This is just out from the FFIEC:
The FFIEC Median Family Income (MFI) Report for 2019 is now available

so what do we do with this...does it help define what is LMI?

Tract level income divided by MSA or state-wide non-metro MFI determines your tract income level category. Tract level income estimates haven't been published yet, but are due out by the end of the month.

Actually the annual MFI is used to calculate the income class of borrowers in a particular year (in this case, 2019). It is not used to determine tract income class. What is used to determine tract income class is another median family income for each MSA and non-MSA adjusted every 10 years for the census and every 5 years for the ACS. In between, if a MSA, MD or statewide non-MSA has been affected by a change in the composition of a MSA or MD the median family income of the area (the denominator in the computation) the MFI is adjusted to the 10 year or 5 year interval (the decennial census or the 5-year ACS, whichever is being used at the time).

I know the foregoing is probably news to most people, not to mention confusing, but it is the way things work as applied by the Agencies.

Len, can you confirm this for me.

I beleive the tract level income categories all stay the same, but the dollar amounts will change based on the changes to the FFIEC MFI Report. So for example.

For 2018 tract XXXX.XX:
2018 Estimated MSA MFI: $62,500
Tract MFI %: 80%
2018 Estimated Tract MFI: $50,000


For 2019 tract XXXX.XX:
2019 Estimated MSA MFI: $70,000 <---- This is what just came out.
Tract MFI %: 80% <---- This only changes every 5 years after census and ACS (or if MSA changes)
2019 Estimated Tract MFI: $56,000 <---- This will change based on the new estimated MSA MFI

Is that right?

Also, does anybody know when the FFIEC geomap will be updated?

The Estimated Median Family Income for MSA's and statewide non-MSA's is changed every year which is why we call it the Annual EMFI. The 2019 number is for use when determining the applicant income class for HMDA applications processed during 2019. If you notice the FFIEC table has another field, the 2015 MSA/MD Med Family Income. That is the number that is used as the denominator (against the 2015 tract median income) to determine tract income class (notice it is adjusted to the year 2015 which was put into effect by the FFIEC in 2017, just like the 2010 Census was put into effect by the FFIEC 2 years later, in 2012).

So the number to use for tract income class determination is the 2015 data field not the 2019 data field. The only areas where that should have changed for 2019 would be those areas where the MSA or MD changed and the statewide areas that would be affected by changes to the MSA or MD. In that case the FFIEC determines the MFI for the year 2015 and publishes it in the year 2019 in which the MSA or MD composition changed.

You may have to read this a few times to understand correctly. It only took me about 20 years before I understood what was going on!

As far as the FFIEC geocoding, that probably will not change till they publish the 2019 complete demographic file which is expected late this summer (end of August I would guess).

Len S,

We have a metropolitan division area that has just changed. Now merged with another. They were always in the same MSA but no median family income is ever reported for that MSA only MD. Am I correct in believing this change will be retroactive to 1/1/19? That the geocoding will need to be adjusted before filing in March?

I only noticed the change when looking up the median incomes. I do not know when they announced the changes to the MD.

Also reading you posts, we are in the process of purchasing low to moderate income loans, the way I read your comments is that those loans should have no issues in qualifying as LMI loans because they use the 2015? Do I have that correct

Please advise.

Any thoughts anyone?

@WABComply, if you collected documentation at the time that the loan was reviewed for CD purpose then I would say that the change is not retroactive. Examiners have indicated both from the OCC and FRB that if whatever governmental tool that you are using leads you to believe it qualifies at the time, and sufficient documentation was gathered that you would be fine.

Also your last statement is confusing, are you saying the loans are from 2015?

Sorry for not responding sooner, I've been away for a few days.

When you are located in a MD the annual EMFI for the MD trumps the EMFI for the MSA when determining the income class of borrowers in a given year. The MSA/MD and non-MSA list published by the FFIEC will be the file that is used for all of 2019, not just for post publication.

Another MFI also is used for determining tract income class, but that MFI is for the year that all tract income classes were changed. The last year is 2015 which became official beginning in January 2017. So the FFIEC will establish the MFI for the MSA or MD or statewide non-MSA for 2015 and will issue the tract list with income class based on that calculation. Notice in the MSA list published a week ago by the FFIEC has the "2015 MSA/Md Med family income" field - that's what will be used for determining the tract income class. But the other field, the 2019 FFIEC Est. MSA/MD Median Family Income is what is used to determine borrower income class during all of 2019.

Hope that is clear. Maybe it's time for a webinar about this stuff.

Thanks all. I myself just got back in from vacation. Still wishing I was on it

Len,

Your explanation is exactly how I thought it worked.

However, in the recent FFIEC update last week, we had several census tracts change income level. For example, when comparing the 2018 census report for a county in which there were changes to the 2019 census tract for the county, the only thing that was different was the 2019 estimated median family income. The MSA, or in this case, non-MSA status, did not change. So the income level of the census tracts APPEAR to have changed based on the 2019 estimated MFI update. I thought the income level of the tracts would be based on the 2015 estimated MFI until the next ACS update after 2020. Where am I getting confused?

I think you are commingling 2 different issues.
1. The 2015 tract MFI won't change for any census tract until the 2020 census is put into effect by the FFIEC (likely in 2022).
2. The tract income class may change due to changes in the MSA or MD or statewide non-MSA which will affect the divisor (MSA/MD or statewide non-MSA ). Even if the MSA/MD did change and the divisor was changed the tract income class may not change. Why? Because the tract ratio may have gone from 30% to say 35% which means even though the computation affects the ratio, it may not have done it enough to change the tract income class.

BTW, we are making available free 2019 Census Tract Reference Books. Simply visit the GeoDataVision website to download whatever state's books you would like.

Thirty-three states plus the District of Columbia and Puerto Rico had tracts whose income class was changed.
Seventeen states had no tracts with income class changes
The states with the most tracts whose income class changed were New York and New Jersey
A total of 1,190 tracts had their income class changed from 2018 to 2019
A table identifying the states and the count of tracts with income class changes is available at the GeoDataVision website.