冰上滑行

GMO Quarterly Letter January 2004 Jeremy Grantham, Chairman Skating on Thin Ice Last Year’s Forecasts, Errors and Lessons Learned currency, outperformed the S&P handsomely: GMO This time last year my quarterly letter was titled Emerging Markets +70.2%, GMO International Small “Predictive ‘Entrails’ and The Case for a Time-out in the Companies +67.4%, and GMO International Intrinsic Great Bear Market?” This was a good heads-up for any Value +43.5% versus an S&P at +28.7%. (The two bears like us, but, to be fair, I thought the direction of the strategies on which asset allocation had the heaviest market was a fifty-fifty proposition. I was obviously overweights were also our best relative performers: surprised how well global markets did. The economy, I Emerging Markets at +13.1% and International Small suggested, would recover, but with occasional Companies at +13.7%! But that, as Will Rogers would disappointments. As it turned out, the economy was have said, is only why we bought in the first place.) We stronger than expected in everything but job creation also had the right bets within our fixed income position, where it remains weak. In hindsight, we were fortunate which was about neutral in total: we had bets against the that fixed income had done so well in the bear market that benchmark US Bond (+4%) in favor of GMO Emerging it had become overpriced, causing us to reduce our Debt (+36%), GMO International Bond (+27%), and substantial overweighting down to neutral early last year, GMO Currency Hedged International Bond (+9%). with the funds moving into larger foreign and emerging Consequently, our GMO Global Balanced Asset country stock positions. However, by staying neutral in Allocation Strategy and similar portfolios outperformed fixed income, we missed incremental returns; our internal their benchmarks by 6.5% and even held the S&P 500 for rules do not allow us to overweight equities when they the year, despite our benchmark neutral 35% fixed are overpriced whatever our short-term ‘entrails’ may income position. predict. Where the year was tough for GMO was in US equities The lesson learned, though, from 2003 is very clear: where we struggled to hold the benchmarks. Neither never, ever, underestimate the desire of an momentum nor value worked well, and our quality bias, administration to be re-elected or the substantial relative to other value approaches, particularly hurt in a cooperation that the Fed will typically provide. 2003 saw year in which low quality did exceptionally well. In the as much combined stimulus as has ever been delivered, end, we just won in our small cap growth (+0.8%) and and since all actions have consequences, the typical third small cap value (+0.3%) strategies and just lost in our year Presidential Cycle’s bullish and speculative spin was large cap value strategy (-0.1%). Our large cap growth bigger and better than normal. And, rightly or wrongly, I portfolio was down by 1.5%, and U.S. Core by just over have become more of a devotee of politics as an 2%. (It is cheating to say this, but both have reduced last intermediate-term influence on the market (see attached year’s shortfalls by over 60% in the first 2 weeks of this Special Topic). year.) All of our equity strategies, and U.S. Core in particular, have a strong bear market bias, and these slightly negative results were typical of our performance 2003 and GMO’s Performance in very strong years over the 23 years of our US quantitative investing.

In most cases, though, you are not shot for what you don’t own but what you do own, and the part of our Our fixed income portfolios had a strong year with prediction that we got very right was the ordering of International Bond up over 8% to its benchmark and performance. We bet on foreign developed, particularly Emerging Country Debt having another remarkable year, emerging equity and small cap, which, helped by finishing nearly 11 points ahead!

