A New Public Sector Comparison Tool: Empirical Evaluation in the Healthcare Sector

Size: px
Start display at page:

Download "A New Public Sector Comparison Tool: Empirical Evaluation in the Healthcare Sector"

Transcription

1 A New Public Sector Comparison Tool: Empirical Evaluation in the Healthcare Sector Miguel Gouveia 1 CLSBE, Universidade Católica Portuguesa Pedro Raposo CLSBE, Universidade Católica Portuguesa June, 2012 ABSTRACT Public sector comparators (PSC) offer official estimates of how much a project would cost if it were carried out by the government, versus production by public sector entities. Comparisons may suggest a benefit of a public private partnership (PPP), due to competitive bidding by potential private partners, which leads to various economic advantages. However, most PSCs require complex processes and offer little transparency; this study instead proposes a general approach that provides an external benchmark for the PSC, using public data, and illustrates the approach in the healthcare sector. Estimates of the PSC of two hospitals built using PPP in Portugal, compared against the official PSC values, reveal that the official construction PSCs were close to the upper limits of the prediction confidence intervals but the operations PSC were very different, whether well above the upper limit or close to the lower limit. The bidding process for one project produced a winning bid substantially below the PSC, so competition partially mitigated the upper bias in the PSC estimate. Keywords: PPP, Public sector comparator, Cost function JEL codes: D24, H43, L32, L33 1. INTRODUCTION There are multiple reasons to use public private partnerships (PPP), ranging from the benevolent belief that PPP are more efficient than traditional procurement or direct public sector production to a more cynical view that PPP shelter governments and politicians from accountability, because they can rely on off- budget financing or provide rents to powerful interest groups. Regardless of their impetus though, it is necessary to gauge the potential efficiency advantage of PPP by comparing the costs to society when governments carry out projects in a traditional way versus developing an otherwise similar project using PPP. This public sector comparator (PSC) is a key parameter of the 1 Corresponding author. Miguel Gouveia; tel: ; fax: E- mail addresses: ucp.pt (P.S. Raposo), ucp.pt (Miguel Gouveia).

2 PPP process that defines whether there is any economic advantage in forming a partnership, according to its comparison against the outcomes of a competitive bidding process. In practice the PSC also serves as a parameter in the bidding process, such that it establishes a ceiling for private firms bids. A low PSC puts more pressure on firms to minimize costs; a high PSC makes a project more attractive to potential private partners. In turn, the official PSC values are often the subject of controversy and accusations that they have been generated by black box processes that are too complex and insufficiently transparent. It would be useful to have clear terms of comparison to clarify the official values of the PSC. This study aims to provide an external view of the PSC, using public data to obtain magnitude estimates. We add to existing literature by presenting an initial empirical analysis of the value of the PSC used in two actual PPP projects. Specifically, we gather data from the Portuguese National Health Service (NHS). Portugal has extensive experience using PPP for roads, water management, energy production, and distribution, dating back to the 1970s (OPPP, 2011). Furthermore, Portugal exhibits the greatest use of PPP among European countries, as measured by the ratio of expenditures on PPP to gross domestic product or overall government budget (Moreno, 2010). Such projects more recently have extended into health care settings, including the construction and, in some cases, operation of hospitals integrated with the NHS. We concentrate on the first two PPP hospital projects, in Braga and Cascais, and provide estimates of their PSC using publicly available data, then contrast our estimates with the official values for PSCs. Because we are dealing with both construction and operation costs, the analysis relies on two distinct sets of data. The first, small data set covers construction costs and basic characteristics of hospitals built by the government using traditional procurement. The second data set instead includes outputs and operating costs for almost all hospitals in the NHS. We thus can estimate construction and operation cost functions that summarize and quantify public sector cost structures. We use data about the hospitals characteristics, the output of clinical services defined in the PPP contracts, and the estimated cost functions to generate our PSC estimates. Our results show that the official PSCs for construction costs in both cases are close to the upper limit of our prediction confidence intervals. The PSC for operations differ though: In Braga, they were well above the upper limit of our prediction, whereas in Cascais, they were close to the lower limit. The bidding process in Braga led to a winning bid that was substantially below the PSC, proving that competition at least partially addressed the upward bias in PSC estimates. In Cascais, press reports indicate the hospital s financial difficulties. Perhaps the PSC was too low, and the even lower bid that won the contest now is suffering from a winner s curse. 2. PUBLIC SECTOR COMPARATOR A PSC is usually an aggregate of many partial estimates. For example, in engineering projects, these estimates might combine unit costs, developed through past experience, to estimate separate costs for a multitude of the components that go into complex structures. Input requirement lists can be long; the relevant market prices may not be obvious; and even the appropriate amounts for some inputs may be unclear, ranging 2

3 from specialized labor to energy and raw materials. Similar problems arise for complex services, such as the provision of healthcare. The public sector costs may not be explicit, such as when a project involves the use of scarce land already in the public domain or relies on earmarked funds from public sector capital budgets. Overall then, arriving at a PSC estimate is a long, complex, and non- transparent process. In such a setting, government officials could easily introduce biases into an estimate, pushing it above or below the accurate value. For example, government officials might prefer a PSC that is too high, because they want to favor private sector interests or ensure that a PPP project that is less socially desirable but politically convenient moves ahead, particularly if it can be done off- budget. These cases offer examples of strategic misrepresentation (Flyvbjerg, 2008). On the other side, a PSC might be too low because the official hopes to extract larger surpluses from private partners. However, this latter example is usually not a great cause for concern, because nothing prevents governments from using bidding processes that set a ceiling on private sector bids below the PSC. 2 Instead, it appears optimal to design a bidding process in which a strategically chosen ceiling maximizes the gains from using PPPs. Thus an optimal ceiling choice should be a function of the PSC, as well as government s beliefs about the distribution of efficiency levels in the private sector. Given these temptations for governments to manipulate the PSC, combined with the margin for unintended error or bias, we see value in deriving a methodology that can provide reasonable estimates of the PSC at low costs, using publicly available information. The best solution might be to use historical data and econometrics to obtain a statistically based estimate, rather than relying on costly procedures and expert opinions. A similar philosophical approach to problem solving is suggested by Meehl (1954) and Tversky and Kahneman (1974). Beyond the hypothetical political reasons for biases, we also acknowledge that mistakes can occur naturally; public officials make mistakes because to err is human. Kahneman and Tversky (1979) note the planning fallacy, which states that people are prone to error, and committees of people can produce estimates that are very far from reality. Flyvbjerg (2008) explains the dangers of relying solely on engineering expertise to generate project forecasts and documents a series of errors to avoid in public projects mistakes that are not due to poor data or models but instead take the form of systematic biases, such as the optimism bias, and do not tend to disappear with experience. 3. METHODOLOGY use hedonic regressions to model the costs of infrastructure and estimates of the hospital cost functions. Thus we can calculate the expected costs of the new clinical services in hospitals, classified according to operational characteristics (i.e., number of inpatient stays, outpatient visits, outpatient services, day surgeries, emergency care). Hedonic models do not replace traditional cost methods, because they use little 2 In some situations, such as when governments or key decision makers change, new authorities might not feel a sense of ownership of the PPP project, or they may not work to their political advantage. In these cases, officials could try to shut down the process expediently by imposing an artificially low PSC. 3

4 information. However, they support a synthesis of modeled and quantified historical data and are convenient for carrying out predictions, as we attempt to do. methodology thus reflects the hospitals characteristics, defined in the PPP contract. We first estimate the model econometrically using historical data about costs and characteristics. Then, we predict the mean and build a confidence interval for the PSC, to predict the cost to the government to build and operate the two focal hospital projects. These projects each comprise two components: hospital construction and hospital operation over a period of time. In parallel with the official PSC, which were published separately, we present separate estimates for each component. 4. DATA We searched for historical data on public hospitals and collected information on costs, project characteristics, and multiple size measures. For clinical services, we used data related to the operational costs of two types of National Health Service (NHS) hospitals: administrative sector institutions (SPA) and government- owned corporations (EPE). Our data also reveal the main production lines: inpatient admissions, outpatient visits, outpatient services, day surgeries, and emergency care. Finally, we gathered data about the average complexity of hospital care, measured by a case mixed index. Data on infrastructure costs are very scarce. Ideally we would prefer to have information about many cases of constructing and equipping new hospitals, each with clear, unique characteristics (e.g., construction area, total number of beds, number of intensive care beds, number of operating rooms, existence of transplant units). In practice though the available data were well below our expectations. Specifically, we obtained data about 10 hospitals, and the available measures common across all hospitals were the built area (m 2 ) and number of beds, as we detail in Tables 1 and 2. Table 1: Relevant dates, beds, and built area Hospital Bid Opening Date Date of delivery / Number of Inauguration Beds Fernando da Fonseca Hospital 24- Jul May ,948 Built up Area (m 2) Hospital N ª Sr ª da Graça - Tomar 09- Feb Aug ,202 Pedro Hispano Hospital Apr Apr ,279 Matosinhos Hospital Santo Andre - Leiria 03- Jun Apr Hospital S. Teotonio - Viseu 23- Oct Jul ,697 Hospital Santa Maria da Feira 11- Sep Jan ,405 Hospital Barvalento Algarve 11- Aug Apr ,005 Hospital Centre Cova da Beira 03- Sep Dec ,342 Vale do Sousa Hospital 07- Jan Aug ,743 Rainha Santa Isabel Hospital - Torres Vedras 18- Apr Sep ,493 4

