A panel analysis of temperature and sales in Utah

In this paper, I document that seasonal temperatures have significant effects on Utah's economy, both at the aggregate level and across different industries. I find that temperature increases in the fall have a particularly pervasive effect on sales. A 1° Fahrenheit temperature increase in the fall reduces sales by $3.418 million.

A PANEL ANALYSIS OF TEMPERATURE AND SALES IN UTAH by Reeves Coursey A Senior Honors Thesis Submitted to the Faculty of The University of Utah In Partial Fulfillment of the Requirements for the Honors Degree in Bachelor of Science In Quantitative Analysis of Markets and Organizations Approved: Scott Schaefer, PhD Thesis Faculty Supervisor Mike Cooper, PhD Chair, Department of Finance Scott Schaefer, PhD Honors Faculty Advisor Sylvia D. Torti, PhD Dean, Honors College May 2020 Copyright © 2020 All Rights Reserved Abstract In this paper, I document that seasonal temperatures have significant effects on Utah’s economy, both at the aggregate level and across different industries. I find that temperature increases in the fall have a particularly pervasive effect on sales. A 1° Fahrenheit temperature increase in the fall reduces sales by $3.418 million. ii Contents 1 Introduction 1 2 Data 4 3 Model and Main Results 11 4 Industry Analysis 16 5 Conclusion 24 A Appendix 26 1 1 Introduction In 2020, a NASA and National Oceanic and Atmospheric Administration (NOAA) analysis revealed that the 2010s was the warmest decade on record (NOAA, 2020). While increasing global average temperatures continue to attract attention from the media and public, economists interested in this phenomenon have one key question: How will temperature change affect economic growth? Previous literature on this topic documents that increased temperatures negatively affect economic growth and lists mechanisms such as decreased labor productivity and increased natural disasters. This paper investigates the relationship between temperature and economic growth in Utah, which provides an interesting case study due to Utah’s widely varying climate divisions and economic reliance on outdoor activities and production. A large and growing body of literature has documented the global and national effect of temperature on growth. Dell, Jones, and Olken (2009) shows that a 1° C temperature increase in developing nations could reduce economic growth by 1.1% per year. Numerous other papers investigate this issue as well e.g. (Bowen, Cochrane, & Fankhauser, 2012; Diffenbaugh & Burke, 2019; Lanzafame, 2014). However, a smaller body of work is dedicated to the relationship between temperature and growth in developed countries. Graff Zivin and Neidell (2014) shows that increased temperatures decrease labor supply in the United States and Cachon, Gallino, and 2 Olivares (2012) documents that high temperatures decrease the productivity and performance of workers. A recent paper by Colacito, Hoffmann, and Phan (2016) finds that rising temperatures could reduce economic growth in the US by one third over the next century. To date, no study has quantified the effects of temperature and economic growth in Utah specifically. A study of this kind is especially important because Utah’s average temperature has risen 3.32° F in the last fifty years, making it the fifth fastest warming state in the nation (Central, 2019). Furthermore, Utah’s geography and distance from the ocean will only exacerbate the future effects of global warming, and the effects on Utah’s economy will be felt in many different ways. For example, Utah ski resorts accounted for $1.43 billion, or nearly 1% of Utah’s economy in 2016 (Leaver, 2018). Lazar and Williams (2010) considers the impact of climate change on Park City Mountain Resort ski operations. They estimate that by 2075 the ski season will be shorter by approximately one month, which has the potential to drastically affect the industry. Burakowski and Magnusson (2012) shows that the number of skier days in Utah are 14% lower during low snowfall years than high snowfall years, and that Utah loses $87 million and over 1000 jobs during bad snow years. Temperature changes affect Utah’s economy in more subtle ways as well. For instance, higher temperatures increase the rate at which ozone is formed, and Utah is especially susceptible to ground level ozone due to its geography. As a result of increased ground-level ozone, Utah citizens will 3 experience negative effects on their respiratory and cardiovascular systems Air Pollution and Public Health in Utah (2016). Additionally, EPA (2016) predicts that construction crews and other outdoor occupations in Utah will have to alter their hours to avoid the increased heat . Observing trends in sales and temperature data at the city level, I find no effect of temperature on sales in Utah at the annual level. These results align with previous literature which found a small and statistically insignificant relationship between temperature and economic growth (Dell, Jones, & Olken, 