All in all, it was a very satisfactory year for GMO’s (particularly China’s), a deflating dollar, and, quite performance in such a strong up year for the markets, and possibly, ensuing trade wars. perhaps even more successful on the asset growth front, with our asset base growing from $23 billion to $54 The outlook for 2005 and 2006 looks about as bad as it billion, incomparably the strongest year in our history. could get as the house-cleaning phase of the political Our emphasis going forward will be on very controlled cycle (half of all first and second years are down!) meets growth. Half of our asset base is in products that are the potentially dangerous excess debt and the badly closed to new accounts or in very limited growth mode. overpriced market in which the hiding places of March We expect to limit or completely stop the growth of the 2000 – REITs, bonds, small cap, and value – have all great majority of our funds in the next 5 years if they done far too well for their own good. We will try to continue to attract assets. I write this touching wood. protect accounts as best we can, upgrading quality in Thank you for your confidence in us – we will do our individual funds, and for asset allocation, using the new very best to justify it. GMO Quality Fund for US equities and quite possibly a new GMO Real Assets Fund, for which the go-ahead decision has still not been made. Reluctantly (chewing The Outlook for 2004 glass might be a better description), we will also move As a new devotee of the Presidential Cycle, I think we back into overweighting fixed income funds, even if they can count on a high probability of a relatively stable stock are still overpriced. We will take little unnecessary risk market year. The value of the market, which on a 1-year there, and see this more simply as a necessary haven. basis never matters as much as sensible investors would But, all this still lies at least a few months out in the like, simply does not matter at all in year three, but in future. year four, it has an almost normal effect and obviously at 25 times normal profit margins, this is a very overpriced market. However, the stimulus program was profound What to Do in 2004? and the economy has responded and this momentum is an The outlook for 2004 is not bad, but the market is very obvious positive. The January Effect (which also does overpriced and all predictors look bad for next year and not matter at all in year three) reverts to giving its usual, the year after. useful signal this year. On average, value stocks (or low growth) also have their best year in year four, perhaps just The purist’s investment position is clear: the market is making up for year three’s overdoing it with growth overpriced and investors should duck! The problem with stocks. My conclusion is that the economy, profits, and this strategy, as we have all painfully learned over and the market are likely to do quite well or better in the first over again, is that an overpriced market can really run, half, and less well in the second half, as the stimulus runs and this overpriced market has quite a lot going for it in out and the overpricing of the market is felt. the near term. The pragmatic position, which I hope is not too greedy, is that the political period of moral hazard (The January Effect is the tendency of the market’s is still with us and careful and balanced risk taking is the performance in January to predict the next 11 months. A percentage shot. Our current battle plan is to leave weak January is over three times more likely to be everything untouched through January and then slowly followed by below average performance than is a strong reduce our risk in some individual funds and allocation January, for example. Even market themes have a good through June to about neutral. (Remember, we have very January record, particularly value, size, and quality; for large positions in some aggressive asset classes: example, if small stocks are weak in January, they are emerging equity and small cap international). After June, over 2½ times as likely as normal to underperform for the the plan would be to steadily batten down the hatches, rest of the year.) reaching maximum conservatism by year-end. All of which would be subject to revision based on changing The Longer-Term Outlook: ‘A Black Hole’ data and possible loss of nerve.