5 Table 2: Costs of construction, building supply (medical equipment, IT and others), and other ( ). Hospital Wining bid Total additional TOTAL WORK Equipment Costs Other Costs TOTAL COST HOSPITAL Fernando da Fonseca 27,985,760 28,712,786 56,698,546 27,082,097 7,593,111 91,373,754 Hospital Hospital N ª Sr ª da Graça - 23,379,089 7,995,586 31,374,675 13,134, ,993,560 Tomar Pedro Hispano Hospital - 20,878,463 31,305,856 52,184,319 15,759,209 4,898,870 72,842,398 Matosinhos Hospital Santo Andre - 19,377,081 26,999,700 46,376,781 19,092,681 3,514,463 68,983,925 Leiria Hospital S. Teotonio - Viseu 35,916,527 8,471,624 44,388,151 17,941,257 3,560,358 65,889,766 Hospital Santa Maria da 29,527,599 6,716,868 36,244,467 5,495,304 3,226,452 44,966,223 Feira Hospital Barvalento Algarve 22,255,577 4,046,785 26,302,362 19,340,360 2,471,975 48,114,697 Hospital Centre Cova da 24,792,889 6,183,159 30,976,048 16,677,238 2,452,952 50,106,238 Beira Vale do Sousa Hospital 36,565,003 6,617,562 43,182,565 27,858,761 2,707,040 73,748,366 Rainha Santa Isabel Hospital - Torres Vedras 16,500,231 4,975,009 21,475,240 13,224,807 1,281,691 35,981,738 Systematic information was available regarding the operating costs of public facilities and key production measures. For each public hospital, ACSS annually discloses the operating costs and volumes of inpatient admissions, day surgeries, outpatient visits, outpatient services, and emergencies. Thus, we obtained operating cost data for the period related to 58 hospitals or hospital centers, which we followed for an average of 7.8 years. 3 We detail these data in Tables 3 and 4. Table 3: Descriptive statistics: Hospital activity Variable Observations Mean Standard Deviation Log (Total Cost) EPE Log (Inpatient Episodes) Log (Day Surgeries) Log (Emergencies) Log (Outpatient Visits) Log (Outpatient Services) Log (Case Mix) Table 4: Hospital activity by year Year Means Log (Total Cost) EPE Log (Inpatient Episodes) Log (Day Surgeries) Log (Emergencies) Log (Outpatient Visits) Log (Outpatient Services) Log (Case Mix) See the Appendix for the list of hospitals

6 Although the data about the operating costs of NHS hospitals offer some precision to the estimates, we lacked comparable access to information about various PPP contracts. To estimate reference values that could serve as PSCs, we used information related to two specific cases of new hospitals in Cascais and Braga. 5. ECONOMETRIC ESTIMATES OF INFRASTRUCTURE COSTS Using the data from Tables 1 and 2, we estimate regressions and develop Figures 1 and 2. Figure 1 shows the relationship between the number of beds and construction costs; Figure 2 depicts the relation between equipment costs and the hospital s built area. Figure 1: Costs of construction and number of beds lthough the number of observations in Figure 1 is relatively small, the regression offers a good fit to the data. 4 In contrast, in Figure 2 we find an "outlier" an extreme, misaligned observation. At Feira Hospital, equipment costs are well below the statistical norm. Therefore, we excluded Feira Hospital from the sample used to run the regression to predict equipment costs. By doing so, we can effectively predict equipment costs using the construction area of the hospital, better than with the number of beds. Thus, we use the area regression to predict equipment costs. 4 See the Appendix for the results of all regressions. 6

7 Equipment Costs PORTIMÃO BEIRA T.NOVAS TOMAR FEIRA V.SOUSA LEIRIA MATOSINHOS AM-SINT VISEU Area (1,000 sq meters) Figure 2: Cost of equipment and hospital area. built a statistically significant model to explain "other costs" according to the built area of the hospital, which then enabled us to estimate a statistically significant regression to explain the total costs per built area. However, we also needed to correct the predicted cost values by the consumer price index (CPI) for the health sector. In Figure 3 we present these values for several recent years, such that the data values collected from the PSC for hospitals in Braga and Cascais indicate 2005 and 2006 prices. For comparability, we corrected them by factors of and 1.378, respectively Figure 3: Consumer price index for the health sector. 6. RESULTS 6.1. Estimating the Cost of Infrastructure 7

8 We predict the costs of new hospitals in Braga and Cascais as if they were built by the public sector. To compare these predictions with the PSC actually used, we tested two estimation methods. All regressions reported here were carried out in logarithmic form, which not only ensures appropriate functional specifications but also helps minimize heteroskedasticity problems. Method 1 relies on a single, univariate regression, 5 in which the total cost of construction and equipment reflects the built area of the hospital. High collinearity prevents us from using the variables area and number of beds simultaneously in the regression; we therefore choose built area, because it is the best variable in terms of goodness of fit. Method 2 instead uses three partial models. A first submodel explains the cost of the building using the number of beds (the best specification found). A second submodel explains the cost of equipment by built area, and the third submodel explains other costs (e.g., overruns, extra charges), again as a function of the built area. This method provides the final costs as the sum of the predictions of each submodel. As an aggregation of forecasts of different regressions, Method 2 does not support the construction of the confidence interval using standard techniques. In this case, the problem must be solved using bootstrapping predictions, which build a distribution of estimates from the construction of multiple samples, taken with replacement from the original sample available. Furthermore, Method 2 offers slightly better fit with the historical data compared with Method 1 (lower mean squared error of prediction). However, for completeness, we provide the results from both methods. Specifically, we generated forecasts and a confidence interval (95%), then multiplied all values by the CPI- Health to generate the estimates in Table 5. Table 5: Infrastructure cost estimates for hospitals in Cascais and Braga ( million) Prediction Mean Estimate 95% Confidence Interval Lower Limit Upper Limit Official PSC Method 1 (model Cascais (P 2005) area in m 2) Braga (P 2006) Method 2 (partial models with bootstrap) Cascais (P 2005) Braga (P 2006) The estimated costs suggest that the values of the public comparators are greater than the average estimates we derived using historical construction data (even after adjusting by CPI- Health). In Method 1, the PSC fall within the 95% confidence interval, whereas with Method 2, they are above the upper limit. Of course, these estimates use limited data, depend on CPI- Health, and do not account for the teaching hospital status of Braga all of which could justify higher infrastructure costs. 5 See the Appendix for the detailed results. 8

9 6.2. Estimating the Cost of Clinical Services There are many available functional forms for econometric estimates of the relationship between costs and outputs. One of the best known and most commonly used is the translog, which takes the following form for our study context:! ln! =! +!!!"!! + 0.5!!"!"!!!"!! + h! h +!!!!!!!!!!!! h!!!!!!!!!, where C refers to total costs; Y is the output vector measured by inpatient admissions, outpatient visits, outpatient services, emergency episodes, day surgeries, and case mixed index; and H and T are dummy variables for hospitals and years, respectively. We provide the statistically significant results from this regression in Table 7; the independent variables offer good predictive power. Table 7: Estimates of the regression of the logarithm of total operating costs Log (Total Cost) Coefficient Robust Std. Deviation t P> t Indicator EPE Year Year Year Year Year Year Year Log (Episodes) Log (Amb Surgery.) Log (ER) Log (Consultations) Log (Hosp. Day) Log (Case Mix) Log (Epis.) * Log (Case Mix) Log (Hosp. Day) Log (Epis.) * Log (Consultations) Log (surgical) * Log (ref.) Log (Hosp. Day) * Log (Consultations) Constant Notes: This regression used 458 observations, with 58 groups. The overall R- square is

10 Cascais, Clinical Services Using the preceding regression estimates, we compared the model prediction for the hospital in Cascais with its PSC. The PPP in Cascais entailed operating the old hospital for two years, then moving into the new hospital and operating it for eight years. We generated two predictions for these operating costs. First, we used the cost function for an average public hospital. Second, we turned to the cost function estimated for the former hospital in Cascais, or an historical Cascais fixed effect, which generates higher than average costs. The results, in Table 8, reveal that the official PSC for the public hospital were below the average prediction and near the lower limit of the forecasting interval. These estimates thus indicate greater cost pressures on clinical services than we found for the infrastructure. Table 8: Cascais Comparison Against Average Hospital Estimated Average Lower Limit Upper Limit PSC First 2 years Last 8 years Present Value for 10 years Against Cascais in the past Estimated Average Lower Limit Upper Limit PSC First 2 years Last 8 years Present Value for 10 years Clinical Services The PPP in Braga required operating the old hospital for one year, moving into the new hospital, and operating there for nine years. Using the same methodology we applied to the Cascais Hospital, we compared the predictions of the cost of PSC against the average of all hospitals and the former Braga Hospital costs (Table 9). The comparison revealed values well above the upper limits of our prediction, in contrast with the situation in Cascais. However, as we noted previously, these estimates fail to take into account that Braga s new hospital is a teaching hospital. Table 9: Braga Comparator Against Average Hospital Estimated Average Lower Limit Upper Limit PSC First year Last 9 years Present Value for 10 years Against Braga in the past Estimated Average Lower Limit Upper Limit PSC First year Last 9 years Present Value for 10 years