2012). However, different approaches have been used to investigate this relationship. Colacito et al. (2016) disaggregated United States GDP and temperature data seasonally and found that temperature increases in the summer have a significant negative effect on economic growth, and temperature increases in the fall have a positive but less substantial effect. They speculate that the different signs on these effects suggest that the previous work has masked the impact of temperature on growth by ignoring the heterogeneous effects of different seasons. They show that the summer effect dominates the fall effect and conclude that the net effect of rising temperatures negatively affects the US economy. To avoid this masking effect, I follow the methods of Colacito et al. and disaggregate my data into the four seasons. I use sales data provided by the Utah State Tax Commission, reported separately by year, quarter, and season. I collect weather data from the Western Regional Climate Center, which pro- 4 vides average monthly temperatures and total inches of precipitation from seven different climate divisions in Utah. Finally, I regress quarterly sales against mean seasonal temperature and total precipitation with city and year fixed effects. I document that seasonal temperatures have significant effects on Utah’s economy, both at the total sales level and across different industries. I find that temperature increases in the fall have a particularly pervasive effect on sales. A 1° F temperature increase in the fall results in a $3.418 million decrease in sales. Industries with outdoor activity or working areas that are hard to air condition such as manufacturing and retail are especially affected. The rest of the paper is laid out in the following way. Section 2 describes my data sources. Section 3 describes my model and methods, including my baseline panel regression and regressions broken up by industry. Section 4 contains the industry analysis and explains potential economic mechanisms for my findings. Section 5 contains my conclusions, which includes future climate projections for Utah and their effect on sales. 2 Data This section describes my data sources and aggregation process from the Western Regional Climate Center and the Utah State Tax Commission. 5 Figure 1: Utah: Average Annual Temperature 2.1 Western Regional Climate Center Data An analysis of historical weather data in Utah shows a quickly warming state. Central (2019) shows that the nation’s fastest-warming cities all lie in the southwest and finds that Utah is the fifth fastest-warming state. They report that Utah has experienced an average temperature increase of 3.32° F in the last fifty years. Using Western Regional Climate Center Data, I document that Utah’s average annual temperature rose roughly 1.25° F from 1998-2017. A time trend of average annual temperature can be found in Figure 1. Temperature is defined in Fahrenheit and precipitation in inches throughout the paper. Additionally, Utah is divided into seven divisions that each experience unique 6 Figure 2: Climate Divisions of Utah climate patterns and temperatures.1 Figure 2 maps these divisions. I collect average monthly temperature and total monthly precipitation from each climate division for the twenty-year period from 1998-2017.2 A table contain1 The National Ocean and Atmospheric Organization (NOAA) divides Utah into the following divisions: Western, Dixie, North Central, South Central, Northern Mountains, Uinta Basin, and Southeast. 2 The WRCC works with NOAA, the National Weather Service, and other NOAA research institutes to provide division-specific climate data and research. Each climate division’s weather data is a weighted average of NOAA weather station data that the WRCC 7 ing average annual temperature for each climate division can be found in Appendix A.1.2. I find that each climate division roughly follows the same temperature pattern, a sign that global climate patterns dominate local climate effects in Utah. This can be seen in Figure 3. However, each division experiences different magnitudes of temperature change. The three climate divisions in Southern Utah, which include South Central, Southeast, and Dixie, all experienced a larger average temperature increase than other divisions and Utah as a whole. It follows that certain cities are going to experience greater temperature increases than others. Furthermore, each city is going to respond to temperature increases differently. For instance, an increase in winter temperature in Park City, a ski town, is likely to decrease sales. However, an increase in winter temperature in Springdale, the gateway city to Zion National Park, is likely to see sales increase due to increased willingness by tourist to be outside. Each city in Utah has these unique, unobservable characteristics that are time invariant which can be controlled for with panel methods. City-specific temperatures allow for across city temperature and sales differences to be controlled for in the regression model, eliminating one source of omitted variable bias3 . decided provides the best data for that division. 