Transportation Performance Management

State Highway Safety Report (2021) - Michigan

The following provides a summary of the Highway Safety Improvement Program's (HSIP) safety performance measures and State safety performance targets. As per the Safety PM Final Rule, States are required to set annual safety performance targets in the HSIP annual report for the number of fatalities, rate of fatalities per 100 million vehicle miles traveled (VMT), number of serious injuries, rate of serious injures per 100 million VMT, and number of non-motorized fatalities and serious injuries. The safety performance targets are based on 5-year rolling averages. States have the flexibility to use the methodology they deem most appropriate when establishing safety performance targets. FHWA encourages States to review data sets and trends and consider factors that may affect targets. The safety performance targets should be data-driven, realistic, and attainable and should align with the performance management framework and legislative intent.

A State Department of Transportation (DOT) has met or made significant progress towards meeting its safety performance targets when at least four of the five safety performance targets established under 23 CFR 490.209(a) have been met or the actual outcome is better than the baseline performance. The baseline performance is the 5-year average ending with the year prior to the establishment of the target.

The Basis for Target and Additional Comments are provided by the State in their HSIP Annual Report and have not been edited by FHWA. Any questions about individual State reports should be directed to the respective State DOT. For additional information about each State's HSIP, the complete reports are available at https://highways.dot.gov/safety/hsip/reporting.

More information and resources on Safety Performance Management are available at https://highways.dot.gov/safety/hsip/spm/safety-performance-management-safety-pm-overview.

All State data used to populate the State Highway Safety Reports for 2021 are available for download at https://www.fhwa.dot.gov/tpm/reporting/state/tpm_dashboard_data.zip.

• 

  Number of Fatalities

• Number of Fatalities	2015	2016	2017	2018	2019	2020	2021	2022	2023
  Annual	967	1,065	1,031	977	986	1,086	1,136
  5-Year Average					1,005.2	1,029.0	1,043.2
  Target (5-Year Average)							968.6	1,065.2	1,105.6

• Basis for Number of Fatalities Target

  • 2021 Target
  • 2022 Target
  • 2023 Target

  To determine a forecasted value for the five-year rolling average for the first four measures, the decision was made to use the change model created by UMTRI used for establishing previous targets. UMTRI predicts 886 fatalities in CY 2020, and 967 in 2021. The change model predicts change in fatalities from the previous year based on several predictors. This log-change regression model is tied closely to whatever happened recently, so it cannot diverge very far from the current time unless we predict many years out into the future. In the future, the change model predicts a steady (slow) decrease in fatalities. The dataset is a set of differences from one year to the next within the state, expressed as a percentage of the previous year. Thus, the predictors can influence exposure and/or risk. The count model, however, directly predicts counts so it could diverge from observed by a lot if the patterns change in the real world. Based on known factors the count model shows a steady increase in fatalities through 2025. As this is not what is expected the change model was selected in developing the targets. This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.

  To determine a forecasted value for the five-year rolling average for the first four measures, the decision was made to use the change model created by UMTRI used for establishing previous targets. UMTRI predicts 967 fatalities in CY 2021, and 1065 in 2022. The change model predicts change in fatalities from the previous year based on several predictors. This log-change regression model is tied closely to whatever happened recently, so it cannot diverge very far from the current time unless we predict many years out into the future. In the future, the change model predicts a steady (slow) decrease in fatalities. The dataset is a set of differences from one year to the next within the state, expressed as a percentage of the previous year. Thus, the predictors can influence exposure and/or risk. The count model, however, directly predicts counts so it could diverge from observed by a lot if the patterns change in the real world. Based on known factors the count model shows a steady increase in fatalities through 2025. As this is not what is expected the change model was selected in developing the targets. This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.

  To determine a forecasted value for the five-year rolling average for the first four measures, the decision was made to use the change model created by UMTRI used for establishing previous targets. UMTRI predicts 1065 fatalities in CY 2022, and 1105 in 2023. The change model predicts change in fatalities from the previous year based on several predictors. This log-change regression model is tied closely to whatever happened recently, so it cannot diverge very far from the current time unless we predict many years out into the future. In the future, the change model predicts a steady (slow) decrease in fatalities. The dataset is a set of differences from one year to the next within the state, expressed as a percentage of the previous year. Thus, the predictors can influence exposure and/or risk. The count model, however, directly predicts counts so it could diverge from observed by a lot if the patterns change in the real world. Based on known factors the count model shows a steady increase in fatalities through 2025. As this is not what is expected the change model was selected in developing the targets. This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.

