Monthly Investment Commentary

October is often a bad month for equities, and this month was no exception. The S&P 500 fell 1.7%, small-caps lost 3.1%, foreign stocks dropped 3.4%, and REITs were down 2.4%. In fact, every asset class was in the red this month, with the Lehman Aggregate Bond Index losing 0.8% and PIMCO’s Developing Local Markets Fund (which can be thought of as emerging-markets cash) was down 0.5%. Commodity futures were the worst-performing asset class for the month, as the Dow Jones-AIG Commodity Index lost 6.3%. The good news is that our equity managers had an almost uniformly strong month versus their benchmarks, resulting in outperformance for our model portfolios in an otherwise tough month for financial assets. Our models are also ahead of their benchmarks year-to-date, and our long-term record remains very strong.

Attractive Valuations for Equities

As we mentioned last month, we believe equity valuations have become fairly attractive. On average, the market hasn’t done much over the past two years (as of September 30), with the S&P 500 gaining a total of just 13%, while at the same time reported earnings have increased by 72%! More often than not, buying opportunities in the past have come in the form of noticeable sell-offs, but this time around things are different: earnings have improved markedly, while prices have moved comparatively little. To us, how we got here is less important than what the end result is, and the end result is that right now valuations look very good.

Of course, things are rarely that simple, and this situation is no exception. Valuations based on current data may look good, but we believe investors are facing a higher-than-average level of risk. Depending on how these risks play out, future fundamentals may not be quite so good, and perhaps quite a bit worse. We’ve talked about most of these risks at length many times, but some of the bigger ones include:

• Our current account deficit could lead to a recession-inducing decline in the dollar.
• An oversized federal budget deficit could lead to higher interest rates, thereby putting a damper on growth.
• Housing prices, while not in bubble territory at the national level, are still pretty high and may be in dangerous territory in certain regions. Even a slowdown in prices could slow the economy quite a bit.
• At the same time that we have longer term deflationary concerns, inflation has spiked up, liquidity is still abundant and the possibility of stagflation is greater than it has been in quite some time.

We’re also probably late in the economic cycle, meaning that earnings and growth are more likely to slow than accelerate, and at some point could decline. Telling signs include consumer confidence, which has spiked down and is at the lowest level since the early 1990s. Of particular concern, profit margins—by virtually any measure—are very high, and as we’ve mentioned before, there is some evidence suggesting that declines in margins from peak levels often result in a big drop in earnings. This is notable given the fact that earnings are far above their long-term trend line. When this happens earnings tend to revert back to trend and often overshoot.

Like all other investors, we do not have a crystal ball. So a key part of our job is to figure out what level of risk is priced into the market and make a judgment as to whether we believe that is rational. What we can say is that the market currently looks cheap to us, at least on the surface, and this suggests that the market is discounting a higher-than-average level of risk. The tougher question right now is whether the market is over-discounting risk, presenting us with a good buying opportunity.

At least a couple of well-known strategists believe that the market is actually overvalued: P/E multiples—while considerably lower than they were during the market bubble in early 2000—are still well above their historical average, and indeed are still at or above the peaks reached in prior periods. The degree to which this is true varies from index to index, but we generally concur that in absolute terms, P/E ratios are higher than their long-term average.

Our valuation analysis, though, is based on the belief that interest rates matter to valuations. Interest rates are a key variable used by many financial analysts in valuing assets. Interest rates (in combination with a risk premium), are used to determine a discount rate that is in turn used to calculate the present value of an asset or future cash flow. Mathematically speaking, a lower interest rate means an asset is discounted less than it would be with a higher interest rate. In other words, the asset is worth more in present terms if interest rates are lower, which in a valuation context means a higher P/E ratio. So while P/E ratios are indeed high relative to history, they look much more reasonable when interest rates are taken into consideration.

There is another, even simpler reason why interest rates should affect stock multiples. Lower interest rates mean lower yields on bonds, effectively making them less appealing relative to other assets such as stocks. It stands to reason, then, that demand for competing investments will be higher, which will push their prices higher. Here’s a simplified example: If a bond is yielding 6%, one would hope to get a comparable yield on a competing asset like stocks. So if the earnings yield on stocks is 6% (the earnings yield being the inverse of the P/E), the choice might seem like a wash. However, if bond yields went down to 5%, investors would be less attracted to bonds, and stock prices would change to reflect the fact that they are now relatively more appealing. If stock prices in our simplified example rose such that the 6% earnings yield became a 5% earnings yield (keeping stocks on par with bonds) the P/E would go from 16.7 to 20. Interest rates go down, P/Es go up. This is not an immutable law, and the two do not move in lockstep all the time, but for the two reasons mentioned above, we think this is a logical assumption. The accompanying chart goes back to the late 1970s, but the data going back even further—prior to 1960—tells the same story.

Our Approach to Valuations

There are two angles we take when assessing valuations. The first is based on known data. In other words, we take the world as it is right now—earnings, interest rates, risk premiums, etc.—and crunch the numbers. As we mentioned above, on this basis the market looks cheap: our valuation model shows the market as 17% undervalued as of this writing (with the S&P 500 at 1214), and this is probably conservative for two reasons. Our model uses normalized earnings (the average of four years’ trailing reported earnings and one-year forward estimates), which still includes a few quarters from the massive earnings depression from late 2000 through 2002. It also stipulates a minimum value of 5% for the 10-year Treasury, which has the same effect as an increased risk premium when the 10-year is yielding less than that. The ubiquitous Fed Model shows the S&P 500 at nearly 30% undervalued, even when using a 5% 10-year Treasury. The Fed Model is probably an overly simplistic way of looking at things, but even if it’s “off” by a considerable amount, it would still be telling a similar story to our own model: either the market is very undervalued or it’s discounting a big meltdown. As such, it’s possible that this known-data method is actually understating the degree of undervaluation.

