Table 1: Distribution Revenue at Current Rates. PowerStream Consolidated 2009 OEB Approved PS South 2009 Actual 2010 Actual 2011 Actual

Transcription

1 EB-0-0 Exhibit C Schedule Page of Filed May, 0 0 THROUGHPUT REVENUE OVERVIEW The components that derive revenue at current rates are identified in Exhibit C,, Schedules to. PowerStream has applied current approved rates to the test year customer and sales forecast in order to derive the test year distribution revenue. At current approved rates PowerStream s distribution revenue including smart meter incremental revenue rate riders ( SMIRR ) is $,0, which is a.% increase over 0. In 00 and 0, PowerStream filed applications EB and EB 0-0, respectively, for recovery of costs associated with the installation of smart meters. The Board approved an annual incremental revenue requirement of $,,000 in 00 and $,0,000 in 0. This incremental distribution revenue generated by these mini-rebasing applications has contributed to year over year variances of approximately.0% annually from 00 to 0. The year over year changes are identified in Table. Table : Distribution Revenue at Current Rates 00 OEB Approved PS North PowerStream Consolidated 00 OEB Approved PS South 00 Actual 00 Actual 0 Actual 0 Bridge Year 0 Test Year Total Distribution Revenue $,, $,, $,0, $,, $,, $0,, $,0, % Change Year over Year.%.%.%.% $ Change Year over Year $,,0 $,0, $,,0 $,, 0 The schedules included in this Exhibit outline and describe PowerStream s load, customer, and distribution revenue forecasts. The load forecast methodology and assumptions are described in detail in Exhibit C,, Schedule. PowerStream s purchase forecast is based on a linear regression model. The load forecasting model relates monthly historical purchases to monthly weather conditions (measured in cooling-degree-days ( CDD ) and heating-degreedays (HDD)), and Ontario Gross Domestic Product ( GDP ) as a proxy for service area customer growth and economic activity. The values for Ontario GDP for 0 and 0 are. and. percent, respectively based on publicly available publications by six major banks as of January, 0. Further adjustments for projected Conservation and Demand Management ( CDM ) reductions and estimated distribution losses are made to derive distribution sales.

2 EB-0-0 Exhibit C Schedule Page of Filed May, 0 0 PowerStream s customer forecast is derived based on historic trending by rate class. The customer forecast methodology is described in detail in Exhibit C,, Schedule. Customer growth is slowing from historic levels to approximately.0% in both the bridge and test years. PowerStream has peaked in terms of high growth single family developments and therefore residential customer growth is beginning to reduce as the availability of green field development becomes less. In addition, economic factors in recent years have contributed to the slower pace of growth for all classes. Table summarizes the 0 test year forecast inputs that are used to derive distribution revenue and illustrates the year over year changes in distribution sales (kwh and kw) and customer growth. Table : Distribution Sales (kwh and kw) and Customers 00 OEB Approved PS North PowerStream Consolidated 00 OEB Approved PS South 00 Actuals 00 Actual 0 Actual 0 Bridge Year 0 Test Year Consumption, KWH,0,,,,0,0,0,,00,,,0,,,,,0,,,,0 Demand, KW,0,0 0,00,,,,,,0,,,,, Customer Count,,, 0,,, 0, Variance Analysis (units) 00 vs vs vs. 0 0 vs. 0 Consumption, KWH,,0 0,0,,0,,0, Demand, KW,,,,0 Customer Count,,,, Variance Analysis (%) 00 vs vs vs. 0 0 vs. 0 Consumption, KWH.% 0.% 0.% 0.% Demand, KW.%.0%.% 0.% Customer Count.%.%.%.% 0 Year over year data analysis is showing that consumption per customer is trending lower which may be attributable to a variety of factors including the 00 economic recession and slow economic recovery, CDM initiatives, smart meters and an increase in general knowledge regarding energy pricing, general efficiencies regarding appliances and home/business construction and a shift from single family dwellings to building intensification. As a result, PowerStream is experiencing lower consumption per customer and a tempering in distribution revenue growth. Figure below illustrates the sales per customer trend. The data are adjusted or normalized for weather.

3 EB-0-0 Exhibit C Schedule Page of Filed May, 0,000 0,000,000 0,000,000 0,000, Figure : Normalized Energy Sales per Customer (00 0) PowerStream anticipates that the impacts of CDM and energy efficiencies will persist as the Province continues to pursue this initiative. As a result, various pressures either economic or industry related are contributing to lower trends in consumption, customer and distribution revenue growth.

4 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 0 LOAD FORECAST Introduction In the 00 Electricity Distribution Rate ( EDR ) Application, PowerStream used a load forecasting methodology which was developed in-house using a SSPS software platform. This methodology performed reasonably well and was accepted by the Board in its last cost of service application EB-00-0 In addition to use for rate filing purposes, this load forecast model was used for setting revenue targets in the annual budgeting process between 00 and 0. Table below shows 00-0 forecast energy purchases compared to the 00-0 actuals. Table : Forecast vs. Actual Energy Purchases Forecast Actuals Variance, MWH Variance, % 00,00,,, (,0) -.% 00,,,,,0.% 0,,,,0 (,0) -.% Note: 00 - PowerStream South, PS Consolidated 0 Given that PowerStream continues to strive to improve its load forecasting methodology, PowerStream explored the ability to forecast class-specific loads, as suggested by the Board in 00, EB-00-0 Draft Rate Order, Schedule H, Section.. Class specific sales models were not nearly as strong statistically as the total purchase model. There is significant variation in monthly billing data as it reflects bi-monthly readings for residential customers. In addition, the data includes various billing adjustments and the historic data set is too short (00-0) in order to normalize this variation. As a rough guide, there is a possibility of overfitting the data if there are less than ten data points per coefficient estimated. Therefore, the decision was made to continue working with a modelling approach using total monthly energy purchases. PowerStream explored various forecasting options and tools available in the market. As a result, PowerStream selected a comprehensive energy forecasting tool, MetrixND supported by Itron Inc. PowerStream is confident that this forecasting tool will assist in providing greater

5 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 efficiencies in deriving its load forecast results based on MetrixND focus for use in the energy sector. MetrixND is a comprehensive energy forecasting tool that can be used for all energy forecasting applications with market share of over 00 users from 0 utilities and energy companies primarily in North America. The following is a list of other Canadian users of MetrixND software: 0 Alberta Electricity System Operator BC Hydro Enersource Hydro Mississauga Enmax Power Corp Hydro One Networks Inc Hydro Ottawa Hydro Quebec Independent Electricity System Operator Manitoba Hydro New Brunswick Nova Scotia Power Ontario Power Generation TransAlta Union Gas 0 Modelling Approach PowerStream has adopted a relatively straightforward approach for forecasting short-term energy purchases. PowerStream s purchase forecast is based on a linear regression model. PowerStream is cognisant that Conservation and Demand Management ( CDM ) is a key initiative in the Province of Ontario since the enacting of the Green Energy and Green Economy

