Issues of reliability of validity of IMPLAN

Comments

49 comments

  • Avatar
    IMPLAN Support
    Hello Kei, Here are a few forum posts that might be helpful for helping you to discuss your choice and IMPLAN data: http://implan.com/V4/index.php?option=com_kunena&func=view&catid=84&id=11639&Itemid=35#11639 http://implan.com/V4/index.php?option=com_kunena&func=view&catid=80&id=11276&Itemid=35#11279 http://implan.com/V4/index.php?option=com_kunena&func=view&catid=84&id=9334&Itemid=35#9335 As regards the choice of the validity of the U.S. data you chose to the region you are in, we seem to recall that you put a lot of thought into the relative GDP's and the Sectoral breakdown difference between the IMPLAN OECD model and the U.S. region. We would recommend discussing why you felt that this was the best choice for your study. Please let us know if you have any additional questions.
  • Avatar
    ckeiwei
    I am trying to estimate the local economic impacts by collecting tourist's expenditures based on 18 IMPLAN industries (i.e. hotel, transport etc). I was told that I could modify Hawaii data to reflect my study region (a small tourism town in Malaysia). However, it subjects to validity and reliability claims since Hawaii data gives higher GDP compare to my study area. Also, the data provided in Hawaii are based on US data (e.g higher household income), it might sound unrealistic for my study area because all information I am going to collect are base on Malaysia data. Thus, i am curious whether i can use Hawaii data to estimate the local impact for my study area? Please give me some suggestions since I am going to collect the expenditure data soon. Thank you
  • Avatar
    IMPLAN Support
    Hi Kei, Here is some information on income that may help you in describing what many see as surprising differences between what is sometimes reported in IMPLAN, and the definition of income that they are more accustomed too. Average Personal Income will always be higher than median money income that Census reports. For an explanation [url=https://www.google.com/url?q=http://implan.com/v4/index.php%3Foption%3Dcom_docman%26task%3Ddoc_download%26gid%3D134%26Itemid%3D60&sa=U&ei=5p84UoDcLOTB2QX10IHgBQ&ved=0CAcQFjAA&client=internal-uds-cse&usg=AFQjCNHgbWWqlAb2AHLdCaB-bkw9KKOEsw]see this document[/url]. The personal income is specific to the region data. Indirect Business Taxes, Proprietor Income and Other Property Income is based on BEA GSP that is lagged one year to IMPLAN data so we have to project it. Employee compensation is based on current year data. You may use Hawaii data for your project. The data (the picture of the economy) is used to create the impact projecting multipliers. The expenditure data that you collect will be applied to the multipliers and will have no bearing on the creation of those multipliers. As mentioned earlier, your choice of Hawaii as you data region was made, as we recall, on the basis of a broader spectrum of Sectors and the fact that the data was more current than our Malaysian data offering. Your validation therefore would be based on these and any other reason that you used in determining the Hawaii data set was a better fit for you study. While the Hawaii regional data set will represent an island economy with a large number of imports and will thus be similar in this sense to Malaysia, if will certainly have variances from your true geography that may skew the results. One other thing to remember is that the size of industry does not affect the Multiplier, even though this may seem counter intuitive. It is the underlying relationships( i.e. local purchasing ability, output per worker, labor income per worker relationships) that drive the impacts. While Hawaii will no doubt under represent some aspects of your Malaysian region, and likely over represent others, you might given these considerations make an argument that without overall better data sources, and with controlling, if you were able to, any of the various industry relationships like production per worker or income per worker, that you have created an adequate approximation. Please let us know if you have additional questions.
  • Avatar
    ckeiwei
    Thank you. Do you mean the production per worker or income per worker will be given through the data (tourist expenditure) I collected? So far, it is very hard to get the information about business sales and taxes :( . The average income for labor in my study area is about USD150-500 per month. Some large oversea business operators were not willing to disclose their sales and taxes and I remembered one dive operator became angry when I asked him about sales and taxes. What I am going to do is to estimate the economic impacts of a diving tourism destination. So, I wish to know what kind of information I need to collect in order to be inputted into IMPLAN analysis. Is there any other information (e.g. secondary data) that I need to collect other than the main tourist’s expenditure data? Thank you.
