The Standard & Poor’s 500 has returned +18% year to date through the second quarter of 2019. This was the best first half showing for the S&P in 22 years. This performance stands in contrast to the darker mood that prevailed seven months ago when the market completed a painful 20% sell-off. We’d like to report that these returns year-to-date are the result of improvements in earnings growth expectations and the macro-economic outlook. While that would be a true feel good story, it would not be accurate. Returns this year have been driven almost entirely by higher valuations linked to shifts in sentiment around trade policy and monetary policy.

The Federal Reserve recently made a U-turn on interest rate policy from hawkish to dovish, as US growth has shown evidence of slowing and inflation softening. This shift in expectations for Fed rate cuts—vs further interest rate hikes—has been reflected in 10-yr Treasury yields plummeting from over 3% to 2%. Trade tensions between the US and both China and Mexico have also ratcheted down as proposed tariff increases have been shelved for the time being as work towards new agreements continues. Optimism around these policy shifts has led to a rise in valuations, pushing the S&P’s 2019 price-to-earnings (P/E) multiple up from 15.5x at the end of December to 18.0x currently. This increase in valuation accounts for the lion’s share of year-to-date returns in stocks, since expectations for earnings have improved only about 1% since the beginning of 2019, so the run up in stocks has largely been about investors willingness to pay more for the same level of S&P earnings.

This ‘valuation-led’ run in stocks is a good jumping off point for us to review investment return expectations and how we think about the components of total return in stocks we hold for clients. As we’ve highlighted in the past, S&P total returns have averaged just shy of 10% annually over the past 90 years, so a +18% first half return could be viewed in the context of a strong ‘year.’ As we consider the next 3-5 years, we view P/E—or the valuation investors are willing to pay for stocks as only one component of future investment returns. As an investor in stocks, your ‘total return,’ or how you make money (or lose money)—is always a function of three components:

  1. earnings growth;
  2. dividends (yield);
  3. the change in P/E ratios—what people are willing to pay for #1 and #2.

In simple math: your P/E multiplied by your EPS growth + dividends = your investment return in the year ahead. However, P/E is the least predictable component of these three because in the short term, it is the factor most impacted by changes in interest rates and swings in investor emotions which can alternate between extremes of fear and greed. For example, the 20% selloff in stocks in 4Q18 was not the result of changes in estimates of dividends or in earnings. It was fear—reflected in a quick and very sharp blow to perceptions of risks which fed into an adjustment in what investors were willing to pay for stocks. No one can forecast swings in investor fear and greed (volatility) with any accuracy. We can however more accurately forecast earnings growth over the next 4-5 years through in-depth research of individual companies. We can also increase our confidence level of hitting those EPS targets by focusing on high quality businesses that have the kind of strengths that make earnings more predictable.

Are there other conclusions to be drawn in looking backwards at these individual components of stock returns? Is the market currently cheap or expensive? On an absolute basis, the market is trading close to 18x 2019 EPS estimates vs a longer-term average P/E of 16x. On this basis, the market looks modestly overvalued. However, we also know that of these 3 components, P/E is the most influenced by interest rates. How does this 18x valuation compare in this context?

2019 06 Figure 1.png

Based on almost 60 years of data in the table above, 18x earnings, while certainly not cheap in a historical context, is also not absurdly out of whack with history in relation to where stocks normally have been valued at current levels of interest rates. The much bigger question surrounds earnings growth estimates into 2020 and whether we can have confidence in those expectations. However, what this chart also illustrates is that there is a direct relationship between the level of interest rates and what people are willing to pay for stocks. If we could obtain say 14% yields on a 10-yr US Treasury Note as you could in the early ‘80s, you can bet we’d pay far less than 18x for stocks.

If we break down S&P 500 returns into its components and look at what has occurred in the past 50 years are there any conclusions that can be drawn in terms of future returns?

2019 06 Figure 2.png

The first observation is that cash dividends are always a positive element in returns. Historically dividends have provided about a third of investors’ returns from investing in stocks (though in the past decade dividends have shrunk to 15-20% of total return as management teams allocate more capital for share buybacks). Even during the Great Financial Crisis, with few exceptions (Financials/GE we’re looking at you), companies did not cut their dividends. In fact, many continued to raise them. The second observation is that negative earnings growth is rare. Companies grow over time, or they disappear. S&P 500 EPS growth has averaged about 6.7% per year for almost a century. Earnings do not grow every year, or even every five years, but historically, the trend line is up, and the long-term rate of growth appears not to have slowed despite wars, inflation, deflation or major recessions. The third observation from this table is that P/E does matter to prospective returns over the longer run but primarily when it is at extremes. Note at the end of 1999 when the market was valued at 30x during the infamous tech-bubble? Ensuing returns for the next 5 years averaged -2.3% per year with the contraction in P/Es costing investors nearly -9% annually between 1999 and 2004! Conversely, with P/Es at only 8.8x at the end of 1984, returns averaged +20% over the next 5 years with a +8% annual benefit coming from rising P/Es.

