Introduction
.
Tactically varying allocations to different assets can significantly improve returns; however, it is notoriously difficult to do well for most people. Asset rotations are tied to economic and market cycles; thus, the key to timing asset rotations is an understanding of where one is in the cycle and where it is going. This is easier said than done, since unfortunately, the cycle is driven by a myriad of variables. There is so much information hitting investors each day that it is easy to get lost in the noise. Therefore, to map the economy, I previously introduced, in Positioning the Cycle, a cycle model (figure 3). Assets behave somewhat predictably based on phases of this model.
.
Figure 1 shows returns of several types of risk-on/risk-off asset pairs: stocks versus bonds, small versus large stocks, cyclical versus defensive sectors, high yield versus government bonds, commodities versus gold, etc. Figure 2 shows returns over the prior 15 years. Consistent with how assets normally behave during phases of the model (figure 3), risk-on assets outperformed over the last year. Stocks outperformed bonds, high yield bonds bettered government bonds, cyclical, growth, and small stocks have led the way, gold lagged other commodities, and emerging markets have outperformed. However, over the last month, defensive assets have started to outperform, which is in line with the cycle regime shifting to slowing (discussed later).
.
The cycle has been running pretty hot (figure 3) and appears poised for a turn. Due to incomplete economic data at the end of 2Q18, it is still too early to be certain of whether it is moving from improving– the green line which shows the economy is above average and getting better – to slowing – the blue line where the economy is above average but worsening; however, the likelihood of the turn down is rising. The improving phase is associated with a risk-on market environment where stocks outperform bonds, etc; however, “slowing” may take on different defensive asset return characteristics (discussed later).
.
.
The cycle model (denoted Composite (Z-Score Method) in the remaining figures) is highly correlated with GDP and changes in GDP (figure 4). Thus, asset rotations are correlated with the economy since the cycle model is correlated with GDP. Turning points in economic growth and asset returns are more likely at extremes in the model, which measures the (1) consumer environment, (2) business sector, (3) credit conditions, and (4) monetary policy. Credit and policy are illustrated by the x-axis, financing, in figure 5 and consumer and business are on the y-axis, spending. The recent move to slowing has been due to a decline in financing as a result of changes in policy.
.
.
Asset Rotations
.
Asset rotations are predictable if one understands the economic and market cycle. While it is very difficult to forecast conditions one year ahead, one can still surmise the approximate point of the cycle and adjust risk-taking accordingly.
.
.
For instance, figures 6-8 show pairs of returns between stocks and government bonds (figure 6), high-yield bonds and government bonds (figure 7), and small-cap value versus large-cap growth stocks (figure 8). In each case, the perceived safe asset annual returns are subtracted from the perceived risky asset one-year returns. These relative returns are correlated – positively or negatively – with the cycle model.
.
Stocks versus Bonds
.
Stocks (Russell 3000) have outperformed bonds (Bloomberg Barclays US Aggregate Government Bonds) (figure 6) over several years as the cycle has been rising and strong. However, since the beginning of 2018, the gap has narrowed. The Russell 3000 is only up 3% for the year, which is still ahead of government bonds (-1%), but the outperformance is much less than in 2017 (21% versus 2%) (figure 2). This makes sense because as the cycle is prolonged, the risk of a slowdown and ensuing recession rises. Corporations get hammered during recessions as sales and earnings drop, while government bonds perform well as interest rates decline after rising (like now) during the recovery.
.
While figure 6 shows annual returns versus the model, figure 9 shows monthly returns. Monthly returns provide a snapshot of a point in time along the economic cycle, whereas an annual return could span a contraction and recovery.
.
.
Stocks anticipate growth about six months ahead (see figure 36 of Positioning the Cycle). As the economy moves from improving to slowing – the current environment (figure 2) – monthly returns between the Russell 3000 and government bonds tighten (figure 6) – like recently. Stocks outperform government bonds to the greatest degree when the cycle is turning – economic growth is below average but getting better (see second column (Russell 3000) and last columns (government bond returns) of figure 9). At this time, news is still negative. Many people anchor expectations on the past, herd, and are overwhelmed by daily negative news, so they miss signs of the turn and err by continuing to position portfolios defensively. Conversely, when conditions have been positive for a long time – similar to the last couple years – stock returns tend to wane. However, this occurs during good times (GDP for 2Q18 is expected to grow at the highest rate in a long time and EPS growth for the S&P 500 is 20%+), when people frequently overlook signs of deterioration and position portfolios too aggressively. To get the asset call correct, one needs to anticipate turns in the cycle, which are more likely at extreme readings of the cycle model (now).
.
.
