Sandalwoods are one of the world’s most valuable trees used to harvest fragrances and oil.
Unfortunately, they take 20 years to mature!
Contrast this with paper somniferum or better known as opium poppy farming. Poppies are easy to cultivate, grow in almost any soil and within 120 days produce opium, the source of morphine and heroin.
Stock investors often behave like either sandalwood or poppy farmers.
But unlike poppy farmers that focus on production and not self-consumption, certain investors gorge and intoxicate themselves with cortisol (fear) and dopamine (greed). We believe that the current market backdrop is lending itself to greed narcotics.
Howard Marks’ eloquently said, ’we may never know where we are going, but we’d better have a good idea where we are’’. And we think where we are is engraved in the annals of history. This article reflects George Santayana’s warning, ’those who cannot remember the past are condemned to repeat it’.
2025 has features from the 1970’s and the 1990’s
The geopolitics and economics of the 1970’s look spookily like 2025 and the AI boom has features of the tech mania of the late 1990’s.
The economics of the 1970’s versus 2025
The 1970’s carried the scars of a world order that was changing. The Vietnam war led to higher US government debt, the Fed raised interest rates in the late 1960’s to quell inflation, which led to a stock market collapse in 1969, and the United States ended Bretton Woods in August 1971, effectively suspending the convertibility of US dollars for gold.
Subsequent oil shocks in 1973 and 1979 fueled gloom with growing unemployment and stubbornly high inflation. From 1970 to 1975 US GDP growth averaged 2.5%, albeit spotty, inflation averaged 5.7% and unemployment averaged 6.3%.
Fast forward to 2025, the past 4 years has seen solid US GDP growth at 3.5%, with forecasts closer to 2% in the three years ahead. Unemployment is rising (3.4% in January 2023 to 4.3% in August 2025), inflation remains sticky at around 3% having been as low as 2.3% in April and inflation expectations are on the up.
Both periods are moments of change – about turns in long-standing economic orders, Bretton Woods in 1971, protectionism rather than free trade in 2025.
The geopolitics of the 1970’s versus 2025
Critically, both periods reflect moments when a president arm wrestled the chair of the Federal Reserve to adjust monetary policy for political gain. In 1971 Richard Nixon appointed Arthur Burns as Fed Chairperson, manipulating him to cut interest rates to ensure Nixon’s reelection in 1972. Burns reluctantly capitulated, cut rates, and catalyzed an inflationary pulse that led to 8.5% average inflation between 1975 and 1980.
Trump is taking Nixon bullying to a whole new level in 2025. His objective is to control the committee of the Fed, there are seven governors and as president, he can nominate the chair and two vice chairs that have 4-year terms. The senate then confirms these nominations. Jerome Powell’s term ends on the 4th of May 2026.
Michelle Bowman, nominated by Trump as a vice chair, took office on the 9th of June 2025. Philip Jefferson is the other vice chair, a Biden nominee whose term ends in September 2027. Trump is currently in a legal battle to have Lisa Cook removed as one of the seven governors. If he achieves this, he will have nominated the majority at the Fed and his goal to drive down interest rates to fuel economic growth, might materialize!
Nixon’s influence over Burns is unambiguous evidence of the pitfalls of interfering with the delicate arts of monetary policy, but the risks to such a ploy in 2025 due to elevated levels of US government debt, are seismically larger than they were in the 1970’s.
Elements of market behaviour in 2025 that mirror the late 1990’s
The US market in 2025 is showing signs akin to the dot.com period from 1996 to 1999.
In 1999, the top thirty stocks in the S&P constituted 42% of the capitalization of the index against 54% today. Making matters worse, most of this concentration is within the largest ten companies. Normally, the top stocks in the S&P account for 20% of the market capitalization of the index.
In both 1999 and 2025, the theme of market concentration is technology – currently, the eight largest companies in the S&P are all IT businesses, in 1999, five of the ten largest stocks in the index were tech businesses.
The number of stocks within the S&P 500 index that outperform the index annually tends to be around 40%. In a good year it rises well above 60% but has fallen below 30% only three times since 1974 (since we have data) – in 1980 (34%), in 1998 and 1999 (mid 20%), in 2023 (29%) and 2024 (19%). The consequence of this narrow leadership is that these companies have attracted outsized amounts of capital at the expense of other areas of the market – this lends itself to the concept of crowding – think of a dual carriage highway for capital to get in, that could turn into a goat track when that capital tries to get out!
