Blog
Who said inflation was transitory? YouTubers' calls, checked
By Callsight · Published Oct 10, 2026 · Updated Oct 10, 2026
Figures as of Nov 4, 2026. They change as new windows settle.
Transitory was the word the Federal Reserve, chaired by Jerome Powell, used for rising inflation in 2021, and Treasury Secretary Janet Yellen used it too. In 2021, the four finance YouTube channels Callsight tracks made four counted calls on where US inflation was going; at one year, all three that said it would rise are Correct (base rate 47.1%) and the one that said it would fall is Incorrect (base rate 53.1%). As of Nov 4, 2026, across all 52 of their counted inflation calls, 30 of the 43 decided as Correct or Incorrect at one year are Correct, against a base rate of 49.5% on the same calls; 17 of those are 2022 calls that it would fall, which one turn in the rate decided together.
Who said inflation was transitory?
The Federal Reserve’s policy committee, then chaired by Jerome Powell, used the word in its statements. On Apr 28, 2021, its statement said inflation had risen largely because of transitory factors; on Nov 3, 2021, it still expected the factors behind the rise to be transitory; on Dec 15, 2021, the word was gone. At the Treasury, Secretary Janet Yellen told reporters after a G7 meeting on Jun 5, 2021 that she believed the rise reflected transitory factors.
Inflation kept rising. Measured the way Callsight’s rule measures it (the consumer price index against the same month a year earlier, as first released), it went from 1.3% in January 2021 to 7.0% in December 2021 and peaked at 9.0% in June 2022. The Bureau of Labor Statistics’ own headline for June 2022, before seasonal adjustment, was a 9.1% rise over 12 months, the largest since the year to November 1981. The most recent figure, for September 2026, is 3.7%.
This article asks the same of four finance YouTube channels: what they said about US inflation, in their own words, and how each call turned out.
What did finance YouTubers say about inflation in 2021?
Callsight counts four calls on the direction of US inflation that these channels made in 2021: three said it would rise and one that it would fall. At one year, all three that said it would rise are Correct (base rate 47.1%); the one that said it would fall is Incorrect (base rate 53.1%).
The one 2021 call that said inflation would fall, made on May 17, 2021, between the Fed’s April and November statements:
“by the end of the year the prices of things will start to come back down personally I believe the latter”Reading when said
5%
May 17, 2021
Reading at close
8.54%
May 17, 2022
Change
+3.54 pp
Inflation stood at 5.00% at the start of the window and 8.54% at May 17, 2022, a change of +3.54 pp (CPIAUCSL, first releases). The call was that it would fall. Moves smaller than 0.5 pp are treated as no change.
30 days
Within band
90 days
Within band
180 days
Incorrect
1 year
Incorrect
The words set a deadline, which the card dates to Dec 31, 2021, and the call is Incorrect there. The headline is still the one-year window. At one year the call is Incorrect, as it is at 180 days; at 30 days and 90 days it is Within band; at 3 years and 5 years it is Correct.
The same video holds a second counted call, said 66 seconds later and filed as a call that inflation would rise: its card. At one year it is Correct. It is among the calls whose reading is questioned below.
The first 2021 call that said it would rise:
“My thinking is that most likely we will end up seeing higher inflation, we'll see slightly higher interest rates, and that will have an impact on the real estate and stock market.”Reading when said
2.65%
Mar 22, 2021
Reading at close
8.65%
Mar 22, 2022
Change
+6.00 pp
Inflation stood at 2.65% at the start of the window and 8.65% at Mar 22, 2022, a change of +6.00 pp (CPIAUCSL, first releases). The call was that it would rise. Moves smaller than 0.5 pp are treated as no change.
30 days
Correct
90 days
Correct
180 days
Correct
1 year
Correct
No deadline was stated; the headline is the one-year window. At one year the call is Correct, as it is at 30 days, 90 days, 180 days, 3 years and 5 years.
Did anyone predict the inflation of 2021 and 2022?
