Measuring Economic Development Results
Category: Local...
Posted Tuesday, February 10, 2004 at 09:14 PM [GMT-6]

On January 25, 2004, in a letter to the editor of the Midland Reporter-Telegram, Barry Horseman (2nd VP of the Midland Development Corporation) took to task those who questioned the value of the results obtained thus far by the MDC in its economic development efforts. The local bloggers over at Jessica's Well quickly jumped on Mr. Horseman's claims, specifically this one:

...MDC-approved projects returned more than $5 million in new local wages and capital investment on a total investment of approximately $1 million in total program expenses. I think most people would find a 5-to-1 return on investment in 17 months very attractive...

The discussion centers on whether the "local wages and capital investment" generated by ED-related expenditures/incentives is an appropriate measurement of the success. While I happen to agree that this particular measurement is of dubious value in answering the question (and the reference to a 5-to-1 return seems particularly inappropriate), I was actually more interested in knowing if there is a widely accepted methodology or computation for quantifying the success of ED projects.

With this in mind, I emailed a local banker pal and asked for his insights. He took the weekend to think about it, and then emailed me a link to a research paper entitled "The Net Effects of Large Plant Locations and Expansions On County Employment" (PDF file; size - 471kb). This paper was written in August, 2003, by Kelly D. Edmiston, PhD, an assistant professor in the Department of Economics, Andrew Young School of Policy Studies at Georgia State University.

This paper is pretty readable for the layperson, although it also provides plenty of statistical analysis to satisfy those who insist on seeing the numbers behind the findings. If you are interested knowing more about the implications of "economic development" projects, I recommend reading this paper.

However, I've taken the liberty of providing some excerpts below, to whet your appetite.

First, Dr. Edmiston points out that most economic impact studies don't do a very good job of evaluating the likely effects of plant locations or expansions.

While these studies do a reasonably good job of predicting the gross economic impact of a single firm location or expansion decision, they tend to neglect the simultaneity that exists between a single firm's location decision and the broader set of all firm start-up, relocation, expansion, and contraction decisions. This means that they may do a relatively poor job of estimating the net impact of firm location decisions, although it is this net impact that is of most concern.

The paper goes on to discuss "forces for agglomeration and dispersion," which determine the extent to which new business is likely to be attracted or repelled after a major plant relocation or expansion. This is an interesting section, but doesn't have a lot of immediate relevance to our local situation, given that we haven't seen any new business of such significance as to call into play the forces described in the paper.

However, an interesting footnote to this section is an analysis of actual jobs created compared to the number initially announced. For newly located plants, actual employment ended up being only 86.2% of the announced totals. Expansion-related jobs were a little more accurately predicted, with the actual number being about 92.5% of announced. I'll let you draw your own conclusions about why the actual numbers don't seem to fulfil the announced ones.

The results of the study get more interesting - and potentially relevant - when considering the impact on county-wide employment.

For expanding firms, there appears to be no statistically significant effect on county employment in the initial year of the expansion and the year following, but impacts are positive and statistically significant for years 2 – 4. ... An expansion yields approximately two workers on net for every new employee.

...

For newly locating firms, there appears to be no statistically significant effect on county employment in the first four years of operation, and only a relatively minor impact of 0.456 in the fifth year of operation. ... This result is broadly consistent with Fox and Murray's (2000) work, which finds no impact of new firm locations on local employment (they did not estimate multipliers).

And then...

After five years, each 100 new employees hired by a new or expanding firm results in a net gain of only 29 workers to the resident county. This result suggests that calculations of the cost of development incentives for new firms may be underestimated by 71.5 percent. ... My findings suggest that the cost of incentives per new worker on net, which is the critical figure, is 251 percent greater than is typically estimated.

But, wait...there's more...

The cumulative population impact over five years for a newly locating firm is -0.722, however, meaning that for every 100 workers hired, the county of the location loses 72 residents. To the extent that industrial enterprises make their surrounding environment less attractive as a residential location due to pollution, congestion, etc., I would not be surprised to see a small decrease in population in the resident county, but the magnitude of the population impact is somewhat surprising...

The results suggest that while expansion of an existing site may result in a minor boost to population from incoming workers, the negative externalities associated with occupying new facilities is a repelling force in county population dynamics.

The author then addresses the impact of relocations and expansions on other quantifiable factors, including tax rates:

The coefficient on sales tax rates is negative, but statistically insignificant, in both the employment and population equations. Although the property tax rate was found to have a negative (and marginally significant) effect on population, the effect on employment is statistically zero.

It's not all bad (or neutral) news, however. Expanding the retail sector, for example, has good significant positive effects:

As expected given earlier results, new firm locations and expansions generally had a negative impact on population in the resident county, with the exception of retail enterprises. The average cumulative five-year impact for new and expanding firms in the retail sector was 2.658, suggesting that new and expanding retail enterprises attract residents in relatively large numbers.