High-Yield versus Government Bonds
.
Risk-on (high-yield) and off (government bonds) in fixed income can also be predicted based on the position of the cycle.
.
High-yield companies tend to be saddled with high-levels of debt and have difficult business and/or industry conditions. As the economy moves from declining – below average conditions and getting worse – to turning, investors in the high-yield space breathe a sigh of relief and these bonds outperform as company fundamentals improve. However, as the cycle ages, outperformance and absolute returns wane. This year, high-yield is flat (0%), barely outperforming government bonds (-1%) (figure 2). High-yield’s monthly return advantage over government bonds is only about 0.3% during the improving phase, versus 1.2% during turning, and essentially zero during the slowing phase (figure 9). While high-yield has a yield advantage – hence its name – spreads to government bonds tend to widen during the late stages of the economy and drive down bond prices (figure 11). High-yield bond spreads are quite narrow; however, spreads on investment grade bonds have recently widened modestly. Plus, rates on government bonds tend to rise as inflation picks up (figure 13) which follows the cycle model by about 22 months (figure 12). The inflation rate is rising.
.
.
Small-Cap Value versus Large-Cap Growth
.
Figures 8 and 9 show small-cap value stock (Russell 2000 Value) returns relative to large-cap growth stock (S&P 500 Growth) returns – two opposites – and how they vary based on the cycle. For a more detailed explanation of the value and growth cycle, see When Do Value and Growth Outperform? Small-cap value performs best early during the recovery and even during the declining phase. Later in the cycle, as growth slows, growth stocks tend to perform best as markets are driven by momentum and the most hyped growth stocks (e.g., FAANG stocks). Remember the following: when overall growth is expected to become scarce (i.e., slowing follows the improving phase), it is finally worth paying up for firms with high rates of growth.
.
.
Small-cap value stocks tend to be lower quality firms, and as a result, they are cheaper on P/E, P/CF, P/B, and P/S (figure 14) than large-cap growth stocks. Quality is defined based on ROE, margins, past sales and earnings growth, and expected EPS growth. Just as low-quality bonds outperform high quality bonds (government bonds) during the turning phase, small-cap value outperforms large-cap growth. Small-cap value stocks, which had the most difficulty during the slowdown, have the most to recover, and an improving economy “lifts all ships.” Furthermore, small-cap value has much more exposure to the non-growth cyclical sectors such as financials, energy, industrials, and materials (figure 14). These industries have the most to recover during an economic expansion. Finally, small-cap value firms tend to have more domestic sales (82.8%) than large-cap growth (56.7%) (figure 14), so it is not surprising that they perform better as the US economy turns since other countries may not be rising in sync.
.
However, as noted above, small-cap value’s performance also improves (versus large-cap growth) during the declining phase. Here are two possible reasons. First, small-cap value stocks are cheaper than large-cap growth stocks, so they have less room to decline if the market falters (Russell 3000 returns are lowest during the declining phase (see second column of figure 9)). Second, interest rates may fall during the declining phase (figure 13), and small-cap value has more exposure to utilities – a defensive sector – that is highly negatively correlated with interest rates. When interest rates rise, people rotate out of safe utility stocks to even safer now higher yielding government bonds, and vice versa. Real estate, which small-cap value is also significantly over-weighted, is negatively correlated with interest rates. Small cap has outperformed this year – the recent decline in the 10-year bond interest rate, a higher dollar, higher tariffs, and lower tax rates all benefit small caps more than large-cap stocks.
.
Other Assets
.
Also, over the last year, oil prices have soared (last two columns of figure 2), which is normal during the turning and improving phases (figure 9). Likewise, emerging markets tend to outperform domestic markets during risk-on phases (turning and improving). Emerging markets have been quite weak this year, which is typical during the slowing phase (figure 9). Rising US rates (emerging market debt rates take their cue from US rates) and talk of trade wars could be hurting these markets. This year’s higher oil prices and a stronger dollar (oil is priced in dollars) also squeezes emerging market consumers and business in countries that are net importers of oil and have dollar-denominated debt.
.
Sector Rotations
.
Figure 15 shows sector returns over the cycle, and equal-weighted composites of different sectors over the cycle phases. There are several cyclical composites: (1) traditional cyclicals that exclude information technology which is also growth, (2) cyclicals including technology, and (3) cyclicals only including capital intensive sectors (energy, industrials, and materials). The two defensive indices include: (1) one with consumer staples, health care, utilities, and telecommunication services, and (2) one without health care since it also has growth characteristics. The only growth composite includes health care and information technology.
.
The cycle has been positioned as improving since October 2016, with the most recent data pointing to slowing. This has significant implications for positioning portfolios.
