At first glance, the August consumer price index (CPI) report, versus the fever raging over the past few months, didn’t look catastrophic. The headline reading of 8.3% showed a tiny increase of 0.1% that, combined with a similar decline for July, marked a two-month flattening for prices. That’s a big downshift from the superhot trend for the first half of 2022, a span where the CPI was surging at an average of 1% a month, or a 12% annualized clip. But a CPI that continued to trend sideways wasn’t the jubilant news investors expected. The July data had stoked optimism that the Fed’s interest rate hikes were already taming the beast, the peak was past, and consumer prices would meaningfully fall in August, sending the U.S. on a glide path to the Fed’s 2% target and creating space for the central bank to ease up. When Sept. 13 data disappointed, the Wall Street bulls stampeded in retreat, sending stocks to their steepest one-day selloff in over two years.

In reality, the outlook for future months is worse than the “no increase” August numbers that so rattled investors. Going forward, we’re likely to see the CPI return to the pattern of serial increases. The reason? Housing costs, the biggest force by far in moving the CPI, are already rapidly raising Americans’ living expenses. But because of the way they’re measured, a large portion of those increases aren’t yet counted in the index. “That component is growing and will continue to grow,” says Campbell Harvey, a finance professor at Duke University. “It’s inflation that’s already happened, but isn’t yet reflected in the CPI. That’s part of the reason that future inflation will be elevated and persistent. The view that the story is just about supply chain and political risk, and that we’ll quickly get back down to 2% or 3%, is misguided.”

The CPI’s methodology makes it highly probable the number will stay high in the months to come

Harvey distinguishes between two factors that will determine future CPI readings. The first are what he calls the “mechanics,” or how the CPI is calculated. Those levers pretty much set the numbers for the next few months in stone, and they aren’t pretty. The second are the “structural” factors dominated by the rising tide of housing and rental costs that will increasingly swell the readings over a longer horizon.

On the mechanics, Harvey explains that the change in year-over-year inflation depends on two things: The size of the month-over-month increase 12 months ago, and the rise in the current month. When the increase in the current month is lower than that of a year earlier, the annual CPI “observation” must drop. That explains the dramatic fall in July that ignited high hopes. Prices rose that minuscule 0.1%, and since the bump in July of 2021 was a much bigger 0.48%, the index fell big-time from 9.1% in June to 8.5%. Likewise, in August the 0.1% increase was less than the 0.30% gain in August of 2021, triggering a lesser fall to 8.3%. Put simply, the bigger month-over-month increase a year ago, the more the CPI will decline if the current month is flat. Those mechanics explain the decrease from 9.1% to 8.3% from June to August.

But 8.3% is still a huge and troubling number. So what do the mechanics tell us about where the CPI is headed in September, the last reading before the November elections? Let’s make the hopeful assumption that inflation runs at a 3% annual rate in September. That would add 0.25% to prices over August. Since the increase in September of 2021 was an almost identical 0.27%, inflation would remain precisely the same at that alarming 8.2%. Using the even rosier forecast that prices stay flat at August levels in September, the number isn’t much better at 8.0%. Bad readings are baked in for months to come. If prices rise at that 3% annualized rate through December, the CPI that month would register 7.6%, almost four times the Fed’s target.

Housing costs will keep inflation elevated for a long time

The Bureau of Labor Statistics methodology for measuring housing costs explains why they’re not fully reflected in the current index, and will keep it high for a long time. “Shelter” carries by far the biggest weight in the CPI at 32%, dwarfing No. 2–ranked food at 13%. The BLS uses two measures for the expense of “shelter,” based on surveys of 50,000 residences. The first is “rent of primary residence.” It’s what apartment dwellers or renters of single family homes are paying now, in the month that’s being measured. For homeowners, the BLS doesn’t use home prices, which have waxed at nearly 20% overall in the past year. Instead, it deploys “owner equivalent rents.” To get those numbers, the BLS takes what renters are now paying for apartments, and adjusts those payments for the size, age, renovations, and neighborhoods of homes to calculate what owners would pay to rent their abodes. Indeed, the growing institutional single family home rental market is a good gauge for that calculation.

That methodology accounts for the big lag between what people are paying right now on their current apartment leases, or a hypothetical lease for a single family home, and what a family pays right now for a new lease—in other words, the current, real-time cost of housing. For the past year, the Zillow rent index has been showing double-digit increases. As of August, the year-over-year number was 12.3%. Early this year, CPI dwelling cost increases were running at an annualized 3% and 4%. In August, the number was around 7%. “Here’s the difference between today’s housing costs and the CPI number,” Harvey explains. “If you signed a 12-month lease in October of 2021, your rent’s locked in until this October. New rents are rising in double-digits, but you’re not paying the extra costs. Then you get a new lease in October, and your rent jumps, say, 10%. You suffer that hit all at once.” Harvey says that those rolling increases, as more and more leases are renewed at much higher rents, will close the gap between the current CPI numbers and much higher rents in the marketplace, pushing the index higher.

A combination of rising rents and food prices will keep inflation entrenched

Housing costs are abating somewhat. In recent months, the Zillow increases are closer to around 10%. Still, that’s a lot higher than the 7% the CPI’s been reporting in the past couple of months. At one-third of the index, a 10% rise in “rents” equates to an annualized rise of 3% in the CPI, and for Harvey, recurring, built-in increases of that size are likely to last for up to a year. The rub is that energy, the famous force that’s held overall CPI constant in the past few months, is unlikely to make the same index-crunching contribution in the future. “Gasoline accounts for only 5% of the index, but it almost solely accounts for the slightly negative overall reading in July and August,” says Harvey. In August alone, gas fell 10.2%. “I don’t see that we’ll get the same weight of collapse again in the price of gas,” he adds.

On the other hand, food prices increased by an annual rate of almost 10% in August, on top of a jump of 13% in July. “I don’t see any indication that the rise in food prices is abating,” says Harvey. “The index may start rising again, unless we get another steep decrease in energy prices.” By mid-2023, Harvey predicts, inflation will still be running significantly above the Fed’s comfort level, at well over 4%. The main factor likely to keep the CPI rising is the catchup in those undercounted housing costs. That’s the spoiler the optimists are missing.