6 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Act, 00. As a result, CDM programs will impact future electricity loads. Therefore, a considerable amount of time was spent determining a robust, effective and accurate methodology for measuring the expected impacts of CDM programs on future loads in order to ensure that the load forecast reflects this change from historical levels. Three commonly used forecast methods, explored were: Method : Forecast using actual load (without any CDM adjustment); Method : Incorporate CDM impacts as an explanatory variable in the regressions equation; and Method : Add back historical CDM impacts to the actual load and then forecast forward. 0 Various purchase models under each method were specified and assessed. Table compares the model results. Table : Purchase Models Comparison Model Statistics Method Actual Purchases Method CDM as a variable Method Gross Actual Purchases Adjusted R.0%.0%.0% MAPE.%.0%.% AIC..0. BIC..0. While the statistics are comparable across the three methods, PowerStream concluded that Method is the most robust and technically sound and it produces a reliable and accurate load forecast. PowerStream has adopted Method and has grossed up the historical load based on reported CDM results. In order to estimate the gross load utilizing Method, the following steps were performed: 0 Step : Derive historic total electricity volume reductions resulting from CDM initiatives using data from following sources:

7 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 0 a. Historic Ontario Power Authority ( OPA ) programs (source: OPA Report, Section..0 of Chapter of the Board s Filing Requirements for Transmission and Distribution Applications, dated June, 0); b. rd Tranche LDC programs (source: PowerStream and former Barrie Hydro Annual CDM Reports for 00-00); c. 0-0 CDM targets each licensed distributor must, as a condition of its license, meet its respective CDM targets as established by the Board (source: EB-00-0, EB-00-0). Table below presents the historic volume reductions by source. Table : Historic CDM Savings (kwh) 0 CDM Targets 0-0 Total CDM Savings Year OPA Programs rd Tranche 00 0,0, 0,0, 00,,,00, 0,,0 00,0,,, 0,0,0 00,0,,, 0 0,, 00,, 0 0,, 00,, 0 0,, 0,, 0,,000,, The gross forecast assumes some level of embedded natural conservation. The scope and rate of natural conservation cannot be measured but may be driven by such factors as relative price effects, industrial plant growth and productivity improvements, incremental technology improvements, changes in the economy that reduce energy intensity, old energy-consuming assets being replaced with new and more efficient technologies, and the availability and performance of energy management measures. There is insufficient evidence to determine how each of these factors impacts the load forecast. Step : The historic loads were grossed up by CDM savings. Table below provides a summary of historic actual load, CDM savings (as per Table ), and historic actual load grossed up by CDM (Gross Load) by year.

8 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Table : Historic Actual Load, CDM Savings and Gross Load 0 Actual MWH CDM Savings MWH Gross Load MWH Year 00,,0 0,,0 00,, 0,, 00,,0 0,,0 00,0,,,, 00,0,0,,, 00,0,,0,, 00,, 0,0,, 00,,,,0, 00,,,,, 0,,0,,,0 Step : Develop gross purchases forecast using grossed-up historic values for load. A resulted gross purchase forecast is adjusted by projected CDM reductions. PowerStream supports the Provincial Government s CDM initiatives and is currently delivering CDM programs funded by the OPA. By 0, the cumulative planned energy savings from new CDM targets is 0 GWH, with a peak savings of MW for PowerStream. PowerStream has forecasted that these targets will be met over the period of 0-0 as shown in Table. Table shows the adjustments to be made to the gross purchases forecast to account for CDM reductions resulting from the historic OPA programs and CDM targets. Table : Future CDM Savings OPA Programs CDM Targets 0-0 Total CDM Reductions CDM Targets Allocation % Year 0A,,,,000,, % 0F,,,,000,, % 0F,0,,,000,, % 0 0,,0,,000,,0 % 0-0,, 0,00,000,, 00%

9 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 0 Modelling Process and Weather Normalization PowerStream s energy purchase forecast is based on a linear regression model. Distribution sales/consumption is derived from purchases based on an adjustment for estimated distribution losses. Distribution consumption is then allocated to the rate classes based on historical billing trends. For those rate classes that use kw consumption as a billing determinant, sales for these customer classes are then converted to kw based on the historical volumetric relationship between kwh and kw. Below are the details of the modelling process: The energy purchases forecasting model relates monthly historical purchases to monthly weather conditions (measured in cooling-degree-days ( CDD ) and heating-degree-days ( HDD )), and Ontario GDP as a proxy for service area customer growth and economic activity. The following historical monthly data were used as inputs into the model: monthly system load (i.e. purchases) grossed up by CDM data for January 00 to December 0 weather data: HDD and CDD; Real Gross Domestic Product (GDP) growth index for Ontario. 0 The forecast is then derived by using the estimated model (i.e. estimated parameters) to predict monthly purchases for projected GDP and normal CDD and HDD. The total gross energy purchases forecast is adjusted to account for the impact of CDM. The net energy purchase forecast is allocated to rate zones (i.e. PowerStream South and PowerStream North) based on the -year average for the 00-0 period. The allocation between rate zones is done to determine distribution at current rates as historical line losses have been approved at separate levels prior to harmonization in 0. Table provides a comparison of the forecasted, actual and weather-normalized purchases GWHs over the past ten years and presents the 0-0 forecasts (not adjusted by CDM). In accordance with the Filing Requirements, PowerStream has also provided a 0 forecast assuming twenty-year normal weather conditions.

10 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Table : Total System Purchases, GWH Year Actual Gross Model Predicted Variance, Actual to Predicted, % Weather-Normal (WN) Actual Gross Variance, WN Actual to Predicted, % 00,,0 0.0%, -.% 00,, -.0%,0-0.% 00,,00 0.%,.% 00,, -0.%, -.% 00,, 0.%, 0.% 00,,0-0.%, -.% 00,, 0.%,.% 00,0, -0.%,.% 00,, 0.%, 0.% 0,, -0.%, -0.% 0 Bridge - Forecast,0 0 Test - Forecast - Normalized 0-year, 0 Test - Forecast - Normalized 0-year, Figure graphically displays actual vs. predicted load for the 00-0 period.