11 6.3. General Discussion Table 10 contains the overall results of the estimates for two scenarios: one in which the costs are given by the lower estimates (both construction and operation) and another in which the costs are given by the higher estimates. Table 10: Overall results of estimates for Braga and Cascais Prediction Interval 95% Confidence ( million) Estimated Average Lower Limit Upper Limit PSC High Cascais Infrast Cascais Services Total Cascais * * Braga Infrast Braga Services Total Braga * * Low Cascais Infrast Cascais Services Total Cascais * * Braga Infrast Braga Services Total Braga * * * Asymptotic approximate values. We find that the official PSCs for construction costs in both cases were close to the upper limit of our prediction confidence intervals but that the PSCs for operations were very different: In Braga the PSC for clinical services was well above the upper limit of our prediction. In Cascais the same PSC was closer to the lower limit of our prediction. Because the operational costs of the clinical services are much larger than the infrastructural costs, they dominate the final results. Therefore we find that the overall PSC for Braga is well above the upper limit of our estimates, whereas for Cascais, the overall PSC is within the prediction confidence interval, and near its lower limit. 7. CONCLUSIONS The results show that the PSC for infrastructures ranged above the average of the estimates based on historical data. The PSCs for clinical services seem substantially different though. In the case of the new hospital in Braga, the official PSC value is clearly above average, and even above the upper limit of the forecast range. In the case of Cascais, the PSC value is below average and close to the lower limit of the forecast range. As it turns out, the bidding process in Braga led to a winning bid substantially below the PSC, proving that competition at least partially reduced the upward bias in the PSC estimate. This is good news, because competition in this case seems to have worked. 11

12 The case of Cascais is also interesting. The firm that won the Cascais bid apparently is suffering losses, and the press has reported that it is even trying to sell its concern and extricate itself from the PPP contract (Santos, 2011). The low PSC initially defined, by comparison with our estimates, makes these results not entirely surprising. REFERENCES Flyvbjerg B. (2008), Curbing optimism bias and strategic misrepresentation in planning: Reference class forecasting in practice. European Planning Studies, Vol. 16, No. 1, Kahneman, D & Tversky, A (1979). Prospect theory: an analysis of decision under risk. Econometrica, Vol. 47, No. 2, Meehl, P E (1954). Clinical versus statistical prediction: A theoretical analysis and a review of the evidence. Minneapolis, MN: University of Minnesota Press. Moreno, C (2010). Como o Estado gasta o nosso dinheiro. Caderno, Lisboa. OPPP (2011). Relatorio das PPP em Portugal. Observatory on public private partnerships (OPPP), Lisbon, Portugal. Santos, A (2011). Hospital de Cascais é um sarilho. Expresso. July 9. Tversky, A, & Kahneman, D (1974). Judgments under uncertainty: Heuristics and biases. Science, Vol. 185,

13 STATISTICAL APPENDIX This appendix reports the technical details about the methodology used. Table A0. Data Base Hospital CEdif CEquip OutCus Custos Camas Area Ano AMADORA-SINTRA Nª SRª DA GRAÇA - TOMAR PEDRO HISPANO - MATOSINHOS SANTO ANFRÉ - LEIRIA S. TEOTÓNIO - VISEU SANTA MARIA DA FEIRA BARLAVENTO ALGARVIO PORTIMÃO HOSPITALAR COVA DA BEIRA VALE DO SOUSA (INCLUI A PSIQUIATRIA) RAINHA SANTA ISABEL TORRES NOVAS Cascais Braga Costs in Millions. Table A.1. Regression of the Cost of Construction reg LCEdif LCAM Source SS df MS Number of obs = F( 1, 8) = Model Prob > F = Residual R-squared = Adj R-squared = Total Root MSE = LCEdif Coef. Std. Err. t P> t [95% Conf. Interval] LCamas Constante Note that the time variable did not prove statistically significant in any of the regressions presented in this appendix. Forecasts for the logarithms of the costs of the hull and Braga Threshold Limit forecast Hospital Inf Sup Hospital Forecast Limit Inf Limit Sup IC 95% IC 95% Cascais Braga Note: STAT command used was "predictnl LCEdhat = predict (), ci (lled uled)" Table A.2. Regression of the equipment costs (excluding H Feira) reg LCEquip LArea if CEquip>5.5 Source SS df MS Number of obs = F( 1, 7) = 6.41 Model Prob > F = Residual R-squared = Adj R-squared = Total Root MSE = LCEquip Coef. Std. Err. t P> t [95% Conf. Interval] LArea Constante Forecasts for the logarithms of the costs of the equipment in Cascais and Braga Hospital Forecast Limit Inf Limit Sup IC 95% IC 95% Cascais Braga

14 Table A.3. Regression of Other Costs reg LCOu LArea Source SS df MS Number of obs = F( 1, 8) = Model Prob > F = Residual R-squared = Adj R-squared = Total Root MSE = LCOu Coef. Std. Err. t P> t [95% Conf. Interval] LArea Constante Forecasts for Other Costs of logarithms in Cascais and Braga Hospital Previsão Limite Inf Limite Sup IC 95% IC 95% Cascais Braga Table A.4. Regression of Total Costs reg LCustos LArea Source SS df MS Number of obs = F( 1, 8) = Model Prob > F = Residual R-squared = Adj R-squared = Total Root MSE = LCustos Coef. Std. Err. t P> t [95% Conf. Interval] LArea _cons Forecasts for the logarithms of the total costs in Cascais and Braga Hospital Previsão Limite Inf Limite Sup IC 95% IC 95% Cascais 4, , , Braga 4, , , Note: The exponential average of the residuals is , so there was no need to use a correction in the forecast, because the estimator is "smeared." The list of hospitals whose data were used to estimate operating costs appears in Table A.5. Table A.5. Hospitals and medical centers EPE hospitals in

15 1 CH Alto Ave 2 CH Alto Minho 3 CH Baixo Alentejo 4 CH Coimbra 5 CH Cova Beira 6 CH do Barlavento Algarvio 7 CH do Porto 8 CH Lisboa Norte 9 CH Lx Central 10 CH Lx Ocidental 11 CH Médio Ave 12 CH Médio Tejo 13 CH Nordeste 14 CH Setúbal 15 CH Tâmega e Sousa 16 CH Trás- os- Montes e Alto Douro 17 CH VN Gaia / Espinho 18 H de Nossa Senhora do Rosário, - Barreiro 19 H do Espírito Santo - Évora 20 H Garcia de Orta, - Almada 21 H Infante D. Pedro, - Aveiro 22 H S. Sebastião, - Vila da Feira 23 H S. Teotónio, - Viseu 24 H Santa Maria Maior, - Barcelos 25 H Santo André, - Leiria 26 H São João - Porto 27 HD da Figueira da Foz, 28 HD de Santarém, 29 IPOFG - CRO de Coimbra, 30 IPOFG - CRO de Lisboa, 31 IPOFG - CRO do Porto, 32 ULS de Matosinhos, 33 ULS Norte Alentejano Outros Hospitais - SPA 1 CH da Póvoa do Varzim/Vila do Conde 2 CH das Caldas da Rainha 3 CH de Cascais 4 CH de Torres Vedras 5 H Amato Lusitano - Castelo Branco 6 H Curry Cabral - Lisboa 7 H da Universidade de Coimbra 8 H Distrital de Águeda 9 H Distrital de Faro 10 H Distrital de São João da Madeira 11 H Reynaldo dos Santos - Vila Franca de Xira 12 H São Marcos - Braga 13 H Sousa Martins Guarda 14 H Bernardino Lopes de Oliveira - Alcobaça 15 H Cândido de Figueiredo - Tondela 16 H Distrital de Pombal 17 H do Litoral Alentejano - Santiago do Cacém 18 H do Montijo 19 H Dr Francisco Zagalo - Ovar 20 H José Luciano de Castro - Anadia 21 H Nossa Senhora da Assunção - Seia 15