  • Avatar
    IMPLAN Support
    Hi Kei, The information we provided to the last post was related to what you could do to make the model more regionally specific, not necessarily the data that you need to bring to the model. You can trigger the model with sales, Employment, or income. But you can also enter several categories. So for example if you were able to determine that an industry had 1MM currency units of production (or it's equivalent dollar value since you are using a U.S. model) and they employed 45 workers, the you could enter both of those items thus customizing the Output per Worker value from the underlying Hawaii relationships. Likewise you can, and might want to consider, making some adjustments to the income per worker based on what you discussed in this last post. Those are both simple ways that you make the Hawaii data set more like your region, if you can get the appropriate information. If you happen to be able to get more information about taxes and profits there are ways that you can make these modifications to the model as well, but this is much harder data to obtain, as you have pointed out, and requires more sophisticated techniques in the model. Any time you can gather additional data information, that will always benefit your study, but it is not essential. There are just some significant variances between Malaysia and Hawaii as you have pointed out, that if you are able to account for these, will enable you to provide additional validation to your study. Please let us know if you have any additional questions. IMPLAN Support Team
  • Avatar
    ckeiwei
    I am worry that i will be questioned about the validity since I am using Hawaii data set. For the profits, i think it will be a big challenge to get the data. If i can get the profits of the business entities, i will be able to get the taxes since they need to pay a specific amount of business taxes.Do you mean that I should also collect the employment data? I guess I can get the employment information since there are not many direct tourism business entities. But, how about those indirect tourism business entity? ie seafood restaurant, mini market?
  • Avatar
    IMPLAN Support
    Hi Kei, Again, you don't have to collect anything more than sales. Anything above and beyond that will just allow you to make the data more specific. If you are concerned that Hawaii really won't work for you, you may want to think about a different region. We do have a 2005 Malaysia data set available, or if you feel that one of the U.S. territories might work we do have U.S. Territory data as well. The U.S. Territory data is 2009 and it includes the 440 Sectoring. Please let us know if you have any additional questions. IMPLAN Support Team
  • Avatar
    ckeiwei
    Do you mean the 2005 Malaysia IO table? I can use the US data as long as the region is small and related to diving tourism industry. Therefore, I think Hawaii gives a better fit to my study. However, is there any tourism islands in US territories that I can use so that the results will not vary too much. Thank you.
  • Avatar
    IMPLAN Support
    Hi Kei, Yes we are referencing the 05 I/O table we discussed with you earlier. The U.S. Territories are Marianna Islands, Guam, Puerto Rico, Virgin Islands and American Samoa. You would need to decide if one of these would be a better fit than Hawaii for you. However, it sounds like you are comfortable with your original choice of Hawaii, so we would recommend using your same reasonings for that to justify your choice to your audience/critics as well. In the end the data region choice is a decision you need to make and to be comfortable defending. We can only go so far as to say that the data is a good reflection of the geography and year that it represents, because it is founded in raw data collected for that region and year. Please let us know if you would be interested in another data region, or if you are comfortable staying with Hawaii.
  • Avatar
    ckeiwei
    I am planning to use Hawaii data set because I don't need to purchase a new one. If there is a better data set that can reflect my study region, i may consider of buying it since my study area is a very small diving tourism coastal town compare to Hawaii, a well established tourism destination and the size of Hawaii is far Bigger. The boundary i set for my region is 5km radius only. I talked to my Professor today, he suggested me to use IMPLAN sector as a reference but ask more details about expenditure categories since I can aggregate the sector. E.g divided the expenditure on diving services to diving courses, diving fees and etc. because tourist might not know what diving services cover.
