April 26, 2016

Paper review: “Incentives for promoting Battery Electric Vehicle (BEV) adoption in Norway”

Electric and hybrid vehicles accounted for more than 60% of the new cars in Norway last month, which fortifies the country’s position as the global frontrunner of vehicle electrification.  This level of enthusiasm among consumers could be enabled by the multiple incentives offered by Norwegian government: (1) vehicle purchase tax exemption, (2) value added tax (VAT) exemption, (3) vehicle license fee reduction, (4) road tolling exemption, (5) free parking on municipal public parking, (6) free ferry ticket and (7) access to bus lane. A paper published recently in Transportation Research Part D analyzed the effectiveness of these incentives for promoting BEVs and how different groups of consumers respond to different incentives based on a survey of BEV owners in Norway.

1. A quick summary of the paper

The paper is based on  a survey of 3400 BEV owners. To identify the importance of the incentives for their decision of purchasing their BEVs, in the survey the respondents were asked to rate the importance of the incentives with ten levels (1-not important at all, 10 –extremely important.) To identify the critical incentives for each respondent’s purchase decision,they were asked whether they would still have bought the BEV if a certain incentive was not offered. The following figure shows the distribution of critical factors and the mean values of the importance ratings of each incentive.

• Importance ratings and critical factors

It shows that exemptions from purchase tax and VAT were critical incentives for more than 80% of the respondents: upfront price reduction has been the most powerful incentive for promoting BEV adoption in Norway. Vehicle license fee reduction and road tolling fee exemption are critical for about 50% of the respondents. Free parking, bus lane access and free ferry tickets are not as important incentives, but critical to some respondents.

• Combinations of critical incentives

16% of the respondents reported that none of these incentives were critical for their purchase. 9% reported that only one single incentive was critical. Among the seven incentives, purchase tax exemption still dominates, but exemption from road tolling and access to bus lane account for substantial portions of BEV owners’ only decisive factor. Most respondents reported multiple critical incentives, with half identifying 4 or more as shown in the following figure.

• Model the types of incentives – the target groups of the incentives

The researchers used factor analysis to group the incentives into three categories in this paper: reduction of fixed costs (RFC) including purchase tax exemption, VAT exemption and vehicle license fee reduction; reduction of use costs (RUC) including free parking, free ferry ticket and road tolling exemption; and priority to infrastructure (PRI), as in free access to bus lanes. To find out the potential target groups of these three incentive types,  logistic regressions models were estimated, as shown in the following table:

The positively significant predictors of being in RFC: being male, above 45 years old, Tesla buyers, most recent buyers (less than 1 year). Income levels do not significantly predict belonging to this target group, suggesting that among the BEV owners, RFC incentives are important for all income-level buyers.

The positively significant predictors of being in RUC: higher education (College/University), lower income , respondents living in or near the city of Trondheim(this might be because road tolling is intensive in Trondheim), not Tesla owners

The positively significant predictors of being in PRI: Female buyers, younger buyers, higher income, earlier years, not Tesla owners, higher education and those in neighboring communities to Oslo (this might be because of the heavy traffic to/from Oslo during rush hours).

2. Evaluation 

The data of this analysis were well suited to the  purpose of examining the role of incentives for promoting BEVs. Being one of the few studies that quantitatively measure the effectiveness of the incentives in a market with relatively high penetration rate, this paper provides some valuable insights: (1) the up-front cost reduction is the most powerful incentive for promoting BEVs (2) reduction of use cost, such as tolling exemption is a critical incentive for those with relatively lower income (3) priority to infrastructure is a critical incentive for those with relatively higher income (4) the geographic characteristics are important when it comes to the efficiency of an incentive policy. However, there are still several unsatisfying elements in this analysis:

(1) Overreaching in conclusions: Do incentives favor high income group? 

According to the results of the logistic regression model for the RFC group (table 8), income is not a significant predictor, suggesting that among the sampled BEV owners, RFC incentives are important for all income-level buyers. The authors took this as evidence against the prominent argument that BEV incentives only favor those in relatively higher income groups.  However, this is a stretch since the survey sample itself is limited to  BEV owners, who tend to have higher incomes. In the paper, the authors compared the sample (3384 BEV users) and new car users (2012-2014 models), showing that the proportion of low income among new car users (33%) is much higher than that of the EV users (10%).

(2) More important question: do these incentives induce rich people to buy more cars? 

Earlier research points out that these incentives could have encouraged people with high income to buy a second car, which brings more cars to the road system and causes congestion in bus lanes and parking difficulties in the city center. This question is not answered in this paper, but could potentially be addressed quantitatively by examining the difference of vehicle ownership of BEV buyers in different income groups.

(3) More discussion could be added: what if one/several of the incentives are removed in the future? 

One prominent discussion about the EV incentives in Norway is that the high cost of the incentives could not be offset by the reductions in CO2 emissions, which is why there has been doubt on how long Norway can sustain its ecosystem of strong incentives for PEV market. Analysis and discussion on what will happen when some/all of the incentives were removed would be an interesting – this could be potentially analyzed quantitatively by this paper but was not.

It is stated in the paper that 16% of the respondents reported that none of the incentives was critical for their decision of BEV purchase, but  these observations were deleted from the analysis later. It would be an interesting analysis and discussion to find out the characteristics of the respondents who did not buy the BEV because of any incentives.

 

(4)  Choice of model 

At the end of the discussion part, the authors talked about the low explanatory power of the models. The possible reasons identified by the authors were:

• Important variables were not included in the models,  such as attitudinal variables (e.g. environmental values)
• Large proportion of the respondents (60%) purchased their BEVs within last year.
• “It is simply difficult to identify general patterns among respondents”

Environmental awareness could be a significant predictor of the purchase decisions, but it is unclear why it would influence BEV buyers’ evaluation of incentives.

The paper does not explain the methodology of the model very clearly, but the authors appear to have used three one-vs-rest binary logistic regression models instead of one multinomial logistic regression to model categorical dependent variables (RFC, RUC, and PRI).  A multinomial logistic model imposes the constraint that all the predicted probabilities add up to 1 by estimating a joint model instead of a stratified model, so it generates smaller standard errors than separate binary logistic models. A mixed logistic model might also help capture the heterogeneity and improve the model’s goodness of fit.

Even though the paper has its limitations, it does offer a very important takeaway message: even though up-front cost reduction is a critical incentive for promoting BEVs,  there is a high proportion of consumers respond to other types of incentives. Multiple incentives may be necessary to sway many consumers to adopt EVs.