11 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Figure : Gross Actual vs. Predicted (kwh).0e Purchases KWh Model Predicted kwh.0e kwh.0e.0e.0e JAN 00 NOV SEP JUL MAY MAR JAN NOV SEP JUL MAY MAR The load forecast model was populated with the actual energy purchase data from January 00 through December 0. Table, below, provides historical actual (grossed up by CDM) and historical normalized annual energy purchased data for PowerStream. The heading normalized actual shows the purchases adjusted to reflect normal weather conditions. PowerStream considers normal weather conditions to be the average of the weather characteristics for the ten-year time period, 00 to 0. PowerStream normalizes energy purchases using a use per degree methodology. This methodology uses the weather-related coefficients in the regression equation to estimate normalized volumes. The difference between actual and normal degree-days is determined. The weather related coefficients are applied to that difference to derive weather-sensitive volume. Actual volumes are adjusted by the weather sensitive volume.

12 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 The formula is: Normalized Volume = Actual Volume (Actual HDD or/and CDD Normal HDD or/and CDD) x Corresponding Regression Coefficient Table : Historic Annual Energy Purchases Grossed up by CDM (GWH) Actual Gross Normalized Actual Gross Change % Change Compounded AVG Growth Year 00,, 00,,0.%.% 00,,.%.% 00,,.%.% 00,,.%.% 00,, 0.%.% 00,,.0%.% 00,0, () -.%.0% 00,,.%.% 0,, 0.%.% Average % Average % Table provides the same information for actual historic loads not adjusted for CDM. Table : Historic Annual Energy Purchases (GWH) 0 Normalized Actuals Change % Change Compounded AVG Growth Year Actuals 00,, 00,, 0.%.% 00,,.%.% 00,0,0.%.% 00,0,.%.0% 00,0, 0.%.% 00,, 0.%.% 00,, (00) -.% 0.% 00,,.%.% 0,, 0.%.% Average % Average % Until recently, PowerStream had relatively strong sales growth experiencing both strong population and economic growth. Between 00 and 00, normalized actual purchases

13 EB-0-0 EB 0 Exhibit C Schedule Page 0 of Filed May, 0 averaged.% annual growth. However, since 00 there has been a trend of dampened sales with the most significant decline in 00. This is likely attributable to a variety of factors which include natural conservation, a focus on CDM initiatives and most pointedly the global economic slowdown. Between 00 and 0, normalized actual purchases averaged 0.% annual growth. Figure graphically depicts variances between actual and weather-normalized energy purchases for 00 to 0. Figure : Consumption Variance between Actuals and Weather-Normalized Energy Purchases, 00 0 (GWH),000,000,000,000, Actuals Normalized Actuals Model Specifications The purpose of a multiple regression equation is to predict a single dependent variable from multiple independent variables. Many variables (e.g., electricity prices, changes in gross domestic product, per capita incomes, employment levels, population and weather patterns), and the interactions among these variables, may affect overall electricity purchases. Given the complexity of load forecasting, the task is to find a specific set of explanatory (independent) variables that reflect PowerStream s circumstances and that can be used to generate the most accurate load forecast.

14 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Different economic drivers were tested using different model specifications, as well as a stepwise regression technique. Stepwise regression is a procedure that adds and deletes one independent variable at a time. The decision to add/delete a variable is made on the basis of whether that variable improves the accuracy of the model. The variables listed in Table were used as initial inputs for the purpose of regression analysis. Table : Initial Set of Explanatory Variables 0 Dependent Variable Y Monthly Energy Purchases (KWh) Independent (Explanatory) Variables X Heating Degree-days X Cooling Degree-days X Real Gross Domestic Product for Ontario X Real Gross Domestic Product for Toronto CMA X Real Personal Income per Capital for Toronto CMA X Population (York Region and Barrie) X Simple Trend X Monthly Peak Hours X Manufacturing GDP for Toronto CMA X 0 Population for Toronto X Total Empoyment for Toronto CMA X Manufacturing Empoyment for Toronto CMA Several monthly models of energy purchases were specified, estimated and tested to derive the energy purchases forecast. The statistical software generated the coefficients that were used in the variables suitability assessment. The detailed results of the model testing are presented in Table 0. Model, using Ontario GDP as a proxy for service area customer growth and economic activity, was selected as the most accurate.

15 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Table 0: Evaluation of Alternative Forecast Drivers Model Model Model Model Model Model Model Constant 0,,0,,,,,,0,,0 (,00,),, Independent Variables HDD0, 0,,,,0,, CDD,0,0,0,,0,0,0,,0,,0,,0, Ontario GDP Index,, GDP for Toronto,, Population (York Region, Barrie) 0,0 Toronto population 0, Manufacturing GDP for Toronto Non-Manufacturing GDP for Toronto Total Empoyment for Toronto,0, Manufacturing Employment for Toronto Non-Manufacturing Employment for Toronto Real Income for Toronto,0, Peak Hours,, Simple Trend, Feb (,,) (,,) (,0,0) (,0,0) (,,) (,,) Apr (,,0) (,0,0) (,,) (,,0) (,,0) (,,) Aug-0 (0,,0) (,,) (0,,) (,,0) (,,) (,,) Oct-0,,,0,,, May-0 (,0,) Jul-0,,,, Model Statistics Adjusted R-Squared.0% 0.0%.0%.0%.0%.0%.0% SEE,,,00,0,,,,,0,,,,, F-Test DW

16 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Figure presents graphically Ontario GDP with its values indexed to.0 in 00. Actual gross purchases and the moving average of those purchases have also been indexed so the economic driver and purchases can be illustrated on the same graph. This graph demonstrates that the selected economic driver tracks actual purchases relatively well. Figure : Indexed GDP Variable against Gross Actual Load and Moving Average purchases The selected model included the following variables: 0 Ontario GDP: Ontario Real GDP Index; HDD0: Monthly HDD with a base of 0 degrees; CDD: Monthly CDD with a base of degrees; Feb: Binary variable for the month of February;

17 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Apr: Binary variable for the month of April. 0 The most significant independent variable for the model is GDP, actual values for which (00 00) were obtained from Statistics Canada. The forecasted values of Ontario GDP are based on a survey of long-term forecasts prepared by six major chartered banks of Canada (as of January, 0). Heating Degree Days ( HDD ) are summations of negative differences between the mean daily temperature and the 0 C base; Cooling Degree Days ( CDD ) are summations of positive differences from the C base. The number of HDDs influences electricity use for space heating, while the number of CDDs influences electricity use for space cooling. The HDD variable also picks up some of the increased lighting load that results from shorter winter days. PowerStream uses the degree days count for the Toronto Lester B. Pearson International Airport Data Point as published by Environment Canada as this provides the most updated data on a monthly basis. The appropriate basis for defining the HDD and CDD temperature breaks was determined by evaluating average monthly purchases against average monthly temperature. Figure shows this relationship. Figure : Purchases vs. Average Temperature.0E.0E HDD Base 0 CDD Base Purchases.0E.0E.0E