3 I use a data set that contains the coordinates of each city center in Utah. I then use GIS software to match city center coordinates to the climate division shape file to which they belong. 8 Figure 3: Average Climate Division Temperature However, Colacito et al. (2016) explains that models investigating the relationship between temperature and economic growth often find no significance at the annual level due to the opposite effect of different seasons. As such, I analyze quarterly patterns in climate and sales. I define winter as January through March, spring as April through June, summer as July through September, and fall as October through December. By defining seasons in this way I am able to closely align them with the standard fiscal quarters. I present average seasonal temperatures in Figure 4. Additionally, a table of average annual seasonal temperatures can be found in Appendix A.1.1. As seen in the figure and table, the seasons do not experience sim- 9 Figure 4: Average Seasonal Temperature ilar annual temperature trends. For instance, in 2008, spring was slightly hotter than the average spring, while winter was much colder than the average winter. Disaggregating annual data seasonally allows for within-season variation to be analyzed and conclusions to be made regarding the unique relationship between temperature and economic growth in each season. 10 2.2 Utah Tax Commission Data I collect data on taxable sales in Utah from 1998-2017 from the Utah State Tax Commission.4 This database provides data on sales in each city, broken up by year, quarter, and industry. Appendix A.2.2 contains a full description of what constitutes taxable sales. I show the time trend of average annual sales in Figure 5 and present a summary statistics table for this data in Appendix A.2.1. As seen in Figure 5, sales in Utah have risen dramatically since 1998. Because of this, year fixed effects must be included in our regression, or it else it would appear that rising temperature causes sales to increase, as they are positively correlated with each other. I also collect data on taxable sales by industry. These industries follow the North American Industry Classification System (NAICS). The Utah Tax Commission also includes four additional categories, which include private motor vehicle sales, special event sales, occasional/nonclassifiable sales, and prior period payments and refunds. I drop prior period payments and refunds, as the recorded sale date does not align with the actual sale date and temperature at the time of the sale. A full description of these extra categories can be found in Appendix A.2.3. Finally, all reported sales in the paper are in 1000s of dollars. 4 https://tax.utah.gov/econstats/sales/quarterly 11 Figure 5: Average Annual Sales 3 Model and Main Results In this section, I present my empirical model and results. I first run a basic fixed effects regression of sales on temperature with no additional controls. I then include year fixed effects to control for temperature and sales both exogenously rise with time and to. Finally, I add a control for precipitation. The regression with year fixed effects and a precipitation control is the main specification of the paper. 12 3.1 Basic Regression I first run a basic fixed effects regression with no controls: Yi,t = β1 T i,t + α i + ² i,t (1) where Yi,t and T i,t denotes annual sales and average temperature in city i and year t and α i denotes city fixed effects. I then run this same specification again but for each season: Yi,t,s = β1 T i,t,s + α i + ² i,t,s (2) where Yi,t,s and T i,t,s denotes annual sales and average temperature in city i, year t, and season s ∈ {S pring, Summer, Fall,W inter }. These results are presented in the first column of Table 1. Unsurprisingly, I find extremely significant positive results in these regressions. Without year fixed effects, the model shows that the temperature rise causes sales to rise, when actually temperature and sales both rise over the twenty-year panel irrespective of each other. So, I move to a model with year fixed effects. 13 3.2 Basic Regression with Year FE I run the basic regressions again, but now include year fixed effects: Yi,t = β1 T i,t + α i + α t + ² i,t (3) Yi,t,s = β1 T i,t,s + α i + α t + ² i,t,s (4) where α t ∈ {1997, 1998, ..., 2017} denotes year fixed effects. I present these results in the second column of Table 1. Now, the true effect of temperature on sales begins to show. However, I find a significant negative relationship between precipitation and temperature, and I hypothesize that precipitation has an impact on Utah sales as well. If this is the case, omitted variable bias is present in this specification. As such, I run regression 3 and 4 again but include a control for precipitation. 