Most of the calls made before 2022 said inflation would rise. Before 2022, the record holds 11 counted calls on the direction of US inflation: eight said it would rise and three that it would fall. At one year, of the eight that said it would rise, seven are Correct and one is Incorrect (base rate 47.2%); all three that said it would fall are Incorrect (base rate 52.7%). Most came from one channel: Graham Stephan made ten of the 11.
Those counts say less than they seem to. The rule checks only the direction of the yearly rate: a call that inflation would rise is Correct once the rate climbs by more than 0.5 pp, whether the words foresaw a small rise or a very large one. The calls of 2020 and 2021 were also checked over windows ending in 2021 or 2022, so the same surge decided them together.
Did anyone predict inflation would fall in 2022?
Most of the calls made that year did. In 2022 the channels made 23 counted calls on the direction of US inflation: 18 said it would fall and five that it would rise. At one year, of the 18 that said it would fall, 17 are Correct and one could not be checked (base rate 51.0%); all five that said it would rise are Incorrect (base rate 48.9%). Joseph Carlson made seven of the 18 that said it would fall.
The 2022 calls were made within months of one another, about the same published figures, over one-year windows that overlap, so the turn in the rate after June 2022 decided most of them together. The calls that it would fall also matched what Fed officials projected: in June 2022 their median projection had inflation, on the PCE measure the Fed targets, falling from 5.2% in 2022 to 2.6% in 2023.
One call from each side, chosen by the rule at the end of this article:
“i'm still in the viewpoint i still maintain the same viewpoint that i think there's a very good chance inflation will be lower by the end of 2022 than at its peak i think it will be much lower”Reading when said
8.54%
May 27, 2022
Reading at close
4.06%
May 27, 2023
Change
−4.48 pp
Inflation stood at 8.54% at the start of the window and 4.06% at May 27, 2023, a change of −4.48 pp (CPIAUCSL, first releases). The call was that it would fall. Moves smaller than 0.5 pp are treated as no change.
30 days
Within band
90 days
Within band
180 days
Correct
1 year
Correct
The words name a date, the end of 2022, which the record did not register as the call’s deadline, so the headline is the one-year window. The windows at 180 days and 1 year end in November 2022 and May 2023. At one year the call is Correct, as it is at 180 days and 3 years; at 30 days and 90 days it is Within band. Its window at 5 years is still open.
“so what this means is that inflation is likely to surge over the winter and this will force central banks around the world to continue raising interest rates aggressively”Reading when said
8.25%
Sep 19, 2022
Reading at close
3.61%
Sep 19, 2023
Change
−4.64 pp
Inflation stood at 8.25% at the start of the window and 3.61% at Sep 19, 2023, a change of −4.64 pp (CPIAUCSL, first releases). The call was that it would rise. Moves smaller than 0.5 pp are treated as no change.
30 days
Incorrect
90 days
Incorrect
180 days
Incorrect
1 year
Incorrect
The words name a span, the winter of 2022 to 2023, which the record did not register as the call’s deadline, so the headline is the one-year window. The windows at 90 days and 180 days end in December 2022 and March 2023. At one year the call is Incorrect, as it is at 30 days, 90 days, 180 days and 3 years. Its window at 5 years is still open.
Four of the five 2022 calls that said inflation would rise, and five of the 18 that said it would fall, are among the calls whose reading is questioned further down.
How accurate were these inflation calls overall?
Of the 52 counted calls, 25 said US inflation would rise and 23 that it would fall. The other four show no direction: each was recorded as naming a level that the figure for the month the words were said in had already passed, so the rule refused to judge them and they are Not verifiable. At one year, the 52 stand at 30 Correct, 13 Incorrect, 3 Within band, 1 Pending and 5 that could not be checked.