18 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 0 As Figure shows, cooling load begins where average temperature is above degrees and heating load can be seen with average temperature below 0 degrees. For purposes of PowerStream s load forecast, weather is not forecasted. Weather inputs are based on monthly normal HDD and CDD data. The decision was made to move from traditional 0-year to 0-year (00 0) weather time series for defining normal weather. The 0-year average has been used in other electricity distribution rate applications in recent years as an acceptable approach for weather normalization. Looking at the details, the 0-year time series weather data is also more representative of the general weather trend of milder winters and warmer summers. Winters in PowerStream s service area are generally mild with annual HDDs averaging, from 00 through 0. The extremely cold winter of 00 was followed by relatively mild winters through 0 with very mild winter in 00. From 00 through 0, HDDs have ranged from, in 00 to,00 in 00. The general trend has been downward, i.e. winters generally are getting warmer. Summers in PowerStream s service area are generally hot and humid with average annual CDDs of for the period 00 through 0. The cool summer in 00 was followed by extremely hot summer in 00 and 00 and again, by unseasonably cold summer of 00. From 00 to 0, cooling degree-days have ranged from in 00 to in 00 with the general trend upward, i.e. summers generally are getting warmer (see Figure ).

19 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Figure : Historic HDD & CDD, 0 0,00 00, HDD,00, CDD Calendar factors, such as the number of days in a month or seasonality, tend to influence energy use. In order to incorporate these effects two binary variables were introduced to the model. The binary variable for February is used to capture an effect of a shorter month and binary variable for April is used to capture the effect of seasonality. The load forecasting model, using GDP, HDD, CDD, and two shoulder month variables, has tracked historic experience quite well in terms of both levels and peaks. Moreover, it captures the historical pattern of energy purchases with respect to economic and weather conditions. Figure shows the selected equation s ability to capture historic monthly energy purchases. It shows the historic time series (Purchases kwh) and presents the current forecast (Model Predicted kwh).

20 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Figure : Monthly Actual vs. Predicted Energy Purchases Forecast (kwh) The selected model specifications are summarized in Table. Table : Summary of Monthly Load Forecast Regression Model Dependent Variable: Monthly Energy Purchases grossed up by CDM Form: Multiple Regression Sample: 0/00 - /0 Included observations: 0 Degree of Freedom for Error: 0 Variable Coefficient t-statistics Sig. (Constant),, % Real GDP,,. 0.00% CDD,0, % HDD0, % Feb (,,) (.) 0.00% Apr (,,0) (.) 0.00% Adjusted R-squared.% MAD,,0 Standard Error of regression,0,0 MAPE.% F-test. Durbin-Watson statistics.

21 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 As shown, the model variables are all statistically significant at the % level of confidence and the model fit is strong with an adjusted R-squared of.%, and an in-sample Mean Absolute Percentage Error ( MAPE ) of.%. From the statistical perspective, the Ontario GDP-based model explains purchases exceptionally well. Regression coefficients generated by the model were used to predict future energy purchases. Coefficients describe the average amount of change to be expected in purchases given a unit change in the value of the particular independent variable while holding other variables constant. Combining the results of the coefficient table into a regression equation, the monthly purchases are expressed as: 0 Monthly kwh =,,0 + (,,*Real GDP) + (,0,*CDD) + (,0*HDD) + ((,,)*Feb) + ((,,0)*Apr)

22 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 The key results of the monthly energy purchases forecast are summarized in Table. Data from January 00 to December 0 was used to help select the model and to estimate its parameters. Forecasts are made for time periods beyond the end of the available data. To estimate the average energy purchases for any particular combination of predictor variable values, the values of the predictor variables are simply substituted in the estimated regression equation itself. Table : Monthly Gross Energy Purchases Forecast (kwh) Month Model Coefficient kwh Purchases GDP CDD HDD Feb Apr,,0,,,0,,0 (,,) (,,0) Jan- 0,, Feb-,, Mar-,0, Apr-,, May-,, Jun-,0, Jul- 0,0, Aug- 0,, Sep-,0, Oct-,00, Nov-,, Dec-,, Total 0,0,, Jan-,, Feb-,0, Mar-,, Apr-,00,... 0 May- 0,, Jun-,, Jul-,, Aug-,, Sep-,0, Oct- 0,0, Nov-,0, Dec-,, Total 0,,,0

23 EB-0-0 EB 0 Exhibit C Schedule Page 0 of Filed May, 0 Table presents gross actual and normalized gross energy purchases for 00 through 0 and forecasts for 0-0. In 0 the total weather-normalized energy was, GWH. In 0 the total weather-normalized gross energy for PowerStream amounted to,0 GWH, an increase of.%. For the 0 Test Year, the forecast predicts a.% decrease from 0. Table : Annual Gross Energy Purchases (GWH) 00 to 0 Year Actuals Gross Normalized Actuals Gross Growth Rate (GWH) Growth Rate (%) 00,, 00,,0.% 00,,.% 00,,.% 00,,.% 00,, 0.% 00,,.0% 00,0, - -.% 00,,.% 0,, 0.% 0 Forecast,0.% 0 0 Forecast,.% To evaluate the model performance the last two years of actual purchases were held out of the estimation period (Jan 00 Dec 0). Predicted purchases were then compared with actual purchases for this period. The two-year out-of-sample MAPE is.% and the average forecast error is 0.%. The following analysis compares the out-of-sample forecast outcomes to a reasonable expectation for outcomes of forecasts generally. Forecasts will normally vary from actual ( error ), either higher or lower, and it is reasonable to expect that the load forecasting methodology is unbiased, if the average error of many forecasts (the Mean Percentage Error ) is close to zero. Table provides a summary of the outcomes of forecasted gross energy purchases compared to actual energy purchases for the period January 00 to December 0. Column ( WN Actual Gross ) is the weather-normalized actual electricity grossed up by CDM for PowerStream in 0. Column ( Model Predicted ) is the forecasted annual energy

24 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 purchased. Column and Column (Error and Error % respectively) is the percentage difference between the actual outcome and the forecast. This percentage error is expressed as a fraction of the weather-normalized actual load. Table : Energy Purchases Actual vs. Forecast (GWH) 0 Year WN Actuals Model Error Error Gross Predicted % Jan-0 0.% Feb-0 00.% Mar-0 0.% Apr-0 0.% May-0 () -0.% Jun-0 0.% Jul-0 0.% Aug-0 00.% Sep-0 0 () -.% Oct-0 0.% Nov-0 0 () -.0% Dec-0 0.% Jan-.% Feb- 0 () -0.% Mar- 0.% Apr- () -.0% May- 0 () -.0% Jun- 0.% Jul- () -0.% Aug- 0 0.% Sep- 0 () -.% Oct- 0.% Nov- 0 0 () -.% Dec- 0 0.%,, 0.% PowerStream has performed due diligence testing of its load forecast methodology using both internal and external resources. The evaluation and validation process included analytical assessment of the forecast results, one-step-ahead forecasts to actual, statistical measures, residual analysis and external review. PowerStream has determined that its current methodology produces reasonable forecasts for the test period.

