How Much Does Successful PBF Cost?

Benjamin Loevinsohn's picture
January 22, 2013

There’s been considerable discussion about how costly PBF is and whether it provides value for money. How much money per capita is required for it to be successful? The common wisdom is that it takes about an additional 3 dollars per capita per year in a low-income country for PBF to be successful, but the justification for this amount is not particularly strong. There are very few studies which compare different levels of financing, so it’s difficult to know what would be required. Given all the other factors that can influence the success of PBF besides the pure incentive effect or the amount of money available at the health facility level, it’s not possible to be definitive about the amount needed. It may be that a dollar or $1.50 per capita per year might have a similar effect to $3. It may be that the smaller amounts are 80% as effective as investing twice as much.

In some places the signaling function of PBF may be what is most important and signaling can be accomplished at relatively low cost. It’s possible that the incentives that have been used in some low-income countries, which have amounted to more than 30% of health workers’ take-home salaries, are unnecessary and that smaller amounts could be just as useful. One commonly-considered factor is the size of PBF incentive provided to health workers as a proportion of their total earned income (not just their government salary). There is considerable variation in this proportion and it does not appear to determine the success of the PBF scheme. For example, in Zambia the PBF incentives amount to less than 15% of health worker’s salaries and it is not obvious that Zambia has been less successful than other PBF schemes which provide quite a bit higher incentive as a proportion of health worker earnings.
 It’s interesting to note that in OECD countries, incentive schemes for improving performance, particularly quality of care, have used incentives amounting to less than 5% of providers’ earnings. This may explain the difference in the results observed in OECD countries compared to low-income countries.  It may also suggest that below some threshold, performance bonuses simply fail to get the attention of health workers.
One possible way of looking at this issue of how much effective PBF really costs would be to include different arms in randomized trials that have different levels of financing. This would provide a sound empirical basis to understand what results those different levels of financing achieve.
 Another way of studying this issue might be to use some of the regression discontinuity that arises from the fact that districts or provinces have discreet remoteness bonuses. If the remoteness bonuses in two contiguous provinces are different, there will probably be some health facilities in the provinces that are similarly remote and have similar characteristics but receive different amounts of remoteness bonus. It would be interesting to see whether the difference in financing between such facilities leads to differences in productivity or overall performance.
Another useful way of thinking about how much PBF really costs is to ponder the effectiveness of negative incentives. For example, it could be that the pay of health workers could be reduced by, say 30%, and that they could gain back more than 30% if they performed particularly well. Thus failure to perform moderately would actually lower their total earnings. This has not been widely explored, partly because of the political connotations to this, however for most countries’ ministers of finance this might be a real opportunity. If they could improve performance at little or no additional cost, that would suggest a more cost-effective approach. To some extent Turkey has tried this approach.
Our understanding of the true cost of effective PBF is at an early stage, but it could be an important contribution to the question of sustaining PBF in the long run.


Add comment

This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.