16 22 H Nossa Senhora da Conceição - Valongo 23 H São Miguel - Oliveira de Azeméis 24 H São Pedro Gonçalves Telmo - Peniche 25 H Visconde de Salreu - Estarreja Table A.6. Estimation of the translog cost function, global xi: xtreg lcustot epe i.ano lepisodios lciramb lurgencias lconsultas lhdia licm_drgs lepi_icmdrgs lhdia2 lepi_cons lcons_cir lcons_hdia, fe robust Fixed-effects (within) regression Number of obs = 458 Group variable: id Number of groups = 58 R-sq: within = Obs per group: min = 5 between = avg = 7.9 overall = max = 8 F(19,57) = corr(u_i, Xb) = Prob > F = (Std. Err. adjusted for 58 clusters in id) Robust lcustot Coef. Std. Err. t P> t [95% Conf. Interval] epe _Iano_ _Iano_ _Iano_ _Iano_ _Iano_ _Iano_ _Iano_ lepisodios lciramb lurgencias lconsultas lhdia licm_drgs lepi_icmdrgs lhdia lepi_cons lcons_cir lcons_hdia _cons The next two tables contain the STATA code used to produce estimates of the hospital operating costs. Note that on average, SPA hospitals are 57% of the sample, and SPC hospitals make up the remaining 43%. Table A.7. Procedures for comparing the expected costs with the PSC - Cascais 1. Without fixed effects a) 8 years xi: xtreg lcustot epe i.ano lepisodios lciramb lurgencias lconsultas lhdia licm_drgs lepi_icmdrgs lhdia2 lepi_cons lcons_cir lcons_hdia, fe robust adjust epe=0.57 _Iano_2001=0 _Iano_2002=0 _Iano_2003=0 _Iano_2004=0 _Iano_2005=0 _Iano_2006=0 _Iano_2007=1 lepisodios= lciramb= lurgencias= lconsultas= lhdia= licm_drgs= lepi_icmdrgs= lhdia2= lepi_cons= lcons_cir= lcons_hdia= , ci level(95) b) 2 years 16

17 xi: xtreg lcustot epe i.ano lepisodios lciramb lurgencias lconsultas lhdia licm_drgs lepi_icmdrgs lhdia2 lepi_cons lcons_cir lcons_hdia, fe robust adjust epe=0.57 _Iano_2001=0 _Iano_2002=0 _Iano_2003=0 _Iano_2004=0 _Iano_2005=0 _Iano_2006=0 _Iano_2007=1 lepisodios= lciramb= lurgencias= lconsultas= lhdia= licm_drgs= lepi_icmdrgs= lhdia2= lepi_cons= lcons_cir= lcons_hdia= , ci level(95) 2. With fixed effects a) 8 years xi: reg lcustot epe i.ano lepisodios lciramb lurgencias lconsultas lhdia licm_drgs lepi_icmdrgs lhdia2 lepi_cons lcons_cir lcons_hdia i.id, robust adjust epe=0.57 _Iano_2001=0 _Iano_2002=0 _Iano_2003=0 _Iano_2004=0 _Iano_2005=0 _Iano_2006=0 _Iano_2007=1 lepisodios= lciramb= lurgencias= lconsultas= lhdia= licm_drgs= lepi_icmdrgs= lhdia2= lepi_cons= lcons_cir= lcons_hdia= if id==12, ci level(95) b)2 years xi: reg lcustot epe i.ano lepisodios lciramb lurgencias lconsultas lhdia licm_drgs lepi_icmdrgs lhdia2 lepi_cons lcons_cir lcons_hdia i.id, robust adjust epe=0.57 _Iano_2001=0 _Iano_2002=0 _Iano_2003=0 _Iano_2004=0 _Iano_2005=0 _Iano_2006=0 _Iano_2007=1 lepisodios= lciramb= lurgencias= lconsultas= lhdia= licm_drgs= lepi_icmdrgs= lhdia2= lepi_cons= lcons_cir= lcons_hdia= if id==12, ci level(95) Table A.7. Procedures for comparing the expected costs with the PSC - Braga 1. Without fixed effects a) 9 years xi: xtreg lcustot epe i.ano lepisodios lciramb lurgencias lconsultas lhdia licm_drgs lepi_icmdrgs lhdia2 lepi_cons lcons_cir lcons_hdia, fe robust adjust epe=0.57 _Iano_2001=0 _Iano_2002=0 _Iano_2003=0 _Iano_2004=0 _Iano_2005=0 _Iano_2006=0 _Iano_2007=1 lepisodios= lciramb= lurgencias= lconsultas= lhdia= licm_drgs= lepi_icmdrgs= lhdia2= lepi_cons= lcons_cir= lcons_hdia= , ci level(95) b) 1 year xi: xtreg lcustot epe i.ano lepisodios lciramb lurgencias lconsultas lhdia licm_drgs lepi_icmdrgs lhdia2 lepi_cons lcons_cir lcons_hdia, fe robust adjust epe=0.57 _Iano_2001=0 _Iano_2002=0 _Iano_2003=0 _Iano_2004=0 _Iano_2005=0 _Iano_2006=0 _Iano_2007=1 lepisodios= lciramb= lurgencias= lconsultas= lhdia= licm_drgs= lepi_icmdrgs= lhdia2= lepi_cons= lcons_cir= lcons_hdia= , ci level(95) 2. With fixed effects a)9 years xi: reg lcustot epe i.ano lepisodios lciramb lurgencias lconsultas lhdia licm_drgs lepi_icmdrgs lhdia2 lepi_cons lcons_cir lcons_hdia i.id, robust adjust epe=0.57 _Iano_2001=0 _Iano_2002=0 _Iano_2003=0 _Iano_2004=0 _Iano_2005=0 _Iano_2006=0 _Iano_2007=1 lepisodios= lciramb= lurgencias= lconsultas= lhdia= licm_drgs= lepi_icmdrgs= lhdia2= lepi_cons= lcons_cir= lcons_hdia= if id==8, ci level(95) b)1 year xi: reg lcustot epe i.ano lepisodios lciramb lurgencias lconsultas lhdia licm_drgs lepi_icmdrgs lhdia2 lepi_cons lcons_cir lcons_hdia i.id, robust adjust epe=0.57 _Iano_2001=0 _Iano_2002=0 _Iano_2003=0 _Iano_2004=0 _Iano_2005=0 _Iano_2006=0 _Iano_2007=1 lepisodios= lciramb= lurgencias= lconsultas= lhdia= licm_drgs= lepi_icmdrgs= lhdia2= lepi_cons= lcons_cir= lcons_hdia= if id==8, ci level(95) 17

The Public Sector Comparator of PPP: An empirical evaluation in the Healthcare Sector

The Public Sector Comparator of PPP: An empirical evaluation in the Healthcare Sector The Public Sector Comparator of PPP: An empirical evaluation in the Healthcare Sector Miguel Gouveia 1 CLSBE, Universidade Católica Portuguesa Pedro Raposo CLSBE, Universidade Católica Portuguesa May,

More information

Patient Classification System based on Dependency of Nursing Care

Patient Classification System based on Dependency of Nursing Care based on Dependency of Nursing Care Cristina Duarte Paulino Lisbon October 2008 Challenge Provide high quality care Contain the costs More flexibility from de organisations More rational use of the existing

More information

Department of Economics Session 2012/2013. EC352 Econometric Methods. Solutions to Exercises from Week 10 + 0.0077 (0.052)

Department of Economics Session 2012/2013. EC352 Econometric Methods. Solutions to Exercises from Week 10 + 0.0077 (0.052) Department of Economics Session 2012/2013 University of Essex Spring Term Dr Gordon Kemp EC352 Econometric Methods Solutions to Exercises from Week 10 1 Problem 13.7 This exercise refers back to Equation

More information

MODEL I: DRINK REGRESSED ON GPA & MALE, WITHOUT CENTERING

MODEL I: DRINK REGRESSED ON GPA & MALE, WITHOUT CENTERING Interpreting Interaction Effects; Interaction Effects and Centering Richard Williams, University of Notre Dame, http://www3.nd.edu/~rwilliam/ Last revised February 20, 2015 Models with interaction effects

More information

DETERMINANTS OF CAPITAL ADEQUACY RATIO IN SELECTED BOSNIAN BANKS

DETERMINANTS OF CAPITAL ADEQUACY RATIO IN SELECTED BOSNIAN BANKS DETERMINANTS OF CAPITAL ADEQUACY RATIO IN SELECTED BOSNIAN BANKS Nađa DRECA International University of Sarajevo nadja.dreca@students.ius.edu.ba Abstract The analysis of a data set of observation for 10

More information

Interaction effects between continuous variables (Optional)

Interaction effects between continuous variables (Optional) Interaction effects between continuous variables (Optional) Richard Williams, University of Notre Dame, http://www.nd.edu/~rwilliam/ Last revised February 0, 05 This is a very brief overview of this somewhat

More information

August 2012 EXAMINATIONS Solution Part I

August 2012 EXAMINATIONS Solution Part I August 01 EXAMINATIONS Solution Part I (1) In a random sample of 600 eligible voters, the probability that less than 38% will be in favour of this policy is closest to (B) () In a large random sample,