25 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 CDM Adjustment The load forecast as described above does not take into account the impacts on energy purchases arising from CDM programs undertaken by PowerStream customers. The gross load forecast is a forecast of the expected level of electricity purchases that would occur over the specified period in the absence of any CDM initiatives. The forecasted gross energy purchases are further adjusted to reflect CDM reductions. The CDM reduction breakdown by year is shown in Table. Table : Energy Conservation Savings: Historic and Proposed 0 Year OPA Programs rd Tranche CDM Targets 0-0 Total 00 0,0, 0,0, 00,,,00, 0,,0 00,0,,, 0,0,0 00,0,,, 0 0,, 00,, 0 0,, 00,, 0 0,, 0,, 0,,000,, 0,, 0,,000,, 0,0, 0,,000,, 0 0,,0 0,,000,,0 The results show that for 0, GWH will be saved. Accordingly, the energy purchases would decline by about.% relative to the gross forecast. In absolute terms, this is a reduction in 0 from, GWH to, GWH as shown below in Table. Weather-normal forecasted net values are derived by subtracting CDM reductions from weather-normal forecasted gross values.

26 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Table : 0-0 CDM Reductions to Forecast Year Actual Gross CDM Reduction Actuals WN Actual Gross WN Actual Net Growth, % 00,,, 0,,,0,,,0,,,0 00,,, 0,,,0,,,00,,,00.% 00,,, 0,,,0,,,0,,,0.% 00,,,00,0,,0,,0,,0,0,,0,0.% 00,,,,,,0,0,0,,0,00,,,0.% 00,,0,,0,0,0,,00,,,0,,, 0.% 00,,, 0,,,,,0,,,000,,, 0.% 00,0,,,,,,0,0,,0,0,,, -.% 00,,,0,,,,,0,,0,0,,,.% 0,,,,,,,0,0,,0,0,,, 0.% 0 Bridge,,,0,,,,,0 0.% 0 Test,,,,,,,, 0.% 0 Test - Normalized 0-year,,,0,,0,,00, -0.%

27 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Figure shows the gross and net forecasts graphically. Figure : Gross vs. Net Forecast,000,000 Net_of_CDM_Load WN_Gross_Load Forecast_WN_Gross_Load Forecast_Net_Load,00,000,00,000 Value,00,000,00,000,000,000,00, Allocation of Purchases by Rate Zone Since the distribution rates for PowerStream rate zones are not yet harmonized, revenues at current rates for each rate zone are required to be calculated separately, using the 0 PowerStream North and PowerStream South approved distribution rates and forecasted loads. In order to derive forecasted loads by rate zone PowerStream used a three-year average of actual loads for each territory for 00-0 periods. Based on the analysis, on average.% of the total load is allocated for PowerStream South and the remaining.% for PowerStream North. Table below provides the details of the allocation percentages.

28 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Table : Energy Purchases Percentages by Rate Zones (00-0) Derivation of Demand (KW) Month PS South PS North PS Consolidated PS South, % PS North, % Jan-0 0,,,0,0,0, 0.%.% Feb-0,,,,00,0, 0.%.% Mar-0,,,,0,, 0.%.% Apr-0,,,,0,,.%.% May-0,,,,00,,.%.% Jun-0 0,0,0,,00,,.0%.0% Jul-0,,,,,,.%.% Aug-0,,,,,,.0%.% Sep-0,,,,,,.%.% Oct-0,,0,0,0,,0.00%.00% Nov-0,,,, 0,, 0.%.% Dec-0,,,,,, 0.%.% Jan-0,,,0,0 0,0, 0.%.% Feb-0,,,,0,,0 0.%.% Mar-0,,,, 00,0,.0%.% Apr-0,,,,0,00,.%.% May-0,0,,0,,,0.%.% Jun-0,,,0,,,.0%.0% Jul-0,,,0,,,.%.% Aug-0,00,00,0,,0,.%.% Sep-0,,,,0,0,.%.0% Oct-0,0,,,0,0,.%.% Nov-0,0,,,0,0,.%.% Dec-0 0,,00,,,0, 0.%.0% Jan-,,,00,,,0 0.%.% Feb-,,0,,0,,0 0.%.0% Mar-,,0,,,,.%.% Apr-,,,0,,,.%.% May-,0,0 0,0,,0,.0%.% Jun-,,,,,,.%.% Jul-,,,,,,00.%.% Aug-,,0,, 0,,.%.% Sep-,,,,,,.%.% Oct-,,,,0,,.%.% Nov-,0,,0,0,,.%.% Dec-,0,,0,,0,0 0.%.% AVERAGE.%.% The 0 energy purchases forecasts for each rate zone are composites of monthly kwh forecasted volumes for all rate classes. Estimated total losses are subtracted from these forecasts to determine the distribution sales forecast. This distribution sales forecast is apportioned to various rate classes based on the historical relationships between energy and demand by rate class obtained from billing data for each service territory.

29 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 There are different billing determinants for various classes: Residential and GS<0kW customers are billed based on kwh units, whereas charges for other Commercial Accounts (GS>0, Large User, TOU, Street Lighting and Sentinel) are based on kw units. The historical relationship between kwh and kw for each rate class is used to translate forecasted kwh to kw for these accounts. Tables through show the historic (three-year average) billed energy (kwh) allocation, by rate class, and a ratio of historic kw to historic kwh, by rate class, as an average for the period 00 through 0 for each rate zone.