Chart 1: Percentage of S&P 500 stocks outperforming the S&P 500 over the calendar year
Source: S&P Dow Jones Indices
In both the late 1990’s and again over the last couple of years, the market has bifurcated with respect to technology businesses. Then, like now, start-ups and unprofitable new listings are being richly valued and anticipated to grow quickly. Against this, are larger, established businesses, generating elevated levels of profitability trading at premiums to the rest of the stock market. Optically, these businesses are appropriately valued, accounting for their growth rates. Below we show how the largest software companies have grown their profitability at a rate to justify their market pricing.
Chart 2: AI Cloud infrastructure and platforms (META, ADBE, MSFT, GOOGL, AMZN and ORCL)
Source: Bloomberg and Northstar as at 31 August 2025
Two examples of AI based businesses trading on P/E ratings exceeding one hundred are Duolingo and Palantir. These 100 plus P/E businesses rival the valuations of tech listings of the 1990’s such as Pets.com, eToys, Excite@Home and Webvan, that are no longer in existence.
Based purely on a simple P/E metric, the mega tech stocks in 2025 are optically, not as expensive as they were in 1999. As shown in the graph below, Microsoft traded at 61 times in 1999; it is at 37 times today. Similarly, Qualcomm reached a P/E of 278 times versus sixteen times today, Cisco 148 times and now twenty-four times, whereas Oracle reached 104 times versus fifty-four times in 2025. As mentioned above, what we have seen since the dotcom burst is that the mega-cap technology survivors have delivered profit growth to justify their share price valuations, as stock performance has in fact been, on the whole, poor and only really rewarded investors over the last couple of years!
Explanation of the graph below: it shows the performance of tech stocks from 1991 until their highest price point, just before the market crash in March 2000. These are the graphs on the left of the slide. The data above the graph shows the stock performances and P/E’s of each company at the height of the bubble. The performance and ratings of those same stocks are shown on the right hand side of the slide, with the starting point being the peak of the market in 2000. Their annualized returns since the 2000 peak and each stock’s change in P/E over time is also shown – from 2000 until today. What should be evident is that investors only really started making positive returns (with the exception being Amazon) on the technology companies that they held through the dotcom crash, 16 years later in 2016.
Chart 3: Dot com: Spectacular returns (earnings + rerating) … but then muted
Source: Bloomberg and Northstar as at 31 August 2025 (? indicates no earnings)
A final, but crucial point on valuations is that the earnings levels for technology companies could currently be cyclically high. Consequently, companies are more expensive than they look at first glance. AI is new and untested technology with hordes of disciples and few disbelievers. A land grab has ensued with gargantuan capital expenditure that faces a real risk of subpar prospective returns on capital.
My last remarks on market similarities between the two periods comes down to price performance. The Nasdaq doubled its value in the three years between May 1996 and May 1999. It then doubled again between May 1999 and March 2000!
Fast forward to August 2016 and the Nasdaq doubled in value from then to January 2023. It then doubled in value again over the past two years and eight months.
How does it all come together?
Markets are currently facing the confluence of two factors, a tectonic global technological shift, with a populist and interventionist political order. This is fertile territory for extreme hype and bubble-like market events.
If Trump drives down interest rates, the risk to markets is that easy money will propel an enormous spike in prices for riskier assets. The probability of such an event is elevated.
How is Northstar navigating these tricky times?
We value every business using realistic assumptions for future growth and valuations.
Based on the reasons we explained above, the potential for certain sectors and specific companies to experience blow-out share price performances similar to the dotcom period in the months or years ahead is plausible. But the subsequent returns from these stocks over the decade post the market crash of 2000 demonstrates clearly that investors would have been better off never participating in the market euphoria in the first place.
Our global portfolios are reducing exposure to overvalued technology companies and investing in sectors and businesses where we have calculated that capital will appreciate over a 5-to-7-year period. Very often these are the least sexy areas of the market!
Source: S&P Dow Jones Indices