Introduction
.
Understanding the position in the cycle is paramount. That point determines returns and risks. Future risks are highest when returns are strong during the latter stages of the cycle, but investors frequently ignore them to their detriment. And the cycle is getting old. Perhaps we have one to two years to go in this expansion, but if it is longer bubble conditions will likely continue to materialize. While tax cuts may spur short-term economic growth, higher growth could be counteracted by the Fed; therefore, tax cuts may just add debt that will need to be dealt with later. Inflation may not be dead; it may just be in hibernation. If it awakens, rates will probably rise quicker than the market expects and this is more than an outlier risk. 2018 earnings growth is expected to be robust – even higher than 2017. However, my earnings expectation model suggests that 2019 will not repeat 2018’s high rate of growth, and the market anticipates earnings trends by about six months.
.
Late Stage Economy
.
Many to most people find predicting the economy to be notoriously difficult. However, too many people go about analyzing the economy in a haphazard manner. With a proper process for evaluating the myriad of variables that impact the economy, we can at least pinpoint with some accuracy where we are in the cycle. This is critically important to understand risks, returns, and for asset allocation decisions (to be discussed in a forthcoming Coach Investing piece).
.
Figure 1 shows the typical phases of a cycle. Early cycle is characterized by business retrenchment, loss of jobs, restrictions on credit, and an accommodative Fed. During mid-cycle, policy normalizes, debt ratios improve as businesses recover and begin to expand, and the large consumer economy revives. Late cycle, the inflation rate rises as both business and consumer spending continue to grow (but possibly at a slower pace) and banks remain willing to lend. However, the Fed begins to restrict growth. Today’s environment reflects late cycle conditions.
.
Making the Model
..
The economic model includes 18 components (see figure 2 and appendix). Two measure monetary policy, four are credit factors, four are business variables, and eight are based on the consumer. Policy is dictated by the real Fed Funds rate and money supply growth in relation to economic growth. A Fed Funds rate that is above inflation is restrictive. Similarly, M2 growth that is less than the GDP growth indicates that the Fed is attempting to slow the economy. Credit is based on growth and level of debt, and loan officers’ willingness to make loans. These variables tend to rise as the economy progresses from early to late stage, except increase in willingness to make consumer installment loans slows late in the cycle. The business sector’s health is measured by industrial production growth, the ISM manufacturing Purchasing Managers Index, capacity utilization, and capital expenditures growth. Utilization, cap ex, IP growth, and the ISM rise as the cycle progresses. Consumer health is based on actual spending (retail sales, savings rate, and housing), a consumer confidence survey, and income (unemployment and income growth). Consumer confidence, retail sales growth, and housing starts tend to peak prior to recessions. Unemployment and the savings rate bottom just before recessions. Income growth is somewhat volatile, but generally tends to rise in mid cycle before slowing late.
.
.
To combine variables with different scales into one index, the economic model uses Z-scores, where each component’s current value is compared to its three-year average and then standardized by its standard deviation. The standardization helps determine whether a value is a true outlier (a hot or cold reading on the economy). A three-year period is about half of an economic cycle. The median economic expansion since the 1970s was about six years and the average contraction was 12 months. A high Z-score, where the current reading is much above its average, represents an overheating, and vice versa. As long as the current reading is not too hot versus its three-year average, economic growth can persist since this implies the economy is growing within its means (e.g., ratios of variables such as debt to GDP are not getting out of whack).
.
.
The Economic Model Reads Late Cycle
.
The economic model indicates (see figures 2, 3, and 4) that the cycle is above average and near overheating (Z-score of about 0.5 over the last few months). As figure 2 shows, all business indicators are above average. Policy is split, with money supply growth indicating an overheating and the real Fed Funds rate at neutral. For credit, loan growth is still lackluster, but debt to GDP is rising and willingness to make loans is high. Consumer factors such as retail sales growth and housing starts growth have slowed to below average. Disposable personal income growth is also only average. Although, other indicators for credit and consumers are strong.
.
.
The model is a coincident indicator; however, this does not mean it is not useful. Extreme readings tend to reverse (see figures 3-5). Readings above 0.75 are uncommon, and normally occur prior to recessions. Thus, with a reading of about 0.50, it appears this expansion still has room to run, but the end is much nearer than the beginning. From these levels, a recession is normally at least a couple years off; although, this cycle is already long. If this expansion runs longer than two more years, I expect bubble conditions to continue to develop (excess debt, spending, financial markets, etc.) and the cycle to end badly. I expect that the longer and stronger the cycle the more likely a bubble develops as people forget risks.
