the past year has been driven by both fundamental and financial market variables; however, it was led by equity markets, credit markets, and business developments (figures 13-16). All measures for equities are up and the reading is 91%, the second highest level over at least the last 17 years (the highest level of 93% was in March 2000 at the peak of the internet/stock market bubble and in June 2014). Every credit variable has also risen. Business surveys have risen substantially, but the lending indicator is down. Perhaps lending will pick up as recent surveys of credit tightening have improved and higher interest rates will recuperate banks. While business investment growth appears to have bottomed, its sentiment reading is still low. However, business investment has been held back by energy, and the sector is improving as the rig count is up 66% from a year ago. Plus, Federal Reserve surveys of capital spending intentions are hooking higher.
Figure 13-16: Sentiment was driven higher over the last year by all categories, and was led by equity, credit, and business indicators
Source: Spellman, FactSet.
All consumer readings are up, led by strong gains in the AAII surveys. Confidence surveys are up substantially as well, and the stock-bond flows sentiment reading has improved. Alternatives have lagged behind as its indicators are mixed. Currency is essentially unchanged, commodity markets have had mixed changes, complacency as measured by volatility has risen, and safety ratios have fallen (sentiment has risen).
Financial market sentiment is at 71%, approaching the high of 82% in June 2014. On the other hand, fundamental sentiment is at 66%, quite a distance from its high of 89% in January 2004. Thus, it appears that financial market sentiment is at least slightly ahead of fundamental developments – financial markets are pricing in improvements. On the other hand, this also implies that sentiment readings could continue to advance, which is good for returns, if financial market indicators hold and fundamentals improve. For comparison, the current readings of 71% financial and 66% fundamental are similar to 70%/70% at the peak of the internet bubble. However, the stock market topped before the financial crisis at only 51%/38%.