煤矿中的金丝雀*

GMO Quarterly Letter April 2005 Jeremy Grantham, Chairman The Canary in the Coal Mine* & Letters to the Investment Committee III Traveling around the world for my first time – 19 flights my trip, I was happily quoted in the Australian Financial in 35 days – mostly on business, I could not help noticing Review as saying that the Australian residential real estate a few things. First, you should travel east to west. Yes, market could be the canary in the coal mine – that is, a it’s 10 or 12 hours longer flying time, but your jet lag harbinger of bad things to come for a lot of us. And it kicks in at 6 p.m. at your new location when nobody may be. Sydney house prices rose earlier, faster, and fur-cares. Going the other way, I found myself continually ther than any other. Australia also raised its rates earlier trying to talk to a group of analysts or clients at 4:15 a.m. and further than England, where rates were in turn raised on my internal clock. And as soon as you adjust, you earlier and have climbed further than the U.S. And both move on again. of these foreign markets have overwhelmingly floating rate mortgages so it would reasonably be expected that More importantly, a traveler could not help but notice the effect of higher rates would impact prices faster. how grimly expensive things are everywhere, seen through dollar eyes. When you have a bet against the dol- Well, GMO’s office in Sydney is on an old wharf, and lar (who doesn’t?) it is disturbing and serves to underline right next to it is another long, converted wharf, at the end my points from last quarter’s letter that were attached in of which is a splendiferous apartment bought 6 months a Barron’s article. First, that the dollar is probably not ago for $3 million and now converted into two still pret-overpriced on any fundamental basis. Second, though, ty splendid units, now for sale for $2.5 million combined. that no one seems to sound convincing on the topic of The official numbers confirm that Sydney prices are well currency valuation. And third, yes, what are you going to off their highs, although as yet far from a real bust. do long term in the face of a 700 billion dollar annual Nearer to home, London prices are also flat to off a little accumulation of foreign obligations but bet against the – finally – in most, but not all, districts. Real estate has a dollar? I said last quarter that the only thing I really lay long history of lagging stock market breaks and had every awake sweating about was the fear of a quick 15% move reason to do even better than normal this time, as global in the dollar’s favor. History says that this kind of interest rates were cut and money and credit made so upward move is routine, even if the real trend is still available. What was new this time, though, was the down. Well the ink on Barron’s had barely time to dry degree to which home owners in these Anglo-Saxon mar-when the dollar set off on a 6.5% run against the euro kets increased the size of the mortgages as they refi-from its low. Anyway, trust me, Sydney and Auckland nanced, supporting their rapidly growing consumption in are expensive, Tahiti is very expensive, and as for the face of only modest increases in income. London, has anyone worked out how the locals can afford Increasing paper wealth from houses in fact more than to live there? offset the negative wealth effect from the stock market The other thing you’d have to try to avoid noticing is the declines of 2000 through 2002. By making home owners attention given to house prices in the three English speak- feel wealthier, it also helped suppress the need to save, ing markets … and who would not give attention to these and all three countries saw their personal savings rates house prices? New Zealand is up over 20% in the last 2 drop and hold at unprecedentedly low rates, despite the years. The whole of the north of England moved up a aging population problem. In this way, delusions caused staggering 40% in a 12-month period a year or two before by stock market paper wealth were followed by similar that, and it is rumored that Sydney has pushed San Francisco out of the champion’s spot in the “who’s got * After 40 years in America, I'm still apparently using Englishisms. Miners used to take a caged canary into the coal mines, which would oblige by drop-the lowest return to renting your house?” contest. During ping dead from odorless gases before the miner holding the cage did.

delusions driven by paper housing wealth. Alan Exhibit 1 Greenspan may be impressed by the increases in house- United Kingdom: Home Prices as a Multiple of hold financial strength caused by higher house prices, but Average Earnings any believer in replacement value cannot be impressed, and I am certainly one of them. (See January’s Special 6.0 Topic titled “Letters to the Investment Committee II – Replacement Cost: The Bedrock of Value”.) When the 5.5 price of a chemical company’s stock doubles, I always urge focusing on its more or less unchanged replacement 5.0 2 std. dev. value and not the change in its paper value to twice that.

