Don’t fight the Fed…or the ECB, the BOJ, etc. Central banks have cast a spell over stocks this year, but their magic alone isn’t what keeps the rally going.

Despite the recent stumble of technology stocks, the S&P 500 is still above where it was on Feb. 19, when fears about Covid-19 shook investors. An oft-cited cause for this flash recovery is low interest rates: With inflation-protected Treasurys now yielding around a record-low minus 1.3%, the S&P 500’s earnings yield of above 4% looks irresistible. Tech firms with high growth potential should gain even more from rock bottom long-term rates.

Investors appear to understand that their companies won’t make much money over the next few years if a selloff in dividend futures is to be believed. According to a report by the Bank for International Settlements this week, it lends credence to the idea that markets are pricing stocks on the basis of a return to normal over the longer term.

A fundamental tenet of textbook finance is that the value of an asset comes from all its future cash payments, discounted by the return that investors would get by placing cash in a safe instrument instead. So far, that seems to fit the facts: When the “discount rate" is as low as it is now, it may not be worth selling shares to buy low-yielding Treasurys, even if executives have explicitly warned that dividend payments will be scant.

But investors shouldn’t take this too far.

Utility stocks point to this rally not being exclusively about discount rates. These steady income payers usually trade in lockstep with bonds, but they have severely underperformed the S&P 500 in recent months. It looks like the market has reasonably decided that utilities were overbought coming into 2020 and that low rates aren’t enough to change that. It could well happen in other sectors too.

To be sure, based on historical precedent, the spread between utilities’ earnings yields and bond yields, known as the “equity risk premium," is wide enough to suggest they can stage a comeback. Still, this premium remained elevated for years after the 2008 crisis.

For the rest of the market, trying to justify asset values based on risk-free interest rates gets even harder because, even though the equity risk premium is now historically high, it doesn’t appear to hover around any given level.

The Credit Suisse Research Institute suggests that over a very long timespan—between 1910 and 2019—high premiums did tend to narrow, and vice versa, but only with the hindsight benefit of knowing what was “high." At each point in time, using past data to determine whether the spread was high enough to trigger a stock rally wouldn’t have worked.

Textbook finance has always struggled to explain why equity risk premiums are so high and hard to predict. Research published in June by hedge fund AQR Capital Management even argues against the popular notion that the underperformance of lower-valued sectors against the likes of tech has to do with a low level of rates, though it does admit to finding some connection with rates being cut. Central banks have done so over the past decade, but they can’t go much lower from here.

The lesson may be that uncertainty about the future prevents rational discounting of cash flows at any given rate. Instead, both numbers are often concocted to justify investment decisions that already have been made. The prospect of a more digital post-pandemic economy, if sustained, can well justify the optimism fueling this year’s 25% rise in the Nasdaq Composite. Low rates on their own are a flimsy argument.