.
The economic model normally peaks at a Z-score of about 1.0, but recessions dive deeper (Z-score of about -2.0). This suggests that downturns are short but volatile, while recoveries are long and moderate. Also, before a recession occurs, the economic index tends to decline quite abruptly. Unfortunately, this means that when the economy deteriorates into a recession everything turns down and perhaps violently. Conversely, during expansions, this also means that economic variables may not necessarily move in unison or they could be rising slowly. Figure 7 shows a typical gradual drop and rise in number of variables that have above median Z-scores (i.e., the hit rate) as the economy deteriorates and improves.
.
.
The latest trend is up from readings of below zero during 2015 and 2016 as oil prices declined and we experienced recessionary conditions in the more capital intensive and cyclical sectors of the economy. Often, when the reading falls below zero it continues down and an overall recession ensues. The other two exceptions since the early 1970s when a negative Z-score did not result in a recession were in the mid-1980s and mid-1990s. The economy experienced soft landings (moderate growth) both times after a period of higher real GDP growth.
.
.
The rise in the Z-score over the past year is primarily due to improving business conditions (all variables are up) and a more restrictive Fed (see figures 8, 10, and 11). Still, the Fed is not tight by historical standards. Money supply growth versus GDP growth is at its median level and the Fed Funds rate is still low. Figure 9 shows an equal weighted composite of financing variables (policy and credit) compared to spending variables (business and consumer). Versus two years ago, financing is essentially unchanged, and most of the rise of the economic index is due to spending (figure 9). The dots on figure 9 represent six months prior to the start of past recessions. Notice that these dates are equally dispersed between easy and difficult financing, but they are normally associated with spending problems.
.
The expansion just passed its 102nd month, whereas the average length since the 1970s was 70 months. Real GDP is up 20.3%, versus an average of 23.3% and median of 20.3% (figure 12). On the other hand, the stock market has posted its second-best returns of an expansion over this period, which has driven up net worth to GDP to record highs and boosted consumer spending (figures 13-15).
.
.
Rising Debt Has Risks
.
While net worth is rising, debt is also increasing. Quantitative easing has driven up the stock market, but it has also made borrowing much easier as there was always a willing lender. It is ironic and somewhat alarming that the financial crisis – a debt crisis – was “fixed” by making it easier for people to borrow. The result is that the system has more nonfinancial debt, even as a percent of GDP, now than before the financial crisis (figure 17).
.
Total US debt to GDP is only slightly better than before the crisis, and the distribution of the debt has shifted (figure 16). Financial institutions are much better off (figure 16 and 19) and consumers have approved (figure 20); although, deleveraging for both appears to be over. On the other hand, the government and business portion of debt have grown (figures 16 and 18). As a percent of profits, business debt looks ok (figure 18); however, profits are nearer a cyclically high than the average.
.
.
Typically, at this stage of the economy when unemployment is low, the Federal deficit to GDP is in better shape (figures 21-2). While the new lower taxes may help boost economic growth in the short-term, it will increase the annual deficit by a fifth or by about 1% of GDP to 4.8% of GDP. The deficit bottomed at 3.3% of GDP in December 2015.
.
.
This higher debt situation is happening during a time of increasing uncertainty. QE is ending; the supply of Federal debt is increasing while the lender is pulling back. The Fed is trying to normalize rates. Over the long-term, the ten-year Treasury bond typically yields a percent or two above the inflation rate (figure 23). This suggests that rates can easily rise another percent. A 1% increase in rates, which raises interest expense the US pays on its debt, could make the deficit to GDP climb another 1% to nearly 6% of GDP. This is probably unsustainable as this rate is greater than GDP growth and will leave the country more and more indebted with less ability to pay (i.e., GDP) as the debt to GDP ratio grows. In good times, we need to run lower deficits to make way for higher deficits during recessions.
.
This higher debt to GDP is sustainable for economic health as long one does not have to pay off the debt when it is due and if lenders are willing to refinance it at the original rate. Or, it works if GDP growth improves and the economy has more ability to pay off the debt when it is due. Betting on these scenarios is risky. If they do not materialize, then more debt now which increases spending and GDP in the short-term just borrows from future spending and growth. We may boost short-term growth at the expense of long-term growth.