3.3 Main Specification I regress sales on average temperature and total precipitation and I include city and year fixed effects. I first regress sales on average temperature at 14 the annual level: Yi,t = β1 T i,t + β2 P i,t + α i + α t + ² i,t (5) where P i,t denotes average precipitation in city i in year t. I then run the same specification but for each season: yi,s,t = β1 T i,s,t + β2 P i,s,t + α i + α t + ² i,s,t (6) where P i,s,t denotes average total precipitation in city i, season s, and year t. These regressions are the main specifications for the paper. The results from each regression can be found in the third column of Table 1. The change in the temperature estimates suggests the omission of a precipitation control caused a positive bias in regressions 3 and 4. I propose the positive bias is due to precipitation and temperature being negatively correlated, as well as precipitation and sales being negatively correlated. If this is the case, the large magnitude of the bias in the spring and summer regressions suggests that changing precipitation levels due to climate change could also have the possibility of substantially affecting sales in Utah. I present the full results of regression 6 in Table 2. I only find significant results on the precipitation estimate in the spring. However, I find that all of the estimates are negative, and the effect is extremely large in the spring and summer at -$5,729 and -$5,206 million respectively. Given that EPA (2016) and others predict that precipitation will increase in Utah due to climate 15 change, the relationship between precipitation and sales is a pertinent one for Utah. Further analysis is recommended and saved for future research. At the annual level, I find that temperature increases positively affect sales. Specifically, a 1° F increase in temperature increases sales by $39,802 a year. While this result is statistically significant, it is economically insignificant. If temperature is increased by a full standard deviation, sales are predicted to rise by .0004 standard deviations of sales. See Appendix A.2.1 for a summary of the sales data. This is consistent with previous literature which either finds a statistically insignificant effect at the annual level (Colacito et al., 2016) or others which find a small effect due to the competing effects of different seasons. However, when I run the main specification for the paper at the seasonal level, the coefficients tell a clearer story of the relationship between temperature and sales. I find that a 1° F temperature increase in the summer decreases predicted sales by $2.196 million and that a 1° F temperature increase in the fall decreases sales by $3.418 million. While still small, the economic significance of this estimate is much larger than at the annual level. A one standard deviation increase in temperature decreases sales by .12 of its standard deviation. The positive sign on the spring coefficient, along with the large standard errors on the winter and spring coefficients, may contribute to the small and positive coefficient on the regression at the annual level. 16 Table 1: Main Results Effects of Temperature on Sales in Utah Winter (1) 77.26*** (21.17) 588.4*** (2) 77.26*** (21.17) -852.1 (3) 39.80** (16.03) -865.8 (184.1) (530.4) (637.8) Spring 1,276*** 664.5 87.25 (218.7) (548.8) (573.8) 1,344*** -399.7 -2196* (344.5) (746.9) (1,253) 1,495*** -3,270** -3418** (327.3) (1,434) (1,591) X X X Annual Summer Fall Year Fixed Effects Precipitation Observations 23,261 5,876 5,889 5,892 5,600 This table presents the estimate on the temperature variable. The first column is a basic panel regression with no controls. Column 2 add years fixed effects, and column 3 adds year and precipitation fixed effects. The final column presents the observations included in each specification. I have precipitation data for all observations, so adding precipitation and year controls does not cause any data points to be lost. The time sample is 19982017. Temperatures are in degrees Fahrenheit. Standard errors are in parenthesis and are clustered by city. ∗ ,∗∗ ,∗∗∗ represent significance at the 10%, 5%, and 1% percent levels. 4 Industry Analysis In this section, I show how temperature change affects different industries in Utah. Industry classifications are not consistent over the 20 year panel, so I use the 2013-2017 five-year segment. Furthermore, the Utah State Tax Commission only provides industry specific data for cities that are in the top sixty-five for sales in Utah. This reduces the number of observation in each 17 Table 2: Main Specification Temperature and Precipitation Whole Year Winter Spring Summer Fall Temperature 39.80** -865.8 87.25 -2,196* -3,418** (16.03) (637.8) (573.8) (1,253) (1,591) Precipitation -553.0* -80.77 -5,729*** -5,206 -681.2 (290.0) (885.7) (1,946) (4,273) (906.5) 23,261 0.927 5,867 0.931 5,889 0.928 5,892 0.925 5,600 0.936 Observations R-Squared regression to 325. I run the following regression: Yi,s,t, j = β1 T i,s,t, j + β2 P i,s,t, j + α i , +α t + ² i,s,t, j (7) This is the same specification as regression 6, the main specification of the paper, except that each