And how much easier it is to see this effect with the house you live in: the price of my house in Boston may have 4.5 tripled in 10 years, but the flow of real service that it 1 std. dev. offers me is clearly unchanged. It protected me from the 4.0 wind and the rain exactly as well 10 years ago as it does now, in fact, slightly better back then since it’s now older. Avg.

(Yes, the replacement cost for the 20% that is land is the 3.5 least simple of all asset classes, since clearly they don’t make any more land at any cost, but for the record, land 3.0 too is mean reverting and has in Europe a several hundred year record of rising at about the rate of the GNP.) But 2.5 apart from land, it is very much easier to understand that 19 52 56 60 64 68 72 76 80 84 88 92 96 00 04 the changes in housing values around replacement value are paper events, than it is to appreciate the same point for Sources: Nationwide, Office of National Statistics (UK), GMO stocks where analysts always justify every price increase more than 1 standard deviation below it. This is about as in the market by proposing some equal and offsetting extreme an event as the recent March 2000 U.S. equity increase in future prospects. Perhaps as much as 10% of bubble. This is remarkable, since March 2000 was the these perceived increased virtues for stocks are real, but I largest deviation from trend in modern U.S. equity histo-doubt it. Also, the increased price of an owner-occupied ry. I have absolutely no doubt that the consequences will house is clearly not contributing to the owner’s ability to be similar, and that U.K. housing will also return to trend service his increased mortgage! or below.

The increased role that housing price increases have We beg your indulgence for Exhibit 2 – recent Sydney played in sustaining the expansion of credit and con-house prices – because neither our Sydney nor U.S. office sumption in the last several years has made us realize that could come up with a decent series longer than 25 years. we have not put enough research effort into this area and Still, it does capture the spirit of the exercise and indi-we are attempting to address this. cates the potential for a major decline. Exhibits 3 and 4 Replacement cost for housing is a messy and difficult show off the great advantage of the size and diversifica-way to approach this problem, but fortunately as a long- tion of the U.S. housing market, where side by side on the term reader of The Economist I am well aware that house two coasts we have something close to a bubble versus prices are also mean reverting around a trend line multi- something near normal through most of the rest of the ple of household income. Our quants in our London country. Exhibit 3 uses median income to median house office reviewed the data and found, to our mild surprise, price ratio and indicates that U.S. average house prices that for the last 45 years this trend seems to show no clear are probably only about 25% over trend. The March 3 upward trend in this ratio as we would have guessed; Economist has a series based on house rentals that indi-Brits seem willing to spend only the same fraction of their cates about a 30% overpricing. A conservative compro-income on housing over time – although they certainly mise would be to assume a one-third overpricing that have a lot of fun roller coasting around this flat trend. would require a 25% decline to get back to trend. Exhibit 1 shows the data and it makes a very worthy Exhibit 4, in contrast, shows the situation in Boston – exhibit for the Bubble Hall of Fame: on the modern data, selected entirely at random to help persuade my wife to current housing prices in the U.K. are over 3 standard sell our house and pay rent – which reflects a more deviations above trend, having as recently as 1995 been Sydney-like rise to 52% over trend, again all in the last GMO Quarterly Letter – April 2005 2 Exhibit 2 Exhibit 4 Sydney: Home Prices as a Multiple of Boston: Median Home Prices as a Multiple of Australian Household Income Household Income 5.5 7.0 6.5 2 std. dev. 2 std. dev.

5.0 6.0 5.5 4.5 1 std. dev. 1 std. dev.

5.0 4.0 4.5 Avg. Avg.

4.1 How painful would a correction be? To start with 2 std. dev.

London, we could look at the last bubble that built up in 3.9 the late ’80s and broke in the late ’90s. Viewed with hindsight from today’s 3 sigma (standard deviation) 3.7 1 std. dev. event, the 1990 event looks like a bona fide member of the GMO bubble club where we draw the boundary at 2 3.5 standard deviations, which seems (at least to us) to be reaAvg. sonable although it is quite arbitrary. A 2 sigma event is the kind that would occur on the upside (i.e., a bubble) 3.3 every 40 years if the data were a normal distribution.