.
Furthermore, when conditions are good, as they are today, lower deficits or even surpluses help the government build a very important policy tool to use in future recessions – expansionary fiscal policy. This tool is more readily available if the debt level is manageable before it is used since fiscal expansion involves increasing leverage. Using this tool now, by lowering taxes and increasing the debt burden when the economy appears to be chugging along fine with falling unemployment and moderate inflation, means we may not have it at our disposal later. This may raise risks considerably during the next recession. Today’s expansionary policy may also propel the economy sooner toward euphoria and excesses that normally end badly. Later, when the economic engine may need a tow or jump start from government spending, the government may also have a flat tire and a bad battery.
.
.
To combat possible excesses during the late stage of this expansion, the Fed will need to raise rates. The Fed may have to implement restrictive monetary policy further and quicker than the financial market expects due to the economic boost from expansionary fiscal policy, and this could be disruptive. However, normalization of rates, which the Fed has wanted to do for years, does have one important positive impact. It will allow the Fed to lower rates further during the next recession. Thus, new debt from lower taxes may not stimulate healthy growth if the Fed counteracts it with higher rates. It just leads to higher debt, which hurts long-term growth and may result in short-term euphoria and harmful conditions and bring the end of this economic expansion closer to today.
.
Again, this is all occurring during an uncertain time. The great QE experiment is coming to an end. The Fed is reducing its balance sheet and Europe and Japan could be next. It appears the market is ignoring this risk (see figures 24-27) given that stocks are at all-time highs, credit spreads are at lows, and the ten-year Treasury is up but still below its normal margin above inflation. However, the yield curve has flattened which generally occurs prior to the end of expansions. Optimism could unravel quickly if there is a mishap such as an acceleration in inflation or earnings problem.
.
The Inflation Model is Hooking Up
.
Inflation has been tamed throughout this economic recovery. There are many possible explanations for this:
.
- Velocity of money fell as capital was tied up as banks deleveraged
- Technology gains reduced costs
- The Internet increased competition and buyer power
- Deflation was imported from China
- Low wage growth
- Demographics (ageing population leads to lower demand)
- Low capacity utilization/output gap
- Productivity growth during early to mid-cycle is strong
- Dollar strength
- Low oil prices
- Debt constrained spending
- NAFTA and the impact on competition
.
.
About all of the factors that drove inflation down are beginning to or have reversed. Let’s look at each factor listed above. Velocity of money (lending) may pick up now that banks have deleveraged. Technology gains and the Internet will continue to impact prices, but keep in mind that we have had these impacts for the last 20 years. The current Amazon effect is strong and I believe it will continue, but it is not the first Internet driven effect to limit US inflation. The gap between the cost of producing goods in China and the US has narrowed, and we have had other low-cost producers impact the US before (e.g., Mexico, Walmart). Wage growth is stubbornly low; however, this is partly due to demographic trends. The Atlanta Fed’s wage tracker shows that while the older population’s (55+ years) wage growth is around 2%, wage growth of those between 16-24 years old is above 7%. Of course, the older generation of employees is more advanced in their careers and earn higher wages, so it dominates overall compensation as a percent of the total and its low growth is holding down the overall average (also see this paper by the San Francisco Fed). Still, it seems only a matter of time before overall wages begin to rise. Some firms have announced wage increases following the implementation of the new tax policy. Wage growth tends to correlate negatively with unemployment and positively with capacity utilization with a 12-month lag (see figures 28-9). As capacity utilization nears its cyclical highs, it is more difficult to increase productivity so firms must hire more workers to increase output (and pay current and new employees higher wages to retain and attract them since this occurs as unemployment moves to lows). Capacity utilization has risen, but it has yet to make a new cyclical high, which could explain still moderate wage growth. However, it is heading up. Also, the dollar has weakened, oil prices are up, and NAFTA may be renegotiated. Plus, as noted earlier, business and even consumer deleveraging appear to be over.
.
.
Inflation is a critical driver; as long as inflation is low, the Fed can raise interest rates gradually. A spike in the inflation rate could lead to a quickened pace of interest rates hikes (rate increases are correlated with inflation (figure 30)) and jolt the economy, especially since inflation expectations are still subdued (figure 31).
