Suppose you have annual data on C (average grams of coffee consumed per capita), YD (real per capital disposable income), PC (price index for coffee), PT (price index for tea), and POP (population in millions). You regress C on lnYD (ln is natural log), PC, PT, and POP, obtaining a reasonable R-square but no significant t statistics. What do you suspect is the problem here, and how would you remedy it?