.
We own real assets, generally due to diversification benefits and inflation hedge properties in particular.
The first test of modern real assets began in 2021 and took over two years to see inflation to levels not seen in generations.
A practitioner might ask, “Did the actual assets work as expected during this episode?”
Managers' returns are undoubtedly high variance, and real asset index data, a large market, suggests that real assets could not hedge the 2021-2023 inflation episodes.
In this blog, we review the performance of three representative indices of asset classes that allocators may include in real buckets: S&P Global Infrastructure Index (SPGI), S&P Natural Resource Index (SNRU), and Northern Trust Global Real Real Real ESTATE INDEX (NTGRE), Multi Asset Northern Trust Real Assets Arlocation (NTRAA), and S&P Real Assets Indexes (SP_REAL). It uses a period of surge inflation that began in 2021 and ended in 2023.
For comparison, we include Bloomberg tips (Bbutistr, short for “Tips”), Bloomberg Commodity Total Return (BCTR), and S&P 500 (SPXTR) index. My measure of inflation is the consumer price index (CPI) defined below and the variables based on it. Unless otherwise stated, returns and levels change are monthly. The R code and additional results are listed in the online R markdown file.
What should inflation hedge be?
Most investors probably expect inflation hedges to be compensated for drugs that may impose on the portfolio compared to stocks, at least in the form of returns that withstand changes in price levels.
Asset allocators typically hold potential inflation hedges to a more generous standard. Ask that hedges only positively correlate with inflation. In other words, when price levels rise, inflation hedges are also required.
Both standards have seen actual assets fade during recent inflationary episodes.
Actual assets and inflation in the COVID era
Exhibit 1 points to my main point. Horizontal Axis and Multi-Asset Northern Trust Real Asset Allocation Index Headlines CPI Inflation Changes[1] (Vertical) For inflation in the COVID era, it is defined between January 2021 and December 2023.
The correlation is close to zero and is actually slightly negative (-0.04) due to the optimum lines of normal least squares (OLS). The results are the same in the S&P Real Assets Index. Of course, these results are not important. The sample size (36) is small.
But what's interesting is not the hypothetical test, but the actual value. The return on the wider real set benchmark did not move in the same direction as inflation from 2021 to 2023.
Figure 1. Headline CPI and a wide range of real asset benchmark indexes were not uncorrelated during inflation during the Covid era.
Source: Fred, Ycharts, Author Calculations
Table 1 shows the correlation table. It shows that there was a real asset index return during the inflation period in the Covid era Negatively It is associated with CPI inflation (column 3) in the heading, as well as hints and stocks. Actual assets moved in the wrong direction on average as they changed inflation.
It is also shown in Table 1 It's at the root Inflation: Median average CPI and (16%) trimmed average CPI calculated by the Federal Reserve Bank of Cleveland. These proxies of sustained inflation (captured by the latest macrophilips curves) generally associated with rising production gaps or inflation expectations. They are a measure of trend inflation as they exclude supply shocks from various sources (Ball and Mazumder, 2008). And I include the traditional core, or Ex. Another measure of food and energy inflation, inflation trends or underlying trends.
By any of these definitions of trend inflation, the actual assets were less fundamental inflation hedge than the headline inflation hedge during the 2021-2023 inflation episode.
Table 1. Select the correlation between asset classes and inflation measures from 2021 to 2023 (n = 36).
| ntraa | SP_REAL | SPGI | snru | Tip | BCTR | ntgre | SPXTR | |
| Median_cpi | -0.3 | -0.34 | -0.17 | -0.21 | -0.35 | -0.3 | -0.35 | -0.33 |
| trimmed_mean_cpi | -0.2 | -0.23 | -0.11 | -0.11 | -0.26 | -0.11 | -0.23 | -0.28 |
| CPI | -0.03 | -0.07 | -0.01 | -0.02 | -0.17 | 0.03 | -0.04 | -0.09 |
| core_cpi | -0.17 | -0.15 | -0.14 | -0.16 | -0.08 | -0.09 | -0.14 | -0.17 |
| headline_shock | 0.11 | 0.09 | 0.06 | 0.08 | -0.01 | 0.17 | 0.12 | 0.06 |
Source: Fred, Ycharts, S&P Global, Author Calculations
Finally, we define heading shocks in a normal, modern way. The difference between headlines and underlying inflation is that the underlying inflation proxy is the median CPI. The results are variables that show episodes of supply shock inflation and development, as shown in Figure 2.
Figure 2. Headline shocks are positive, as in 1990 and early 2020s, and unfavourable or negative, as in the mid-1980s.
Author's calculations, Fred's source
Actual assets respond slightly better to shock headlines than underlying inflation. The coefficients for actual asset variables are generally higher than the coefficients in the broader stock market (SPXTR and TIPS). Extend the sample to the longest general period (2016-2024, n = 108) to reinforce these conclusions (Table 2).
Table 2. Select the correlation between asset class and inflation measures over the longest common period (2015-12/2024, n = 109).
Source: Fred, Ycharts, S&P Global, Author Calculations
Using this long dataset, inflation beta can be calculated in traditional ways by regressing the returns of CPI inflation (using OLS). These betas are statistically and economically insignificant, as shown in Table 3. The median CPI regression results worsen for actual assets. Coefficients are incorrect signs (more negative online supplements).
Table 3. inflation beta estimates and its uncertainty (n = 109).
* R-squared is zero in each case.
Source: Fred, Ycharts, S&P Global, Author Calculations
Investors will probably be less interested in correlation or beta than the actual (or lacking) actual assets performance during the inflation episode. Here, the story is also disappointing for anyone who expects inflation protection from actual asset classes during the COVID inflation period. As shown in Chart 3, of the actual assets, only natural resources (SNRU, light green line) have grown more cumulatively than CPI inflation (orange line), but just barely. Of the broad set of indexes considered, only products can “beat” inflation.
Figure 3. Cumulative growth, 2021-2023.
Source: YCHARTS, S&P Global, Author Calculations
Actual assets failure
Since at least the 2000s, actual assets and inflation protection strategies have been equipment for sophisticated asset pools. After decades of dormancy, high inflation has resurfaced in 2021. The institutional investors probably felt ready. But they may have been disappointed instead.
debate among economists whether Covid inflation is the result of supply shocks, demand shocks, or both (see, e.g. Bernanke and Blanchard, 2023; Giannone and Primiceri, 2024). The “truth” may take years to reveal.
Asset allocators can now draw conclusions to the extent that the index used in this article represents manager returns and future actions of actual assets during the surge in inflation. When inflation arrived, the actual assets failed.
reference
Ball, LM and Mazumder, S. (2019), “Non-puzzling behavior of median inflation”, Nber Working Paper, No. 25512
Bernanke, B. and Blanchard, O. (2023), “What causes inflation during the US pandemic era?”, Nber Working Paper, 31417.
Giannone, D. and Primiceri, G. (2024), “Drivers of post-pandemic inflation”, Nber Working Paper, No. 32859
[1] https://www.northerntrust.com/united-states/what-we-do/investment-management/index-services/index-performance/equity/real-assets-a location-index