Yes, we know the real world actually has more outliers 3.1 than a normal distribution, that is to say, its distribution is ‘fat tailed’ and therefore there will be more 2 sigma 2.9 events than there ‘should’ be. But like arcane sports rules, these rules although arbitrary are the same for 2.7 every asset class. In any case, even in our world of fat 19 75 77 79 81 83 85 87 89 91 93 95 97 99 01 03 tailed distributions, we have only found 28 good exam-Sources: National Association of Realtors, U.S. Census ples of previous bubbles including: stock markets around Bureau, GMO the world, currencies, and commodities. As frequent 3 Quarterly Letter – April 2005 GMO readers know only too well, I am patiently waiting for the in family income. In contrast, for the Bostons of the current 28th bubble, the S&P 500, to go all the way back country, of course, the outcome could be a much more to trend – about 750 versus today’s 1150.

It fell to with- painful decline of 25% even after allowing for any offset-in 10% of trend in 2002, but still no cigar. But, as also ting income gains. Unfortunately it must also be admit-often mentioned by us, all the other 27 identified bubbles ted that the badly overpriced cities of the East and West did indeed move all the way back to (or below) the trend Coasts do have a disproportionate share of both the media that existed prior to those bubbles forming. and the financial industry, so that their pain will probably be misrepresented as more significant than it really is in The U.K. housing data in Exhibit 1 shows, in addition to GNP terms, which might exaggerate the depressing effect the current mega bubble, two prior substantial bubbles on “animal spirits”. The key point in the U.S., though, is that both fully mean reverted.

An interesting point, that in the recent 3-year stock market decline all the stock though, is that the 1990 bubble, although held responsible market wealth lost by the median family holding stocks (rightly or wrongly) for considerable damage to the U.K. was more than offset by a 21% advance in house prices. economy as it broke, did not involve as much pain getting This favorable circumstance seems extremely unlikely to back to trend as you might think because the trend line reoccur this time. The inevitable 30% to 40% decline in average income was also rising rapidly. It rose for two stock assets necessary to get to fair value, accompanied reasons. First, inflation was high by today’s standard, by flat to down housing prices, will pose substantially averaging 4% a year for the 5-year decline, and despite greater risks for consumer spending than last time.

And GNP problems, average real incomes did surprisingly leveraging housing debt was such an easy, effective, and well. Nominal income gains for the 5 years totaled about unthreatening way to allow consumption to keep growing 30% – about twice what you could expect today in the despite the quite modest gains in household incomes in U.S. under recent conditions. So although the U.K. house all our three Anglo-Saxon markets. There is likely to be price to income ratio had to fall by 24% to get to trend very little increase of this debt if prices stop rising and from a peak of 4.7 times to a trend line of 3.6 times, the rates are even modestly higher, and there might be some typical home owner’s house was more or less the same modest, but painful, reversal. The best reasonably likely nominal price at trend as it had been at the peak! outcome in the U.S. is that a moderate stock market (4.7 × 1 = 4.7, 3.6 × 1.30 = 4.7) decline in the next 2 years – my ‘dreary’ forecast – could All the real pain from the decline of house prices then be accompanied by up to 1 more year of average house came not from the move to trend, but from the typi- prices rising, for the U.S. housing market has lagged the cally disturbing tendency for down cycles to over cor- other countries and has some good potential for catch-up rect.

Prices bottomed at 2.9 times income, or another in certain regional markets. This lag might be expected 20% below trend. Today, though, in the U.K., the because our house prices have risen less and our rates price/income ratio would have to fall by 37% to merely rose later. But, by this time next year, time would really get to trend, and today’s lower inflation and lower income seem to be running out for our U.S. housing semi-bubble. growth will cushion far less of the decline, perhaps only It also seems very likely that by then the housing markets half. Once again any overrun would inflict additional in the other two countries will have completely run out of pain. And any unexpected help from accelerating infla- steam. tion in reducing the ratio would in the U.K. come with an Another Small Canary with Apologies for a Late equally unexpected sting in the tail: their floating mort- Warning gage rates would of course rise with inflation, leaving There has been a small cap effect in the past in the sense most people worse off, all things considered, than if infla-that an annually rebalanced portfolio of small stocks has tion stayed low.

