by Asset Allocation Committee

One factor we have been tracking is the recent behavior of retail money market funds.  We have noted that households began building money market funds about the time that the equity market peaked and U.S. trade policy began to turn toward protectionism.  In the coming months, money market funds continued to rise even as the S&P 500 made new highs.  However, as equity markets fell in Q4, money market funds rose rapidly, with the pace of the increase rivaling what we saw in 2007-08.

This chart shows retail money market funds and the S&P 500.  The gray shaded area shows the 2007-09 recession.  The orange shaded areas show periods when money market funds fell below $920 bn.  When money market funds fell to those levels, equities tended to stall.  In recent weeks, the pace of increases in retail money market funds has slowed but still continues to rise, even with the market’s recovery.  It may be difficult for equities to move much higher without retail participation.

In light of this analysis, which we have been discussing on a regular basis, another thought emerged—what is the role of cash in a portfolio?  To examine this issue, we decided to look at cash instruments in the holdings of households.

It turns out that cash has a complicated history.

This chart shows the percentage of deposits in household financial assets.  From the early 1950s into the early 1980s, the level of deposits generally rose.  This rise was probably due to a number of factors.  During this period, there was a steady rise in household income; at the same time, the financial system was heavily regulated.  Regulations limited households’ flexibility in investing.  However, by the mid-1980s, Regulation Q, which put caps on deposit rates, was removed.  And, the financial services industry greatly expanded the products available and improved the logistics of investing.  Financial deregulation increased the access to borrowing for households; thus, the need to save for purchases diminished.  Income inequality also rose after 1979, which likely concentrated saving into fewer households.  With more liquidity, these high income households were more likely to look for other investment opportunities.  Essentially, the bull market in stocks that began in the early 1980s and ended in 2000 was supported by a more than 10% point decline in the share of deposits in household accounts.

Household deposits function as an investment, a way to smooth out spending and a flight to safety instrument.  Thus, one would expect there to be a close relationship between consumption and deposits and a rise in deposits relative to consumption during periods of crisis.  In fact, these features do exist, but there was a significant break in the relationship starting in 1990.

The chart on the left shows the time series of personal consumption and household deposits; the one on the right shows a scatterplot with regression lines.  Note that after 1990, the relationship curve shifted to the right, meaning that fewer deposits were required to support consumption.  Deposits rose sharply in 2005, even faster than consumption.  That was when the housing crisis began and there was clearly a flight to safety.

Since 2005, there has been a steady increase in deposits relative to consumption, which would suggest a generalized increase in fear relative to the period from 1995 to 2005.

This chart shows the ratio of deposits to consumption.  In 2000, deposits fell below 70% of consumption. The ratio rose after that and has currently reached 90%.  Note how fast the ratio rose after housing peaked in 2005—there was a clear flight to cash that culminated in the Great Financial Crisis.  That level of fear has persisted, shown by the high level of deposits relative to consumption.

In conclusion, as the first chart shows, we have seen a lift in money market funds.  This does suggest an elevated level of fear in financial markets.  However, the recent rise should be viewed within the context of overall cash holdings in households.   Essentially, households are holding high levels of deposits relative to consumption, suggesting rather high levels of caution already.  We still believe that retail money market funds need to decline in order to see equities rise this year; however, it also seems that households are not as complacent as they were in 2005 and therefore the likelihood of a financial crisis is probably not very high.

View the PDF

by Asset Allocation Committee

The Financial Accounts of the United States, formerly known as the Flow of Funds Report, was released last week.  It is a plethora of information about the state of the economy.  Below we discuss the charts we find most noteworthy.

First, here is the saving balance by sector.

The tax cut has led to an increase in business saving and a wider fiscal deficit.  Household and foreign saving was essentially unchanged.  We are watching to see if the administration’s trade policy reduces the trade deficit; if it does, then foreign saving, the inverse of the trade deficit, will decline and require higher saving from the other three sectors to offset that saving decline.

Second, the share of national income going to capital and labor were mostly stable.

Since the end of communism in the early 1990s, capital has been taking a steadily larger share of national income.  In each expansion, the capital share has made a higher peak.

Third, the drop in equity markets adversely affected household net worth.

Net worth relative to after-tax income took a dive in Q4; the recent recovery in equity markets should reverse this dip, but we do note that in previous cycles such declines tended to signal that the business cycle was coming to a close.

Fourth, owners’ equity in real estate finally reached 60%, a level which we consider normal.  This was the level the market was at pre-1995, when the government eased the rules on home ownership.  During the real estate crisis, equity plunged as falling home prices collided with excessive mortgage debt.  From a financial perspective, this suggests the residential real estate market has overcome the trauma of the Great Financial Crisis.[1]

Finally, household deleveraging is continuing, although the pace has slowed.