  
  

• Data Sources:
  Fatalities: 2015-2020 Final FARS, 2021 FARS Annual Report File
  Targets: 2020-2022 Michigan HSIP Annual Reports

• 

  Fatality Rate (per 100 million VMT)

• Fatality Rate (per 100 million VMT)	2015	2016	2017	2018	2019	2020	2021	2022	2023
  Annual	0.99	1.07	1.01	0.95	0.97	1.25	1.17
  5-Year Average					0.998	1.050	1.070
  Target (5-Year Average)							0.982	1.098	1.136

• Basis for Fatality Rate Target

  • 2021 Target
  • 2022 Target
  • 2023 Target

  VMT values have been predicted for CYs 2019, 2020 and 2021. VMT estimates for CY 2020 are reduced due to COVID-19. Using the fatal injury values, along with the respective predicted VMT, the forecasted fatality rates are 1.040 for CY 2020, and 0.945 for CY 2021. This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.

  VMT values have been predicted for CYs 2020, 2021 and 2022. VMT estimates for CY 2020 are reduced due to COVID-19. Using the fatal injury values, along with the respective predicted VMT, the forecasted fatality rates are 0.945 for CY 2021, and 1.051 for CY 2022.This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.

  VMT values have been predicted for CYs 2020, 2021 and 2022. VMT estimates for CY 2020 are reduced due to COVID-19. Using the fatal injury values, along with the respective predicted VMT, the forecasted fatality rates are 1.051 for CY 2022, and 1.136 for CY 2023.This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.

  
  

• Data Sources:
  Fatalities: 2015-2020 Final FARS, 2021 FARS Annual Report File
  VMT: 2015-2021 FHWA Highway Statistics Series, VM-2 Table
  Targets: 2020-2022 Michigan HSIP Annual Reports

• 

  Number of Serious Injuries

• Number of Serious Injuries	2015	2016	2017	2018	2019	2020	2021	2022	2023
  Annual	4,865	5,634	6,084	5,586	5,629	5,433	5,979
  5-Year Average					5,559.6	5,673.2	5,742.2
  Target (5-Year Average)							5,533.6	5,733.2	5,909.2

• Basis for Number of Serious Injuries Target

  • 2021 Target
  • 2022 Target
  • 2023 Target

  The model predicts 4,960 serious injuries in CY 2020, and 5,409 in 2021. While serious injuries have fluctuated over the past several years, the linear relationship of the ratio of serious injuries and fatalities (A/K) going back to 2003 is still evident. However, this trend suggests a greater reduction in serious injuries than being observed. Therefore, a linear model using the last eight year of data was used which projects a flattening pattern. This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.

  The model predicts 5,409 serious injuries in CY 2021, and 5,673 in 2022. While serious injuries have fluctuated over the past several years, the linear relationship of the ratio of serious injuries and fatalities (A/K) going back to 2003 is still evident. However, this trend suggests a greater reduction in serious injuries than being observed. Therefore, a linear model using the last eight year of data was used which projects a flattening pattern. This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.

  The model predicts 5,673 serious injuries in CY 2022, and 5,909 in 2023. While serious injuries have fluctuated over the past several years, the linear relationship of the ratio of serious injuries and fatalities (A/K) going back to 2003 is still evident. However, this trend suggests a greater reduction in serious injuries than being observed. Therefore, a linear model using the last eight year of data was used which projects a flattening pattern. This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.