  • Avatar
    IMPLAN Support
    Hello Kei, The total values of production, and Value Added factors will not affect your Multipliers because the magnitude or size of the industry is not important to the Multipliers. Instead what is important is the relative ratios (e.g. Value Added: Intermediate Expenditures, Output per Worker, Labor Income per worker). So the size of the various industries is not what is important, but rather those ratios within the industry(s). For example, suppose the output and employment for the grain milling sector were $1 million and 100, respectively. The output per worker is then $10,000. Then, when you run a $1 million impact on the sector, you will get 100 workers. If, on the other hand, you reduced the grain milling sector’s output to $500,000 and re-balanced, the employment would be reduced to 50 and the output per worker remains the same at $10,000. So if you now run a $1 million impact on the sector, you will still get 100 workers. This is why we are suggesting to get as much data as possible to try to 'localize' these ratios. As regards the size or your region. Hawaii may well have larger Regional Purchasing Coefficient values than are appropriate and may show industries that you know are not local. If this is your concern, please let us know and we can show you how to adjust the RPC values to reflect what you know to be true of your local economy, but we will be unable to help you to determine if local industries that do exist contribute to local purchasing or to what degree. You will either need to leave these at the model estimate or find a way to use your knowledge of the area to make an estimation. The Regional Purchasing Coefficient is part of the Multiplier calculation, so making changes to this will affect your results. But any changes you make you will want to document and be comfortable justifying. Please let us know if you have any additional questions. IMPLAN Support Team
  • Avatar
    ckeiwei
    What is the difference between scenic and sightseeing transportation and support activities for transportation and transport by land/water? I understand transport by land refers to taxi, airport transfer. How if tourist taking boat to nearby islands for snorkeling activity? It falls under transport by water or scenic and sightseeing transportation and support activities for transportation? Next, the difference between travel arrangement and reservation services and Tourist guide services? Will it be overlapping since travel arrangement sometimes might include tourist guide services. I will travel to my study site to run pre-test for the expenditure categories by this week. Hopefully i can detect more industries and expenditure categories that might be useful to me.
  • Avatar
    IMPLAN Support
    Hi Kei, There are definitely some overlap between these Sectors. The best way to determine what works best to meet your needs is through the Help>Sector Search function in the software. The Sector search can be sorted by clicking the Industry Code column. You can then scroll down through the industry codes until you find the codes you are interested in examining. The associated NAICS descriptions will show you the types of Activities that will report into each category. Both transport by water and scenic sight-seeing have water based transport associated to them, so we would recommend deciding which Sector, overall, seems most appropriate. Another way to do this, if two Sectors seem too close using the method above, is the use the Explore> Social Accounts> Balance Sheet (Tab), and select View By: Industry Balance Sheet and the Commodity Demand tab. If one industry purchases another industry as one of it's primary commodities (based on sorting by the Gross Absorption column). If one industry consumes (absorbs) a the other industry as one of it's primary inputs you will typically want to choose that industry, as it will stimulate purchases of the other industry. For example, Sector 383 Travel Agency service purchases as one of its primary inputs, air transport. So in choosing this Sector you are also reflecting the purchase of airline services. Please let us know if you have any additional questions.
  • Avatar
    ckeiwei
    I flew to my study site this morning to pre-test the tourist expenditure questionnaire. Why I couldn't upload my document? I noticed that majority of the travel agencies, dive operators and large resorts are belong to foreign-owned and some are from other states. How IMPLAN captures the leakages using the expenditure data?
  • Avatar
    IMPLAN Support
    Hi Kei, What were you trying to upload? By definition, if the dollars are spent at local stores, the Direct impact is local and you will not want to change the default setting of 100% local. As regards the purchase dollars you assign to those stores and services there are a couple of different things that occur. In regards to retail stores, the only portion of the sale that is kept by the model for examining local impacts are the dollars of payroll to the stores employees and the materials and services purchased to operate the store (electricity, air conditioning, janitorial services and supplies, etc). In addition to this, many Sectors make a purchase of Sector 381 Management of Companies and Enterprises. This is basically a payment made to home offices. You can verify if the Sectors you are interested in are making this payment by going the Explore> Social Accounts> Balance Sheet (Tab), and select View By: Industry Balance Sheet and the Commodity Demand tab. You will choose the industry that you are considering modeling and then scan through the purchases made by that industry for Sector 381. You will also notice that this same sheet includes a values for Gross Absorption, Regional Absorption and RPC. These values represent the total amount of a commodity that is needed, the total amount locally available, and the amount of the commodity that is purchased from local suppliers. The Regional Purchasing Coefficient is the basis for how we determine what values will remain in the region and how much will leak (i.e. if you saw a purchase of Sector 3381 (a payment to corporate) with a Gross Absorption of 0.931% and the Regional Absorption 0.745% than 80% of that payment to corporate would be local for that Sector however if the Gross Absorption was 0.931% and the Regional Absorption was 0.00% than the RPC would 0 and all the corporate payment would leak from the region). Please let us know if this addresses your concern or if you have any additional questions.