This chart shows household debt as a percentage of after-tax income.  From the early 1980s into 2005, household debt rose steadily.  The Great Financial Crisis led households to lower their debt levels relative to income.  Although there is no generally accepted level that signals when deleveraging has been achieved (we would prefer around 80%), the continued decline is both good and bad.  It is good because the reduction in debt is supportive for household balance sheets.  However, growth will tend to be slower during periods of deleveraging.

Overall, the report does show the tax bill affected business saving and government dissaving.  The recent market decline was reflected in the slide in household net worth.  Capital is still gaining on labor; the housing market is now on a more solid footing compared to a decade ago.  And, household balance sheets are improving.  Overall, we view the report as consistent with steady, slow growth.

View the PDF

---

[1] This is only true for the financial system.  Many communities are still struggling with the aftermath of the crisis.

by Asset Allocation Committee

The Federal Reserve is experiencing a crisis of sorts.  For years, policymakers have used the Phillips Curve as a guide to policy.  The Phillips Curve postulates that there is a tradeoff between inflation and unemployment. Essentially, to quell inflation policymakers need to raise rates to create unemployment.  The basic idea is that unemployment represents capacity; when the unemployment rate falls below some level, sometimes called the “natural rate of unemployment,” capacity constraints develop and inflation rises.  Theory has developed around the Phillips Curve.  For example, the Non-Accelerating Inflationary Rate of Unemployment (NAIRU) suggests an equilibrium unemployment rate; inflation falls above this rate, and inflation rises below this rate.

However, if NAIRU is natural, it doesn’t appear to be stable.

This chart shows the natural rate of unemployment, calculated by the Congressional Budget Office.  As is evident, it has increased and declined over time.

The lack of inflation given the low level of unemployment has led the FOMC to consider if another model might work better.  If the reaction of the economy to low unemployment has changed, then policymakers may need to keep rates lower for longer to avoid the risk of raising rates too soon or too much.  Doing either could needlessly shorten a business cycle.  If inflation reacts more slowly to falling capacity, it might make sense to use a different way to adjust the policy rate to inflation.  Currently, the FOMC uses a 2% target on core PCE inflation.  In other words, based on the yearly change in the core PCE index, the FOMC attempts to keep inflation around a yearly 2% change.  Although the Fed attempts to suggest 2% isn’t a ceiling, it is generally treated as one by the financial markets.

Of the current 17 members of the FOMC, nine are Ph.D. economists.  Of that group, five have advocated for a reexamination of inflation policy.  None of the non-Ph.D. members have advocated for a change.  Thus, the drive to make a change appears to be coming from the “technical” side of the shop.  At present, there are three ideas being explored to address the inflation policy issue.

1. Raise the inflation target: The official inflation target is relatively new. Although the Fed privately agreed to an inflation target in 1996, Chair Greenspan insisted it was to be kept secret.[1]   It wasn’t until 2012 that the FOMC made the 2% target official.[2]  The working definition of inflation control is a rate low and stable enough that inflation concerns are not part of the economic process.  In other words, if I am buying something, I neither accelerate the purchase (a reaction to higher expected prices) nor wait (a reaction to lower expected prices).  If the interest rate that creates equilibrium has fallen, one way to achieve that rate, on a real basis, is to simply tolerate higher inflation.
2. Use a moving average of the yearly change in inflation: The use of the core rate, the rate excluding food and energy, is designed to smooth out the changes that can be triggered by weather or geopolitical events. However, even items outside of food and energy can be affected by short-term events.  Changes in tax rates at the state and local level can raise prices for a short time.  In the past, price wars in mobile phone plans led to lower core prices.  If the goal is to control inflation over the medium term, a smoothed series might make more sense.[3]
3. Target the level of prices, not the rate of change: The FOMC uses the core PCE deflator as its policy tool. Instead of aiming for a 2% rate of change, the committee could target a level instead.  That way, if price levels were above trend for an extended period, monetary policy would be tightened until the index falls back to trend.  In the case of price levels being below trend, the Fed may move slowly to raise rates until prices return to trend.  The advantage of this model is that the FOMC would not have to react to a simple jump in prices after a period of low prices.

All three options have pluses and minuses.  Raising the inflation target is the easiest to understand.  But, there is concern that if the target appears flexible then inflation expectations could become unanchored.  The risk is that once the target is changed, the Fed could find itself facing political pressure to keep rates steady or cut them even with rising inflation to suit short-term political goals.  Moving average models will reduce the likelihood that the Fed would react quickly to inflation impulses and make the policy rate steadier.  However, by design, moving averages will tend to delay easing and tightening.  In the present circumstance, it appears attractive but it will tend to slow rate reductions going into an easing cycle.  Targeting the level of price is attractive in that it takes previous price levels into account, but it will be hard to explain to the public how it works and what to expect.  And, the calculated trend can be sensitive to initial conditions.  In other words, if you start your model in a high inflation period, it can affect the deviation from trend.