30 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Table : PowerStream South Historic kwh Allocation by Rate Class (00 0) Year Residential GS <0 kw USL GS>0 kw TOU Large User Street- Lighting Sentinel Total 00.%.% 0.%.% 0.0% 0.% 0.% 0.0% 00.00% 00.%.% 0.%.% 0.% 0.% 0.% 0.0% 00.00% 00.%.% 0.%.% 0.% 0.% 0.% 0.0% 00.00% 00 0.%.% 0.%.% 0.% 0.% 0.% 0.0% 00.00% 00.%.% 0.%.% 0.00% 0.% 0.0% 0.0% 00.00% 00.%.% 0.%.% 0.00% 0.% 0.% 0.0% 00.00% 0.00%.% 0.%.% 0.00% 0.0% 0.0% 0.0% 00.00% Average 00-0.%.% 0.%.% 0.00% 0.% 0.% 0.0% 00.00% 0.%.% 0.%.% 0.00% 0.% 0.% 0.0% 00.00% 0.%.% 0.%.% 0.00% 0.% 0.% 0.0% 00.00% Table : PowerStream South Historic Relationship between Billed kwh and kw Demand by Rate Class (00 0) Year GS>0 kw TOU Large Users Street- Lighting Sentinel 00 0.% 0.% 0.% 0.% 0.% 00 0.% 0.% 0.% 0.% 0.% 00 0.% 0.% 0.% 0.0% 0.% 00 0.% 0.% 0.% 0.% 0.% 00 0.% 0.00% 0.0% 0.% 0.% 00 0.% 0.00% 0.% 0.% 0.% 0 0.% 0.00% 0.0% 0.% 0.% Average % 0.00% 0.0% 0.% 0.% 0 0.% 0.00% 0.0% 0.% 0.% 0 0.% 0.00% 0.0% 0.% 0.%

31 EB-0-0 EB 0 Exhibit C Schedule Page of Filed May, 0 Table 0: PowerStream North Historic kwh Allocation by Rate Class (00 0) Year Residential GS <0 kw USL GS>0 kw TOU Large User Street-Lighting Sentinel Total 00.%.% 0.% 0.% 0.% 00.00% 00.%.% 0.%.% 0.% 00.00% 00.%.0% 0.0% 0.% 0.0% 00.00% 00.%.% 0.0%.% 0.% 00.00% 00.%.% 0.%.% 0.% 00.00% 0.%.% 0.0%.% 0.% 00.00% Average 00-0.%.% 0.%.0% 0.0% 00.00% 0.%.% 0.%.0% 0.0% 00.00% 0.%.% 0.%.0% 0.0% 00.00% Table : PowerStream North Historic Relationship between Billed kwh and kw Demand by Rate Class (00 0) Year GS>0 kw TOU Large Users Street-Lighting Sentinel 00 0.% 0.% 00 0.% 0.% 00 0.% 0.% 00 0.% 0.% 00 0.% 0.% 0 0.% 0.0% Average % 0.0% 0 0.% 0.0% 0 0.% 0.0%

32 EB 0 Exhibit C Schedule Page of Filed April, 0 The overall forecast process is illustrated in Figure below. Figure : Load Forecast Process Flowchart Government Economic Variables Past weather Major Inputs Billing statistics Historic Energy Purchases Customer additions CDM Programs Development of Baseload Forecast Prepare Forecast Develop Model Evaluate Model Generate Forecasts CDM Adjustment Allocation by Rate Class Residential GS<0 GS>0 Large User Others Development of Probable Forecast Sensitivity Analysis GDP CDM Management Review Review & Approval Documentation

33 EB-0-0 Exhibit C Schedule Page of Filed May, 0 0 CUSTOMER FORECAST In order to determine the fixed distribution charges, PowerStream requires customer counts by class. PowerStream s derives the forecast of new customers based on historic averages. The three year average is used for the Residential and General Service < 0kW classes. A five year historic average is used for the General Service > 0kW class. The economic slowdown has impacted the General Service > 0kW class such that the three year average is not considered representative of future growth patterns. Overall, the total number of customers for 0 is expected to be.% higher than 0. PowerStream has been experiencing reduced growth trends in its service territory over recent years. The peak growth period was in -00 and averaged.% in customer growth rates, followed by a more moderate growth rate of.% over the periods. The 00-0 growth rates averaged only.%. Table below summarizes historic and forecast growth rates. Table : Historic and Projected Customer Additions Growth Rates ( 0) Period F 0F Growth Rate.%.%.%.%.%.%.% Figure below presents the historic and projected total customer additions in the graphical format. Figure : Historic and Projected Total Customer Additions (00 0) 0,000,000,000,000,000,000,000,000,000, F 0F

34 EB-0-0 Exhibit C Schedule Page of Filed May, 0 Residential customer additions in PowerStream s service territory have been relatively flat over the recent three year period. This is attributed to the economic slowdown, the introduction of Harmonized Sales Tax on new homes and the limited availability of land for new subdivisions. Figure below shows historical and forecasted net residential additions in the graphical format. Figure : Historic and Projected Residential Customer Additions (00-0),000,000,000,000,000,000,000,000, F 0F 0 Commercial and Industrial customers and their respective loads are typically known only when the connection is requested. It is difficult to forecast or anticipate the general service customer rate class required for revenue billing purposes in a proposed commercial area. PowerStream considers the best method to forecast future General Service < 0kW growth to be a three year historical average. The General Service >0 kw customer additions forecast is based on a five year average, since there has been significant volatility in the historic data for the last three years due mainly to the economic slowdown. Figure below shows historic and projected customer additions for General Service <0kW, General Service >0kW and Large User rate classes. PowerStream currently bills one large user based on its customer specific approved rates however it does have another customer that is greater than MW that it proposes to move to the large user class in 0.

35 EB-0-0 Exhibit C Schedule Page of Filed May, 0 Figure : Historic and Projected General Service Customer Additions (00-0) GS<0 GS> F 0F Table below summarizes the 0 bridge and 0 test year customer additions by the three major customer categories. Table : Net Customer Additions Year Total Res GS<0 GS>0 00,, 00,, 00,, () 00,0, () 0,,0 0F,, 0F,, The detailed forecast of customers by rate class is presented in Table.