More information

Marginal Effects for Continuous Variables Richard Williams, University of Notre Dame, http://www3.nd.edu/~rwilliam/ Last revised February 21, 2015

Marginal Effects for Continuous Variables Richard Williams, University of Notre Dame, http://www3.nd.edu/~rwilliam/ Last revised February 21, 2015 Marginal Effects for Continuous Variables Richard Williams, University of Notre Dame, http://www3.nd.edu/~rwilliam/ Last revised February 21, 2015 References: Long 1997, Long and Freese 2003 & 2006 & 2014,

More information

MGT 267 PROJECT. Forecasting the United States Retail Sales of the Pharmacies and Drug Stores. Done by: Shunwei Wang & Mohammad Zainal

MGT 267 PROJECT. Forecasting the United States Retail Sales of the Pharmacies and Drug Stores. Done by: Shunwei Wang & Mohammad Zainal MGT 267 PROJECT Forecasting the United States Retail Sales of the Pharmacies and Drug Stores Done by: Shunwei Wang & Mohammad Zainal Dec. 2002 The retail sale (Million) ABSTRACT The present study aims

More information

The average hotel manager recognizes the criticality of forecasting. However, most

The average hotel manager recognizes the criticality of forecasting. However, most Introduction The average hotel manager recognizes the criticality of forecasting. However, most managers are either frustrated by complex models researchers constructed or appalled by the amount of time

More information

MEDIGRAF - TELEMEDICINE SUSTAINABILITY 2013

MEDIGRAF - TELEMEDICINE SUSTAINABILITY 2013 MEDIGRAF - TELEMEDICINE 2013 Framework Medigraf IT platform is a solution totally developed by PT to suppress an identified need of health care professionals; Populations living in isolated or disadvantaged

More information

Statistics 104 Final Project A Culture of Debt: A Study of Credit Card Spending in America TF: Kevin Rader Anonymous Students: LD, MH, IW, MY

Statistics 104 Final Project A Culture of Debt: A Study of Credit Card Spending in America TF: Kevin Rader Anonymous Students: LD, MH, IW, MY Statistics 104 Final Project A Culture of Debt: A Study of Credit Card Spending in America TF: Kevin Rader Anonymous Students: LD, MH, IW, MY ABSTRACT: This project attempted to determine the relationship

More information

Failure to take the sampling scheme into account can lead to inaccurate point estimates and/or flawed estimates of the standard errors.

Failure to take the sampling scheme into account can lead to inaccurate point estimates and/or flawed estimates of the standard errors. Analyzing Complex Survey Data: Some key issues to be aware of Richard Williams, University of Notre Dame, http://www3.nd.edu/~rwilliam/ Last revised January 24, 2015 Rather than repeat material that is

More information

Lecture 15. Endogeneity & Instrumental Variable Estimation

Lecture 15. Endogeneity & Instrumental Variable Estimation Lecture 15. Endogeneity & Instrumental Variable Estimation Saw that measurement error (on right hand side) means that OLS will be biased (biased toward zero) Potential solution to endogeneity instrumental

More information

Nonlinear Regression Functions. SW Ch 8 1/54/

Nonlinear Regression Functions. SW Ch 8 1/54/ Nonlinear Regression Functions SW Ch 8 1/54/ The TestScore STR relation looks linear (maybe) SW Ch 8 2/54/ But the TestScore Income relation looks nonlinear... SW Ch 8 3/54/ Nonlinear Regression General

More information

João Saraiva Mendes & Ricardo Rosa Santos arquitectos associados

João Saraiva Mendes & Ricardo Rosa Santos arquitectos associados João Saraiva Mendes & Ricardo Rosa Santos arquitectos associados João Saraiva Mendes 1974 2007 Works exclusively in his own works and projects 2001-2006 Collaboration with Manuel Salgado at Risco firm

More information

GETTING STARTED: STATA & R BASIC COMMANDS ECONOMETRICS II. Stata Output Regression of wages on education

GETTING STARTED: STATA & R BASIC COMMANDS ECONOMETRICS II. Stata Output Regression of wages on education GETTING STARTED: STATA & R BASIC COMMANDS ECONOMETRICS II Stata Output Regression of wages on education. sum wage educ Variable Obs Mean Std. Dev. Min Max -------------+--------------------------------------------------------

More information

Please follow the directions once you locate the Stata software in your computer. Room 114 (Business Lab) has computers with Stata software

Please follow the directions once you locate the Stata software in your computer. Room 114 (Business Lab) has computers with Stata software STATA Tutorial Professor Erdinç Please follow the directions once you locate the Stata software in your computer. Room 114 (Business Lab) has computers with Stata software 1.Wald Test Wald Test is used

More information

Median and Average Sales Prices of New Homes Sold in United States

Median and Average Sales Prices of New Homes Sold in United States Jan 1963 $17,200 (NA) Feb 1963 $17,700 (NA) Mar 1963 $18,200 (NA) Apr 1963 $18,200 (NA) May 1963 $17,500 (NA) Jun 1963 $18,000 (NA) Jul 1963 $18,400 (NA) Aug 1963 $17,800 (NA) Sep 1963 $17,900 (NA) Oct

More information

Correlation and Regression

Correlation and Regression Correlation and Regression Scatterplots Correlation Explanatory and response variables Simple linear regression General Principles of Data Analysis First plot the data, then add numerical summaries Look

More information

ESTIMATING AVERAGE TREATMENT EFFECTS: IV AND CONTROL FUNCTIONS, II Jeff Wooldridge Michigan State University BGSE/IZA Course in Microeconometrics

ESTIMATING AVERAGE TREATMENT EFFECTS: IV AND CONTROL FUNCTIONS, II Jeff Wooldridge Michigan State University BGSE/IZA Course in Microeconometrics ESTIMATING AVERAGE TREATMENT EFFECTS: IV AND CONTROL FUNCTIONS, II Jeff Wooldridge Michigan State University BGSE/IZA Course in Microeconometrics July 2009 1. Quantile Treatment Effects 2. Control Functions

More information

ECON 142 SKETCH OF SOLUTIONS FOR APPLIED EXERCISE #2

ECON 142 SKETCH OF SOLUTIONS FOR APPLIED EXERCISE #2 University of California, Berkeley Prof. Ken Chay Department of Economics Fall Semester, 005 ECON 14 SKETCH OF SOLUTIONS FOR APPLIED EXERCISE # Question 1: a. Below are the scatter plots of hourly wages

More information

Module 6: Introduction to Time Series Forecasting

Module 6: Introduction to Time Series Forecasting Using Statistical Data to Make Decisions Module 6: Introduction to Time Series Forecasting Titus Awokuse and Tom Ilvento, University of Delaware, College of Agriculture and Natural Resources, Food and

More information

EMC meeting No./date 120/July 09. Decision. INSTITUTO POLITÉCNICO DE BEJA ESCOLA SUPERIOR DE TECNOLOGIA E GESTÃO DE BEJA School Address/Contact

EMC meeting No./date 120/July 09. Decision. INSTITUTO POLITÉCNICO DE BEJA ESCOLA SUPERIOR DE TECNOLOGIA E GESTÃO DE BEJA School Address/Contact s submitted to Expert: Name: INSTITUTO POLITÉCNICO DE BEJA ESCOLA SUPERIOR DE TECNOLOGIA E GESTÃO DE BEJA Rua D. Afonso III, n.º 1 7800-050 BEJA PORTUGAL Tel. (+351) 284 311 540 Fax (+351) 284 311 542

More information

HOSPIRA (HSP US) HISTORICAL COMMON STOCK PRICE INFORMATION

HOSPIRA (HSP US) HISTORICAL COMMON STOCK PRICE INFORMATION 30-Apr-2004 28.35 29.00 28.20 28.46 28.55 03-May-2004 28.50 28.70 26.80 27.04 27.21 04-May-2004 26.90 26.99 26.00 26.00 26.38 05-May-2004 26.05 26.69 26.00 26.35 26.34 06-May-2004 26.31 26.35 26.05 26.26

More information

MULTIPLE REGRESSION EXAMPLE

MULTIPLE REGRESSION EXAMPLE MULTIPLE REGRESSION EXAMPLE For a sample of n = 166 college students, the following variables were measured: Y = height X 1 = mother s height ( momheight ) X 2 = father s height ( dadheight ) X 3 = 1 if

More information

Multiple Optimization Using the JMP Statistical Software Kodak Research Conference May 9, 2005

Multiple Optimization Using the JMP Statistical Software Kodak Research Conference May 9, 2005 Multiple Optimization Using the JMP Statistical Software Kodak Research Conference May 9, 2005 Philip J. Ramsey, Ph.D., Mia L. Stephens, MS, Marie Gaudard, Ph.D. North Haven Group, http://www.northhavengroup.com/

More information

Industry Environment and Concepts for Forecasting 1

Industry Environment and Concepts for Forecasting 1 Table of Contents Industry Environment and Concepts for Forecasting 1 Forecasting Methods Overview...2 Multilevel Forecasting...3 Demand Forecasting...4 Integrating Information...5 Simplifying the Forecast...6

More information

The Wage Return to Education: What Hides Behind the Least Squares Bias?