  
  

• Data Sources:
  Serious Injuries: 2022 Michigan HSIP Annual Report
  Targets: 2020-2022 Michigan HSIP Annual Reports

• 

  Rate of Serious Injuries (per 100 million VMT)

• Rate of Serious Injuries (per 100 million VMT)	2015	2016	2017	2018	2019	2020	2021	2022	2023
  Annual	4.97	5.67	5.98	5.46	5.51	6.28	6.18
  5-Year Average					5.518	5.780	5.882
  Target (5-Year Average)							5.609	5.892	6.058

• Basis for Serious Injury Rate Target

  • 2021 Target
  • 2022 Target
  • 2023 Target

  VMT values have been predicted for CYs 2019, 2020 and 2021. VMT estimates for CY 2020 are reduced due to COVID-19. Using the fatal injury values, along with the respective predicted VMT, the forecasted fatality rates are 1.040 for CY 2020, and 0.945 for CY 2021. The annual serious injury rates of 5.822 for CY 2020, and 5.287 for CY 2021. This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.

  VMT values have been predicted for CYs 2020, 2021 and 2022. VMT estimates for CY 2020 are reduced due to COVID-19. Using the fatal injury values, along with the respective predicted VMT, the forecasted serious injury rates are 5.287 for CY 2021, and 5.778 for CY 2022.This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.

  VMT values have been predicted for CYs 2020, 2021 and 2022. VMT estimates for CY 2020 are reduced due to COVID-19. Using the fatal injury values, along with the respective predicted VMT, the forecasted serious injury rates are 5.778 for CY 2022, and 6.058 for CY 2023.This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.

  
  

• Data Sources:
  Serious Injuries: 2022 Michigan HSIP Annual Report
  VMT: 2015-2021 FHWA Highway Statistics Series, VM-2 Table
  Targets: 2020-2022 Michigan HSIP Annual Reports

• 

  Number of Non-Motorized Fatalities and Serious Injuries

• Number of Non-Motorized Fatalities and Serious Injuries	2015	2016	2017	2018	2019	2020	2021	2022	2023
  Annual	761	740	798	740	794	742	698
  5-Year Average					766.6	762.8	754.4
  Target (5-Year Average)							771.2	791.6	743.4

• Basis for Number of Non-Motorized Fatalities and Serious Injuries Target

  • 2021 Target
  • 2022 Target
  • 2023 Target

  Results from the UMTRI model as described (the fatality and serious injury relationship) were also used to generate non-motorized forecasted annual values of 714 for CY 2020, and 799 for CY 2021. This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.

  Results from the UMTRI model as described (the fatality and serious injury relationship) were also used to generate non-motorized forecasted annual values of 799 for CY 2021, and 762 for CY 2022.This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways. The annual forecasted values for CY 2021 and CY 2022 along with the actual values from CY 2018 to 2020 to determine the 2022 Targets (five-year rolling average) are shown in the table. In addition, actual values dating back to CY 2012 are included as part of the determination of the 2020 baseline condition.

  Results from the UMTRI model as described (the fatality and serious injury relationship) were also used to generate non-motorized forecasted annual values of 762 for CY 2022, and 743 for CY 2023.This supports the SHSP by identifying Michigan's key safety needs and guide investment decisions to achieve significant reductions in traffic fatalities and serious injuries on public roadways.
  The annual forecasted values for CY 2022 and CY 2023 along with the actual values from CY 2018 to 2020 to determine the 2023 Targets (five-year rolling average) are shown in the table. In addition, actual values dating back to CY 2012 are included as part of the determination of the 2020 baseline condition.

  
  

• Data Sources:
  Fatalities: 2015-2020 Final FARS, 2021 FARS Annual Report File
  Serious Injuries: 2022 Michigan HSIP Annual Report
  Targets: 2020-2022 Michigan HSIP Annual Reports

Additional Comments

2021 Comments:

The annual forecasted values for CY 2020 and CY 2021 along with the actual values from CY 2017 to 2019 to determine the 2021 Targets (five-year rolling average) are shown in the table. In addition, actual values dating back to CY 2011 are included as part of the determination of the 2019 baseline condition.

Safety Performance Target Assessment

PLEASE NOTE: Each State’s safety performance target assessment is based on its own State-specific target methodology and program philosophy. Therefore, conclusions should not be drawn based only on the information in the Safety Performance Target Assessment Summary table. For example, the State may have set aggressive targets, and not met those targets, while another State may have set more easily attainable targets, and met those targets. FHWA understands that each State’s safety program is unique and therefore does not prescribe a methodology for States to set targets. States have the flexibility to use the methodology they deem most appropriate when setting their safety performance targets.