These census tracts had statistically significant high-hardship clustering in both 2013 and 2019. The pattern is not random: a Global Moran’s I of 0.684 confirms that hardship is spatially concentrated, not scattered. Tracts in this category share infrastructure deficits, limited employment access, and concentrated poverty that reinforce one another across neighborhood boundaries.

Implication: Individual-level interventions alone are unlikely to move the needle. Place-based, multi-sector investment is required.

Persistent hot spots are located mostly in the central and southern areas of Phoenix, making it a corridor of long-term hardship. These areas remained as places of hardship over time, meaning that hardship is likely not random but instead structural.

These tracts were not significant hardship clusters in 2013 but became statistically significant by 2019, representing the spatial expansion of hardship beyond historically distressed cores. This is an early warning signal that hardship is spreading, not contained.

Displacement paradox: Some “improving” tracts nearby may be gentrifying, pushing lower-income households outward into these emerging clusters. Declining hardship scores do not necessarily mean existing residents are better off.

Emerging hot spots are located around Mesa, as well out the outer areas of Phoenix. This could be due to displacement from rising costs in the more central areas, making those with lower-income move outward.

The majority of tracts improved over the decade-long window. However, aggregate improvement masks significant variation: 31.1% of tracts worsened over the same period. The data cannot distinguish genuine economic uplift from population turnover: a tract with a declining hardship index may simply have replaced lower-income residents with higher-income newcomers.

Data limitation: Before drawing conclusions from improving scores, ground-truth verification through community engagement and displacement tracking is essential.

Many tracts did improve on the whole, but the improvement is uneven. Some areas still show signs of worsening or consistently high hardship, meaning that not all areas are are benefiting from the improvement in the same way.

Target: 158 Persistent Hot Spot tracts (17.8% of all Maricopa tracts): spatially concentrated, entrenched hardship confirmed by a Global Moran’s I of 0.684.

Critical areas to target would be central and southern Phoenix, since these areas contain 158 persistent hot spot tracts with high hardship index values. With a Global Moran’s I of 0.684, these areas show high spatial clustering rather than random. I would recommend that Maricopa County (which has divisions in community development) and the City of Phoenix put in place investments that are targeted and location specific such as affordable housing preservation, expansion of transit services, and workplace development initiatives to help long-term structural hardship.

Target: 24 Emerging Hot Spot tracts (2.7% of all Maricopa tracts), new high-hardship clusters not present in 2013, signaling spatial expansion.

Emerging hot spots appear in Mesa and the western part of Maricopa County, likely due to the rising cost of central urban areas displacing those with lower-income. This embodies the displacement paradox, showing that rather than resolving hardship in an area, improvement is due to pushing the hardship out of the area instead. The establishment of a system that gives early warnings, perhaps run by the county or planning divisions, might be able to track rent rates, population changes, and other indications of hardship. This system could be paired as well with interventions like rental assistance to help further prevent the rising hardship levels in these areas.

Evidence base: 53.7% of tracts improved (2013→2023) but 31.1% worsened; Moran’s I = 0.684 confirms strong spatial clustering persists.

With the use of the trajectory map and mobility Sankey, it shows that neighborhood change is not even throughout the region, meaning that monitoring over time is necessary. A possible solution would be making a cross-sector data coordination system that involves housing, transportation, and social service agencies that can track changes within neighborhoods and intervene as issues arise, allowing for earlier, more dynamic intervention rather than making interventions for data gathered periodically, since that data could already be outdated by the time change is implemented.

The baseline EHI consists of 3 measures: Poverty + Unemployment + Income (inv.)

Current index: 7-component EHI: Poverty + Unemployment + Income (inv.) + Renter Burden + Low Ed. Attainment + Food Insecurity (SNAP) + Transp. Disadvantage

Q1: What changed spatially?
Compare Hot Spot tract counts and cluster map patterns between your expanded index and the 3-component baseline. Did adding Renter Burden + Low Ed. Attainment + Food Insecurity (SNAP) + Transp. Disadvantage shift which tracts or corridors are flagged?

After adding the additional variables, the expanded index shows a more widespread distribution of hardship in Maricopa County in comparison to the 3-component baseline. The amount of HH tracts increased from 201 to 267, showing this increase in areas considered to be hot spots, most of which pop up around existing HH cluster areas like Mesa and Glendale. The fact that adding additional variables increased the amount of HH like this shows that those variables portray types of hardship that weren’t captured by the 3-component baseline, so now shows a more broad and encompassing picture of hardship.

Q2: What stayed the same?
Which Persistent Hot Spot areas appear robustly across index specifications? What does consistency across different index compositions tell us about the reliability of hardship diagnoses in those tracts?

With the expanded index, the main areas that show persistent hardship stayed mostly the same. Central and southern Phoenix show large groups of high hardship in both indexes. The amount of persistent HH areas only increased from 150 to 158, showing that these areas are the most entrenched in hardship even with the expanded index. Because of this, the hardship in these persistent areas don’t seem to be dependent on what variables you’re measuring with, but show that the cause of the hardship is more structural, and the areas will show as disadvantaged no matter what. This makes them more reliable areas to target with intervention.

Q3: Policy implications of index choice
If a policymaker targeted place-based investments using the baseline index versus your expanded index, would resource allocation differ? Name specific tracts or geographic corridors and argue which composition better captures the full burden of economic hardship for policy purposes.

The choice in index would make a difference in resource allocation. The components used in the baseline index would focus resources on improving areas with economic hardship like central and southern Phoenix. The expanded index increases the amount of HH tracts from 201 to 267, and instead of just economic hardship, also places focus on housing, food, and transportation instability. This expanded index broadens the area of focus to sections like Mesa and Glendale. While these expanded areas aren’t as severely disadvantaged economically since they weren’t HH areas in the baseline index, including them gives a broader scope and better grasp of how hardship is present, making the extended index the best choice when wanting to know the “full burden” of hardship within the region.