In De­cem­ber 1996, Alan Greenspan then Fed­er­al Re­serve Chair­man, gave a speech at the Amer­i­can En­ter­prise In­sti­tute in Wash­ing­ton D.C. From that speech we now have this fa­mous and time­less quote:

“How do we know when ir­ra­tional ex­u­ber­ance has un­du­ly es­ca­lat­ed as­set val­ues which then be­come sub­ject to un­ex­pect­ed and pro­longed con­trac­tions?”

What fol­lows is not a pre­dic­tion of the next as­set bub­ble, but rather a re­cap of the events that lead to one, be­cause you can on­ly tell a bub­ble is a bub­ble af­ter the fact. It is al­ways in­ter­est­ing to see how to­day’s in­ter­na­tion­al in­vest­ing en­vi­ron­ment lines up with what ob­tained three decades ago.

De­spite Greenspan’s ques­tion and im­plic­it warn­ing, the fi­nan­cial mar­kets heard the warn­ing and climbed any­way. It climbed for more than three years post Greenspan and the Nas­daq more than tripled af­ter that speech be­fore it fi­nal­ly peaked in March 2000. Any­one who sold in 1996 be­cause of the words of the most pow­er­ful cen­tral banker would have spent the next three years watch­ing col­leagues grow rich, look­ing on in en­vy the whole time.

That is the awk­ward truth that needs to be rec­og­nized to­day. The per­son with the best seat in the house, the most in­for­ma­tion, the most au­thor­i­ty, and no rea­son at all to cheer­lead, saw some­thing com­ing and was still wrong about when. He was not wrong about what, he was wrong about the tim­ing, which in mar­kets is the on­ly kind of wrong that mat­ters.

To re­peat, the les­son in the gap be­tween the speech and the mar­ket peak is that a bub­ble is a word that on­ly works in the past tense. While you are in­side one, it does not feel like a bub­ble. It feels like a boom. The two are built in the same way. You have ris­ing prices, con­fi­dent buy­ers, and a sto­ry every­one be­lieves, and you can on­ly tell them apart once one of them has end­ed. The nar­ra­tive fol­lows the price ac­tion not the oth­er way around as many tend to be­lieve.

The dot-com era is the clean­est il­lus­tra­tion we have, pre­cise­ly be­cause the be­liev­ers were not fools and the sto­ry was not false. The in­ter­net did change the world. Pro­duc­tiv­i­ty gen­uine­ly ac­cel­er­at­ed in the sec­ond half of the 1990s and it did pro­duce sig­nif­i­cant lev­els of growth and changed the way we did things.

Mon­ey poured in­to com­put­ing and net­works, and much of the gain was re­al. This is what makes the pe­ri­od so in­struc­tive. The the­sis was right. Most of the com­pa­nies were wrong. And no rule of thumb could sep­a­rate the two while it mat­tered.

It be­gan with Netscape. How many of you re­mem­ber those days? When the brows­er com­pa­ny went pub­lic in Au­gust 1995 and its shares rough­ly dou­bled in short or­der, Wall Street learned some­thing that it milked to the end. The pub­lic mar­ket would pay for in­ter­net growth even when there were no prof­its then or for the fore­see­able fu­ture. Mar­kets were pre­pared to pay for op­por­tu­ni­ty.

The chase

In­sti­tu­tion­al mon­ey chased the next big in­ter­net idea and soon enough the gen­er­al pub­lic joined in. By late 1999, or­di­nary in­vestors made up about one quar­ter of mar­ket ac­tiv­i­ty and the abil­i­ty to place those or­ders on­line meant that the in­ter­net it­self cre­at­ed the fa­cil­i­ta­tion of the ma­nia which it was fu­el­ing. You may see par­al­lels in that cy­cle to­day with ar­ti­fi­cial in­tel­li­gence.

Cap­i­tal was be­ing con­vert­ed in­to noise, and the noise was be­ing read as proof.

Take the time to no­tice how this worked. Each part fed the next. High­er prices made the busi­ness case re­al, which fund­ed more spend­ing, which pro­duced more rev­enue growth, which ap­peared to jus­ti­fy the high­er prices. An­a­lysts, bro­kers and in­vestors were not stand­ing out­side the ma­chine ob­serv­ing it. They were in­side it, paid to keep it run­ning. Be­ing paid to keep the dance go­ing is why you should not ex­pect the mar­ket top to an­nounce it­self.

So how did it burst?

As you might ex­pect. It isn’t a sin­gle event, but there were what I call tip­ping points. The first chink was when the Fed­er­al Re­serve be­gun rais­ing rates in 1999 and kept go­ing in­to 2000. High­er rates in­crease the cost and the avail­abil­i­ty of cap­i­tal. It low­ers val­u­a­tions and this is lethal to com­pa­nies whose whole val­ue sits in dis­tant, hoped-for prof­its.

