The Federal Reserve Bank of Atlanta’s Commercial Real Estate Market Index (CREMI) provides a quarterly indication of commercial real estate (CRE) market performance at the Core-Based Statistical Area (CBSA) geographic level and across various asset types. A diffusion index, the Commercial Real Estate Market Diffusion Index (CREMDI), is also provided. The Atlanta Fed also publishes quarterly articles highlighting CREMI data.
CREMI is a standardized index, and each CBSA and asset type combination has its own index based on its history. Zero represents the long-term index average of CRE performance in a given geography and asset type. Values above zero indicate from a risk perspective that a market is stronger than its long-term average based on the selected variables. Conversely, values below zero indicate a market that is weaker than its long-term average. Please see the “About CREMI” tab for more details.
The chart below shows the median CREMI score across all CBSAs (black line) and the time series of scores for each by asset type. Filters can be used to toggle selections, including selecting all CBSAs in a given state. Some CBSAs have a shorter data history than others.
This chart shows the distribution of CREMI scores across CBSAs for relevant asset types at a point in time. CREMI covers about 390 metro areas and divisions.
This chart displays CREMI scores for all asset types for a selected CBSA to facilitate comparative analysis across asset types.
This network graph allows users to identify the top three inputs that drive CREMI results for a given CBSA and asset type. Each grey node represents a single CBSA. Hovering over the node highlights the top three inputs to that CBSA’s CREMI score. Contributing inputs are identified by the larger colored nodes, identified in the legend. When a CBSA is highlighted, numbers appear in the input nodes that represent the weight of the input toward the CREMI score for that CBSA. The size of the input nodes indicates how many CBSAs connect to a given node. Multiple filters are available for further exploration.
CREMDI is a diffusion index which represents what percentage of the eight CREMI input variables contribute positively to commercial real estate performance at a point in time. Each variable’s contribution is measured by quarter-over-quarter changes. A value of zero means all variables had a quarter-over-quarter change in a direction that negatively affected CRE, while a value of 100 indicates all variables changed in a direction that positively affected CRE. A value of 50 indicates an even split between positive and negative changes.
The chart below shows CREMDI scores for CBSAs. Users can filter by CBSA and asset type.
This chart averages CREMDI scores (unweighted) for all CBSAs in a given state and by asset type.
The Federal Reserve Bank of Atlanta’s Commercial Real Estate Market Index (CREMI) provides a quarterly indication of commercial real estate market performance at the Core Based Statistical Areas (CBSA) geographic level and across various asset types.
CREMI is standardized so that the long-term average of commercial real estate performance is centered at zero. Values above zero indicate a market that is stronger compared to its long-term average. This is a positive performance indicator for that CBSA. Conversely, values below zero indicate a market that is weaker compared to its long-term average. This is a negative performance indicator. The farther market trends are from zero, the stronger or weaker they are.
Specifically, CREMI represents the standard deviation (SD) of commercial real estate performance around the long-term average. Given this, one would expect 68 percent of values for a given geography and asset type to be distributed between +1 and -1 SD, 95 percent distributed between +2 and -2 SD, and 99 percent distributed between +3 and -3 SD. Extreme values can and do occur, particularly in times of economic stress (the 2008 Great Recession) or rapid economic recovery (the 2021 COVID-19 recovery).
To provide further clarity, five scenarios are described:
Note: Holistic CRE market analysis is nuanced and requires context and information from various sources, some of which may not be easily quantified or captured in CREMI. Meaningful trends observed in CREMI are meant to prompt users to look into specific markets in detail rather than draw wide-ranging conclusions about that market.
CREMI is derived from multiple commercial real estate asset type performance data and local-level economic data through use of a dynamic factor model (DFM). As part of data setup, each variable’s sign is adjusted so that the level or change of variables that contribute positively to commercial real estate performance receive a positive sign and variables where the level or change contributes negatively receive a negative sign. Once the variables’ signs are adjusted, they are standardized (mean = 0, standard deviation = 1) and the first principal component is calculated. The reported CREMI is a Kalman filtered smoothed version of the first principal component. The R package "DFMs"is used for the CREMI model.