iteration is run only using data from industry j. 4.1 Affected Industries I find significant results in 12 industries. These results can be found in Table 3. This table also reports the estimate’s share of average seasonal industry sales, along with the industry’s average percent of total Utah sales. I find that three industries have been most affected by rising temperature: Nonstore Retailers, Manufacturing, and Occasional/Non-Classifiable. These industries each constitute a significant portion of Utah sales and their es- 18 timates are relatively high compared to average seasonal sales numbers in their industry. In the Nonstore Retailers industry, I find that a 1° F increase in temperature decreases sales in the spring by $516,700, in summer by $1,691,000, and in the fall by $1,161,000. I do not find a significant effect in the winter. At almost 2% of Utah’s economy, the Nonstore Retailers industry represents a significant portion of Utah’s economy, which makes the pervasive temperature effects in this industry particularly concerning. Nonstore Retailer sales include sales from door-to-door solicitation, in-home demonstration, selling from portable stalls, distribution through vending machines, and other nonstore categories. The inverse relationship between sales and temperature in this industry is backed by previous literature, which finds that high temperatures negatively affect workers in interface areas, which are outdoor areas or areas that are hard to cool, such as portable stalls (Cachon et al., 2012). While increased outdoor temperature decreases worker productivity in the Nonstore Retailers industry, the negative temperature effects also likely come from changed consumer behavior. Luepker et al. (1996) finds that increased temperature reduces consumers’ willingness to shop outside and Griffit and Veitch (1971) finds that increased temperatures reduced people’s willingness to be around others, which would cause people to shop less. 19 In the manufacturing industry, I find that an increase of 1° F increases sales by $507,000 in winter, reduces sales by $1,488,000 in the summer, and reduces sales by $619,000 in the fall. I find no significant effect in spring. These findings align with previous literature, which finds that high temperatures reduce manufacturing worker productivity (Sudarshan, Somanathan, Somanathan, Tewari, et al., 2015) and that the relationship between temperature and manufacturing productivity is shaped like an inverted parabola (Cai, Lu, & Wang, 2018). This means that at low temperatures, temperature increases make manufacturing workers more productive, and they keep getting more productive as temperatures rise. However, at a certain point, the temperature becomes a detriment to the workers, and productivity begins to quickly fall as temperatures continue to increase. The positive effects of increased temperature in the winter are possibly explained by the average temperature in that season. With an average temperature of 39° F, winter has the lowest average temperature, leading to the belief that rising temperatures in winter have a positive first-order effect on sales and that summer and fall are both on the downward-sloping portion of the inverted parabola. Due to current global warming rates, we could see the coefficient on the winter estimate become negative as temperatures climb to the peak of the inverted parabola. On average, manufacturing constitutes 4.3% of Utah’s economy, which makes losses in this industry substantial and harmful. Finally, I find that the Occasional or Nonclassifiable industry is particularly 20 affected by temperature increases. I find that a 1° F increase in temperature decreases sales by $604,700 in the spring, by $2,957,000 in the summer, and by $1,560,000 in the fall. Economic analysis of this industry is difficult, as this industry primarily includes businesses that did not report NAICS codes or incorrectly reported NAICS codes. Further analysis is needed to see if there is a common type of business that fails to report or incorrectly reports its NAICS code. As the coefficient on the fall estimate from the main specification is the most significant and has the greatest effect on sales, I report all of the fall estimates from equation 7 in Appendix A.3.1. I only find nine significant negative effects in the fall (Finance and Insurance, Manufacturing, Nonstore Retailers, Occasional/Nonclassifiable, Retail-Electronics and Appliance Stores, Retail-Health and Personal Care, and Utilities). I also find that twenty-two of the thirty-four industries have negative coefficients on temperature. I believe there are two reasons for these imprecise results. The first is that I only use a five-year data segment, severely limiting the number of observations in each regression. Furthermore, the only sales data in these regressions are from the sixty-five largest sales cities in Utah, further reducing the number of observations and potentially causing selection bias. Further analysis is needed to completely interpret