Then the doubts ar­rived. In March 2000, the soft­ware firm Mi­croS­trat­e­gy ad­mit­ted it would have to re­state its earn­ings, and the stock fell. Ques­tions were be­ing asked and an­swers were not as eas­i­ly forth­com­ing. If this com­pa­ny’s re­port­ed growth was not re­al, whose was?

The cir­cle

Then came the mo­ment I re­mem­ber like it was yes­ter­day. Days lat­er the courts ruled against Mi­crosoft on an­titrust mat­ters, strik­ing at the in­dus­try’s an­chor. Sud­den­ly there were shares freely avail­able, buy­ers were los­ing their con­vic­tion and things be­gan to hap­pen. My per­son­al les­son is that it takes the nar­ra­tive around a re­al bell weath­er stock to move the mar­ket func­tion­al­ly in­to re­verse.

What had been a vir­tu­ous cir­cle mov­ing for­ward start­ed to run back­wards, and it ran fast. Falling prices trig­gered mar­gin calls, which forced sell­ing, which drove prices low­er still. The win­dow for new flota­tions slammed shut. Ven­ture cap­i­tal funds, starved of ex­its, hoard­ed cash for the com­pa­nies they al­ready owned and stopped feed­ing new ones. By 2002, ven­ture in­vest­ment had col­lapsed to one fifth of its in­ter­net peak. The NAS­DAQ bot­tomed that Ju­ly at 1,290, down rough­ly 74 per cent from its high.

The dam­age did not stay on the screens. The tele­coms com­pa­nies that had bor­rowed to lay ca­ble for all that promised traf­fic found the traf­fic, and the fi­nanc­ing, gone. Work­ers were laid off, or­ders were can­celled, and the US econ­o­my slipped in­to re­ces­sion in the spring of 2001. This is my sec­ond les­son. It wasn’t just the tech­nol­o­gy but the in­fra­struc­ture be­hind the tech­nol­o­gy that took a hit.

The pro­duc­tiv­i­ty fig­ures that had once flat­tered the sto­ry stopped co­op­er­at­ing. A bub­ble, it turned out, had nev­er been just a price. It was a fi­nanc­ing sys­tem, and when the price fell the whole sys­tem fell with it. The in­ter­net rev­o­lu­tion sur­vived. The first wave of in­vestors who bet on it did not.

Here is what I would take from all this, and to say up­front, it is not a method for spot­ting the next bub­ble.

The re­al sub­stance of the dot-com sto­ry is that the warn­ing signs that look so ob­vi­ous now, the ab­surd val­u­a­tions, the prof­it­less IPOs, the crazy bor­row­ing, were all plain­ly vis­i­ble at the time and were all ex­plained away by in­tel­li­gent peo­ple with good ar­gu­ments. Greenspan saw it and was three years ear­ly. The scep­tics with cap­i­tal saw it and bought any­way. The en­tire ap­pa­ra­tus of pro­fes­sion­al fi­nance looked at the same da­ta we now read as a flash­ing alarm and de­cid­ed, with con­vic­tion, that this time the growth was re­al enough to car­ry the price.

Some were even right about the tech­nol­o­gy, they ar­gued that we were in a struc­tur­al shift and that “this time it’s dif­fer­ent”. They were ru­ined re­gard­less.

If the smartest, best in­formed, best paid peo­ple in the mar­ket could not re­li­ably tell a boom from a bub­ble while stand­ing in­side it, the ra­tio­nal re­sponse is not to as­sume you will do bet­ter. It is to build a port­fo­lio that does not re­quire you to. This is the unglam­orous case for di­ver­si­fi­ca­tion.

Di­ver­si­fi­ca­tion is not, at its heart, a way to earn more. It is an ad­mis­sion, made in ad­vance and in writ­ing, that you can­not see the top com­ing and do not in­tend to try. It will feel, dur­ing every boom, like a small tax on your con­vic­tion, and that is ex­act­ly its val­ue. It is the sin­gle de­ci­sion that sur­vives be­ing wrong.

The in­vestor who con­cen­trat­ed in In­ter­net stocks in March 2000 need­ed a fore­cast to stay safe. The in­vestor who held a spread of as­sets across in­dus­tries and re­gions did not. One was bet­ting on his own fore­sight. The oth­er had qui­et­ly arranged his af­fairs so that fore­sight was op­tion­al. Over a long enough life, op­tion­al is the on­ly sort of fore­sight worth re­ly­ing on.

So the next time you feel cer­tain that a price is jus­ti­fied, that the sto­ry is sim­ply too good and too re­al to end bad­ly, re­sist the urge to set­tle the ques­tion of whether you are right. You prob­a­bly will not know if you’re right un­til the an­swer is al­ready ap­par­ent. Ask the bet­ter ques­tion in­stead. If I have read this com­plete­ly wrong, what hap­pens to me? When the hon­est an­swer is “not very much”, you have learned the most valu­able les­son from the dot-com bub­ble.

Ian Nar­ine is a fi­nan­cial con­sul­tant who prefers his bub­bles in a bev­er­age. Please send your com­ments to ian@ian­nar­ine.com