To provide additional information on trends in commercial real estate performance, the Commercial Real Estate Market Diffusion Index (CREMDI) represents the percent of variables where the quarter-to-quarter change is generally considered a positive for performance. Ranging from 0 to 100 percent, a value of 0 means all variables had a quarter-to-quarter change in a direction that negatively affects commercial real estate, while a value of 100 indicates all variables changed in a direction that positively affects commercial real estate. A value of 50 percent indicates an even split between positive and negative changes.
To help identify potentially longer-term trends in the data and reduce quarter-to-quarter volatility, a smoothed version of the CREMDI is also provided. This is simply the diffusion index with a locally estimated scatterplot smoothing (LOESS) regression applied.
Please submit any questions to cremi.support@atl.frb.org
The commercial real estate and economic measures used are all related to specific CBSAs and are obtained through public and nonpublic data sources. All variables were chosen based on their importance in reflecting commercial real estate market performance, reporting frequency, and on their availability across all CBSAs. Specifically, variables with type “CRE” in the table below are sourced from CoStar Group, and the data provider does not allow the Federal Reserve Bank of Atlanta to share them externally.
| Asset Type: Office | |||
| Variable Name | Type | Transformation |
|
| Occupancy Rate | CRE | Level |
|
| Net Operating Income Index | CRE | Log Change (YoY) |
|
| Market Cap Rate | CRE | Level |
|
| Asset Return | CRE | Percent Change (YoY) |
|
| Unemployment Rate | Economic | Level |
|
| Office Employment | Economic | Percent Change (YoY) |
|
| Industrial Production | Economic | Percent Change (YoY) |
|
| CPI | Economic | Percent Change (YoY) |
|
| Asset Type: Multifamily | |||
| Variable Name | Type | Transformation |
|
| Occupancy Rate | CRE | Level |
|
| Net Operating Income Index | CRE | Log Change (YoY) |
|
| Market Cap Rate | CRE | Level |
|
| Asset Return | CRE | Percent Change (YoY) |
|
| Unemployment Rate | Economic | Level |
|
| Households | Economic | Percent Change (YoY) |
|
| Industrial Production | Economic | Percent Change (YoY) |
|
| CPI | Economic | Percent Change (YoY) |
|
| Asset Type: Retail | |||
| Variable Name | Type | Transformation |
|
| Occupancy Rate | CRE | Level |
|
| Net Operating Income Index | CRE | Log Change (YoY) |
|
| Market Cap Rate | CRE | Level |
|
| Asset Return | CRE | Percent Change (YoY) |
|
| Unemployment Rate | Economic | Level |
|
| Retail Employment | Economic | Percent Change (YoY) |
|
| Industrial Production | Economic | Percent Change (YoY) |
|
| CPI | Economic | Percent Change (YoY) |
|
| Asset Type: Industrial | |||
| Variable Name | Type | Transformation |
|
| Occupancy Rate | CRE | Level |
|
| Net Operating Income Index | CRE | Log Change (YoY) |
|
| Market Cap Rate | CRE | Level |
|
| Asset Return | CRE | Percent Change (YoY) |
|
| Unemployment Rate | Economic | Level |
|
| Industrial Employment | Economic | Percent Change (YoY) |
|
| Industrial Production | Economic | Percent Change (YoY) |
|
| CPI | Economic | Percent Change (YoY) |
|
| Asset Type: Hospitality | |||
| Variable Name | Type | Transformation |
|
| Occupancy Rate | CRE | Level |
|
| Average Daily Rate (ADR) | CRE | Log Change (YoY) |
|
| Market Cap Rate | CRE | Level |
|
| Asset Return | CRE | Percent Change (YoY) |
|
| Unemployment Rate | Economic | Level |
|
| Hospitality Employment | Economic | Percent Change (YoY) |
|
| Industrial Production | Economic | Percent Change (YoY) |
|
| CPI | Economic | Percent Change (YoY) |
|
CREMI Data (.csv) This file contains CREMI scores for each CBSA and asset type over time.
CREMDI Data (.csv) This file contains CREMDI scores (raw as well as smoothed) for each CBSA and asset type over time.
Weights Data (.csv) This file contains CREMI input weights as of the most recent quarter for each CBSA and asset type. The weights indicate how much each input variable contributes to the CREMI score for a given CBSA and asset type at a point in time.