these results, but my initial conclusions are that climate change will affect a wide variety of industries in Utah. I believe this is the case for two reasons. First, previous 21 literature Colacito et al. (2016) finds that climate change is going to affect nearly every industry in the United States. Second, the significant negative estimate on the fall coefficient from the main specification of this paper, combined with finding that nine of the industry fall regressions have statistically significant negative estimates, leads me to believe I have correctly estimated the sign on many of the insignificant results, but suffer from imprecision. If this is the case, Utah’s economy as a whole is going to feel the effects of climate change, rather than just isolated industries. 22 Table 3: Industries with Significant Estimates Effects of Temperature on Sales in Utah by Industry and Season Industry Estimate % of Average Industry Sales Arts, Entertainment, & Recreation QTR 3 -314.7* .20% Industry % of Total Utah Sales 1.172% (188.5) Construction QTR 2 -203.6** .10% 1.401% .15% 0.420% .50% 0.430% .10% 0.420% .10% 3.610% .23% 4.001% .10% 3.833% (100.7) Finance & Insurance QTR 2 -83.5** QTR 3 -294.6*** (38.35) (93.81) QTR 4 -67.04* (39.38) Manufacturing QTR 1 507.1** (211.8) QTR 3 -1,488** QTR 4 -619.4* (732.3) (343.0) Mining, Quarrying & Oil & Gas Extraction QTR 4 385.2* .2% 0.860% (213.6) Nonstore Retailers QTR 2 -516.7*** .40% 1.352 .30% 1.377% .72% 1.780% (158.8) QTR 3 -1,691** (708.0) QTR 4 -1,161** (441.6) This table uses Utah State Tax Commission data separated by industry. The time frame is 2013-2017. The second column on this table presents the temperature estimate and standard errors from 7. The third column of this table divides the estimate by the average seasonal sales of the industry. The fourth column reports the average seasonal share of total Utah sales the industry makes up. 23 Table 3: Industries with Significant Estimates Effects of Temperature on Sales in Utah by Industry and Season Industry Estimate % of Average Industry Sales Industry % of Total Utah Sales .30% 1.734% 1.2% 1.896% .52% 2.162% Retail-Clothing and Accessories QTR 4 -207* .01% 3.585% Occasional or Nonclassifiable QTR 2 -604.7** (283.8) QTR 3 -2,957* (1,675) QTR 4 -1,560** (781.0) (115.5) Retail-Electronics & Appliances QTR 3 -710.9*** .33% 1.636% (236.3) QTR 4 -543.4** .20% 2.001% Retail-Health and Personal Care QTR 1 89.04** 0.00% 0.973% (225.3) (44.17) QTR 4 -134.5** 0.00% 1.001% (56.85) Retail-Sporting Goods, Hobby, Music, & Book Stores QTR 1 -116.7* 0.00% 1.930% (68.47) Utilities QTR 4 -1,273* .24% 3.191% (695.8) This table uses Utah State Tax Commission date separated by industry. The time frame is 2013-2017. The second column on this table presents the temperature estimate and standard errors from 7. The third column of this table divides the estimate by the average seasonal sales of the industry. The fourth column reports the average seasonal share of total Utah sales the industry makes up. 24 5 Conclusion The main findings of this paper show that historical temperature increases have negatively affected sales in Utah in the summer and winter months. Additionally, this effect is widespread- many industries in Utah saw sales decrease from rising temperatures. These results tell a frightening story for the future of Utah. Utah is the fifth fastest-warming state and Frankson, Kunkel, Stevens, and Easterling (2017) shows that Utah is projected to continue this trend throughout the 21st century. I present their findings in Figure 6. Figure 6: Note: Reprinted from (Frankson et al., 2017) 25 The low end of their projections see Utah’s average temperature rising by over 2° F in the coming century, while the high ends sees Utah’s average annual temperature rising by over 16° F by 2100. And while Utah and its citizens cannot do much to alter the global climate and emissions, it can prepare for the economic hardship that will come from rising temperatures. These results can inform political and industrial leaders so Utah can make the necessary changes to adapt to the warming world. 26 A Appendix A.1 Western Regional Climate Center Data A.1.1 Average Seasonal Temperature Table 4: Average Seasonal Temperature Year Winter Spring Summer Fall 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 34.34 36.53 36.20 33.83 31.05 35.97 33.29 35.55 33.29 33.91 30.60 34.54 32.34 31.68 35.79 30.23 36.40 39.39 34.95 36.33 52.90 53.44 59.80 58.41 58.84 56.92 57.61 55.15 59.29 58.80 54.28 55.95 54.05 53.10 59.38 57.41 57.09 57.61 58.30 57.58 69.95 67.56 70.16 70.84 70.01 71.35 67.62 68.80 68.31 70.84 69.81 69.71 69.46 69.82 70.54 69.56 69.28 69.43 68.90 69.68 38.32 41.97 37.15 39.59 38.11 40.56 38.99 40.34 38.66 38.91 40.20 36.31 40.24 38.44 40.76 36.66 41.59 39.25 41.51 41.67 This table presents average seasonal temperatures for Utah. The degrees are in Fahrenheit. 