An anonymous reader pointed me to this 2020 article by John Bradbury (Kennesaw State University), published in the journal Contemporary Economic Policy (ungated earlier version here), on the impacts of state-level movie production incentives in the US. This is a topic I have blogged about before, and something I retain an interest in (not least because our government is very keen on providing large subsidies to foreign movie production firms).
Bradbury collates the data on movie production incentives (MPIs) over the period from 2000 to 2015, during which time:
...44 states enacted MPIs, seven states ended their programs, and four states suspended their programs temporarily... Several states also expanded and reduced incentives for filming over time...
That provides sufficient variation to determine the relationship between incentives and economic growth, especially as the variable of interest (MPIs) could be defined as a continuous variable. However, instead:
MPI is equal to zero or one, where one reflects the state having any active state-funded film incentives in place during any part of the year of observation or it denotes a certain incentive threshold for in-state spending on film projects...
So, essentially this is an analysis at the extensive margin of MPIs (comparing states with, and without, MPIs), or at the intensive margin at particular thresholds. Bradbury does note (in a footnote) that using a continuous measure of MPIs "did not produce meaningfully different estimates", but that just makes me wonder why those results weren't presented as the primary results. To deal with the potential endogeneity of MPIs and economic growth (which would arise, for example, if characteristics of states that affect economic growth also affect whether the state offers MPIs), Bradbury employs an instrumental variables approach. Specifically, he instruments for MPIs using the age of the state's film commission and whether neighbouring states have MPIs. Those seem like sensible instruments, and they are good predictors of MPIs and pass the usual tests for weak instruments in the first stage of the estimation.
As an outcome variable, Bradbury looks at income per capita (in levels and growth rate), as well as Gross State Product (GSP) per capita (the state-level equivalent of GDP per capita, again in levels and growth rate). He also looks at GSP specific to the film industry alone. In the main results, he finds that:
Estimates of having any MPI program are not statistically significant for any measure of economic performance, which is not indicative of a large multiplier that enhances economic development. Estimates of the economic impact of MPI implementation on the film industry are also insignificant, which does not indicate an industry-specific effect.
So, not only is there no effect of MPIs on the economy overall, there is not even a statistically significant effect on the film industry. Bradbury then goes on to look at different types of incentives and different levels of incentives, and there is no clear pattern of significant positive effects on the economy (the few statistically significant effects he finds would probably not survive an adjustment for multiple hypothesis testing, and to highlight them would simply be a case of cherry-picking of results).
So, in addition to not paying off in terms of tax revenue, it seems that movie production incentives don't pay off in terms of economic growth. You might wonder why these incentives don't pay off. There are a number of potential reasons. Perhaps movie production crowds out other economic activities. However, if that were the case then Bradbury would have found an increase in GSP in the film industry, and there would have been an offsetting decline in other industries. Alternatively, perhaps movie production (at some level) would have happened in the state anyway, and the incentives simply become a transfer from the state government to the movie production firms. On that point, Bradbury notes (in discussing an earlier study) that there may be:
...a “crowding out” effect due to many states offering incentives in a zero-sum game, where film production companies move (or threaten to move) to competitor states unless appropriate incentives are offered. If many states offer incentives, then film companies ultimately end up filming in their desired location (where they would have located absent subsidies) with a significant subsidy.
That strikes me as likely to be the case. If your state was the only state offering an incentive, then it is likely to attract movie productions that would have otherwise occurred elsewhere. However, if most states are offering incentives, then Tiebout competition kicks in, and the movie production firms simply play off states against each other. Only the movie production firms benefit from that situation.
Bradbury concludes that:
...it appears that MPIs divert tax revenue to the film industry from other economic sectors (public and private) without generating corresponding economic growth, which calls into question the popular use of film incentives to promote economic development...
Film incentives represent another example of an economic development strategy that fails to promote growth, which provides further evidence that industry-specific incentives to lure business offer an uncertain path to economic improvement.
I think we would do well to take this as a warning against these incentives, and we certainly shouldn't be overly concerned about losing productions to other countries.
No comments:
Post a Comment