A damned if you do, damned if you don’t beaten large stocks by a wide margin since 1925 (2½% a situation if ever there was one. This point is more relevant year ±). A less well known factor is that over 100% of now in the U.S. than it has been before because of the sud-this outperformance has been concentrated (on average) den recent rise towards 40% in the use of floating rates. in the month of January. Even less well known is that Australia and New Zealand would both be in the same when small caps have a bad January, it is a very bad sign boat as the U.K., but the U.S. will obviously be less bad. for their performance for the year. (Now he tells us.) Even if the ratio here were to over correct by 10%, that Since 1965, poor Januaries for small cap are followed by would only be a drop of 25%; if the decline took 5 years, average underperformance of 7.2%, and they have out-all but a few percent would be offset by even modest rises performed only 15% of the time!

GMO Quarterly Letter – April 2005 4 Germany to the Rescue!? of the 12 great bubbles that form some of the bedrock of Germany has looked like a basket case for so long now our belief in mean reversion, followed by our Exhibit 1 that it is only fair to point out that several things have from this letter on U.K. real estate bubbles. I must say, improved. First, through lower inflation they have amongst the Austrian-Monetarist squabbles, our data did become more competitive and largely worked their way seem clear cut. We and other believers in mean reversion out of a badly and painfully overpriced currency position would have seen 1929 as an asset class bubble for it at the start of the euro. Second, although their economic looked at the time like a 2½ sigma event. Roger Babson reforms have been disappointingly slow, there has been famously did predict a severe decline, but of course 2 progress and it continues. Third, and relevant to our years too early, to which I can only say, “Been there, done topic, their debt leverage at all levels has not been that!” The contrast between their theoretical arguments increased like the English speaking Anglo-Saxons; they (Monetarists, Austrian, etc.) and our simple data driven 2 have not increased their housing debt, and house prices sigma analysis led me to think about the few points that themselves have not increased. Indeed, German house we believe are beyond reasonable argument. prices are where they were in the late 1970s in real terms! What Do We Really Know?

This relative strength in both debt and housing may give „ Data is data, and all bubbles defined as 2 sigma events Germany an unexpected substantial reprieve relative to us have indeed broken. They broke regardless of the should we get into a credit crisis any time in the next few steps that were taken during and after the bubble. Bad years, which seems quite likely – I would think at least a monetarist policy may have caused the Great one in three shot. It is probably worth adding that while Depression and good policy may have let us down some other European countries have had some house gently after 2000 (we shall see), but both were clear price escalation – Holland and more recently France – asset bubbles and both broke. The economic and mon-and some debt expansion, none come close to the potenetary environment was different for all 28 bubbles, but tial risk of the three Anglo-Saxon countries. all of them broke.

Edward Chancellor and Credit Cycles „ A breaking bubble in an important asset class will def-On my 35-day trip around the world I read a new book by initely affect animal spirits, investing intentions, and Edward Chancellor (whose Devil Take the Hindmost: A consumption. This effect may or may not be offset or History of Financial Speculation was a great success) that postponed by monetary or fiscal moves. analyzes credit expansion, its past role in market and eco- „ The increase in housing prices in several countries this nomic breaks, and the current credit expansion’s potential cycle has definitely facilitated easy credit expansion for making trouble. Chancellor’s book, titled “Crunch and allowed total household credit to go to new highs. Time for Credit,” reviews the major competing schools of „ Housing prices will eventually retreat to trend, and this thought on the dangers of asset class inflation, the power will cause this part of the credit expansion to stop and of monetary measures to balance them, and the possible quite possibly to reverse. role of animal spirits; in short, the Monetarists, „ Higher average asset class prices since 1995, led by ‘Austrians’, and Keynesians.

I must say I thought it was stocks and followed by bonds and real estate, definite-excellent – dense and informative and continuously sum- ly made people feel richer, spend more, and save less. marized. (I even read a few pages on the beach at „ We have in the U.S., U.K., Australia, and a few other Huahine off Tahiti, a considerable hurdle to clear in terms countries definitely saved less at the household and of distractions.) What impressed me was that nothing corporate pension level than is necessary for good per-was agreed on, certainly not the causes of the 1929 crash sonal retirement plans. or the following depression. And some views were as hard to understand as the Efficient Market Hypothesis of „ This shortfall will be revealed when asset classes French, Fama, Malkiel, et al. For example, Chancellor revert to normal, and a great majority of savers will be reports that Milton Friedman and the Monetarists basical- forced to realize that their nest eggs are inadequate. ly maintain that 1929 was not a bubble, just the market’s „ When this is revealed there will be a lot of broken reflection of good times, just as Irving Fisher maintained hearts. in early 1930.