36 EB-0-0 Exhibit C Schedule Page of Filed May, 0 Table : Customers by Rate Class 00 Board Approved 00 Board Approved 00 Actuals 00 Actual 0 Actual 0 -Bridge 0 - Test Residential,0,, 0,, 0,, GS Less Than 0 kw,,00, 0,0 0, 0,, GS 0 to, kw,0,,,,, GS 0 to, kw Legacy Large Use Unmetered Scattered Load,,,,,0, Sentinel Lighting Street Lighting Connections,0,0,, 0,,,0 Street Lighting Customers Total Customers,0,0 0,,,,0 0, Total Connections,0,0,, 0,,,0 TOTAL,,, 0,,0,,

37 EB-0-0 Exhibit C Schedule Page of Filed May, 0 DISTRIBUTION REVENUE The year over year comparison of PowerStream s distribution revenue is summarized in Table, below. At current approved rates, PowerStream s revenue requirement including smart meter increment revenue rate riders ( SMIRR ) is $,0, which is a.% increase over 0. The 0 and 0 revenue amounts were calculated by applying current rates (December, 0 and May, 0 for 0 and May, 0 for 0) to the forecast sales and customer numbers. Table : Distribution Revenue at Current Rates 00 OEB Approved PS North 00 OEB Approved PS South 00 Actuals 00 Actual PowerStream Consolidated 0 Actual 0 Bridge Year 0 Test Year Fixed and Variable Charge,,,,0,0,0 0,,,,,,,,00 Transformer Credit (,) (,,) (,,) (,,) (,,) (,,) (,,) Distribution Revenue w/o SMIRR,,,,,0,,, 0,,,,,, % growth Year over Year.%.%.%.% SMIRR Revenue,0,,0,0,, Total Distribution Revenue,,,,,0,,,,, 0,,,0, % growth Year over Year.%.%.%.% 0 0 Year over year variances in distribution revenue based on weather normalized sales for the period 00 to 0 are mainly attributable to growth of PowerStream s customer base. Weather normalized sales have tapered in recent years which can be attributable to various external and uncontrollable industry or economic factors. Some of these factors include slow economic growth due to 00 recession and subsequent slow recovery, Conservation and Demand Management ( CDM ) initiatives, and energy price increases. In addition, as a result of low inflation rates, Incentive Regulation Mechanism ( IRM ) adjustments in 0 were fairly negligible which contributed to lower growth in base distribution rates (0.% for the South and 0.% for the North). The IRM adjustment remained low in 0, 0.% for both South and North service areas which continues to mitigate the distribution revenue increases from base rates. The distribution revenue at current rates in 0 is forecasted to trend lower as PowerStream continues to pursue CDM initiatives in order to meet its licence obligations. In 00, PowerStream filed an application (EB-00-00) for recovery of costs associated with the installation of smart meters for the South rate zone. The Board approved an annual incremental revenue requirement of $,,000. In 0, PowerStream had completed its Smart Meter program and filed an application (EB-

38 EB-0-0 Exhibit C Schedule Page of Filed May, ) for final recovery of costs associated with the installation of smart meters for both the Barrie (former Barrie territory) and South rate zones. The Board approved an annual incremental revenue requirement of $,,000 for Barrie and $,,000 for the South. This incremental distribution revenue generated by these mini-rebasing applications has contributed to year over year variances of approximately.0% annually from 00 to 0. During the IRM period (00-0), when capital additions were made, distribution revenue from IRM and customer and load growth alone do not provide the appropriate level of funding for the capital outlay. PowerStream recovers revenue based on a fixed and variable rate methodology. The fixed revenue component is derived based on a customer forecast and the variable revenue component is derived based on a sales forecast. PowerStream has applied current approved rates to the test year customer and sales forecast in order to derive the test year distribution revenue. A summary of the consumption, demand and customer count is outlined in Table. Table : Energy Sales, Demand and Customers 00 OEB Approved PS South 00 Normalized Actuals 00 Normalized Actual PowerStream Consolidated 0 Normalized Actual 0 Bridge Year 0 Test Year Consumption, KWH,,0,0,0,0,0,0,,,,,0,,0,,,,0 Demand, KW 0,00,,0,0,0,,,0,,,, Customer Count,, 0,,, 0, Note: PS = PowerStream The details regarding forecast distribution revenue are supported by following Tables: Table : Distribution Revenue by Rate Class Table : Demand and Consumption Table : Unit Revenues Table : Customer Count by Rate Class Table : Residential and General Service Classes Average Normalized Consumption per Customer

39 EB-0-0 Exhibit C Schedule Page of Filed May, 0 Table : Distribution Revenue by Rate Class Distribution Revenue, $ Variance Analysis Bridge Year Actual Normalized Actual Normalized Actual Normalized Test Year Normalized Actual Norm vs 00 Actual Norm. 0 Actual Norm vs 00 Actual Norm. 0 Actual Norm vs 0 Actual Norm. 0 Actual Norm vs 0 Actual Norm. $ $ $ $ $ $ % $ % $ % $ %,0,0,, 0,,,0,,0,,,0.0%,0,0.%,,.%,0,.%,,00,0,0,0,,,,,0,0.% 0, 0.%,.0%,.%,,0,,,,,,,0,,.%,0 0.%,0.% (,) -0.%, (,0) -00.0% 0 0 0, 0, 0,0,, (,) -.%,.%,0.%,.%,,,,,,0 0.% (,) -0.%,.%,0.0%,0,,,,,0 0.% (0) -0.% 0.% 0.%,0,,,,0,,,,, 0,.%,.% 0,.%,0.%,, 0,0,,,,,,,00,00,.0%,,0.%,,0.%,,0.%

40 EB-0-0 Exhibit C Schedule Page of Filed May, 0 Table : Demand and Consumption Demand Load (kw) Variance Analysis Actual Normalized Actual Normalized Actual Normalized Bridge Year Normalized Test Year 00 Actual Norm vs 00 Actual Norm. 0 Actual Norm vs 00 Actual Norm. 0 Actual Norm vs 0 Actual Norm. 0 Actual Norm vs 0 Actual Norm kw kw kw kw kw kw % kw % kw % kw % ,,,,0,0,,,0,0,,.%, 0.%,.% (,) -0.% ,0,,,,,.0% 0.% 0.% 0,0.0% ,,,,,0.0% 0.% 0.% 0.%,,,00,,,.0% 0.%,.% 0.%,0,0,0,,,0,,,,,.%, 0.%,0.%,0 0.% Consumption Consumption (kwh) Variance Analysis Bridge Year Actual Normalized Actual Normalized Actual Normalized Test Year Normalized Actual Norm vs 00 Actual Norm. 0 Actual Norm vs 00 Actual Norm. 0 Actual Norm vs 0 Actual Norm. 0 Actual Norm vs 0 Actual Norm. kwh kwh kwh kwh kwh kwh % kwh % kwh % kwh %,,0,0,,0,,,,,,,,,0,,,.0%,, 0.%,,.%,, 0.%,0,,0,0,0,,0,,00,0,,,0,,,0,.%,0,0 0.%,,.%,0,0 0.%,,0,,00,00,,,,,,0,,,,,,.%,, 0.%,,.% (,,0) -0.% ,,,0,,,,,,0,0,00.0%, 0.%, 0.%,,.0%,0,,,,0,,,,, 0, 0.% 0, 0.%,0.%,0 0.%,,,,,,.0%, 0.%,00 0.%, 0.%,,,0,,, 0,0, 0,,,.% 0, 0.%,00.%, 0.%,0,0,0,0,,,,,0,,0,,,,0,0,0.%,, 0.%,0,.%,0, 0.%

