Taking Away the Punch Bowl: Can the Latest IPO Surge be a Warning Sign?

At the time of writing, 4 hours ago JPMorgan raised its S&P 500 target to 8000; this means the S&P 500 will have had an annual return of 16.8%, above its average nominal return of 10% since 1957. Goldman Sachs's research team already raised their S&P 500 forecast to 8,000 back in May, so there is a clear consensus that the market is still bullish. However, are there any reasons to argue the contrary? The short answer is no, but maybe the bubble is only just starting.

Are these a red flag for markets? That's the debate. IPOs are often seen as a late-cycle warning sign if accompanied by other key signals, such as retail involvement. Past bubbles have often been fuelled by new technologies (such as the Dot-Com bubble) and have followed equity issuance and huge capex; however, this is not a huge predictor: Jay Ritter, the Director of the IPO initiative, argues that this signal works only 52% of the time, barely better than a coin flip.

When plotting IPO raw count and highlighting when the stock market has crashed, it is quite hard to see a correlation. If we expect huge levels of IPOs to lead to a crash, we expect the red dots to be closer to the peaks; in fact, it looks like IPOs tend to decrease before the crash. The red dots tend to cluster disproportionately in the troughs and declining phases of IPO volume.

Even when zooming in on the key crashes, it is quite clear that they occur after the peak of IPOs. In fact, this is evident in historical data, as exiting the market when an IPO wave occurs would have left you with several years of gains during the Japanese historic asset price bubble. So, can we predict bubbles by adding a time lag effect? My analysis says no.

To test whether the IPO-crash relationship requires more time to materialize, I ran a regression of SPY (S&P 500) returns on IPO volume at one-, two-, and three-month lags. All three regressions are almost completely flat with even a slight positive result in the 1-month lag state, totally contradicting the hypothesis. With p-values of 0.65, 0.93, and 0.62 for 1-month, 2-month, and 3-month lags, respectively, none of these results are statistically significant, which is in line with current literature. This makes intuitive sense, as if there was an easily predictable pattern, there would be plentiful trading strategies arising from this, and such an edge would be competed away by market participants. Therefore, these results confirm that simple IPO-based heuristics are an unreliable basis for timing the market.

In addition, are we really in an IPO wave? So far, the evidence points towards no. Adding on data from 2026 so far, by raw numbers of IPOs, we aren't anywhere near the peaks seen beforehand. Although this has marked the strongest year since 2021, Own Lamont, Senior Vice President and Portfolio Manager at Acadian Asset Management, describes this as more of a recovery from an IPO drought in the previous years. IPOs haven't recovered from their pre-Internet bubble scale of roughly 300 in the 1980s and 1990s, compared to the median of 100 since.

It is argued that the current recovery in IPO volume is partly just due to the macroeconomic backdrop. The 2022 Fed rate hike led to declining valuations for firms, preventing them from going public. Since the Fed has slowly been easing rates, equity valuations have been rising, coupled with corporate profits and the economy growing. In addition, this supportive backdrop has allowed for private equity investors to take their backlog of companies public. With a record $4tn in unrealized PE portfolio value, now is a good time for them to take advantage. In fact, Ben Snider, Chief US Equity Strategist at Goldman Sachs, has reported that even without AI, this resurgence would likely still be occurring due to the macroeconomic climate.

So, what should we look for in the future? One key thing is IPO valuations. When a company prices multiples higher than their fundamentals, even modest disappointments in growth or earnings can trigger sharp downwards falls. Historically, a key example of this is the dot-com bubble. However, the median IPO last year traded at 5x EV/sales versus 9x in 199 and the median of 4x. They are clearly not that overvalued. In addition, Owen Lamont warns about large first-day pops. During the dot-com bubble, pops of 100% weren't uncommon, far exceeding normal increases of 15-20%. This is something we just haven't seen yet.

Overall, it looks like currently we aren't in a bubble when solely looking at IPO volume. However, if you could predict bubbles based on one metric alone, we'd be able to prevent every crisis. Although as seen before, IPO volume alone isn't a good predictor, we should be cautious if they rise a lot further. Firms are smart: they want to sell equity when equity is overvalued; if we see a large rise, particularly in one industry, it could be a sign that a bubble has begun. In addition, if the warning signs of huge first-day pops and larger multiples arise, maybe have a think. But for now, we can carry on partying. No need to take the punch bowl away just yet.