  • Avatar
    ckeiwei
    I was trying to upload the tourist expenditure categories based on IMPLAN industry.
  • Avatar
    IMPLAN Support
    The forum is limited to certain file attachment types and sizes. If you are looking for help in discerning how to assign them to IMPLAN Sectors we would recommend using the Sector Search function found under the Help menu in the software or by right clicking on the Event field in the Setup Activities window. Please let us know if you have any additional questions. IMPLAN Support Team
  • Avatar
    ckeiwei
    I gone through the search for the sectors and these are what i found. 1.Hotels 2.Other lodging (bed & breakfast, hostels, guest house) 3.Tourist guide services 4. Food and beverage 5. Social drinks (e.g. beer, liquor) 6.Groceries 7.Petrol 8.Diving equipment rental 9.Souvenirs & department store 10.Transport by water (e.g. boat rental) 11.Transport by air 12.Transport by land (e.g. taxi, bus) 13. Scenic and sightseeing transportation and support activities for transportation 14.Telecommunications (e.g. internet) 15. Travel arrangement and reservation services 16. Diving services (e.g. diving course) 17. Laundry services 18. Medical service (e.g. clinic) 19. Dried fish and seafood 20.Pharmacy 21. Hair salons 22.Seaweed and related products 23.Bakery 24.Fruit 25. Massage 26. Others: ________________
  • Avatar
    IMPLAN Support
    Hi Kei, This looks like a reasonable tourist spending pattern, though you haven't specified which IMPLAN sectors you'll be using when you move from your survey into your model using. Be sure when you setup your model to ensure that you are not using sector 319 for any of these, since 319 [i]wholesale trade[/i] pops up in most sector searches. Please let us know if you have any additional questions. IMPLAN Support Team
  • Avatar
    ckeiwei
    I had pre-tested my questionnaire last week. I am planning to use and modify Hawaii data to suit my study area. Can I modify the local study area? It's quite hard for me to provide the gross regional product, total personal income, total employment, number of industries, average household income of my study area as the baseline information of my study area are quite limited.What is the different between default LPP and RPC. Most of the products in my study area (except fish products) are imported from a city situate around 90km away.
  • Avatar
    IMPLAN Support
    Hello Kei, You certainly can edit information in the Model. You can edit information directly in the Events by entering in as many variables as you know of Employment, Industry Sales, Employee Compensation and Proprietor Income. You can also edit the regional background data (if you have more information then the above) through the Customize screens which can be turned on through the File>User Preferences and the Analysis tab>Advanced Modeling> Enable Model Customization. These screens will also allow to modify the RPC values (Trade Flow). LPP alters the amount of Direct Effect that is applied to the Multipliers. For instance we often recommend if you have an airport in your region that you set LPP for the airline ticket costs to 50% because approximately half of the ticket cost is derived by the airport where the plane lands (the other 50% typically goes to the airport where the plane originated). If your survey is capturing expenditures in the Study Area only you will want to leave LPP at 100%. RPC is based on local use of local supply for the region and this is the factor that 'builds leakage into the Multiplier'. So if you know that you have higher leakages for the rounds of Indirect and Induced purchases than the Hawaii data set reflects, then you can change these values, per industry, to reflect the leakage you know for the area. Please let us know if you have any additional questions.
  • Avatar
    ckeiwei
    Is it possible to create a local input output model using IMPLAN?
  • Avatar
    IMPLAN Support
    Hi Kei, Are you asking if you can create a model for a subregion of Malaysia, to match your Study Area? If so this can be done, but there are a number of caveats. For example you would need to be able to provide us all the raw data material necessary to build the model (i.e. Employment by industry for each industry in the region, Output by industry for each industry in the region, Value Added and each of its elements by each industry for every industry in the region, etc.). Then depending on the time of year the request comes in (for example we are currently in the midst of our 2012 U.S. data production so these types of orders would not be able to be looked at until December/January) there can be some lag time between our receipt of the data and when we are able to produce a model. If you are interested in pursuing this, please let us know and we can certainly provide you with a pricing structure for custom model development. Please let us know if you have any additional questions, or if you would like to have a quote. IMPLAN Support Team
  • Avatar
    ckeiwei
    I don't think I have the raw data material (Employment , output, value added) as the area is very small and data are limited or unknown. So far, the only relevant data i have in hand is the Malaysia IO tables 2005. I am not sure whether i can use the Malaysia IO tables to build IO tables for my small study region.