The depression they argue was entirely „ And finally, nothing that Greenspan and his successor caused by lack of monetary stimulus and had nothing to do will prevent this reversion to the mean of asset do with the run-up in debt, speculation, and asset prices. prices, although their actions may have a very substan-While still shaking my head in disbelief, I came across tial and beneficial effect on how badly the economy Chancellor’s use of GMO’s “All Bubbles Break” exhibit fares in an environment of falling asset prices – 5 Quarterly Letter – April 2005 GMO although they have probably run out of asset classes Barron’s as advising, “not to assume because it is at $50 to inflate (except for a potential catch-up leg in some per barrel it will come down.” The December 2010 real estate markets) and may be approaching some future was then $37 and it’s now $50.

I must also admit maximums in debt leverage. that I have no confidence that it will stay over $50, and my view is that we should be extra careful as oil could GMO and the First Quarter burn us badly either way. China has probably changed My first quarter letter predicted a “dreary” rather than the commodities world for decades, and the world may be very painful year, and so far global markets have been very close to pumping as much oil as we ever can, just as trying exceptionally hard to please. In dollar terms, the the U.S. itself is long past its own maximum production. S&P was -2.1%, EAFE -0.2%, U.S. bonds -0.4%, non If that is so, then we must substitute and economize, or U.S. bonds -3.0%, and emerging debt -1.3%. Only good bust, and that would make for exciting times.

Oil may be old emerging equity was up at 1.9%. Now that’s what I that very rare bird – a paradigm shift. Over the years we call dreary! have asked over 2000 professionals for an exception to GMO fund alphas had a slight average lead in fixed our claim that every asset class move of 2 sigmas away income and a respectable start in our hedge funds. On from trend had broken, and not one of the 2000 has ever balance, we were nicely up in U.S. equities, and on a offered an exception! This should be scarier than the fact weighted basis slightly up in foreign equities, thanks to that GMO has tried so hard to find one and failed. But we an over 2% lead in emerging. GMO asset allocation also have always said that intellectually you can imagine a had a good quarter.

I would certainly settle for three more paradigm shift in an asset class price, even if we have quarters like this in all respects. been unable to document one yet in history. Exhibit 5 shows the price of oil and 1 and 2 standard deviation So far oil prices have been worryingly strong and bands. If the new price averages $50 and above, it will European growth worryingly weak, but the U.S., China, look suspiciously like the real McCoy. Chinese growth and emerging have kept going nicely and the dollar has ral- and supply problems could do it. It’s the best possibility lied. All in all, no major land mines, and without one or I’ve seen in my career. But the investment desert is lit-two I believe we will struggle through this year about okay. tered with the bones of those who bet on new paradigms.

Oil The closest we are to a serious land mine is the oil price and I am happy to have been quoted in the January 3 Exhibit 5 Oil Price in 2005 Dollars WTI Crude $ / Barrel 2 std. dev.

1 std. dev.

20 Avg.

GMO Quarterly Letter – April 2005 6 GMO Special Topic April 2005 Jeremy Grantham, Chairman Letters to the Investment Committee III 6 Kinds of Bull It is surprising to me that as the years go by the type and 1. The Dow 36,000 book that came out in 1999 was quality of market analysis in terms of the market’s value typical of several similar books and articles has been so consistently bullish that it is misleading about published in 1929 such as Raskob’s famous half the time. This negative comment is not aimed at the “Everyone Ought to be Rich.” Only at the height substantial and influential number of academics and of bubbles are people credulous enough to buy financial experts who believe in market efficiency, for them, but when the time is right they sell like hot they of course can have nothing to say on this topic. cakes because they reinforce our worst impulses of What I’m really getting at here is the analysis of market wishful thinking, overconfidence, and just plain value you read in the stock market columns of daily greed.