In the longer run, we have to worry about extended debt at all levels, and financial assets and housing that have The Risks in 2004 of Skating on Thin Ice been inflated by too much monetary ease and too much I am confident – but far from certain – that the long-term moral hazard. Unfortunately, we also have to worry problems already discussed, which will affect every about a jobless recovery, rigged exchange rates country, lie out beyond the next 6 months and probably GMO Quarterly Letter – January 2004 2 this year. If things go wrong earlier I think it is likely cheapest segment of equities. It appeals to speculators in that, in dollar terms, the US will do worse than ‘the rest’ speculative market phases, and 2003 was the most collectively, but let’s hope the thin ice holds. broadly based speculative year on record. The local emerging economies are strong, improving, and better The essence of the strength in the third and fourth year of yet, stronger than the developed world. Arjun Divecha’s the Presidential Cycle is the implicit offer by the work suggests that they are not particularly vulnerable to Administration and its ally, the Fed, to try to fix any a financial crisis. Their currencies are neutral to strong, unexpected problems that may turn up. Another potential and their recent market performance is excellent, problem for this fourth year may be that nearly all their encouraging local wealth invested abroad to return to ammunition has already been used up. The interest rates local markets. Finally, US institutions are now clamoring can come down a little, but not much. Taxes can be cut, to invest in this quite illiquid market. What’s not to like? but probably not by much. It would obviously be fiscally In a flat or better US market and without a sudden bolt ridiculous and politically cynical to cut taxes, but out of the blue, like a Chinese crisis, emerging could go administrations really, really want to be re-elected so we straight up 30%, 40%, or more early this year. can’t rule it out. So the real risk is that something goes unexpectedly wrong with the economy and that the help It is argued that emerging market equities have done too offered by the Administration and Fed is seen to be too well. Exhibit 1 shows the longer-term view. Since the little or ineffective. In this event, which I would place at index for emerging markets started in 1988, there have less than one in five or so through June, the market would been three prodigious relative moves – two in favor of probably decline. After June, the odds of a bump in the emerging relative to the S&P 500, but the emerging index road are probably higher, as the incentives from the 2002 still has to rally almost 30% to catch up since inception. and 2003 stimulus programs will be wearing off. Perhaps If Ben Inker and I are correct in our estimates of fair the odds are as high as one in three, but probably no value, and if both indices moved to fair value tomorrow, higher, for the economy has a decent head of steam and a emerging would have closed the gap and be nicely ahead, broad consensus of economists is predicting a strong finally. More to the point, emerging is still under 14 4½% real GNP growth, and once in a while they get to be times last year’s earnings, much cheaper than US right. alternatives. Being a very inefficient asset class where Arjun Divecha has added 6% a year for 10 years doesn’t There is also some chance that the sheer overpricing of hurt either! the market will cause it to fall, despite decent growth in the economy and corporate earnings, but this is probably EAFE Relative to the US the value manager in me trying to take over. The sad The gap in relative value in favor of EAFE has never truth is that although the value of the market is a decent been wider in the last 35 years for which we have decent indicator of the next year’s performance (see Special Topic), value is not as reliable an indicator as managers would like. There are plenty of exceptions where overpriced markets rise; most of these are in year three, Exhibit 1: Volatile, But Emerging Behind and some are in year four, but, for the record, few are in Since Inception of IFC Investable Index years one and two. Over 5 and 10 years, value is a great 2.5 Relative Strength of S&P 500 Relative IF C - 4 9 % predictor, but over 1 year, the Presidential Cycle is S &P 5 0 0 +14 0 % stronger and the January Effect equally strong. And 2.0 unfortunately, short-term ‘noise’ and ‘random events’ have an even greater effect on a 1-year horizon than any IF C 10 7 % 1.5 S &P 5 0 0 +18 % of these three factors.

to IFC Investable Index 1.0 The Picks of the Litter for 2004 IF C +3 0 0 % Emerging Market Equities 0.5 S &P 5 0 0 +9 8 % The star for the next few months seems likely to continue to be emerging markets equity; all it needs, in my 0.0 Dec- 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 opinion, is for the US market to hang in. Emerging is the 3 Quarterly Letter – January 2004 GMO data. Last year in local currency, international developed Forestry equities underperformed the S&P and only outperformed The only relevant data for forestry investing is the for US investors because of the weak dollar. The attractiveness of everything else, since forestry is more or unprecedented advantage in relative value should result less a constant. Enough said. in excellent relative performance over the next several years, although it should be recognized that in absolute The broad range of investment opportunities are terms, EAFE is a little overpriced and in a major decline unattractive, and in the US, about as unattractive as they its correlation with the US makes it unlikely to do well; it have ever been. is, though, likely to go down substantially less.

2004 Predictive Model (see also Special Topic) Non Dollar Investments Nick Nanda, a fairly new arrival to GMO, and I have tried We had an unusually large bet against the dollar last year to put together a rough and ready model to predict 1 year and it was even weaker than we had expected. Today, in ahead and the 11 months after January using three factors: purchasing parity terms, the dollar is about right, but it is the year of the Presidential Cycle, the strength of January, hard for us to imagine that with a trade deficit of half a and the overall value of the market. For this year, the trillion dollars the dollar is likely to make a large, Presidential Cycle suggests a low probability of a major sustained upward move. There is also some risk that decline, and performance in the -1% to +19% range. investors might lose confidence in the dollar and by (This range for the fourth year, shown for one standard bringing money home – in emerging countries for deviation, is only half that of the other 3 years.) Current example – out of dollar deposits, create a self-reinforcing overpricing of the market suggests performance will be in downward cycle. We retain substantial anti-dollar the lower three quarters of this range, or -1% to +14%. exposure, but it must be admitted that once again, this is January, if down, would suggest a -1% to +10% year, and on thinner ice than the case a year ago against the dollar if up, a +5% to +16% year. Our special research topic when value and deficits agreed. this month has a fancier (more over engineered?) model.