  • Avatar
    IMPLAN Support
    Hello Kei, Unfortunately an I/O table isn't a complete picture of the economy, so without the additional detail, there would be little that could be done with that, relative to IMPLAN. If you were to edit by slashing a county's output and value added, it would become unbalanced with its final demands which would foul up the multipliers and the trade flow estimates (even if using the econometric RPCs). Again, another option would be other small economies, like the Northern Marianas model, which is a tiny island economy. Realize, though with an economy like that output per worker for the farm sectors is low, so that farming employment may be a big part of the response - the question being: would the economy have a land base capable of such a response? You might need to limit the farm portion of an impact. To some degree, unless you have the time and ability to collect or find the regional data, you will have to work with a proxy and defend it on the basis of the fact that no better or more precise data will be available.
  • Avatar
    ckeiwei
    I don't think i have the time and ability to collect or find the regional data, or it may be not exist at all. Northern Marianas model sounds similar to my study site as it is a tiny island economy. The data set i have currently are Hawaii, Honolulu, Kauai, and Maui. I am not sure whether these data are applicable to my case.
  • Avatar
    IMPLAN Support
    Hi Kei, CNMI has a total GRP value of $2.4BB, land area is approximately 184 miles and it includes 152 or the 440 possible IMPLAN industries. You can certainly compare it to which island that you have that seems closest or you can just build all the models and see if any seem like they would have a good GDP to represent your region. Again, unfortunately with a lot of these regions an exact regional economy is just not available.
  • Avatar
    ckeiwei
    What is the IMPLAN industry code for social drinks (e.g. beer, liquor), Seaweed sold in the market and Pharmacy? I couldn't search for the sector.
  • Avatar
    IMPLAN Support
    Hi Kei, If the drinks are purchased through a bar or restaurant this would be Sector 413. If they are purchased at a retail store, this would be Sector 324 Retail Food and Beverage. As regards seaweed, the correct retail store would depend on where it was purchased, again 324 is all retail food and 325 is the Sector that includes pharmacies. If you select retail sales you will want to select Gross Retail Sales in the pop-up box. In both these cases if you choose the retail Sector, you may have the advantage of knowing exactly what is purchased (you might need to have some additional break down for alcoholic beverages to apply this methodology), so you can apply a more advanced level of Margining that will capture a more complete picture of the impacts, whereas, using the retail Sectors will be more conservative, you can manually apply Margins to the producing Sectors in this case (71/72/ breweries/wineries/distilleries, and 61 Seafood product preparation manufacturing. To apply Margins by this method you will choose the producing industry Sector, set Event Year, and enter Industry Sales. You will then apply Margins via the Event Options> [url=https://implan.com/v4/index.php?option=com_multicategories&view=article&id=567:567&Itemid=71]Edit Event Properties>Margins> Yes[/url]. You will also want to set the Local Purchase Percentage to SAM Model Value via [url=https://implan.com/v4/index.php?option=com_multicategories&view=article&id=567:567&Itemid=71#lpp]Event Options> Edit Event Properties>Set Local Purchase Percentage> SAM Model Value[/url]. Please let me know if this addresses your concern or if you have any additional questions.
  • Avatar
    ckeiwei
    Tourists buy drinks through bar or restaurant as well as retail store, therefore i have broken down the expenditure categories into Social drinks (e.g. beer, liquor) and Groceries. Do you mean i need to have a break down category for alcoholic beverage instead of putting social drinks? I am afraid tourists might give overestimate price since they usually buy drinks and meal together, unless they keep the receipt. Most of the seaweed purchased by tourists are for own consumption. Seaweed is under groceries, mean i need to aggregate seaweed into groceries sector since they are sector 324? Currently i have 27 expenditure categories in my questionnaire. I am worry my respondents feel annoy when they see so many things to estimate >_< :(

Please sign in to leave a comment.