The Dow at 36,000 would have been close papers and magazines and ‘strategy reports’ from to five times replacement value! Every new $100 brokerage houses. The reasons for the bullish bias form million factory would instantly become worth a complicated issue that I hope to cover next quarter. $500 million, and the whole U.S. would have Here I would like to look at results: the six kinds of eventually drowned not just in fiber optic cable – sloppy bullishness. There is of course a seventh one area that actually got close to five times completely kosher kind of bullishness, and that is when replacement value – but virtually everything else the market is cheap: when it is selling below its trend line as well, as excess capital spending would have price to earnings of 16, and better yet, on a profit margin produced overcapacity which in turn would have level that is also below normal.

1982 is the perfect led to low profits and eventually many more example. The market sold at 8 times badly depressed bankruptcies than we actually had. profit margins: it was simply very cheap and a bullish 2. Jeremy Siegel and his “Stocks for the Long Run” is a outlook was entirely justified. much more serious threat to sensible thinking because Sloppy and therefore dangerous bull market approaches as a Finance Professor who was obviously smart and can, I believe, be divided into six types: persuasive and who had produced useful and extensive stock market data, he seemed like someone who 1. Dow 36,000: the purely ‘PR’ driven nonsense should be listened to, and in the 1998 to 1999 run-up 2. Jeremy Siegel: price doesn’t matter, 7% real returns his thinking influenced many investment committees. by divine right His argument was that stocks had always beaten bonds and had delivered 7% after inflation, and therefore the 3.

Sloppy or no earnings adjustments to market P/E safe bet was to assume the same for the future. I 4. Sloppy growth forecasts debated him several times and used to summarize his argument as “price doesn’t matter” – stocks always 5. The Abby Cohen and Alan Greenspan Show: the win by divine right. Curiously Jeremy agrees that economy is great and therefore the market will do investors get the “earning yield” – the inverse of the fine P/E – so that at 10 P/E you get 10% real a year, and at 6. The ‘Fed Model Effect’: yes P/E ratios are quite high 20 P/E you get 5%. But there is a serious inconsistency but the market is still cheap because interest rates are in his argument. His data showed that the historical so low. 7% real stock return had come from an earnings yield of 7%.

That is to say a historical average P/E of 14 4. Sloppy growth forecasts form probably the widest times. What a coincidence! What he never answered shared optimistic factor. About twice a year I get to in our debates is why in March 2000 at 33 times talk to investment professionals attending 1-week earnings the S&P 500 should not return 3.0% in seminar sessions, and I have often asked them to perpetuity! Price of course is everything, or almost estimate the past long-term growth rate for the S&P everything, for diversification also matters. And 500’s earnings per share and sales per share. The concerning diversification, the idea that stocks always estimates have always been very high relative to win was giving fits to Jeremy’s old PhD teacher, Paul history. Their estimates average in the 4% to 5% range Samuelson, who was quoted in Forbes as saying, “I after inflation, and Lord knows it was hard to find any have students of mine – PhDs – going around the analyst’s estimates as low as 5% in 1999.

The data country telling people it’s a sure thing to be 100% shows that the actual growth is only 1.8% real per year, invested in equities, if only you will sit out the declines. and Rob Arnott and others come up with even lower It makes me cringe!” Me too. The occasional very bad numbers. markets are capable of panicking any committee that 5. Abby Cohen and Alan Greenspan belong to the would be 100% invested in stocks. school that if recent productivity, profit margins, and 3. Sloppy or no adjustments to P/E make evaluations GNP have been strong then: a) they will continue to be of the aggregate market misleading. At the extreme – strong; and therefore b) the market will continue to go which is fairly common – you can read that “the up. Trying to debate these bulls (and most strategists market is selling at 15 times next year’s estimated at brokerage houses fell into this camp in the late operating earnings, which does not seem 1990s) was like talking Swahili to a Russian.

We unreasonable.” This misses the mark on every issue. believers in value and mean reversion talked The analysis fails to make several necessary exclusively about adjustments and fair value and they adjustments: talked about productivity and GNP growth. When I got to speak second I could attempt to trash their case, a. You cannot compare next year’s P/E with but speaking first was a problem so we attempted to historical P/E ratios that are always based on come up with a simple pre-emptive strike. Exhibit 1 trailing earnings. And given the optimistic bias shows the three most commonly used important (and self interest) of brokerage firms, next year’s fundamental factors for the economy. The middle estimates have averaged 11% higher than will column shows the correlation that these three factors actually be recorded. have with the same factors in the following 2 years. b.