41 EB-0-0 Exhibit C Schedule Page of Filed May, 0 Table : Unit Revenues Revenue per Customer, $ Variance Analysis Bridge Year Actual Normalized Actual Normalized Actual Normalized Test Year Normalized Actual Norm vs 00 Actual Norm. 0 Actual Norm vs 00 Actual Norm. 0 Actual Norm vs 0 Actual Norm. 0 Actual Norm vs 0 Actual Norm. $/Customer $/Customer $/Customer $/Customer $/Customer $ % $ % $ % $ % $. $.0 $. $. $. $ (.) -0.% $ (.) -0.% $ (0.) 0.0% $ (.) -0.% $. $. $. $0. $. $ (.) -0.% $ (.) -0.% $. 0.% $ (.) -0.% $,. $0,. $0,0. $0,. $0,0. $..% $ (.) -0.% $. 0.% $ (0.) -0.% $,0. $,. $0,.00 $0,0.0 $,. $,. $ (,.0) -.% $,.0.% $,0..% $ 0,.0.% $0. $. $. $0. $0. $ (.) -.% $ 0. 0.% $..% $ 0. 0.% $. $0. $. $0. $.0 $.0.% $..% $.0.% $ % $.0 $. $. $. $. $..% $. 0.% $.0.% $ (0.) -0.% $. $. $. $. $. $ (.) -0.% $ (.) -0.% $ (0.0) -0.% $ (.) -.0%

42 EB-0-0 Exhibit C Schedule Page of Filed May, 0 Table : Customer Count by Rate Class CUSTOMER COUNT (-months average, Jan. st - Dec. ) Number of Customers (Connections) Variance Analysis Bridge Year Actual Normalized Actual Normalized Actual Normalized Test Year Normalized Actual Norm vs 00 Actual Norm. 0 Actual Norm vs 00 Actual Norm. 0 Actual Norm vs 0 Actual Norm. 0 Actual Norm vs 0 Actual Norm. # # # # # $ % $ % $ % $ % 0,0,, 0, 0,0,0.%,.%,.%,.%,, 0, 0,,.% 0.0% 0.% 0.%,,,,, () -.%.%.% 0.% () -00.0% % 0 0.0% 0 0.0% 00.0%,,,,,.% () -.0% 0.% 0.% 0 0 () -.% () -.% () -.% 0 0.0%,, 0,,,0,.%,0.%,.0%,.%, 0,,, 0,,.%,.%,.%,.% Table : Residential and General Service Classes Average Normalized Consumption per Customer Average consumption (kwh/customer) Variance Analysis Bridge Year Actual Normalized Actual Normalized Actual Normalized Test Year Normalized Actual Norm vs 00 Actual Norm. 0 Actual Norm vs 00 Actual Norm. 0 Actual Norm vs 0 Actual Norm. 0 Actual Norm vs 0 Actual Norm. kwh/customer kwh/customer kwh/customer kwh/customer kwh/customer kwh/customer % kwh/customer % kwh/customer % kwh/customer %,0,,,0, () -.% () -.% (0) -.% () -.%,,,,, (0) -.% () -0.% (0) -0.% () -.%,,,,, () -.% () -.% () -.% () -.%

43 EB-0-0 Exhibit C Schedule Page of Filed May, 0 0 OTHER REVENUE - OVERVIEW Other Revenue is defined as sources of utility revenue other than Distribution Revenue. PowerStream divides Other Revenue, or Revenue Offsets into the following categories: Specific Service Charges, Late Payment Charges, Other Distribution Revenue and Other Income and Deductions. These are the same categories that were used in PowerStream s 00 Cost of Service Application. Table below shows PowerStream s Revenue Offsets from the last Board Approved Cost of Service Rate Application through the 0 Test Year. Table : PowerStream Revenue Offsets ($000 s) PowerStream Barrie PowerStream South PowerStream Combined 00 Board Approved 00 Board Approved 00 Actual 00 Actual 0 Actual 0 Actual 0 Estimate 0 Forecast CGAAP MIFRS Specific Service Charges Late Payment Charges Other Distribution Revenue Other Income and Deductions Total Revenue Offsets,,,,0,0,0,,,,,,,00,00,,,,,,00,0 0,,,0,0,,,, 0,0,,,,,0 A detailed account level breakdown of Other Revenues is provided in Exhibit C-- (OEB Appendix -C).

44 EB-0-0 Exhibit C Schedule Page of Filed May, PowerStream presents the information in this table, as per Chapter of the OEB minimum filing requirements. The historical trend analysis, however is distorted by the accounting changes due to Internal Financial Reporting Standards ( IFRS ) implementation. The 00 Barrie Approved amounts, as well as PowerStream Approved and Actual 00 and 00 amounts are not comparable to the 0 amounts as the 00-0 information is shown in CGAAP and 0-0 is shown in MIFRS. In addition, the 00 PowerStream Board-Approved levels do not include Barrie data, which was merged with PowerStream in 00. The 00 PowerStream actual results include Barrie. The 0 Test Year revenue offsets are $,0,000 or $,000 lower than the revenue offsets in 0 Actual (under MIFRS), the decrease is mainly due to the 0 change in the accounting treatment of gains on work orders, and the decrease in the Other Income, as explained in the detailed variance analysis in Exhibit C,, Schedule. Revenue Offsets are deducted from the Service Revenue Requirement to derive the Base Revenue Requirement. The latter is used to determine distribution rates. The Revenue Offsets are not equal to the Other Revenue as shown in PowerStream s Financial Statements since the Other Revenue line in the Financial Statements includes non-distribution items. In particular, the revenues and expenses of PowerStream s Solar Business are separated from the distribution business and recorded in accounts and 0, as per the OEB s Guidelines: Regulatory and Accounting Treatments for Distributor Owned Generation Facilities (G ) issued on September, 00. The revenues and expenses related to Conservation and Demand Management ( CDM ) activities are also recorded in subaccounts of and 0 and are not included in the amounts to be recovered in the distribution rates and in the revenue offsets. A reconciliation is provided in Appendix, Schedule. CDM activities are discussed in Exhibit C,, Schedule and the Solar Business is discussed in Exhibit D,, Schedule.