Disclaimer: Past performance is not indicative of future results. The views expressed herein are those of Jeremy Grantham and GMO and are not intended as investment advice.

GMO Quarterly Letter – January 2004 4 GMO Special Topic January 2004 Jeremy Grantham, Chairman Predicting the Market on a 1-Year Horizon Part I: Yet More on the Presidential Cycle may be very independent, but since 1932 the combination Exactly a year ago, I reintroduced this topic with the data of the Fed and the Administration has created a summarized in Exhibits 1 & 2. It became, in the second remarkably effective team that has engineered falling half of the year, a training research project for Nick unemployment through to the election and, as probably Nanda, who recently joined us fresh from college. It unintended consequences, rising stock prices in the third seems to have developed a life of its own because the year followed by an incredibly stable stock market in the more we dig, the more it reveals. First of all, it is fourth year with half the normal volatility (see Exhibit 5). compellingly easy to understand and appreciate. We Isnt this exactly what any politician would desire? know that politicians would like to be re-elected. What is Strong, sometimes risky action in year three and then let surprising is how effective they have been at influencing things stabilize with just a little more of the same, the economy to advance their cause. It turns out that perhaps more visible than real to nudge the game along in changes in unemployment have a much stronger a safe way up to the election. Well, believe it or not, they correlation with voting swings than changes in GNP have got this down to a fine art and whether this is growth (see Exhibit 3). And why wouldnt they? GNP teamwork or good fortune the numbers do not reveal. growth is a fairly nebulous concept to the voter in the One of the nicer features of this Presidential Cycle work street, but unemployment is, in contrast, simple and has been finding how logical it all is. A good example is concrete. Anecdotally, people are also well aware of the effect of a new party in power compared to a re-those around them who lose jobs. Well, by trial and error, elected party. Of course, they both have to create elbow or sheer talent, administrations have learned to have room in years one and two to have room to move and unemployment rise on average in years one and two, fall create stimulus in years three and four. But a new party quite rapidly in year three, and drift slowly down to its has someone to blame for the necessary house cleaning. cycle low in the fourth year (see Exhibit 4). Not bad! In Look at what a mess weve been left by those comparison, the shifts in GNP, or productivity, are barely incompetents, which we now have to clean up since, measurable. What is measurable, though, is the effect on unlike them, we are fiscally prudent. Not surprisingly, the stock market. What traditional stimulus there is – therefore, we find that although the old partys actions money supply, tax cuts, and interest rate cuts – is not as drive the market down in years one and two, a new party much as we expected; in fact, it is quite modest and its drives it down more. Equally unsurprisingly, having impact on growth is even more modest. The stimulus, driven it down more, the new party benefits from a bigger though usually modest, has a more powerful effect on the bounce than a re-elected party (see Exhibit 6). market than the economy as funds, which often cannot be effectively absorbed by the economy in the short-run, run There is also some indication that markets do better in off into the financial system. But even this indirect effect general in Democratic terms than for Republicans. This of stimulus is not nearly enough to fully explain the is one part of the puzzle we have not fully digested and market effect. The real driver of the market is what the we do not yet find it convincing. It does, though, come Fed and the administration say: it is about jaw boning with some reassurance on one of our expectations: the and moral hazard. Basically, the powers that be are low volatility in year four that we are counting on is saying to the market that should anything unexpected go substantially lower for Republican fourth years than wrong in the run-up to the election, they will ride to the Democratic fourth years (see Exhibit 7). We have not rescue. If you have ever wanted to speculate, now is the used this input as it seems to be slicing the data pie too time is the sub text, and seldom has this been made thinly, but at worst, it cannot hurt and at best it gives us a clearer than Chairman Greenspan spelling out that rates little safety margin on the low volatility number we are will stay low way into the future. The Federal Reserve using.

The First and Second Years caps are not exceptional, low quality continues to do The first 2 years underperform on average by 5.5% and quite well after a strong year three, but value reverses – 3.8% respectively and fully half of these years are down having had by far its worst stomping on by growth in year in complete contrast to the last 2 years of the cycle. The three – coming back to its strongest year of the cycle at good news is that the January Effect and the value of the 5½% over average. market are both substantially more predictive than in the last 2 years. So, in 11 months we will enter the weakest Part II: The January Effect year of the cycle with (probably) one of the two or three The fact that the January performance has predicted the most overpriced markets in history and the complete balance of the year has been known for years, just like the removal, normally, of any moral hazard.