.
Several of the factors noted above are key to my inflation model (below) and they are all pointing up (figures 33-5). Core inflation tends to be quite volatile, but swings in it can be explained by real GDP growth, oil prices, and the dollar. However, the dollar and oil prices fall out of the model for core inflation. Core inflation is still driven by real GDP growth, wage growth, and unemployment (higher wages and lower unemployment put money in the hands of consumers who spend and drive up prices). Furthermore, capacity utilization is part of the model as higher utilization may drive up wages (noted above) and prices. Currently, the inflation model suggests that CPI all items should be leveling off but still rising, and core CPI is about to pick up (both inflation models are at a moderate level). If the core inflation rate increases, the Fed is more likely to raise rates sooner to slow growth and this may negatively surprise the markets (the Fed Funds futures rates are lower than the Fed’s rate hike expectations).
.
.
Earnings Model Suggests Growth Rate May Peak in 2018
.
Stock returns are driven by earnings and the P/E ratio. E * P/E = P (price). The P/E is driven by expectations of growth and the discount rate. The discount rate is determined by the risk premium and inflation and its impact on interest rates. Rates are currently low and economic risks appear to be ignored as is typical late in the economic cycle. However, risks will not be ignored if inflation rises and induces the Fed to get aggressive or if earnings growth or expected growth falters. Markets normally rise as long as growth is positive; although the odds decrease when the P/E is above normal (see figure 2 of The Stock Market: Where Do We Go From Here?).
.
.
Earnings growth is estimated by consensus to be about 11% for the S&P 500 in 2017, up from easy comps of about flat earnings in 2016. However, consensus expects 2018 growth to accelerate to 16%. The acceleration is probably due to the tax cuts, a lower dollar, and rising oil prices. The sectors with the highest expected earnings growth are energy (59%), which is benefiting from the rise in oil prices, followed by financials (28%), which will benefit from higher interest rates. Materials growth is also expected to be strong (21%). The next two highest earnings growth sectors are also cyclical (industrials at 15% and consumer discretionary at 14%). Thus, 2018 looks set to be a year driven by cyclical earnings and these sectors’ equity returns have also started the year much better than defensive areas (consumer staples, utilities, and telecommunications).
.
The S&P 500 tends to move six months ahead of earnings growth (figure 36). Thus, as long as growth remains robust through the end of the year, the market should do well the first six months of 2018. Although, it is due for a correction (as mentioned in prior work on Coach Investing), plus 2019 numbers look even more difficult for a repeat of strong growth after robust growth in 2017 and cyclically-driven growth in 2018.
.
Earlier, I discussed a macro model for GDP. Borrowing from that model and adding a few other variables, I developed an earnings expectations model that predicts earnings growth 12 months forward. The current reading (figure 37) suggest growth will remain high through 2018, but it is at a level where it normally reverses so 2019 growth is likely to slow.
.
The earnings expectations model (figure 38) includes ten variables. Two are financial market variables: the yield curve and credit spreads. Since bond investors only have limited upside (timely payment of coupon and principal payments) and 100% downside (default), they tend to focus on risks and are normally earlier than stock investors to price in earnings problems. The recent flattening yield curve implies that bond participants expect lower inflation and/or a worsening of the economy (figure 27). Or, low long-term rates could be due to foreign flows into Treasuries (US rates are relatively attractive compared to other countries). However, credit spreads remain tight which indicates an optimistic credit market (figure 26).
.
As unemployment falls, consumers also have more money to spend so this boosts earnings. Higher charge offs may lead banks to curtail lending, which slows growth. Charge offs are low, but they have started to rise (2.3% vs 1.8% two years ago). A low inventory/sales ratio implies a tight supply chain and additional manufacturing to correct it. The I/S ratio has been trending down for two years. The capital expenditure outlook survey correlates well with future cap ex, and the survey has hooked up for two years. Plus, capacity utilization is up so plant expansion is more likely. Cash repatriation due to changes in the tax law may also lead to higher cap ex. A declining dollar (the dollar is down over the last year) and higher CPI lead to higher profits (CPI all items growth is flat over the last year). Higher CPI leads to higher profits if sales and cost rise at the same pace. Profits are often a precursor to investment for growth (e.g., hiring of employees and expanding capital) (see figure 10 of The Expectations Clock: A Model for Cycles and Sentiment).
.