The Wage Return to Education: What Hides Behind the Least Squares Bias? DISCUSSION PAPER SERIES IZA DP No. 8855 The Wage Return to Education: What Hides Behind the Least Squares Bias? Corrado Andini February 2015 Forschungsinstitut zur Zukunft der Arbeit Institute for the

More information

Lab 5 Linear Regression with Within-subject Correlation. Goals: Data: Use the pig data which is in wide format:

Lab 5 Linear Regression with Within-subject Correlation. Goals: Data: Use the pig data which is in wide format: Lab 5 Linear Regression with Within-subject Correlation Goals: Data: Fit linear regression models that account for within-subject correlation using Stata. Compare weighted least square, GEE, and random

More information

Regression Analysis. Data Calculations Output

Regression Analysis. Data Calculations Output Regression Analysis In an attempt to find answers to questions such as those posed above, empirical labour economists use a useful tool called regression analysis. Regression analysis is essentially a

More information

REGRESSION LINES IN STATA

REGRESSION LINES IN STATA REGRESSION LINES IN STATA THOMAS ELLIOTT 1. Introduction to Regression Regression analysis is about eploring linear relationships between a dependent variable and one or more independent variables. Regression

More information

List of accommodations with special EMCDDA rates

List of accommodations with special EMCDDA rates List of accommodations with special EMCDDA rates This overview contains a list of hotels, guest-houses and apart hotels in Lisbon for which the EMCDDA has special rates. In case you are interested in booking

More information

An Empirical Analysis of Insider Rates vs. Outsider Rates in Bank Lending

An Empirical Analysis of Insider Rates vs. Outsider Rates in Bank Lending An Empirical Analysis of Insider Rates vs. Outsider Rates in Bank Lending Lamont Black* Indiana University Federal Reserve Board of Governors November 2006 ABSTRACT: This paper analyzes empirically the

More information

Introduction to Stata

Introduction to Stata Introduction to Stata September 23, 2014 Stata is one of a few statistical analysis programs that social scientists use. Stata is in the mid-range of how easy it is to use. Other options include SPSS,

More information

LATIN AMERICA. A World of Inspiration

LATIN AMERICA. A World of Inspiration Africa EUROPe LATIN AMERICA A World of Inspiration Profile The Mota-Engil Group An international and diversified vision for the future The Mota-Engil Group has a business record of 67 years, marked by

More information

International Statistical Institute, 56th Session, 2007: Phil Everson

International Statistical Institute, 56th Session, 2007: Phil Everson Teaching Regression using American Football Scores Everson, Phil Swarthmore College Department of Mathematics and Statistics 5 College Avenue Swarthmore, PA198, USA E-mail: peverso1@swarthmore.edu 1. Introduction

More information

Managing Staffing in High Demand Variability Environments

Managing Staffing in High Demand Variability Environments Managing Staffing in High Demand Variability Environments By: Dennis J. Monroe, Six Sigma Master Black Belt, Lean Master, and Vice President, Juran Institute, Inc. Many functions within a variety of businesses

More information

Portugal. Sub-Programme / Key Activity. HOSPITAL GARCIA DE ORTA PAR 2007 ERASMUS Erasmus Network European Pathology Assessment & Learning System

Portugal. Sub-Programme / Key Activity. HOSPITAL GARCIA DE ORTA PAR 2007 ERASMUS Erasmus Network European Pathology Assessment & Learning System HOSPITAL GARCIA DE ORTA PAR 2007 ERASMUS Erasmus Network European Pathology Assessment & Learning System UNIVERSIDADE DE AVEIRO PAR 2007 ERASMUS BeFlex Plus: Progress on Flexibility in the Bologna Reform

More information

College Education Matters for Happier Marriages and Higher Salaries ----Evidence from State Level Data in the US

College Education Matters for Happier Marriages and Higher Salaries ----Evidence from State Level Data in the US College Education Matters for Happier Marriages and Higher Salaries ----Evidence from State Level Data in the US Anonymous Authors: SH, AL, YM Contact TF: Kevin Rader Abstract It is a general consensus

More information

The retrofit of a closed-loop distribution network: the case of lead batteries

The retrofit of a closed-loop distribution network: the case of lead batteries 20 th European Symposium on Computer Aided Process Engineering ESCAPE20 S. Pierucci and G. Buzzi Ferraris (Editors) 2010 Elsevier B.V. All rights reserved. The retrofit of a closed-loop distribution network:

More information

THE UNIVERSITY OF BOLTON

THE UNIVERSITY OF BOLTON JANUARY Jan 1 6.44 8.24 12.23 2.17 4.06 5.46 Jan 2 6.44 8.24 12.24 2.20 4.07 5.47 Jan 3 6.44 8.24 12.24 2.21 4.08 5.48 Jan 4 6.44 8.24 12.25 2.22 4.09 5.49 Jan 5 6.43 8.23 12.25 2.24 4.10 5.50 Jan 6 6.43

More information

Testing for serial correlation in linear panel-data models

Testing for serial correlation in linear panel-data models The Stata Journal (2003) 3, Number 2, pp. 168 177 Testing for serial correlation in linear panel-data models David M. Drukker Stata Corporation Abstract. Because serial correlation in linear panel-data

More information

IAPRI Quantitative Analysis Capacity Building Series. Multiple regression analysis & interpreting results

IAPRI Quantitative Analysis Capacity Building Series. Multiple regression analysis & interpreting results IAPRI Quantitative Analysis Capacity Building Series Multiple regression analysis & interpreting results How important is R-squared? R-squared Published in Agricultural Economics 0.45 Best article of the

More information

Econ 371 Problem Set #3 Answer Sheet

Econ 371 Problem Set #3 Answer Sheet Econ 371 Problem Set #3 Answer Sheet 4.3 In this question, you are told that a OLS regression analysis of average weekly earnings yields the following estimated model. AW E = 696.7 + 9.6 Age, R 2 = 0.023,

More information

1. What is the critical value for this 95% confidence interval? CV = z.025 = invnorm(0.025) = 1.96

1. What is the critical value for this 95% confidence interval? CV = z.025 = invnorm(0.025) = 1.96 1 Final Review 2 Review 2.1 CI 1-propZint Scenario 1 A TV manufacturer claims in its warranty brochure that in the past not more than 10 percent of its TV sets needed any repair during the first two years

More information

Executive Summary. Abstract. Heitman Analytics Conclusions:

Executive Summary. Abstract. Heitman Analytics Conclusions: Prepared By: Adam Petranovich, Economic Analyst apetranovich@heitmananlytics.com 541 868 2788 Executive Summary Abstract The purpose of this study is to provide the most accurate estimate of historical

More information

is paramount in advancing any economy. For developed countries such as

is paramount in advancing any economy. For developed countries such as Introduction The provision of appropriate incentives to attract workers to the health industry is paramount in advancing any economy. For developed countries such as Australia, the increasing demand for

More information

Value of IEEE s Online Collections

Value of IEEE s Online Collections Value of IEEE s Online Collections Judy H. Brady, IEEE Aveiro, Portugal February 2013 About the IEEE A not-for-profit society World s largest technical membership association with over 400,000 members

More information

Linear Regression Models with Logarithmic Transformations

Linear Regression Models with Logarithmic Transformations Linear Regression Models with Logarithmic Transformations Kenneth Benoit Methodology Institute London School of Economics kbenoit@lse.ac.uk March 17, 2011 1 Logarithmic transformations of variables Considering

More information

Age to Age Factor Selection under Changing Development Chris G. Gross, ACAS, MAAA

Age to Age Factor Selection under Changing Development Chris G. Gross, ACAS, MAAA Age to Age Factor Selection under Changing Development Chris G. Gross, ACAS, MAAA Introduction A common question faced by many actuaries when selecting loss development factors is whether to base the selected

More information

COMPARISON OF FIXED & VARIABLE RATES (25 YEARS) CHARTERED BANK ADMINISTERED INTEREST RATES - PRIME BUSINESS*

COMPARISON OF FIXED & VARIABLE RATES (25 YEARS) CHARTERED BANK ADMINISTERED INTEREST RATES - PRIME BUSINESS* COMPARISON OF FIXED & VARIABLE RATES (25 YEARS) 2 Fixed Rates Variable Rates FIXED RATES OF THE PAST 25 YEARS AVERAGE RESIDENTIAL MORTGAGE LENDING RATE - 5 YEAR* (Per cent) Year Jan Feb Mar Apr May Jun

More information

COMPARISON OF FIXED & VARIABLE RATES (25 YEARS) CHARTERED BANK ADMINISTERED INTEREST RATES - PRIME BUSINESS*

COMPARISON OF FIXED & VARIABLE RATES (25 YEARS) CHARTERED BANK ADMINISTERED INTEREST RATES - PRIME BUSINESS* COMPARISON OF FIXED & VARIABLE RATES (25 YEARS) 2 Fixed Rates Variable Rates FIXED RATES OF THE PAST 25 YEARS AVERAGE RESIDENTIAL MORTGAGE LENDING RATE - 5 YEAR* (Per cent) Year Jan Feb Mar Apr May Jun