Presidential Cycle. It probably derives its usefulness from the indisputable tendency of most of us to sit around Year Three of the Cycle at the end of each year and worry about taxes, capital Because the Presidential Cycle is clear in its logic and gains, the disposition of portfolios, and, particularly, the clear and powerful in the data, I have become more use of year end bonuses. Its the only time most of us sit impressed than I was a year ago. The year just finished back and ask ourselves, What do we think this next year was a classic year three. It had, to be sure, more actual is going to be like? The good news, is that, like all market stimulus than average and even clearer moral hazard. inefficiencies, it is intellectually interesting to think what This should have caused the year three characteristics to causes them, but it is not necessary.

It is only necessary be even more pronounced than normal, and this is exactly to understand and to measure the consequences of the what occurred. Year three is much stronger than average behavior. There is some evidence that investors are (+9½% since 1932), and this one was almost +20%. It is picking up on the January Effect as the average typically by far the best year for growth, and this one was effectiveness over the last 25 years, although still excellent too. The 30% of stocks with the best growth substantial, is a third less than its original predictive estimates steadily outperformed the market and ended up power (see Exhibit 10). almost 10% ahead. (For the record, price to book type Part III: The Value of the Market proxies for growth have completely broken down under increasingly odd accounting practices. High versus low The correlation coefficient of current market value and growth estimates, however, capture the same historical future performance rises with time, and at 10 years it is a flavor of the old growth/value cycles and unlike price to heart warming .71.

At 3 years, it is still a reasonable .44, book are still working well.) Low quality stocks and but at 1 year, it is down to .26. To find out what small caps have their best year in the third year as befits percentage of the market move this explains, it is the speculative environment that moral hazard produces, necessary to square this, so at 1 year, the explanatory and this third year 2003 was a doozy. Russell 2000 beat power is only about 7%. Another more confidence the S&P by 19% and low quality won by 7%. (Exhibit 8 inspiring way to approach this is to say that since 1927, shows the typical year three data and how 2003 the cheapest 20% of markets have outperformed over the compared.) following year by 10%; that is, if you excluded all year threes, the cheapest 20% beat the remaining 80% by 10 The substantial added strength of year three comes with percentage points.

This sounds heart warming, but has slightly reduced volatility so that only 2 of the last 18 two drawbacks. First, there is a dishearteningly large third years have been down, and those just barely. The variability around this, so that many cheap years decline Presidential Cycle effect in the third year is in fact so and many expensive years go up, and second, the record powerful that our other useful 1-year predictors – the for the last 35 years is down to a 7 percentage points value of the market and the January Effect – have spread (see Exhibit 11). Still, a 7% spread feels a lot absolutely no influence on them; third years simply go more impressive than a .26 correlation coefficient and up. Because of this, we study the other two variables since it is only the tail ends of distributions that induce us excluding all third years. to bet heavily at GMO, you can easily see why this rugged type of analysis – focusing only on the highest The Fourth Year and lowest 20% of a distribution – is usually more useful Year four of the Presidential Cycle has interesting to us than correlation coefficients.

(Year three of the characteristics as shown in Exhibit 9. Its most distinctive Presidential Cycle is, in comparison, a meat cleaver. It and reliable feature is its extremely low stock market adds 9½% and with somewhat lower variability, so that volatility. Its return is, on average, almost normal; small only 2 years out of 18 are down, and those barely.)