Channel by channel, the calls Correct at one year out of those decided there, each with the base rate of the same calls: Graham Stephan 12 of 17 (base rate 49.3%; four of the 17 questioned below), Coin Bureau six of 12 (base rate 49.1%; five of the 12 questioned below), Joseph Carlson eight of eight (base rate 50.4%; three of the eight questioned below) and Andrei Jikh four of six (base rate 49.7%; one of the six questioned below). The samples are small, so a single call moves any of these a long way, and 17 of the 30 Correct calls are 2022 calls that inflation would fall, which one turn in the rate decided together.
| Channel | Said | Calls | 30 days | 90 days | 180 days | 1 year | 3 years | 5 years |
|---|---|---|---|---|---|---|---|---|
| Graham Stephan | It would rise | 9 | 1 of 3 | 2 of 4 | 3 of 5 | 7 of 9 | 6 of 8 | 6 of 7 |
| Graham Stephan | It would fall | 8 | 2 of 3 | 4 of 6 | 5 of 7 | 5 of 8 | 5 of 7 | 1 of 3 |
| Coin Bureau | It would rise | 12 | 0 of 4 | 1 of 6 | 0 of 5 | 2 of 8 | 1 of 5 | — |
| Coin Bureau | It would fall | 4 | 2 of 2 | 3 of 3 | 3 of 3 | 4 of 4 | 4 of 4 | — |
| Andrei Jikh | It would rise | 3 | 2 of 2 | 2 of 2 | 1 of 1 | 1 of 3 | 0 of 2 | — |
| Andrei Jikh | It would fall | 4 | — | 1 of 1 | 2 of 2 | 3 of 3 | 2 of 2 | — |
| Joseph Carlson | It would rise | 1 | — | — | — | 1 of 1 | — | — |
| Joseph Carlson | It would fall | 7 | 4 of 4 | 6 of 6 | 7 of 7 | 7 of 7 | 7 of 7 | — |
| Base rate, calls that it would rise | — | 48.7% (12) | 45.0% (11) | 47.6% (21) | 41.7% (15) | 42.7% (7) | ||
| Base rate, calls that it would fall | — | 50.0% (16) | 53.7% (19) | 51.3% (22) | 57.6% (20) | 57.2% (3) | ||
The same calls by the year they were made, at one year:
| Year made | Said it would rise | Said it would fall | ||||||
|---|---|---|---|---|---|---|---|---|
| Calls | Correct | Incorrect | Base rate | Calls | Correct | Incorrect | Base rate | |
| 2019 | 1 | 0 | 1 | 47.3% | 0 | 0 | 0 | — |
| 2020 | 4 | 4 | 0 | 47.2% | 2 | 0 | 2 | 52.5% |
| 2021 | 3 | 3 | 0 | 47.1% | 1 | 0 | 1 | 53.1% |
| 2022 | 5 | 0 | 5 | 48.9% | 18 | 17 | 0 | 51.0% |
| 2023 | 5 | 0 | 3 | 48.5% | 2 | 2 | 0 | 51.6% |
| 2024 | 3 | 2 | 1 | 46.9% | 0 | 0 | 0 | — |
| 2025 | 3 | 2 | 0 | 46.2% | 0 | 0 | 0 | — |
| 2026 | 1 | 0 | 0 | — | 0 | 0 | 0 | — |
The inflation page lists 109 calls from these four channels, and the record counts 52 of them. Of the rest, 39 are forecasts read out from someone else, listed with their source and never counted. The other 18 are the channels’ own but not scored: 11 whose direction could not be read with confidence and 7 held after a review questioned how they were recorded.
How is an inflation call judged?
An inflation call is judged on the US consumer price index for all urban consumers, seasonally adjusted (CPIAUCSL on FRED), as the change from the same month a year earlier, with each figure as first published rather than as later revised. A call that inflation will rise or fall compares the figure for the month the words were said in with the one for the last month inside the window. A move smaller than 0.5 pp either way is Within band, neither Correct nor Incorrect. A call that names a level is Correct if any monthly figure inside the window reaches it. A call recorded as naming a level that the figure for that month, usually published weeks after the words, had already passed is Not verifiable.