The conclusion here is that total return over longer periods is dictated largely by earnings growth, but in attempting to predict returns over shorter periods, the starting point for valuation does matter. Under “normal” conditions, the stock market has traded roughly between 8x and 20x earnings. This band has been driven primarily by the level of interest rates. Given the S&P is now trading closer to the upper end of this band, investors should neither expect much additional P/E expansion nor fret too much about a modest contraction. If interest rates do rise materially (10-yr US Treasury yield greater than 4%), the market’s P/E would likely fall somewhat.

Another way to state our view is that the burden of total return for stocks is likely to fall more on earnings growth and dividends over the next year or two.


Alphabet (Google’s parent company), Apple, Amazon and Facebook are in the U.S. Government’s crosshairs. On June 12th, the Department of Justice (DOJ) and Federal Trade Commission (FTC) announced that they would begin antitrust probes of the mega cap technology companies. The DOJ will lead the Google and Apple investigations while the FTC will lead the Amazon and Facebook investigations.

Google is no stranger to government regulation. The FTC investigated Google in 2011 and 2012 and determined the following year that its search practices did not violate anti-trust law. The company paid a relatively small fine and agreed to change certain business practices such as allowing marketers to port their ads to other competitors. The company has also been fined by the European Commission three times, including a record €4.3 billion fine for forcing users of the Android operating system to preinstall certain Google apps (such as Google Play and Google Search) in exchange for free use of the system. Google is appealing all fines. The European Commission has the power to investigate, fine and force compliance. It’s then up to the company to appeal. In the present case, the DOJ must first bring a lawsuit to federal court.

Given that Google’s products are generally free, some antitrust experts have claimed that there cannot be consumer harm. There can’t be price gauging when there is no price. The DOJ’s Assistant Attorney General Makan Delrahim has outlined some other arguments as to why consumer harm can go beyond price. One is diminished quality in the form of less privacy. Another is exclusivity or volume discounts which prevent entry of competitors.

We are not experts in antitrust law and will not speculate on the narrow view of price only or Delrahim’s more expansive view of antitrust regulation. However, we do know that nothing is free. Google’s free services are free from a monetary transaction point of view, but not from an exchange of information point of view. Google users provide data and Google sells that data to advertisers.

Google’s website is the most explicit form of this exchange. After your search, you must view, if not click, on the ads before reaching your ultimate destination. Google also places advertisements, such as banner ads, ads within videos or small boxes within text, throughout the web based on your past searches and activity. It’s not dissimilar from a retail rewards card, which collects data on everything you buy at the store. Perhaps this data collection is more intrusive, as it’s essentially a rewards card for all your activity—or perhaps the user is getting access to a wealth of information without a monetary transaction. We would argue that this exchange of information is the price for Google’s product. The harm is debatable.

Google offering Android for free is a similar transaction. Google is exchanging an operating system for access to the phone user and the corresponding data of that user. Phone manufacturers knew this back in 2007 when Android was unveiled, but opted not to develop their own system, and instead relied on a 3rd party. It’s somewhat analogous to being given a Gillette razor, then suing Gillette for not being able to use Schick blades.

Google may not be harming consumers, but we do know it uses its dominant market share to box out competitors. Google has had a greater than 80% U.S. market share for search ads for the last decade or so (see Figure 3). The company has been accused of favoring their own content over competitors such as Yelp or TripAdvisor (take note of what comes up first when searching for a restaurant). Note: both Yelp and TripAdvisor trade at an earnings premium to Google, so investors think they are competing just fine.

2019 06 Figure 3.png

Google also uses overwhelming market share to force advertisers into using its ad exchanges. For example, to buy on YouTube an advertiser must use Google properties such as AdX. Google will also bundle its products to advertisers, making competing products non-competitive. For example, allowing advertisers free access to Analytics 360 (which analyze the ad campaign) in exchange for spending a certain amount on Display & Video 360 (an exchange for placing display, banner and video ads). Most companies look to take share from competitors and offer discounts, but perhaps Google is dominant to the point where other companies can’t compete.

We know Google’s products are not free and they do offer discounts, both actions of any rationally behaving company. We do not know what will come of the DOJ investigation. If the government determines Google violates antitrust regulation, it could lead to action like the General Data Protection Regulation (GDPR) in the European Union which mandates that European visitors be given a number of data disclosures, essentially warning that a site collects “cookies.” We don’t expect this to alter the long term secular trend of consumers spending more time and money on-line. And since this is what drives advertising dollars to the web, we don’t think it will have a material effect on Google’s ad business.