GMO 2 Predicting the Market on a 1-Year Horizon, January 2004 But still, the decline in predictive power in the last 30 Up January years is disturbing to worshipers of mean reversion. Given where the S&P stands currently, it is hard to What is going on here? Is value, surprisingly, being used imagine that January will register as a down month. so much that it is losing its power? Or is short-term Assuming that January ends up at a reasonable +2% we momentum and other noise becoming larger, a likely can expect the overall year to be in the top half of our candidate given the high jinks of the last 7 years. To test range i.e., +9% to +20% or, in other words, the remaining this hypothesis, we looked at the predictiveness of value 11 months of the year will return between +7% and for 3- and 5-year holding periods and were surprised to +18%. While January might give us a positive signal, find that at 3 and 5 years the numbers, although quite there is no escaping the fact that on a value basis this is strong, were again well behind the dazzlingly strong data one of the most overpriced markets in history, so we need of 1932 to 1967 (see Exhibit 12). This raises an to bias our estimate downwards to the lower two-thirds of important, but not too surprising issue: to win on value, our range, which leaves us at +7% to +14%. Adding back you must be prepared for increasing noise and greater our assumption of a +2% January, we are predicting an pain and time on average before you win. overall return of +9% to +16% for 2004.

This does not seem to be a formula likely to have an army The prediction of +9% to +16% does appear a little of strategists lining up to try their luck with long-term optimistic, but we take comfort in the fact that the party value. What else is new? Indeed, this casts an interesting in power is serving out its first term, and, on average light on Keynes famous Chapter 12 of his General years three and four of the cycle do better under a new Theory, which moans of the career dangers of choosing administration than under a re-elected party. Under new long-term value over the relative safety of short-term parties the fourth years of the cycle have returned almost momentum, or the beauty contest. What a wimp! He +15%, and compared to this figure, our estimate of +9% made this point when 1- and 5-year predictions based on to +16% seems appropriate. value were twice as powerful and reliable as they are now! If ever there was data to reinforce our belief that Down January almost all firepower should be kept for outliers, this is it, In the unlikely event that January is a down month, we for it is only at the outer 20% at most of these series that need to move our primary estimate of 2% to +20% there is reliable power, and the outer 10% is surely better. towards the bottom half of this range i.e., -2% to +9%. Even then there is, of course, plenty of pain to go around. So assuming a -1% January, this implies the February to December period should return between -1% and +10%.

Part IV: Putting it all Together Since a down January still leaves the market overpriced, From all the data we have studied these past 6 months, it we need to take a debit for that. So, slicing off the top appears that three factors signal the outcome of the third of our range gives us an estimate of -1% to +6%, market on a 1-year time horizon: which after subtracting the down January, leaves us „ Presidential Cycle predicting -2% to +5% for the entire year. „ January effect Summary „ Value (Price to 10-year trailing earnings) Admittedly, the Presidential Cycle effect lacks sufficient data points to make all our results statistically significant.

Since 1932, the average real return for the S&P in a Also, there is a certain crudeness to our decilization or calendar year has been 9% and the fourth years of the quartilization approach of figuring out the January Effect Presidential Cycle (2004 is a year 4) have averaged about or the impact of value on the market. However, given the the same. Assuming that the characteristically low surprisingly logical nature of our findings, and how volatility of year 4 (11%) continues to hold through 2004, neatly all these factors fit together, our estimates are the S&P should return somewhere between -2% and probably better than ignoring the intermediate term +20% (9% ± 11%). As mentioned earlier, volatility under completely. The actual outcome, though, could be very Republican administrations tends to be about half of what different if the ice breaks. it is under Democrats. While we are not explicitly using this knowledge in our prediction, it certainly reinforces our belief that there is a low probability of a huge upward Disclaimer: Past performance is not indicative of future results. or downward move in the market this year. Lets now The views expressed herein are those of Jeremy Grantham and GMO incorporate the January effect into our estimate. and are not intended as investment advice.

Predicting the Market on a 1-Year Horizon, January 2004 3 GMO Exhibit 1: Market Returns for the Years of the Exhibit 4: Level of Unemployment in the Different Presidential Cycle Relative to Average (1932-2003)* Years of the Presidential Cycle (1948-2003) 12% 6.1% 9.5% 6.0% Average Excess Real Return 6.0% 8% Unemployment Rate 5.9% 4% 5.8% 5.7% 5.7% for the S&P 500 5.6% 0% 5.6% -0.3% 5.5% -4% 5.4% -3.8% 5.4% -5.5% -8% 5.3% 1 2 3 4 1 2 3 4 Year in Election Cycle Year in Election Cycle * The numbers presented here are slightly different from those shown at our 2003 Fall Conference because we have switched from PPI to CPI as a deflator.