The earnings used are ‘operating earnings’not ‘net For example, we are asking if above average profit earnings’. For the market as a whole this should margins predict falling or rising margins and does make no difference since in theory there should be above average GNP growth predict acceleration or as many pleasant surprise write-ups to assets and deceleration? All three factors are in fact negatively earnings as there are write-downs. But in a world where corporations are desperate to beat earnings Exhibit 1 by a penny, accounting weasels are everywhere Correlation and write-downs net of any write-ups have averaged 14% for the last 10 years! (Et tu AIG?) Current Economic Stock Market At GMO we take aggregate operating earnings Economic Factor Over Returns Over and mark them down by 14% – rough justice, but Factor Next 2 Years Next 2 Years on average much better than no adjustment.

Profit Margins -49.0%* -4.1% c. Most importantly, you cannot compare boom GNP Growth -9.4% -5.2% economies with slumps since profit margins are so Productivity Growth -1.5% -10.3% highly mean reverting. Strong economies should have lower than normal P/E ratios applied, and Data: 1950-2003 vice versa. At GMO we normalize by regressing * Correlation between current level of profit margins with future change in profit margins. current profit margins to the long-term average Sources: GMO, Federal Reserve over a sedate 7 years.

GMO 2 Letters to the Investment Committee III, April 2005 correlated with the next 2 years of data. The last fundamentals, which justified even higher stock prices: column shows the correlation of the three factors with about as far from regression to the mean as you could stock market moves (above or below the past average get. of about 7% real). Nor surprisingly, since the 6. The So-Called Fed Model maintains that the fair P/E economic factors themselves are negatively correlated, level depends on the level of inflation and interest the market returns are also negatively correlated, rates. This really is the vampire theory that refuses to spectacularly so for profit margins. This means that die. Modigliani years ago made the point that stocks when Abby, Jeff Applegate, and the rest finish their are real assets that should sell at real replacement value presentation of how good the fundamentals are they and that earnings pass through inflation. If that was not should add … “and therefore we expect below average the case, Brazil would be way under 1 times earnings performance from the market.” Trust me, this is not by now given its historical inflation. what they say.

Analyzing the data, though, is a very muddy job because Alan Greenspan is of course more convoluted but inflation and P/E ratios are coincident. They do move delivers the same implications. In January 2000 he together for behavioral reasons: shifts in inflation disturb said famously that, “the American economy was investors. To make matters worse, since there is often experiencing a once-in-a-century acceleration of some short-term momentum in inflation, i.e., rising innovation, which propelled forward productivity, inflation tends to predict rising inflation, it will also output, corporate profits, and stock prices at a pace not appear to predict rising P/E ratios and stock moves, and seen in generations, if ever.” The internet, which had over a horizon of months this does seem to uphold the “pushed back the fog of uncertainty” for corporations, Fed Model. But it really is an illusion as both series are was his particular pet. “Lofty equity prices,” he said a mean reverting so that throughout history the great few weeks later, “have reduced the cost of capital. The market peaks – 1929, 1965, and 2000 – from which point result has been a veritable explosion of spending on medium-term returns have been poor, have all had low high-tech equipment … And I see nothing to suggest inflation and low rates. Conversely, the market lows like that these opportunities will peter out anytime soon … 1982 that resulted in magnificent intermediate-term especially in the business to business arena,” i.e., the results had very high inflation and interest rates (13% and lofty equity prices would continue. All this within 1 16%, respectively). This medium-term mean reversion is week of the Nasdaq reaching a level from which it reflected in the average 5-year holding periods for would decline 75%, the business to business sub index buying stocks in the highest 10% inflation and rates 95%, and even the S&P 500 falling by 50%! The Fed compared to the lowest 10% since 1925. High rate and Reserve Boss seemed to believe that there was a high inflation periods beat the other end by 2.5% and virtuous cycle in which high stock prices helped 2.4% a year, respectively. Not bad.

Disclaimer: The foregoing does not constitute an offer of any securities for sale. 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.

Letters to the Investment Committee III, April 2005 3 GMO