What if the DOJ decides to break-up Google? Analysts have pushed for increased disclosure by Alphabet to better value its non-search properties such as Gmail, Google Maps/Waze and Google Play. YouTube is the crown jewel of the properties with 2 billion users. Google’s “Other” segment includes two well thought of businesses: Cloud and Waymo. Google’s Cloud business is much smaller than Amazon’s or Microsoft’s but growing. Google is investing in the business, for example hiring SAP’s cloud expert and recently acquiring data-migration company Alooma. Waymo is another potentially highly valued property as the potential to become the default, or one of a few, self-driving operating systems. Waymo has recently partnered with Nissan, Renault and Lyft as the company continues to expand its self-driving capabilities. This indicates Alphabet could be worth more broken up.

Currently, Alphabet trades at a discount to its historical valuation and intrinsic value. We don’t like the government investigation, but view secular trends driving Alphabet’s growth and the non-Google properties as too valuable for Alphabet to not be part of a portfolio.


Many investors own traditional IRA or retirement accounts which received pretax contributions during their careers. These assets should not be viewed in a similar vein as taxable accounts or Roth IRAs since any distributions trigger income taxes. Therefore, the value of these accounts needs to be discounted by the marginal rate at which they will be taxed. For instance, an investor who owns a traditional IRA worth $1 million pays a combined federal and state rate of 30% should only value that IRA at $700,000. This discounting process often causes dissonance since it is difficult for an investor to admit to owning less than the face value of the account.

The IRS allows account holders to convert these traditional plans to Roth IRAs at once or gradually over time. Any assets converted to Roth status are treated as taxable income on federal and state returns.However, any assets converted to Roth status are tax-free for the duration of the account holder’s life and potentially the lives of his or her spouse and children. The conversion process can be likened to ripping a Band-Aid off a scab and enduring a flash of pain (paying higher than normal taxes) but getting the relief afterwards of being beyond the ache of taxes forever.

We recently analyzed the benefits and downsides of a Roth conversion for an investor taking Required Minimum Distributions (RMDs). We assumed a steady market return and level tax rates and contrasted the hypothetical asset values of the Roth IRA after conversion and the traditional IRA plus a phantom taxable account which received the RMDs. This exercise is very difficult to model since it is impossible to determine future investment returns, future tax rates and lifespans. A particularly tricky facet involved discounting the after-tax returns of a taxable account compared to an IRA since dividend, interest and capital gains taxes apply to the taxable account while the Roth IRA is truly tax free.

We expected that converting to a Roth IRA would be the superior strategy. However, our analysis suggested that the difference between asset values post conversion was less than anticipated. The scenario assumed a fairly long period of compounding returns for the Roth IRA to “catch-up” to the traditional IRA that remained unconverted. Regardless, there were a few factors that make converting to a Roth more attractive. These factors include:

Higher incomes/tax rates in retirement: During retirement, most people earn less than during their working years. However, a few individuals and families have higher retirement incomes due to tax deductions they were able to take while working and then large distributions from those tax deferred accounts after age 70. Tax rates themselves may also rise due to future legislation. We believe this scenario is likely since the federal government is running historically high deficits and the federal debt is at an all-time high. State and local governments have also promised high benefits relative to current in flows. Therefore, if future tax rates are higher, it makes sense to convert to a Roth now and pay a lower rate rather than wait and pay higher rates over time since the net result will produce higher investment balances.

Children: If an investor is planning to leave retirement assets to children, converting to a Roth IRA is the more attractive option. Children that receive Roth distributions also pay no taxes on distributions while children who inherit traditional IRAs pay income taxes of those same distributions. Children are required to withdraw the funds over their lifetimes at a lower rate than the original owner since they are much younger than their parents. The longer the period of compounding tax free, the more beneficial the Roth becomes relative to the traditional IRA.

Taxable Estate: Wealthy investors with taxable estates could also benefit from Roth conversions since the value of the estates will be decreased immediately by the income taxes paid after conversion. Lower estate values are a benefit since the estate tax is assessed at a top rate of 40% which is higher than the top Federal income tax rate of 37%. Certain state governments also assess estate taxes such that combined rates can exceed 50%.

Market Downturn: We do not believe that it makes sense to wait for a bear market to convert a traditional IRA to a Roth. Converting to a Roth IRA during a market downturn would allow a superior after-tax return since the income tax would be assessed on a lower valuation. Assuming the market then bounced back after conversion, the tax-free assets would be larger and escape future income taxes to benefit the owner. However, this strategy is not viable since it is impossible to time market downturns and rebounds with any accuracy.

Future legislation outside of tax rates themselves could also influence the decision to convert a traditional IRA to Roth status. For instance, the Secure Act, which recently passed the House of Representatives, limits the period over which children are permitted to withdraw IRA balances to 10 years. This change contrasts unfavorably to current law which allows lifetime withdrawals and makes Roth conversions less attractive. Whether or not to convert to a Roth IRA is very complicated and dependent on each investor’s unique financial circumstances.

2019 06 Figure 4.png