Puerto Rico Employment discrepancy

Comments

9 comments

  • Avatar
    IMPLAN Support
    Hello Joseph, The 2009 County Business Patterns and the 2007 5-year Economic Census were used as the primary sources. County Business Patterns and BLS data are often quite contradictory and apparently this is true for water transportation. QCEW data was used for some relationships for items not covered by the Census (and for wage and salary to employment relationships for Virgin Islands and Puerto Rico). Please let me know if this addresses your concern or if you have any additional questions.
    0
    Comment actions Permalink
  • Avatar
    joetawfik
    Thanks for your response. 1) Looking at CBP (County Business Patterns) It lists an F in the employment column for NAICS 483 in Puerto Rico which refers to 500-999. 2007 Economic Census is about 363 for paid employees. I was wondering what adjustments were done to reach the 181 number? 2) Another question was regarding the social accounting matrix mentioned in my previous post. "It only has one income bracket instead of the usual 9 brackets. I was wondering why that is?" 3) Finally, I tried a couple of simulations using the non-aggregated model and the aggregated model using the 3-digit NAICS code aggregation scheme. The numbers for water transportation are significantly different. The indirect nearly doubles when aggregated due to a huge jump in the government and non NAICS line item. I was wondering what may be causing this huge discrepancy. Thanks,
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hello Joseph, 1. In the 2009 County Business patterns, we see Puerto Rico NAICs 483 is disclosed and has employment of 181. 2. Two reasons, the CES data we used to define spending by income class would not transfer well to island areas (low income would be likely more subsistence based) and #of households by income class is tough to come by. 3. State and local government transportation is highly subsidized - for each dollar of ticket value there is usually more than one dollar of employee compensation by the workers (drives induced) and 3 dollars of intermediate purchases of goods and services (drives indirect). Combining the government enterprise with private transportation changes the indirect as you noted. Please let us know if you have any additional questions.
    0
    Comment actions Permalink
  • Avatar
    Qiang
    As a follow-up regarding the issue related to the original 440-sector model vs. aggregated model for PR, when the 3-digit NAICS aggregation scheme is chosen, it collapses Sector 427-440 into one, but water transportation remains its own sector as in the original model, as is each of the other transportation sectors. So I'm not sure where "combining the government enterprise with private transportation" comes from. Could you tell us a little about what government enterprise sectors and private transportation sectors are combined when we go from the 440-sector model to the aggregated model at the 3-digit NAICS level? Also, with regard to CBP vs. QCEW, my understanding is that for the 50 states and DC, Implan mainly relied on QCEW. QCEW also had employment data for PR. So for PR, could you tell us the rationale behind going with CBP when QCEW is available? Isn't QCEW generally preferred over CBP? Is it not the case for PR? Thanks.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hello Qiang, State and local enterprise ground transportation is included with private ground transportation when you aggregate to 3 digit NAICs. State and local enterprise is not combined with water transportation until you go to 2-digit NAICs. This is why we thought you meant 2-digit NAICs in your previous post. You are correct, the 51 "states" and "US" are based on CEW because they were all represented and could be programmed consistently. The 5 island areas could only be programmed consistently using CBP data, as Guam, American Samoa and Northern Marianas are not covered by QCEW. Please let us know if you have any additional questions.
    0
    Comment actions Permalink
  • Avatar
    Qiang
    Thanks for your explanation regarding the use of CBP data for the 5 island areas. With regard to the aggregation issue, just to make sure I'm following you, is "state and local enterprise ground transportation" Sector 430 in the model? If so, when I look at the 440-sector model, Sector 430 does not show any output or employment for Puerto Rico. Is this because the state and local enterprise ground transportation has already been grouped with private ground transportation (or Sector 336) even in the 440-sector model? If not, then I'm not sure why Sector 430 is zero.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hello Qiang, Excellent question. State and local enterprise employment data was not available for the 5 island areas, so all government enterprise sectors will be 0. The annual Census of government employment is where we collect enterprise data does not cover the Island Areas. If you to customize the data for Puerto Rico, the econometrics will take the new supply into account for the forward linkages that use water transportation, but that will have little effect on the multipliers for water transportation (depending on how much it buys from itself). Please let us know if you have any additional questions, --Implan Support Staff
    0
    Comment actions Permalink
  • Avatar
    Qiang
    I'm still a little confused about the before vs. after aggregation results. I understand you had no state and local enterprise employment data for the 5 island areas. How do I see that state and local enterprise ground transportation is included with private ground transportation when we aggregate to 3 digit NAICS, but the two are not grouped together in the original 440-sector? Is this just something going on behind the scene that the user does not see when doing aggregation? Which model is more accurate, the original 440-sector model, or the 3-digit NAICS model? The impact results are quite different.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hello Qiang, As you know, the more detailed the model, more the accurate the results you will get out it, which is one of the reasons that we typically recommend aggregating results rather than the model itself. State and Local Enterprise Ground Transportation is sector 430 in the 440 Sector model. However, if it is 0, then it will not affect the Indirect effects of the aggregated model. The unaggregated and aggregated comparison was replicated here to try to see the above (since enterprise is 0, that was not the cause). When the Model is aggregated the "transport by water" Indirect and Induced increase by ~50%. Two things are happening. [ol] [li]The aggregation bias tends to increase multipliers the more you aggregate a model.[/li] [li]The pre-estimated trade flows minimize this a great deal, but the island area models have no trade flows so the econometric RPCs are required which are quite affected by aggregation.[/li] [/ol] Hopefully this helps to explain the results you are seeing. Please let us know if you have any additional questions.
    0
    Comment actions Permalink

Please sign in to leave a comment.