.
Clearly, defensive stocks outperform during declining and slowing periods (figures 15 and 16). At these times, overall S&P 500 returns are lower as growth slows and perhaps becomes negative. Thus, defensive sectors’ stable, but often slow growth rate, profiles appear relatively attractive. Plus, during these periods, monetary policy tends to ease (figure 13), and lower government interest rates may lead investors to rotate into these higher yield sectors. As of June 30, the S&P 500 dividend yield was 1.9%, whereas utilities were 3.5% and telecommunication services was 5.7% (source: FactSet sector composition data of the SPDR SPY ETF).
.
.
During periods that are below and above average and getting better (turning and improving), cyclicals perform well (figure 15) as their financial results take off. Annual changes in the cycle model are positively related to one-year relative returns of cyclicals less defensives (broad definitions) (figure 16). Information technology has been one of the best sectors during these periods, which are typically associated with rising markets. Technology outperforms 70% of the time in rising markets (see figure 3 of Thinking About Technology) – much higher than all other sectors – and compares to consumer staples that only outperforms in 24% of rising markets.
.
Defensives have been hammered over the last couple years relative to cyclicals (figures 1 and 16). One of the worst performing sectors over the last couple years is consumer staples, which tends to perform poorly as the cycle model is improving (figure 18). However, it is the best performing sector for the month of June (up 4.5%) as defensives (broad definition) rose 2.6% and ahead of cyclicals (broad definition) that were down 0.3% and the S&P 500 that was up 0.5%. This rotation indicates a move to the slowing phase of the cycle. It is occurring when consumer staples’ relative P/E is at one of its lowest points since 1999 (grey line in figure 19), indicating that investors’ outlook is pessimistic. The sector has faced various headwinds as consumer preferences have changed (e.g., away from carbonated beverages), it experienced declining pricing power relative to costs (figure 20), and as it fell behind due to a generally favorable economy that benefits cyclicals.
.
.
Interestingly, the cyclical sector consumer discretionary also performs well during the slowing phase (it was up 3.6% in June). Perhaps as the economy slows, investors project decelerating inflation. Slowing inflation puts more money in consumers’ pockets to spend on discretionary items, such as retail and restaurants. The annual return of the consumer discretionary sector has the second lowest correlation (-0.09) (information technology is lowest at -0.11) of the sectors to annual changes in core inflation. Consumer confidence, which also likely impacts spending, tends to be high during the slowing phase; although, it may be declining from lofty levels.
.
Financials have historically performed well as the economy has slowed (the sector was down 1.9% in June). During this phase, these stocks may benefit from higher interest rates, positive markets (good trading volume and M&A and IPO activity), positive lending growth, and low delinquencies. However, the recent weak financial performance could be the result of the flattening yield curve – the 10-year interest rate has not kept up with increases in short-term bond rates. This hurts banks’ traditional lending businesses which borrow short and lend long. See the paper Financial Sector Valuation Up So Fundamentals Must Deliver for drivers of the financial sector.
.
Contrary to what may be conventional wisdom, during the declining phase, energy has performed better than most cyclicals, and in line with defensives. Oil prices are not necessarily positively correlated with the model (figure 21). Actually, higher oil prices could contribute to a deteriorating cycle, and vice versa. Thus, energy is somewhat counter-cyclical.
.
Asset Class and Sector Correlations
.
Data in figures 9 and 15 are monthly returns. While this is quite useful as it provides point in time data, asset returns also diverge over longer periods in the cycle. Monthly returns may be noisy, so annual returns should also be considered. Figures 22 and 23 show annual asset and sector return correlations with points of the model. The direction of the model has been up for some time (figure 3) so this has favored assets with high correlations; however, the model may be peaking, or at least it will be more difficult to improve greatly from here. If it declines significantly over the next couple years, one would want to add assets – right now – with low correlations.
.
It may be surprising that equity asset class returns are not more correlated with points in the cycle (figures 22 and 23). The highest correlation is 0.59 (Russell 3000 and Russell 200). The moderately high correlations can be explained because the cycle model does not 100% explain annual growth in GDP (the R2 is still over 50%), and GDP does not 100% explain sales, earnings, cash flow, valuation multiples (e.g., P/E) and other important variables to equity returns. Furthermore, equity markets may at least modestly lead economic conditions (the lead to earnings is about six months (see figure 36 of Positioning the Cycle)). The correlations shown are between past 12-month returns and the current point of the model. If one expects the cycle to slow, then one may expect future 12-month returns for positively correlated assets to weaken, and vice versa..
.
.
Asset Classes
.