More information

DECEMBER 11 BUSINESS MEETINGS LISBON, PORTUGAL ASSOCIAÇÃO DOS PROFISSIONAIS E EMPRESAS DE MEDIAÇÃO IMOBILIÁRIA DE PORTUGAL

DECEMBER 11 BUSINESS MEETINGS LISBON, PORTUGAL ASSOCIAÇÃO DOS PROFISSIONAIS E EMPRESAS DE MEDIAÇÃO IMOBILIÁRIA DE PORTUGAL DECEMBER 11 BUSINESS MEETINGS ASSOCIAÇÃO DOS PROFISSIONAIS E EMPRESAS DE MEDIAÇÃO IMOBILIÁRIA DE PORTUGAL HISTORY: HIGHLIGHTS 1143 1910 1933 1974 1986 Monarchy Republic Estado Novo Authoritarian Revolution

More information

Regression in Stata. Alicia Doyle Lynch Harvard-MIT Data Center (HMDC)

Regression in Stata. Alicia Doyle Lynch Harvard-MIT Data Center (HMDC) Regression in Stata Alicia Doyle Lynch Harvard-MIT Data Center (HMDC) Documents for Today Find class materials at: http://libraries.mit.edu/guides/subjects/data/ training/workshops.html Several formats

More information

Forecasting in STATA: Tools and Tricks

Forecasting in STATA: Tools and Tricks Forecasting in STATA: Tools and Tricks Introduction This manual is intended to be a reference guide for time series forecasting in STATA. It will be updated periodically during the semester, and will be

More information

Exponential Smoothing with Trend. As we move toward medium-range forecasts, trend becomes more important.

Exponential Smoothing with Trend. As we move toward medium-range forecasts, trend becomes more important. Exponential Smoothing with Trend As we move toward medium-range forecasts, trend becomes more important. Incorporating a trend component into exponentially smoothed forecasts is called double exponential

More information

Handling missing data in Stata a whirlwind tour

Handling missing data in Stata a whirlwind tour Handling missing data in Stata a whirlwind tour 2012 Italian Stata Users Group Meeting Jonathan Bartlett www.missingdata.org.uk 20th September 2012 1/55 Outline The problem of missing data and a principled

More information

Portuguese Research Institutions in History

Portuguese Research Institutions in History José Luís Cardoso Technical University of Lisbon jcardoso@iseg.utt.pt Luís Adão da Fonseca University of Porto lfonseca@letras.up.pt The growth in historical research in Portugal over the last two decades

More information

Change-Point Analysis: A Powerful New Tool For Detecting Changes

Change-Point Analysis: A Powerful New Tool For Detecting Changes Change-Point Analysis: A Powerful New Tool For Detecting Changes WAYNE A. TAYLOR Baxter Healthcare Corporation, Round Lake, IL 60073 Change-point analysis is a powerful new tool for determining whether

More information

Team References List of the main services provided by our team

Team References List of the main services provided by our team Municipality of Gouveia Municipality of Mangualde Municipality of Arraiolos Legal counsel, to dissolve the municipal company DLGG-Sport, leisure and culture, and the creation of new municipal company GOUVEIA

More information

Message from the National Coordinator of RPA 3. RPA Report June 1st 2009 May 31st 2010 5. Overall Analysis 7. Hip Primary Arthroplasty 11

Message from the National Coordinator of RPA 3. RPA Report June 1st 2009 May 31st 2010 5. Overall Analysis 7. Hip Primary Arthroplasty 11 Table of Contents Message from the National Coordinator of RPA 3 RPA Report June 1st 2009 May 31st 2010 5 Overall Analysis 7 Hip Primary Arthroplasty 11 Hip Revision Arthroplasty 38 Knee Primary Arthroplasty

More information

Data Analysis Methodology 1

Data Analysis Methodology 1 Data Analysis Methodology 1 Suppose you inherited the database in Table 1.1 and needed to find out what could be learned from it fast. Say your boss entered your office and said, Here s some software project

More information

Case 2:08-cv-02463-ABC-E Document 1-4 Filed 04/15/2008 Page 1 of 138. Exhibit 8

Case 2:08-cv-02463-ABC-E Document 1-4 Filed 04/15/2008 Page 1 of 138. Exhibit 8 Case 2:08-cv-02463-ABC-E Document 1-4 Filed 04/15/2008 Page 1 of 138 Exhibit 8 Case 2:08-cv-02463-ABC-E Document 1-4 Filed 04/15/2008 Page 2 of 138 Domain Name: CELLULARVERISON.COM Updated Date: 12-dec-2007

More information

The Risk Driver Approach to Project Schedule Risk Analysis

The Risk Driver Approach to Project Schedule Risk Analysis The Risk Driver Approach to Project Schedule Risk Analysis A Webinar presented by David T. Hulett, Ph.D. Hulett & Associates, LLC To the College of Performance Management April 18, 2013 2013 Hulett & Associates,

More information

ENERGY STAR for Data Centers

ENERGY STAR for Data Centers ENERGY STAR for Data Centers Alexandra Sullivan US EPA, ENERGY STAR February 4, 2010 Agenda ENERGY STAR Buildings Overview Energy Performance Ratings Portfolio Manager Data Center Initiative Objective

More information

José M. F. Moura, Director of ICTI at Carnegie Mellon Carnegie Mellon Victor Barroso, Director of ICTI in Portugal www.cmu.

José M. F. Moura, Director of ICTI at Carnegie Mellon Carnegie Mellon Victor Barroso, Director of ICTI in Portugal www.cmu. José M. F. Moura, Director of ICTI at Victor Barroso, Director of ICTI in Portugal www.cmu.edu/portugal Portugal program timeline 2005: Discussions and meeting with Ministry of Science Technology, Higher

More information

2. Simple Linear Regression

2. Simple Linear Regression Research methods - II 3 2. Simple Linear Regression Simple linear regression is a technique in parametric statistics that is commonly used for analyzing mean response of a variable Y which changes according

More information

The Numbers Behind the MLB Anonymous Students: AD, CD, BM; (TF: Kevin Rader)

The Numbers Behind the MLB Anonymous Students: AD, CD, BM; (TF: Kevin Rader) The Numbers Behind the MLB Anonymous Students: AD, CD, BM; (TF: Kevin Rader) Abstract This project measures the effects of various baseball statistics on the win percentage of all the teams in MLB. Data

More information

Simple Linear Regression

Simple Linear Regression STAT 101 Dr. Kari Lock Morgan Simple Linear Regression SECTIONS 9.3 Confidence and prediction intervals (9.3) Conditions for inference (9.1) Want More Stats??? If you have enjoyed learning how to analyze

More information

Air passenger departures forecast models A technical note

Air passenger departures forecast models A technical note Ministry of Transport Air passenger departures forecast models A technical note By Haobo Wang Financial, Economic and Statistical Analysis Page 1 of 15 1. Introduction Sine 1999, the Ministry of Business,

More information

Methodology For Illinois Electric Customers and Sales Forecasts: 2016-2025

Methodology For Illinois Electric Customers and Sales Forecasts: 2016-2025 Methodology For Illinois Electric Customers and Sales Forecasts: 2016-2025 In December 2014, an electric rate case was finalized in MEC s Illinois service territory. As a result of the implementation of

More information

Sensex Realized Volatility Index

Sensex Realized Volatility Index Sensex Realized Volatility Index Introduction: Volatility modelling has traditionally relied on complex econometric procedures in order to accommodate the inherent latent character of volatility. Realized

More information

OBJECTIVE ASSESSMENT OF FORECASTING ASSIGNMENTS USING SOME FUNCTION OF PREDICTION ERRORS

OBJECTIVE ASSESSMENT OF FORECASTING ASSIGNMENTS USING SOME FUNCTION OF PREDICTION ERRORS OBJECTIVE ASSESSMENT OF FORECASTING ASSIGNMENTS USING SOME FUNCTION OF PREDICTION ERRORS CLARKE, Stephen R. Swinburne University of Technology Australia One way of examining forecasting methods via assignments

More information

AT&T Global Network Client for Windows Product Support Matrix January 29, 2015

AT&T Global Network Client for Windows Product Support Matrix January 29, 2015 AT&T Global Network Client for Windows Product Support Matrix January 29, 2015 Product Support Matrix Following is the Product Support Matrix for the AT&T Global Network Client. See the AT&T Global Network

More information

Multicollinearity Richard Williams, University of Notre Dame, http://www3.nd.edu/~rwilliam/ Last revised January 13, 2015

Multicollinearity Richard Williams, University of Notre Dame, http://www3.nd.edu/~rwilliam/ Last revised January 13, 2015 Multicollinearity Richard Williams, University of Notre Dame, http://www3.nd.edu/~rwilliam/ Last revised January 13, 2015 Stata Example (See appendices for full example).. use http://www.nd.edu/~rwilliam/stats2/statafiles/multicoll.dta,

