Every mania and bubble share differences and similarities, but the current wave of artificial-intelligence optimism looks a lot like what we experienced with the internet euphoria in the late 1990s.

We apologize, but this video has failed to load.

Try refreshing your browser, or
tap here to see other videos from our team.

David Rosenberg: Sorry folks, bear markets, bubbles and recessions are not things of the past Back to video

It is 100 per cent true that many of the “dot-coms” back then had no earnings and no real business model. But even the bellwethers that did have earnings and real business models got caught up in the World Wide Web mania. Their share prices shot to the moon and then came crashing down to earth, but only after the first major disappointment (God forbid, Cisco Systems Inc. missed its EPS number by a penny, which got the ball rolling) did reality begin to set in.

These megacap tech companies are still around today, but most went through a huge three-year bear market when their stock prices dove 60 per cent or more. You would have been a laughingstock to have predicted this future in the opening months of 2000. And, of course, the internet was transformational in our professional and personal lives, just as generative AI will very likely prove to be. That is not the point.

The real economy does not go through irrational exuberance — investors do

We could have said the same thing about shipbuilders, railways and the Nifty Fifty concept stocks as well, but the financial economy is not the same thing as the real economy. The real economy does not go through irrational exuberance — investors do.

Extreme market concentration

The one thing we do know with certainty is that money is an extremely emotional thing. That is what we are talking about: the extreme emotion of greed and how to make a quick buck in the equity market.

I would agree that valuations were indeed more stretched in the late 1990s. And to reiterate, companies such as Nvidia Corp. are real, not some whippy dot-com stock. But, as I say, the same was true back then with the likes of Microsoft Corp., Intel Corp., Dell Inc. and Cisco Systems.

Even though the valuations are less extreme today, we are still experiencing a phase of top-decile P/E multiples of all time. Sentiment levels are equally extreme, as is market concentration. We are up to 42 per cent of the S&P 500 market cap being dominated by tech and telecom, back to the degree of concentration prevailing in 1999 and 2000.

The market is juiced up on the excitement of the 25 per cent surge in the S&P 500 since November, only one-third of which can be explained by earnings or earnings estimates by the consensus analyst community — if this was just about earnings alone, the S&P 500 would be sitting at 4,800.

The 600-plus-point gap between that earnings-driven reality and where we are today is simply due to the anticipation of what the total addressable market for AI will look like in the future, and that, frankly, is pure speculation. Of that 25 per cent surge in the S&P 500, fully 60 per cent has come from just 10 companies and more than 20 per cent from Nvidia alone. This lopsidedness and degree of frothiness resemble the late 1990s more than any other cycle in modern history.

Bubble of its own accord

We can debate whether this bubble, in all its forms, is as big as it was back then. But it obscures the general point that we are into a gigantic bubble of its own accord. Remember, too, that the late 1990s followed a period of United States Federal Reserve easing and significant liquidity provisioning in the context of the Asian crisis, followed by concerns over the uncertainty surrounding Y2K.

The Fed quickly mopped up that liquidity and, like now, inverted the yield curve. Yet the consensus remained that we would never have a recession again — that the business cycle was defeated. The lags from the prior Fed easings percolated to the peak of the bubble in the winter months of 2000, but then the effects of the tightening reared their ugly heads. We must always consider the lags. Those who threw in the recession towel last year have failed to understand the policy lags.

I get this all the time: “But interest rates are still so low by historical standards.” That simply is not true. The 10-year T-note yield is roughly 180 basis points above the average of both the past five and 10 years; 130 basis points above the average of the past 20 years; and 40 basis points above the average of the past 30 years. Just to set the record straight.

Interest rates do matter

But the most apt comparison is with the average of the past five years, emphasizing that it is the change in interest rates and not the level that matters for growth, and keeping in mind that it takes time for the economy to reset to shifting economic backdrops in both directions. Only in 1969-1970, 1979-1980 and 1982 have we ever seen such a huge gap between the spot 10-year yield and the average over the prior five years. All recessions. Case closed.

I cannot believe the narrative that in the most credit-driven economy of all time, interest rates no longer matter. There is nothing more fundamental to anything in the economy or financial asset valuations than interest rates.

Perhaps that is why Albert Einstein once said compound interest was the “eighth wonder of the world.” Perhaps he knew a thing or two about basic math and that the laws of physics apply to capital markets and how future economic growth rates get discounted. I see no such level of sophistication today, just the classic myopic question of, “Why haven’t you been more bullish on the S&P 500?”

I will tell you with 100 per cent conviction that generative AI and the graphics processing unit (GPU) wave have not changed the relationship between discount rates and the present value of future cash flows; the impact interest rates have on the cost of capital and C-suite decisions on long-term business spending decisions; or the effect interest rates exert on the ability of households to finance expenditures on housing, autos and all other major debt-reliant durable goods commitments.

As was the case in 1999-2000, the peak hubris coincided, like today, with a general belief that the business cycle has been repealed. I can tell you one thing that is exactly the same as was the case at the bubble peak in the first quarter of 2000: the household financial asset mix concentrated in equities (including pensions) is 76 per cent, the same as it was back then. This share tops 60 per cent for the 80-million-strong retiree or near-retiree baby boomer class.

Once again, nobody has taken profits and nobody has rebalanced the portfolio. The cost of protecting against any downside in the stock market is some 40 per cent cheaper than the long-run norm, and yet few, if any, believe they need it in the mistaken belief that bear markets and recessions have become relics of the past.

Discipline and diversification have gone the way of the dodo.

David Rosenberg is founder and president of independent research firm Rosenberg Research & Associates Inc. To receive more of David Rosenberg’s insights and analysis, you can sign up for a complimentary, one-month trial on the Rosenberg Research website.

Bookmark our website and support our journalism: Don’t miss the business news you need to know — add financialpost.com to your bookmarks and sign up for our newsletters financialpost.com.