Here is one potential view of the second option.

Using the Mankiw Rule (which uses core CPI instead of core PCE), we take a five-year average of inflation on the model shown on the right.  The model on the right has better performance; during the 1990s, it did a better job of informing policymakers that rate changes were unnecessary and would have told the Bernanke Fed that policy was too tight going into the 2007-09 recession.  Interestingly enough, both versions suggest the Fed is currently too easy.

Here is another view using the employment/population ratio as a proxy for capacity.

The moving average model gave clearer signals; it would have told policymakers to ease sooner in 2001 (a recession year) and would have signaled to the Bernanke Fed to begin easing sooner than it did before the 2007-09 recession.  Both suggest the neutral rate is below the current rate but the moving average model suggests policy is tighter than the unadjusted inflation model.  Essentially, these views show that the real issue remains the measure of slack, but adjusting the inflation indicator might mitigate the uncertainty surrounding the capacity issue.

Here is one potential view of the third option.

The chart on the left shows the core PCE index, log transformed, regressed against a time trend.  In general, price levels vary to trend.  The chart on the right compares the detrended index to the fed funds target.  Note that the cycles in policy tended to move with price levels except for the current one.  This model would have kept the FOMC from tightening at all in this cycle.

Under current conditions, this reassessment of inflation policy will likely lead to an end of tightening.  Essentially, the considered adjustments will probably discourage further rate hikes by reducing the policy level of inflation.  However, this assessment still depends on the proper measure of slack in the economy, which, so far, has not been resolved by the FOMC.  Nevertheless, we think the proper assessment of this change is dovish for policy.

View the PDF

---

[1] https://economia.icaew.com/opinion/april-2015/the-federal-reserves-battle-with-price-stability

[2] https://www.federalreserve.gov/monetarypolicy/files/FOMC_LongerRunGoals.pdf

[3] The Bank of England has a sort of smoothing process where it targets a 2% rate with a 1% variance.  If inflation rises or falls outside this band, the bank must explain to the Exchequer why this occurred and what it will do about it.  Thus, if the variance is due to a one-off situation, the bank may do nothing.

by Asset Allocation Committee

Our cyclical position on foreign investing remains with a zero allocation; although the committee has not been negative on foreign, our work suggested that the risk/reward compared to small and mid-cap stocks warranted putting more assets in those areas.  However, we are continuing to pay close attention to foreign as an area that may be attractive in the future.

In the past, we have noted that relative performance between foreign developed and U.S. equities is sensitive to the dollar.  In general, during periods of dollar strength, U.S. equities tend to outperform (assuming both are denominated in dollars, which is the case for a U.S. investor).

The blue line on the chart shows the ratio of performance between EAFE and the U.S., rebased to 1970.  When this line is rising, foreign stocks are outperforming.  The red line shows the JPM Dollar Index.  Note that a rising dollar tends to favor U.S. outperformance, while dollar weakness helps foreign market performance.   Although the dollar has remained strong, both on a purchasing power parity basis and a cycle basis, the dollar is extended and should begin to depreciate later this year.

However, in this business cycle, U.S. stocks have generally outperformed even during periods of dollar weakness.  This has led us to look for other factors that might account for this discrepancy.  It appears that the growth/value variation explains at least part of this divergence.

As in the first chart, the blue line shows the relative equity performance.  The red line shows the Russell 3000 Growth/Value divergence; a rising line suggests growth outperformance.  Growth stocks have outperformed in this bull market but are showing signs they may finally be starting to give way to value.  The primary driver of growth/value is the P/E ratio.  Significant multiple expansion isn’t all that likely, although a return to the 18x+ area would not be surprising.

Why would the growth/value divergence affect foreign stocks?  The most likely reason is index construction.  U.S. indices will tend to have a greater weighting toward technology due to the size of that industry in the U.S. economy.  Foreign nations, for the most part, have less dominant tech industries.  With technology being considered a growth sector, a period of strong technology performance would tend to lead to outperformance by growth.  If foreign equity indices have less technology, they would likely underperform.  Consequently, when the growth/value balance shifts to the latter, we would anticipate foreign outperformance.  The table below provides a comparison of sector exposures.

Overall, we continue to monitor the relative performance of foreign compared to other asset classes.  If our risk/reward estimates change later this year, we could consider a return to international.

View the PDF