Finally, Dr. Edmiston's conclusions include this summary (emphasis mine):

The analysis suggests that the employment multipliers associated with these events are less than one for new firm locations, on average; that is, that the net employment effect of a large firm opening or expansion is smaller than the gross employment impact. Because the cost per job of incentives packages generally are calculated on the basis of gross job impacts, we argue that these cost estimates generally are too low by 71.5 percent, on average. The relevant multiplier for expanding firms is considerably larger at 2.0, suggesting that one spillover job is generated for every new employee hired by expanding firms. I find that the impact of large firm locations or expansions on population in the resident county generally is negative for newly locating enterprises and statistically zero for expansions. I find that the population impact is positive for both new and expanding firms in the broader area consisting of the resident county and its contiguous neighbors.

In total my results suggest that the net economic impact of large new firm locations generally are over-estimated, as traditional input-output models generate gross impacts that are always larger than 1.0, sometimes considerably greater, and these gross impacts generally are used for calculating the costs per job of economic development incentive packages. The results suggest that local governments are not likely to receive significant long-term employment or population benefits from large new firm locations, where estimated multipliers are on the order of 0.3, and therefore local governments should be skeptical of typical economic impact studies which encourage the recruitment of these firms with incentives. Large expansions of existing enterprises do have a substantial multiplicative impact on employment in the resident county, however.

There's much more in this study that's worth reading, and it should be considered in its totality. It should be obvious that there are regional factors at work that might make some of the specific conclusions less relevant to our local situation. However, my point in focusing on this study is to raise the question of whether we're using the best possible indicators of success for our ED efforts. Those indicators are accessible, and I hope that the MDC is taking advantage of them. Unfortunately, Mr. Horseman's aforementioned letter doesn't give much assurance that this is the case.

Incidentally, I have contacted (or will, once this is posted) Dr. Edmiston to give him the opportunity to weigh in on my summary of his research, and to point out any apparent misinterpretation of his work. This is a complex area of study, and I'll be the first to admit my relative ignorance regarding its details.

Permalink
Comments

Bingo. Are some methods of measuring the returns better than others? Yes.

But we seem to agree that Mr. Horseman's letter shows that he, at least, is using no serious method of evaluation. Or if he is, we was not willing to share it with us.

Posted by: Natalie at February 10, 2004 10:49 PM

Excellent work, Eric. It's too bad that the average media type is far too dense to understand any of it.

Posted by: Vaughn at February 11, 2004 09:52 AM

Vaughn....I disagree.

They are not dense. They are just on the other side.

Posted by: Natalie at February 11, 2004 10:04 AM

%%%%%%%%%%
All these started with that MDC report submitted to the Midland City Council 1/6. You think you can print that report here so that we can see it firsthand instead of just relying on Stewart Doreen?

Mr. Horseman's complaint is that Doreen did not stress that 1:5 ROI in 17 months which roughly speaking is really only about 1.00:3.53 in 1 yr.

That Report should be on the MDC website, or on the City of Midland Website, or on the mywesttexas.com website without any need for an adobe acrobat program or subscribing to a newsletter.

Unless I see that report and see in that report MDC specifically mention that 1:5 ROI, I would not put any weight on Mr. Horseman's complaint but simply treat that 1:5 ROI as something they perhaps only thought about later between 1/7 and 1/25.

Posted by: The Eurasian at February 11, 2004 02:21 PM

"Vaughn....I disagree. They are not dense. They are just on the other side."

Point well taken.

Posted by: Vaughn at February 11, 2004 03:54 PM

I'm not interested in sniping at the media or the MDC. I'd just like to see some relevant analysis by qualified experts brought to bear on the issue of measuring the return on our taxes, rather than shooting from the hip without much apparent thought. This is a very complicated area, and an acknowledgement to that effect - and some evidence that true experts are being consulted - would go a long way in helping the MDC build some credibility.

Judging by Dr. Edmiston's response to my emailed request to vet this post, I suspect he'd be happy to try to provide answers to specific questions. As he put it, "I am always happy when my research gets noted by practitioners and citizens, because in the end, that is why I do what I do - I want to have a positive impact on policy."

This is also an area which is subject to an amazing amount of political pressure. Without going into specifics, I can tell you that just raising the questions has generated some disappointing reactions from some who should be more open (and I'm not referring to anyone on the MDC, btw).

Posted by: Eric at February 11, 2004 04:15 PM

Re: "Without going into specifics, I can tell you that just raising the questions has generated some disappointing reactions from some who should be more open."

Silly, you should know never to get into an argument with someone over their religion.

Posted by: Natalie at February 12, 2004 01:43 PM

And particularly not with the deity himself...! ;-)

Posted by: Eric at February 12, 2004 01:48 PM
Post a comment [Take your time...we're in no hurry.]
Note: You must provide either a website URL or an email address...but there's no rule that says you have to tell the truth. If you don't want your email address published on this site, just input a website URL, or make up an email address.









Remember personal info?