Germany Data

Hello Elizabeth, The difference in multipliers you are seeing, is not in how they are calculated but in the extent of the indirect and induced effects within a region for a sector. The direct effects are always one. If the sector in the region imports more of its inputs and if more household make purchases outside the region then the multipliers will be smaller. In addition, - Germany is a smaller country than the U.S. and would be expected to have smaller RPC's for many commodities. Another reason is that in the OECD data, Proprietor Income is not separated out from OPI; rather, there is just a single GOS (Gross Operating Surplus) category and it is not endogenized (i.e., it is treated like a leakage like OPI is in the standard IMPLAN data and does not generate induced effects). Thus, multipliers for all OECD countries will be somewhat lower due to this (including the U.S. OECD model). So if you are comparing to a 'normal' (i.e., 440-sector) U.S. model, the multipliers are not comparable because that U.S. model spends PI, which increases the induced effects and thus the Type SAM Multipliers. Hopefully this helps to further explain the differences you are seeing.

Thank you for our answering our first question. While we understand that perhaps the two data sets are not comparable, we have a follow up question regarding the difference in level of magnitude of multiplier effects between OECD data and US data. We will try and be very specific using an example. US Data: If I enter in a new activity for “pharmaceuticals” sales of $1,000,000,000, the Implan output is: ImpactType Employment Output Direct Effect 886.2 1,000,000,000.0 Indirect Effect 3,199.4 727,187,911.6 Induced Effect 4,089.2 601,879,799.5 Total Effect 8,174.8 2,329,067,711.1 The indirect and induced output is 133% of the direct output. Germany OECD Data: If I enter in a new activity for “chemicals excluding pharmaceuticals” sales of €1,000,000,000, the Implan output is: ImpactType Employment Output Direct Effect 2,285.7 1,365,172,982.2 Indirect Effect 8.6 1,817,822.3 Induced Effect 1,032.2 124,384,403.6 Total Effect 3,326.5 1,491,375,208.1 The indirect and induced output is 9% of the direct output. Moreover, the indirect effect for output is very low. Based on your initial response, we do understand that the OECD countries will be somewhat lower, however this difference seems rather extreme and we are trying to understand. Please advise if this disparity in magnitude of multiplier is correct, and if so, can you help us understand the very low indirect effect leading to such differences in multipliers? Thank you very much.

Hi Elizabeth, You are most welcome. A couple of quick questions. Just for verification, are you using the U.S.OECD data so that you have like Sector comparisons or are you using a U.S. total file (with 440 Sectoring) and comparing that to your Germany file? We are wondering if there might be an issue with deflator settings in your Germany Model since the effects that you are reporting for the Direct Output value are larger than the value that you entered as the Industry Sales. This would tend to make use think that you have the Event Year and Dollar Year for View settings are not matched. That said, the Multiplier for Sector 9 matches your results sets. So the overall calculation seems correct. If you could help us out by letting us know if you are comparing two OECD files or an OECD file and 440 U.S. file we would appreciate it, and also if you could let us know what year your U.S. data file is, if it is not OECD we would appreciate it and it will certainly help us to be able to better explain what you are seeing, since we will be working from the same frame of reference.

hi. Thank you for your response. To answer your questions. 1. Yes - we are using US 440 file with total sectoring. 2. Germany data: 2005 OECD Data 3. US data: 2011 440 sectoring. 2. In terms of the deflator setting, I do not think this is the problem. My settings are the same for US and for the Germany data. Moreover, you are correct, I did not, before exporting the data, change the year to the same year as the data, and while that does have an effect on the results the same problem of having a very very small indirect/induced effect persists. Thank you again.

hi. Thank you for your response. To answer your questions. 1. Yes - we are using US 440 file with total sectoring, year: 2011. 2. Germany data: 2005 OECD Data 2. In terms of the deflator setting, I do not think this is the problem. My settings are the same for US and for the Germany data. Moreover, you are correct, I did not, before exporting the data, change the year to the same year as the data, and while that does have an effect on the results the same problem of having a very very small indirect/induced effect persists. Thank you again.