27 A.1.2 Average Climate Division Temperature Table 5: Average Temperature State and Division Year Utah Western North Central South Central Dixie Northern Mountains Uinta Basin Southeast 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 48.88 49.88 50.83 50.67 49.50 51.20 49.38 49.96 49.89 50.61 48.72 49.13 49.02 48.26 51.62 48.47 51.09 51.42 50.91 51.31 49.85 50.75 51.83 51.70 50.30 52.45 50.33 50.68 50.70 51.32 50.00 49.90 49.93 49.05 52.52 48.95 52.25 52.66 51.69 51.86 48.55 49.34 50.19 50.08 48.28 50.77 48.51 49.15 49.32 50.11 47.89 48.10 48.51 47.60 51.30 47.75 50.44 51.34 50.60 50.12 45.93 47.24 48.34 47.99 47.25 48.65 46.71 47.48 47.25 48.07 46.43 46.77 46.49 45.92 48.82 46.15 48.78 48.78 48.48 49.20 57.67 59.75 60.58 60.63 60.29 61.15 59.55 59.70 59.90 60.94 59.45 59.80 59.29 58.43 60.77 58.84 61.06 60.57 60.41 60.99 40.91 41.61 42.34 42.40 40.80 42.70 41.10 41.60 41.49 42.45 39.95 40.56 40.79 39.91 43.40 40.80 42.74 43.68 43.09 43.05 46.75 47.32 48.14 47.95 46.45 48.00 46.70 47.45 47.30 47.37 45.11 46.01 45.91 44.99 49.55 45.00 48.22 48.89 48.18 48.91 52.48 53.15 54.40 53.95 53.17 54.67 52.75 53.67 53.25 54.04 52.23 52.75 52.25 51.90 54.99 51.71 54.14 54.05 53.95 55.07 This table presents average annual temperatures for the state of Utah and each climate division. The degrees are in Fahrenheit. The average temperature in Utah takes an equally weighted average of all seven climate divisions in Utah. 28 A.2 Taxable Sales A.2.1 Taxable Sales Data Table 6: Average Annual Sales Year Sales Standard Deviation Observations 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 108326.7 110962.6 116061.3 119332.0 119321.9 119885.5 129569.7 142819.2 161739.6 172028.1 164744.3 144217.0 147090.4 155222.3 164291.8 169985.6 178922.5 186687.3 194576.7 206823.9 396905.1 396948.7 413728.4 415667.8 405454.8 389359.2 416148.4 446688.7 489992.1 514965.9 484646.7 429681.8 446908.6 472614.6 510408.0 524887.1 547064.3 576663.4 599681.1 633701.8 1,054 1,070 1,079 1,080 1,087 1,085 1,091 1,100 1,111 1,113 1,114 1,127 1,128 1,139 1,160 1,163 1,159 1,160 1,163 1,182 This table presents average annual sales in Utah. Sales are in thousands. 29 A.2.2 Taxable Sales Definition Taxable sales are defined by the Utah State Tax Commission in the following way. “Taxable sales are calculated by summing data from sales and use tax returns from the local sales tax distribution each month and aggregating by various combinations of time period, location and industry categories. Only transactions that are taxable are included in the data (see Utah Code §5912-103 and §59-12-104 for more information on what is taxable and what is exempt). Reported taxable sales numbers are gross numbers and do not include adjustments such as vendor discount or the Tax Commission’s administrative fee. 5 A.2.3 Description of Other Categories Private Motor Vehicle Sales Private motor vehicle sales are taxable sales that occur between two private parties. The purchaser will pay the sales tax at the time the vehicle is titled and registered at the Utah Division of Motor Vehicles. For purposes of these reports, the date of the transaction is assumed to be the month prior to when the sales tax is paid. Special Event Sales 5 https://tax.utah.gov/econstats/sales/about-sales-reports 30 Taxable sales for a special event where a temporary sales tax license is required are reported in this category. A special event is a one-time event or an event that runs for six months or less where taxable sales are made and sales tax must be collected. Special events fall under a variety of situations including sporting events, state and county fairs, festivals, antique shows, gun shows, food shows, art shows, auctions, mall kiosks, swap meets, conventions, hobby shows, concerts, seasonal stands in malls, and other similar events. Prior to 2008, Special Event Sales were not included as their own category but were included in ‘Occasional/Nonclassifiable’. Occasional/Nonclassifiable Occasional/Nonclassifiable includes sales that were unable to be classified in any of the other categories. This category primarily consists of sales for companies that did not report a NAICS code or companies that reported an incorrect NAICS code. Taxable sales categorized in the ‘Out of State’ location are also included in this category. A.3 Industry Analysis A.3.1 Fall Industry Regressions 31 Table 7: Fall Industry Analysis Effects of temperature variation on sales in Utah by Industry from 2013-2017 Industry Accommodation Estimate % of Sales -66.69 2.020% (264.7) Admin. & Support & Waste Management -64.64 0.363% (39.39) Agriculture, Forestry, Fishing & Hunting 7.824 0.020% (12.04) Arts, Entertainment, and Recreation -305.1 0.915% (225.3) Construction -569.0 1.396% (435.7) Educational Services 10.18 0.218% (35.17) Finance and Insurance -67.04* 0.422% (39.38) Food Services & Drinking Places -647.6 7.915% (575.9) Health Care and Social Assistance 14.6 0.231% (38.24) Information 4.992 3.959% (323.5) Management of Companies & Enterprises 3.208 0.038% (30.06) Manufacturing -619.4* 3.834% (343) Mining, Quarrying, & Oil & Gas Extraction 385.2* 0.859% (213.6) Nonstore Retailers -1,161** 