Each call is quoted word for word and linked to the second it was said, then checked at six windows, from 30 days to 5 years; the one-year window is the headline, fixed in advance. A hedged call (inflation could or might rise) that still names a direction is checked like any other. The rules are the same for every channel, and the methodology page gives them in full. Callsight makes no forecast of its own for inflation, and nothing here is advice on what to buy or sell.
What is the base rate for an inflation call?
It is how often inflation moved the called way anyway. For each call, the method takes every past window of the same length that had closed by the day the words were said, back to 1980 in the same series at first release, leaves out the windows in which the rate moved less than 0.5 pp, and counts the share that moved in the called direction.
At one year it averages 47.6% over the 21 decided calls that said inflation would rise, and 51.3% over the 22 that said it would fall: close to an even chance either way. Beside those, 11 of the 21 calls that said it would rise are Correct, and 19 of the 22 that said it would fall.
The base rate has a limit: it does not know where inflation stood when the words were said. It counts every past window since 1980, not only those that began near a 40-year high, so it could understate the odds of a fall for a call made in 2022. The method computes no base rate by starting level, and neither does this article.
Which calls might Callsight have misread?
For this article every counted call was read again on Oct 10, 2026, with the passage around its words, whatever its verdict. In 14 of them, Callsight’s reading, which decides the verdict, may not match what was said. Each link opens the call:
- Filed as a call that inflation would fall, though its words, read in full, point the other way: Coin Bureau on Apr 11, 2022. At one year, as recorded: 1 Correct.
- Words about inflation staying high or prices continuing to rise, filed as calls that the yearly rate would climb further: Graham Stephan on May 17, 2021; Andrei Jikh on Feb 28, 2022; Coin Bureau on Nov 14, 2022; and Coin Bureau on May 26, 2024. At one year, as recorded: 1 Correct and 3 Incorrect.
- Words about one part of inflation (shelter, services, goods or food), judged on the headline rate: Joseph Carlson on Oct 10, 2022; Graham Stephan on Nov 18, 2022; Graham Stephan on Dec 14, 2022; Graham Stephan on May 11, 2023; Coin Bureau on Jul 17, 2025; and Coin Bureau on Mar 30, 2026. At one year, as recorded: 4 Correct, 1 Incorrect and 1 Pending.
- Words about what would be best, or what the market would need, filed as calls that inflation would fall: Joseph Carlson on Jun 15, 2022 and Joseph Carlson on Sep 6, 2022. At one year, as recorded: 2 Correct.
- Words about the next monthly figure, due within days, judged over the fixed windows instead: Coin Bureau on Mar 7, 2022. At one year, as recorded: 1 Incorrect.
Leaving all 14 out would change the one-year count from 30 of 43 Correct to 22 of 30, beside a base rate of 49.6% on those calls. The verdicts stand as recorded. Only a published correction or a new version of the rules can change one, and this article does not. The corrections page says how anyone can ask for one.
What else did these channels predict?
Their calls on a recession, a crash and a housing crash are checked in the article on who predicted the recession, and each channel’s whole record is set beside its base rate in the article on whether finance YouTubers beat the market. Their calls on the Fed’s rate are on the Fed rate page, and each channel’s page lists its calls: Graham Stephan, Coin Bureau, Andrei Jikh and Joseph Carlson.
How these examples were chosen
The cards were chosen on Oct 10, 2026 by a fixed rule, from the counted calls of 2021 and 2022. For each year and direction, the rule takes the earliest call whose words, read in full, forecast which way US inflation would go. A call was passed over if a review or a correction about it was open; if it is among the calls questioned above; if its words gave a hope, a preference, a what-if or a mechanism rather than a forecast; or if they named a date or a span the record did not register as the deadline, unless the windows ending either side of that date give the same verdict.
Figures are computed by code from the same record as the rest of the site. This article was drafted with the help of AI tools and reviewed by a person before publication. It is not investment advice.