Hello Elizabeth, Thank you for providing the information. There are at least two large scale issues with the comparison that you are trying to reconcile: [ol] [li]lack of comparability in Sectoring, data sources and years [/li] [li]and differences in the density of the economies--US has a more dense economy than Germany by virtue of size of its economy and the number of its Sectors.[/li] [/ol] So, the US would/should have a larger Multipliers, i.e,. Indirect and Induced effects than Germany. Both models are national meaning that actual foreign export and import data is used, there is no trade flow estimation involved. The software used to generate Multipliers is the same for both the US and the Germany Models. There is one difference in the data in that Germany (OECD) has a Gross Operating Surplus which combines Labor Income and Other Property Type Income. Therefore in the US Proprietor Income is separate and is used to help drive the Induced Effect. This actually a difference in the way the raw data is provided, and so we have no way of modifying this to meet similar definitions. A couple of additional thoughts: [ol] [li]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. Since many of these ratios may be affected by the aggregating of the data with other chemical manufacturing which perhaps doesn't have the Output per Worker ratios or Value Added contribution of pharmaceuticals, the aggregating of these Sectors together in the OCED data may cause lower ratios. Again this is one of the issues that arises in comparing to non-like data collection schemes (NAICS and NACE and then further converting of the NACE data to OECD Sectors)[/li] [li]RPC is a significant component of the Multipliers. Thus if the pharmaceutical industry has to import a large number of it's good from outside of Germany this will dramatically impact the Multipliers. You can see the values of RPC in the Explore> Social Accounts> Balance Sheet (Tab), and select View By: Industry Balance Sheet and the Commodity Demand tab.[/li] [li] Interestingly, the U.S. OECD Model also does not report any pharmaceutical production for the 2005 year. To help you the get a comparison of the data between Germany and US for Sector 9: [table] [tr] [td]Description[/td] [td]US[/td] [td]Germany[/td] [/tr] [tr] [td]Gross Absorption[/td] [td]66.029%[/td] [td]56.045%[/td] [/tr] [tr] [td]Regional Absorption[/td] [td]56.789%[/td] [td]41.501%[/td] [/tr] [tr] [td]Value Added Proportion[/td] [td]33.971%[/td] [td]43.955%[/td] [/tr] [tr] [td]Labor Income[/td] [td]13.965%[/td] [td]18.293%[/td] [/tr] [/table] [/li] [/ol] Thus in comparing like Models (both OECD and both 2005 for the same Sector) we can see that the U.S. manufacturers in the Sector buy more and buy more within the U.S. than Germany, thus resulting in lower German RPC's for this Sector and lower Indirect Multipliers, but their payments to the Induced component are larger than the U.S.. Hopefully this helps.

Thank you again for your thorough response. We very much appreciate all of your help. We appreciate and understand the differences you highlighted. We understand that there are major differences between the sectoring of the data, the years, and the magnitude of the economy, which make it difficult to compare the data. However, we have a different question regarding how Implan is calculating multipliers using OECD data. We are able to see on the OECD website the source of the Implan OECD data for Germany. More specifically, it seems the employee compensation and output data is pulled from the STAN Input-Output Total, Domestic and Imports table. Our understanding is that output multipliers can be calculated as the sum of the columns in the Stan I-O Inverse Matrix Total table. When comparing these calculations to the Summary Multipliers Implan provides, we see a large difference. Please see the attached excel spreadsheet, which illustrates this difference. Can you help us understand the discrepancy. Again, it seems like the multipliers being generated by Implan are much smaller than our expectations. In particular, the indirect multipliers are very close to 0.

Hi Elizabeth, We are looking into this, but one thing that we think might be the source of the difference is whether or not the imports were removed before the Multipliers were calculated. Page 12-5 of [url=http://www.bea.gov/papers/pdf/IOmanual_092906.pdf]this document [/url]tells of the need to remove imports from the Use matrix before creating multipliers. In your numbers were you accounting for this? Thanks!

Hi - Thank you again for being so helpful as we try to figure out this germany data. We have a follow up question regarding this response. You write about three ratios that effect the multiplier. Are their rules regarding these ratios that cause the multipliers to be higher or lower? In effect, are there general rules that dictate if a multiplier is higher or lower? Thank you.

Hello Elizabeth, Hopefully this will help. In general: 1. Employee Compensation/Output ratio - the higher this ratio the larger the payment to employee compensation. Higher employee compensation means more household income and larger induced effects, therefore higher multipliers. 2. Intermediate Expenditure/Output - the higher this ratio the larger the payment to input providers. Higher payments to input providers means greater indirect effects, therefore higher multipliers. 3. Imports/Output - the higher this ratio the larger the payment to input providers outside the region. This has no impact in the region and is considered a leakage. The higher the leakage to imports the smaller the multiplier. 4. Closing the model - the size of a multiplier is also affected by how the model is closed. If the model is closed with respect to households and other final demand accounts, then the multipliers will be larger than otherwise. However, an additional factor affecting the OECD multipliers in particular, other than having a different sectoring scheme, is that in the OECD data, Proprietor Income is not separated out from OPI. Rather, there is just a single GOS (Gross Operating Surplus) category, which is not endogenized (i.e., it is treated like a leakage and therefore does not generate induced effects). Thus, the induced effects (and therefore multipliers) for all OECD models will be somewhat lower due to this.