That brings us to our second way of looking at valuations: scenario analysis. What would have to happen in the future in order for the S&P to be fairly valued at its current price? There is a nearly infinite number of permutations to this analysis, but here is one example: Using a 5% Treasury yield, normalized earnings would need to decline by roughly 17% from current levels for the market to be fairly valued, and “current levels” still include a meaningful part of the worst earnings decline since the Great Depression. By this measure, the market is pricing in very bearish future earnings. We’ve looked at several points throughout history to help us understand what a real worst-case scenario could look like, and a truly worst-case scenario is not reflected in current prices (although this almost never happens). But in all likelihood things would have to get pretty bad to justify current prices, which leaves us with a nice margin of safety: If a bad scenario comes to pass, it’s at least partially priced in. So any significant declines in stock prices would begin to represent undervaluation even in a pessimistic scenario, suggesting that long-term downside is not too negative. (Of course over short time periods all bets are off—markets can temporarily overshoot and undershoot by large margins.) Importantly, we don’t know if a very negative scenario will occur in the next few years. If the fundamentals don’t significantly deteriorate, returns going forward would likely be better than average, perhaps in the low teens over the next several years.

One very tricky—and very important—piece of the valuation puzzle is evaluating what risk premium we should use. P/E multiplies are one way of observing the risk premium (holding the interest-rate influence constant, high perceived risk = low P/E, low perceived risk = high P/E). One would assume that in a bearish scenario, where earnings are declining sharply, that the risk premium would increase, but this is where the conventional wisdom and the data don’t agree. In every earnings peak-to-trough cycle since 1950, multiples (as measured by the P/E ratio) have actually gone up, not down, during the earnings decline (though it wasn’t always a straight-line move). We’ve heard from stockpickers that this kind of thing happens with cyclical stocks: investors know that an earnings peak isn’t going to last forever, so stock prices don’t go up as much as earnings, therefore the P/E ratio actually decreases as the fundamentals improve. On the flip side, when earnings are at a cyclical low, investors likewise assume that this is part of the business’s natural cycle, prices don’t decline as much as earnings, and presto, P/E multiples actually expand. It turns out that this same phenomenon is remarkably consistent at the overall market level. So even if earnings go down, multiples are likely to stay put, and indeed could easily even go up from the beginning of the peak to the trough; for example, P/Es went considerably higher during the nadir of the 2002 bear market than they did just prior to the peak of the bull market in early 2000. This knowledge helps us look at a wide range of earnings and P/E scenarios, estimate the “fair” value of the market in each scenario, and compare that to today’s price. The result gives us an expected return for the market (not including dividends) in that scenario. Could the risk premium go up in the bad earnings scenario? It’s certainly possible, but that would be in conflict with our observation that P/Es actually go up, not down, during earnings declines.

Where Do We Go From Here

Fat pitches are rarely a slam dunk. There is always some amount of uncertainty. When current fundamentals look bad, for example, it’s easy to overweight them and assume the worst. In the past, market declines have been precipitated either by stretched valuations, a specific catalyst or event, or a combination of the above. In the worst bear markets, where it took several years or more to recover from a big market decline, poor valuations were almost always present before the sell-off began. Right now, however, in our opinion valuations are certainly not stretched, and we are comfortable saying that they are discounting an above-average level of risk. While this doesn’t guarantee we won’t have a bear market, it strongly suggests that a big decline in stock prices would in all probability be followed by a strong recovery at some point down the road. And indeed, stocks may actually be undervalued.

In looking at a wide range of potential scenarios, as always, we were able to see a number of negative outcomes that are not currently priced into stocks. We don’t necessarily view those scenarios as likely, but they are not far-fetched, and this knowledge makes us reluctant to aggressively overweight equities. The cyclical considerations mentioned earlier also give us pause. We’ve spoken with several of the managers we hold in high regard for their stock-picking skills and their intellectual honesty to see if their observations from the field jibe with what we’re seeing at the big-picture level. Their answers ranged from very cautious optimism to more enthusiastic statements about valuations. Generally the views were consistent with our valuation work: overall valuations look good, but not quite at the table-pounding level.

Right now, our balanced portfolios are 3% underweighted to equities by virtue of our 5% commodity futures positions (the other 2% coming from bonds). Balancing what we think are reasonably attractive valuations against a higher-than-average level of risk, we don’t believe it makes sense to be underweighted to equities and are moving back to a full weighting. This will be funded by a small reduction in bonds and commodity futures. A neutral equity weighting may seem conservative in view of our valuation analysis, but this move will result in the models being 3% underweighted to investment-grade bonds, which should be viewed as slightly more aggressive than our neutral weightings, and we believe that is the right place to be. If the market suffers a meaningful sell-off from here we will re-evaluate and consider another move. We still like commodities and the diversification benefit is still solid. But they have had a big move since we bought them on an absolute basis and also relative to equities, so on an expected-return basis the story is not quite as compelling as it was. From a practical standpoint, some of the trades resulting from these changes may be fairly small, and we believe individual investors should use their discretion in weighing the advantages of being “perfectly” allocated versus the costs of making very small trades.

This information is also available at www.stappfinancial.com. If you do not want to receive future e-mail newsletters from Stapp Financial, link to our subscription form, or send an e-mail to bstapp@stappfinancial.com. Before acting on any advice it is recommended to seek appropriate counsel applicable to your individual circumstances.