The most defensive asset is government bonds (second column of figure 22). This makes sense. If the Federal Reserve is proactive in counter-acting movements in the cycle to keep the economy from running too hot or cold, then interest rates should move inversely to the cycle. A decline in interest rates to boost a poor economy means that prices of bonds rise; therefore, bond returns are inversely correlated with the cycle model, and vice versa.
.
Other defensive assets are the dollar and gold (figure 22). Investors historically rotated to gold in times of stress for its perceived store of value. They also rotate to the dollar. As world economies have become more integrated, perhaps the world catches a cold when the US sneezes. However, the US is considered the stronger country, so during a sick world the US is considered the most resilient nation and the dollar outperforms. The dollar is up 4.7% since the end of January; although, gold is down 4.9%.
.
.
If government bond returns are inversely correlated with the cycle and the dollar is as well, this implies that higher government interest rates are associated with a weaker dollar. This is the opposite of what most market pundits have called for over the last few years. They claim that higher US interest rates, due to a stronger economy, should attract foreign flows and drive up the dollar, not result in a weaker dollar. Albeit, the correlation between the dollar and government bonds is still a weak positive 0.08.
.
The two US equity indices that are least correlated are the Russell 2000 Value and S&P 500 Growth at 0.51 (figure 22), and the S&P 500 Growth is quite correlated with points in the cycle. This means that a safer place to hide out if you expect low returns is the Russell 2000 Value index. Remember, I showed earlier that during the declining phase small-cap value outperforms (figure 9). Keep in mind, however, that the period reviewed covers the years following the bursting of the internet bubble, so the data could be skewed in favor of the Russell 2000 Value because of that correction.
.
Emerging markets have the lowest correlation with other broad US equity indices. However, REITs also have a low correlation with other equity indices, except the Russell 3000 Value and Russell 2000 Value. REITs are a large part of the Russell 2000 Value (figure 14). Both assets also have low positive correlations with the cycle model. Thus, they are good places to invest for diversification. I say this with a couple warnings. First, correlations can change over time and in different market environments. During a crisis, correlations between asset classes tend to rise. As world markets have become more intertwined, correlations between emerging markets with the US have probably risen. Also, returns of REITs have been highly influenced by falling interest rates over the last few decades. Remember, REITs pay a high yield to maintain their tax-exempt status, so a drop in government bond rates helped this asset. A decline in rates during a recession makes REITs’ high dividend yield look more attractive, and apartment REITs could benefit if tenants delay upgrading to homes. Although, in a recession, real estate values could decline and rising vacancies in office and industrial and bankruptcies in retail could hurt the sector as well.
.
Large-Cap Equity Sectors
.
Turning to sectors (figure 23), the most positively correlated sectors to the cycle are information technology and consumer discretionary. These sectors make up a large weight of growth indices (nearly 60% of the S&P 500 Growth) (figure 14), which helps explain why the Russell 3000 Growth and S&P 500 Growth indices also have higher correlations to the cycle model (figure 22).
.
The most defensive sector is consumer staples, which as noted earlier, has been pounded over the last couple years. If an investor is worried about the economy and is seeking to be defensive, consumer staples has historically outperformed in 95% of down markets (see figure 3 of Thinking About Technology). It also has the lowest correlation with information technology (0.26) (figure 23), so it could be a safe place to hide if you are worried about FAANG stocks.
.
Health care and utilities also have low correlations with the model. Interestingly, utilities also has a low correlation with consumer staples; thus, these two defensive sectors appear to provide diversification benefits. This compares to health care and consumer staples which have the second highest correlation of any pair of sectors at 0.78, just behind materials and industrials at 0.80 (materials firms perhaps provide the raw materials for industrial companies).
.
Finally, materials and energy also have low correlations with the cycle model (figure 23). This confirms the low correlation of GSCI (Goldman Sachs Commodity Index) and WTI (West Texas Intermediate oil) to the cycle (figure 21). While most investors may shy away from commodities as the cycle turns down, this may not make sense in light of this data. As shown in figures 21 and 22, oil prices may be countercyclical (correlation with model is 0.11), and yearly changes in WTI has a 0.54 correlation with energy sector annual returns – the highest correlation of any asset discussed. WTI is also positively correlated with gold (0.36), GSCI (0.94), and materials stocks (0.31). Perhaps higher commodity costs contribute to higher inflation (see figures 33-4 of Positioning the Cycle), which causes the Fed to become more aggressive (figures 12-13), which then leads to the slow down which hurts market returns but not the commodity sectors that benefit? Commodities are the top two performing asset classes this year (figure 2). Does this imply the end of the cycle is nearer than most people expect?