Exhibit 2: Ratio of Up Years to Down Years for the Years of the Presidential Cycle (1932-2003) Exhibit 5: Market Volatility (1932-2003) 10 25% Average Volatility of the S&P 500 8:1 19.3% 8 20% Ratio of Up Years to 6 5:1 15% 11.1% Down Years 4 10% 2 5% 1:1 1:1¼ 0 0% 1 2 3 4 All Years Year 4's Year in Cycle Exhibit 3: Explaining Vote Swings Using Exhibit 6: Market Returns for Unemployment, GDP and Market Return New Party vs. Reelected Party Votes Sorted by 30% N e w P a rt y Decline in Increase in Market 24.2% R e - E le c t e d P a rt y 1952-2000 Unemployment GDP Gain Average Returns 20% 1 20.1 10.2 0.9 2 2.0 11.8 11.1 13.9% 14.7% 3 1.6 1.7 9.7 for the S&P 500 The numbers above represent the ratio of Electoral College votes won by the incumbent 10% 7.7% party to votes of the other party. Decline in Unemployment: 1 represents those years where 5.6% unemployment reduced the most in the year leading up to the election and 3 represents 4.9% those years where unemployment increased the most. Increase in GDP: 1 represents those 2.0% 0.9% years where real GDP had its biggest increase in the year leading up to the election and 3 represents those years where GDP increased the least. Market Gain: 1 represents those 0% years where the S&P 500 had its biggest gains in the year leading up to the election and 3 Year 1 Year 2 Year 3 Year 4 represents those years where the S&P 500 had its smallest gain.

GMO 4 Predicting the Market on a 1-Year Horizon, Janaury2004 Exhibit 7: Volatilities Relating to Exhibit 10: January Political Parties (1932-2003) Predicting the Following 11 Months 25% S&P 500 Return 19.3% 20% Quintile 1936-1967 1968-2003 Average Volatility 15% 13.7% 1 18.8% 13.9% 11.1% 2 3.0% 6.7% of the S&P 500 10% 7.0% 3 20.0% -4.0% 4 -1.4% -1.2% 5% 5 -3.9% -2.4% 0% All Year 3’s have been excluded All Years Year 4's Year 4's Year 4's 1 represents the Januaries where the S&P 500 experienced its biggest gains and 5 represents Under Under the Januaries where the S&P 500 experienced its biggest losses Democrats Republicans Exhibit 8: Comparing 2003 to an Average Year 3 Exhibit 11: Value as a Short-Term (1-Year) Predictor 40% Year 3 34.3% S&P 500 Return 30% Quintile 1927-2003 1968-2003 Relative Return 1 14.0% 8.0% 20.1% 20% 15.2% 2 9.4% 2.6% 3 5.1% -1.1% 9.5% 10% 6.1% 7.3% 4 0.5% 0.5% 4.5% 1.8% 5 1.4% 4.5% 0% All Year 3s have been excluded S&P 500 Grow th vs. Small vs. Low Quality 2 3 1 represents the cheapest years and 5 the most expensive on a Price/10-year trailing Value Large vs. High earnings basis Quality Exhibit 9: Performance in Year 4 Relative to Average Exhibit 12: Value as a Long-Term Predictor 10% S&P 500 Return 1932-1967 1968-2003 5% Relative Return 2.9% Quintile 3-Year 5-Year 3-Year 5-Year 0.3% 0% 1 63.7% 97.6% 31.7% 74.6% -0.3% 2 56.4% 75.1% 22.0% 31.7% -5% 3 14.2% 61.4% 12.0% 41.4% -5.2% 4 15.2% 35.2% 16.5% 24.1% -10% 5 15.3% 19.6% 20.5% 50.0% S&P 500 Grow th vs. Small vs. Low Quality Year 3’s have not been excluded Value Large 3 vs. High Quality 4 1 represents the cheapest years and 5 the most expensive on a Price/10-year trailing earnings basis 1 S&P 500: 1932-2003 2 Growth vs. Value: 1960-2003 3 Small vs. Large: 1961-2003 4 Low Quality vs. High Quality: 1960-2003 Predicting the Market on a 1-Year Horizon, Janaury2004 5 GMO