1.776% (441.6) Occasional/Nonclassifiable -1,560** (781) 2.163% 32 Table 7: Fall Industry Analysis (Continued) Effects of temperature variation on sales in Utah by Industry from 2013-2017 Industry Other Services- Except Public Administration Estimate % of Sales -62.56 2.521% (193.9) Private Motor Vehicle Sales -83.43 1.280 (172.7) Professional, Scientific, & Technical Services 105.6 1.208% (135) Public Administration 21.47 0.410% (50.82) Real Estate, Rental, & Leasing -383.6 2.140% (276.6) Retail-Build. Material, Garden Equip. & Supplies Dealers -488.6 5.199% (727.4) Retail-Clothing & Clothing Accessories Stores -207* 3.852% (115.5) Retail-Electronics & Appliance Stores -543.4** 2.009% (225.3) Retail-Food & Beverage Stores 266.9 8.445% (1,103) Retail-Furniture & Home Furnishings Stores -55.46 1.944% (125.2) Retail-Gasoline Stations -25.31 1.697% (137.7) Retail-General Merchandise Stores -559.7 12.849% (580.5) Retail-Health and Personal Care -134.5** 1.005% (56.85) Retail-Miscellaneous Retail Trade -199.4 3.173% (210.8) Retail-Motor Vehicle & Parts Dealers -1,140 (1,078) 11.058% 33 Table 7: Fall Industry Analysis (Continued) Effects of temperature variation on sales in Utah by Industry from 2013-2017 Industry Retail-Sporting Goods, Hobby, Music & Book Stores Estimate % of Sales 376.2 2.304% (232.6) Special Event Sales -8.260 0.090% (36.45) Transportation & Warehousing 637.8 0.235% (401.9) Utilities -1,273* 3.191% (695.8) Wholesale Trade-Durable Goods 7.952 6.867% (18.67) Wholesale Trade-Electronic Markets -217.5 0.070% (224.4) This table presents the temperature estimate and standard errors from equation 7 as well as the percent of total sales that industry constitutes from 2013-2017 in the fall 34 References Air pollution and public health in utah. (2016). Retrieved from https:// www.health.utah.gov/utahair/pollutants/O3/ Bowen, A., Cochrane, S., & Fankhauser, S. (2012). Climate change, adaptation and economic growth. Climatic change, 113(2), 95–106. Burakowski, E., & Magnusson, M. (2012). Climate impacts on the winter tourism economy in the united states. Cachon, G. P., Gallino, S., & Olivares, M. (2012). Severe weather and automobile assembly productivity. Columbia Business School Research Paper(12/37). Cai, X., Lu, Y., & Wang, J. (2018). The impact of temperature on manufacturing worker productivity: Evidence from personnel data. Journal of Comparative Economics, 46(4), 889–905. Central, C. (2019). American warming: The fastest-warming cities and states in the u.s. Retrieved from https://www.climatecentral.org/news/ report-american-warming-us-heats-up-earth-day Colacito, R., Hoffmann, B., & Phan, T. (2016). Temperature and growth: A panel analysis of the united states (IDB Working Paper Series No. IDBWP-676). Washington, DC. Retrieved from http://hdl.handle.net/ 10419/146465 (hdl:11319/7654) 35 Dell, M., Jones, B. F., & Olken, B. A. (2009). Temperature and income: reconciling new cross-sectional and panel estimates. American Economic Review, 99(2), 198–204. Dell, M., Jones, B. F., & Olken, B. A. (2012). Temperature shocks and economic growth: Evidence from the last half century. American Economic Journal: Macroeconomics, 4(3), 66–95. Diffenbaugh, N. S., & Burke, M. (2019). Global warming has increased global economic inequality. Proceedings of the National Academy of Sciences, 116(20), 9808–9813. EPA. (2016). What climate change means for utah. Retrieved from https://19january2017snapshot.epa.gov/sites/production/ files/2016-09/documents/climate-change-ut.pdf Frankson, R., Kunkel, K., Stevens, L., & Easterling, D. (2017). Utah state climate summary. Retrieved from https://statesummaries.ncics .org/chapter/ut/ Graff Zivin, J., & Neidell, M. (2014). Temperature and the allocation of time: Implications for climate change. Journal of Labor Economics, 32(1), 1–26. Griffit, W., & Veitch, R. (1971). Hot and crowded: influence of population density and temperature on interpersonal affective behavior. Journal of Personality and Social Psychology, 17(1), 92. Lanzafame, M. (2014). Temperature, rainfall and economic growth in africa. Empirical Economics, 46(1), 1–18. Lazar, B., & Williams, M. W. (2010). Potential impacts of climate change for u.s. wastach range ski areas: Projections for park city mountain resort in 2030, 2050, and 2075. Retrieved from http://arc.lib.montana .edu/snow-science/objects/ISSW_P-023.pdf Leaver, J. (2018). The state of utah’s travel and tourism industry. Luepker, R. V., Perry, C. L., McKinlay, S. M., Nader, P. R., Parcel, G. S., Stone, E. J., . . . others (1996). Outcomes of a field trial to improve children’s dietary patterns and physical activity: the child and adolescent trial for cardiovascular health (catch). Jama, 275(10), 768–776. NOAA. (2020). State of the climate: Global climate report for annual 2019. Retrieved from https://www.ncdc.noaa.gov/sotc/global/201913 Sudarshan, A., Somanathan, E., Somanathan, R., Tewari, M., et al. (2015). The impact of temperature on productivity and labor supply-evidence from indian manufacturing (Tech. Rep.).