More information

From the help desk: Bootstrapped standard errors

From the help desk: Bootstrapped standard errors The Stata Journal (2003) 3, Number 1, pp. 71 80 From the help desk: Bootstrapped standard errors Weihua Guan Stata Corporation Abstract. Bootstrapping is a nonparametric approach for evaluating the distribution

More information

CONTRACT NUMBER: TCP8-GA

CONTRACT NUMBER: TCP8-GA CONTRACT NUMBER: TCP8-GA-2009-234082 Gare do Oriente Interchange Station and Connection with Linha do Norte Railways - Case Study 15th December 2011 Lisbon, Portugal Liliana Magalhães, José Viegas, Rosário

More information

Demand forecasting & Aggregate planning in a Supply chain. Session Speaker Prof.P.S.Satish

Demand forecasting & Aggregate planning in a Supply chain. Session Speaker Prof.P.S.Satish Demand forecasting & Aggregate planning in a Supply chain Session Speaker Prof.P.S.Satish 1 Introduction PEMP-EMM2506 Forecasting provides an estimate of future demand Factors that influence demand and

More information

FORECASTING. Operations Management

FORECASTING. Operations Management 2013 FORECASTING Brad Fink CIT 492 Operations Management Executive Summary Woodlawn hospital needs to forecast type A blood so there is no shortage for the week of 12 October, to correctly forecast, a

More information

operating under private management, but with payments from the public to the private partner much higher than initially contracted.

operating under private management, but with payments from the public to the private partner much higher than initially contracted. BEST PRACTICE UNECE PPP Newsletter Issue 3 PPP hospitals and the Portuguese experience Mr. Rui Sousa Monteiro, Senior Economist PPP Unit, Parpública SA, Portugal Mr. Rui Sousa Monteiro is one of the region

More information

Qi Liu Rutgers Business School ISACA New York 2013

Qi Liu Rutgers Business School ISACA New York 2013 Qi Liu Rutgers Business School ISACA New York 2013 1 What is Audit Analytics The use of data analysis technology in Auditing. Audit analytics is the process of identifying, gathering, validating, analyzing,

More information

ANNEXURE 1 STATUS OF 518 DEMAT REQUESTS PENDING WITH NSDL

ANNEXURE 1 STATUS OF 518 DEMAT REQUESTS PENDING WITH NSDL ANNEXURE 1 STATUS OF 518 DEMAT REQUESTS PENDING WITH NSDL Sr. No. Demat Request No.(DRN) DP ID Client ID Date of Demat Request Received Quantity Requested Date of Demat Request Processed No. of days of

More information

Interaction effects and group comparisons Richard Williams, University of Notre Dame, http://www3.nd.edu/~rwilliam/ Last revised February 20, 2015

Interaction effects and group comparisons Richard Williams, University of Notre Dame, http://www3.nd.edu/~rwilliam/ Last revised February 20, 2015 Interaction effects and group comparisons Richard Williams, University of Notre Dame, http://www3.nd.edu/~rwilliam/ Last revised February 20, 2015 Note: This handout assumes you understand factor variables,

More information

INTENTION TO FLOAT ( ITF ) Press release, 14/01/2014

INTENTION TO FLOAT ( ITF ) Press release, 14/01/2014 INTENTION TO FLOAT ( ITF ) Press release, 14/01/2014 NOT FOR RELEASE, PUBLICATION OR DISTRIBUTION, DIRECTLY OR INDIRECTLY IN OR INTO THE UNITED STATES, CANADA, AUSTRALIA OR JAPAN This announcement is not

More information

1998-2000 Collaboration with Manuel Salgado at Risco firm in Lisbon. 1997-1998 Junior Architect at Expo 98 world Exhibition in Lisbon

1998-2000 Collaboration with Manuel Salgado at Risco firm in Lisbon. 1997-1998 Junior Architect at Expo 98 world Exhibition in Lisbon João Saraiva Mendes 1974 2007 Works exclusively in his own works and projects 2001-2006 Collaboration with Manuel Salgado at Risco firm and development of own works and projects 2001 Partnership with Ricardo

More information

Quantitative Methods for Economics Tutorial 9. Katherine Eyal

Quantitative Methods for Economics Tutorial 9. Katherine Eyal Quantitative Methods for Economics Tutorial 9 Katherine Eyal TUTORIAL 9 4 October 2010 ECO3021S Part A: Problems 1. In Problem 2 of Tutorial 7, we estimated the equation ŝleep = 3, 638.25 0.148 totwrk

More information

HURDLE AND SELECTION MODELS Jeff Wooldridge Michigan State University BGSE/IZA Course in Microeconometrics July 2009

HURDLE AND SELECTION MODELS Jeff Wooldridge Michigan State University BGSE/IZA Course in Microeconometrics July 2009 HURDLE AND SELECTION MODELS Jeff Wooldridge Michigan State University BGSE/IZA Course in Microeconometrics July 2009 1. Introduction 2. A General Formulation 3. Truncated Normal Hurdle Model 4. Lognormal

More information

Poverty Assessment Tool Accuracy Submission USAID/IRIS Tool for Peru Submitted: September 15, 2011

Poverty Assessment Tool Accuracy Submission USAID/IRIS Tool for Peru Submitted: September 15, 2011 Poverty Assessment Tool Submission USAID/IRIS Tool for Peru Submitted: September 15, 2011 The following report is divided into five sections. Section 1 describes the data used to create the Poverty Assessment

More information

Included Tour Features:

Included Tour Features: Northern Portugal Douro Valley To discover Portugal you really need to live it, meet its friendly people, take in the scenery and experience its rich and healthy gastronomy. Driving around the country

More information

Integrated Resource Plan

Integrated Resource Plan Integrated Resource Plan March 19, 2004 PREPARED FOR KAUA I ISLAND UTILITY COOPERATIVE LCG Consulting 4962 El Camino Real, Suite 112 Los Altos, CA 94022 650-962-9670 1 IRP 1 ELECTRIC LOAD FORECASTING 1.1

More information

Multiple Regression: What Is It?

Multiple Regression: What Is It? Multiple Regression Multiple Regression: What Is It? Multiple regression is a collection of techniques in which there are multiple predictors of varying kinds and a single outcome We are interested in

More information

The introduction of new methods for price observations in the Consumer Price Index (CPI) New methods for airline tickets and package holidays

The introduction of new methods for price observations in the Consumer Price Index (CPI) New methods for airline tickets and package holidays Statistics Netherlands Economics, Enterprises and NA Government Finance and Consumer Prices P.O.Box 24500 2490 HA Den Haag The Netherlands The introduction of new methods for price observations in the

More information

Practice Profile. Rui Castro Practice Profile 1

Practice Profile. Rui Castro Practice Profile 1 Practice Profile Rui Castro Practice Profile 1 Table of contents 1. Presentation 2. Curriculum Curriculum vitae 3. Works The practice work Rui Castro Practice Profile 2 1. Presentation Rui Castro is an

More information

Quick Stata Guide by Liz Foster

Quick Stata Guide by Liz Foster by Liz Foster Table of Contents Part 1: 1 describe 1 generate 1 regress 3 scatter 4 sort 5 summarize 5 table 6 tabulate 8 test 10 ttest 11 Part 2: Prefixes and Notes 14 by var: 14 capture 14 use of the

More information

Should we Really Care about Building Business. Cycle Coincident Indexes!

Should we Really Care about Building Business. Cycle Coincident Indexes! Should we Really Care about Building Business Cycle Coincident Indexes! Alain Hecq University of Maastricht The Netherlands August 2, 2004 Abstract Quite often, the goal of the game when developing new

More information

S&P Year Rolling Period Total Returns

S&P Year Rolling Period Total Returns S&P 500 10 Year Rolling Period Total Returns Summary: 1926 June 2013 700% 600% 500% 400% 300% 200% 100% 0% 100% Scatter chart of all 931 ten year periods. There were 931 ten year rolling periods from January

More information

NAV HISTORY OF DBH FIRST MUTUAL FUND (DBH1STMF)

NAV HISTORY OF DBH FIRST MUTUAL FUND (DBH1STMF) NAV HISTORY OF DBH FIRST MUTUAL FUND () Date NAV 11-Aug-16 10.68 8.66 0.38% -0.07% 0.45% 3.81% 04-Aug-16 10.64 8.66-0.19% 0.87% -1.05% 3.76% 28-Jul-16 10.66 8.59 0.00% -0.34% 0.34% 3.89% 21-Jul-16 10.66

More information

Volatility in the Overnight Money-Market Rate

Volatility in the Overnight Money-Market Rate 5 Volatility in the Overnight Money-Market Rate Allan Bødskov Andersen, Economics INTRODUCTION AND SUMMARY This article analyses the day-to-day fluctuations in the Danish overnight money-market rate during

More information