Sexually transmitted infections (STIs) have important consequences for individuals and society. Extensive literature has shown that various individual factors impact STIs. However, much less is known about their structural causes and how they affect sexual behavior and sexual network formation. In the first two chapters of this dissertation, I investigate how sex ratios and ethnic divisions affect sexual activity and the spread of STIs. In the third chapter, I analyze the effect of ethnic-based romantic homophily on STIs. I provide a brief description of each chapter below. Chapter 1. We extend a theory of fidelity in a two-sided economy, and empirically discriminate between different rationales of sexual network formation by testing their implications for how sex ratios affect sexual activity, relationship stability, and the spread of sexually transmitted diseases in men versus women. We use a unique individual-level dataset in combination with census data from England and Wales, a setting where adult women outnumber adult men. Exploiting variation in cohort/ethnicity/region-specific sex ratios as a quasi-natural experiment, we find that a decrease in sex ratio imbalance decreases sexual infidelity and the number of serial partners, and increases the likelihood of safe sex. This in turn reduces the likelihood of acquiring a range of sexually transmitted infections and diseases, including chlamydia, gonorrhoea, genital warts, and herpes. Consistent with the rationale underlying the formation of egalitarian (in)fidelity networks, the effects of the sex ratio on sexual activity are larger for men compared to women, while its effects on sexual diseases are larger for women compared to men. The causality of these effects is established using classical and recent instrumental variables approaches and various robustness checks. For falsification, we show that sex ratios have no impact on several "atheoretical" health conditions, such as Parkinson's disease, chronic lung disease, heart attack, stroke, and diabetes, which do not arise from sexual interactions. Chapter 2. In societies organized around distinct racial and ethnic groups, limited communication between these groups might increase the search cost of sexual partners outside of own group, leading to racially segregated sexual networks and low risks of sexually transmitted diseases. At the same time, because sexual infidelity is more likely to be discovered when the cheated-upon individuals are co-ethnics, individuals in multiracial societies might find it cheaper to select sexual partners from diverse ethnic groups to hide their infidelity, which would lead to large interethnic sexual networks and high risks of STIs. We test these conflicting hypotheses by analyzing the causal effect of neighborhood-level racial diversity on sexual activity and STIs, using unique individual-level data from England, Wales, and Scotland. We find that individuals residing in multiracial neighborhoods have a greater number of sexual partners and are more likely to be infected with a wide range of STIs than their counterparts residing in more racially homogeneous neighborhoods. We use traditional and new instrumental variables approaches and various robustness checks to establish causation. Analyzing mechanisms, we find that within racially diverse neighborhoods, individuals who select sexual partners from diverse racial groups are more likely to be infected with STIs, holding the number of partners and other individual characteristics fixed. For falsification, we conduct a reverse-placebo test showing that racial diversity has no effect on a wide range of health conditions that do not arise from sexual interactions. From a policy perspective, our analysis implies that policies that promote racial and ethnic integration are likely to reduce unhealthy sexual activity and the spread of STIs in racially heterogeneous societies. Chapter 3. A classical hypothesis in social network theory holds that central individuals are more likely to receive and spread information than are their peripheral counterparts. We test this hypothesis in the context of sexual networks and sexually transmitted diseases, using data from the United Kingdom. Romantic homophily - the tendency to select sexual partners with similar ethnic background - is used as a measure of the extent to which an individual is peripheral in a sexual network. We find that more sexually homophilous individuals have a lower risk of sexual infections. This effect is causal, and larger for women, Whites, and heterosexuals.